Graham A.C. Later Mohist Logic, Ethics and Science, 1978

Later Mohist Logic
Ethics and Science
A.C. Graham
The Canons and other later writings of the
school of Mo-tzu, dating from about 300
B.C., contain nearly all that survives of the
logic of ancient China, and its optics and
mechanics, the only organised set of
geometrical definitions, and the only fully
rationalised system of ethics. They repre
sent the high point of abstract rationality
in traditional Chinese civilization, and are
crucial documents for any inquiry into its
achievements and limitations in logic and
science. Unfortunately their formidable
textual difficulties have hitherto made it
impossible to use them with any con
fidence, and English translations of Mo-tzu
have omitted them. Western sinologists
have generally ignored this rich material
with the result that they have been forced
to draw their conclusions about Chinese
logic from the almost negligible remains
of the Sophists.
The present work begins with a general
account of the school of Mo-tzu, its social
basis as a movement of craftsmen, its
isolated place in the Chinese tradition,
and the nature of its later contributions to
logic, ethics and science. The relation of
Mohist thinking to the structure of the
Chinese language is also discussed. The
textual problems of the later writings,
the grammar and style, the technical
terminology, the significance of stock
examples, and the overall organisation of
the documents, are then explored in detail.
With the investigation of these preliminary
questions, the possibilities of interpreta
tion are confined within controllable limits.
The edited and annotated Chinese text
follows, with an English translation and
commentary, a glossary, and a photo
graphic reproduction of the unemended
text from the Taoist Patrology.
LATER MOHIST LOGIC, ETHICS A N D SCIENCE
Later Mohist Logic,
Ethics and Science
A . C . GRAHAM
T H E C H I N E S E U N I V E R S I T Y PRESS
T H E CHINESE U N I V E R S I T Y OF H O N G K O N G
SCHOOL OF O R I E N T A L A N D A F R I C A N STUDIES
UNIVERSITY OF L O N D O N
1978
© The Chinese University Press, 1978
All rights reserved
Published in the Far East by The Chinese University Press,

Shatin, New Territories, Hong Kong, and in Great Britain,
Europe and North America by the School of Oriental and
African Studies, University of London, Malet Street, London
WC1E 7HP.
ISBNs
The Chinese University Press 962-201-142-X
School of Oriental and African Studies 0 7286 0025 0
Set in Monotype Imprint

Printed in Hong Kong by Libra Press Limited
CONTENTS
PREFACE xi
PART I. INTRODUCTION
1/1 The Mohist philosophy 3

1/1/1 The historical background 3
1/1/1/1 Early Mohism 3
1/1/1/2 The metaphysical crisis in the 4th century 15
1/1/1/3 Later Mohism 22
1/1/2 Later Mohist logic, ethics and science ... 25
1/1/2/1 Logic and the Chinese language 25
1/1/2/2 The four objects of knowledge 30
1/1/2/3 Knowing, names and objects, change and necessity 32
1/1/2/4 Description 35
1/1/2/4/1 The art of description in the Canons 35
1/1/2/4/2 The art of description in Names and objects 40
1/1/2/5 Ethics 44
1/1/2/6 The sciences 53
1/1/2/7 The absence of a later Mohist metaphysic 59
1/1/2/8 Relations with other schools 61
1/1/2/9 The loss and rediscovery of the Mohist logic 64
1/2 Textual problems 73
1/2/1 Problems common to the six chapters 73
1/2/1/1 External evidence 73
1/2/1/2 Graphic conventions 76
1/2/1/3 Graphic corruption 81
1/2/1/4 The Sung taboos 85
1/2/2 The Canons and Explanations 87
1/2/2/1 The order of the Canons 87
1/2/2/2 The Canons at Stage 1 89
vi Contents
1/2/2/3 The Canons at Stage 2 91

1/2/2/4 The two parts of the Canons and Explanations 94
1 /2/2/'5 The head characters of the Explanations 95
1/2/2/6 Glosses 98
1/2/2/7 Fragmentation 100
1/2/3 The reconstruction of Expounding the canons and Names
and objects from the Ta-ch'u and Hsiao-ch'u 101
1/3 Grammar Ill

1/3/1 Introduction Ill
1/3/2 Sentence structure 113
1/3/3 'XisY' 116
1/3/4 Pronouns 119
1/3/5 Negatives 123
1/3/6 Distribution 127
1/3/7 Questions 136
1/3/8 Comparison 137
1/3/9 Numerals 140
1/3/10 Quotation ... 140
1/3/11 Logical implication 141
1/3/12 Keywords ... 145
1/3/12/1 Wei m 145
1/3/12/2 Cte* 147
1/3/12/3 Soft 148
1/3/12/4 YiSk 149
1/3/12/5 ErhM 152
1/3/12/6 Yu m 152
1/3/12/7 Yi % 153
1/3/12/8 Yehfe 153
1/3/13 Causative uses 159
1/3/14 Style 161
1/4 Technical terminology 167
1/4/1 The problem of technical terminology 167
1/4/2 Cheng IE, 169
1/4/3 Chienpai^ 170
Contents vii
1/4/4 Chih l h and hsing ft .. 177

1/4/5 Chih ft .. 178
1/4/6 Ch'ing Wf .. 179
1/4/7 Ch'ii%,№,®, ft ... . .. 182
1/4/8 Ch'ilan fll .. 184
1/4/9 Fan ifi. .. 184
1/4/10 Fang $C .. 186
1/4/11 Fu a .. 187
1/4/12 Hai * .. 188
1/4/13 Hsien 5fe, hsien chih 3fcftl and wei k'o chih ^ J f t ..
l .
188
1/4/14 Ku @ and ku #C .. 189
1/4/15 Kuo ® .. 190
1/4/16 Li m .. 191
1/4/17 urn .. 191
1/4/18 Liang M, p'ien M, chouffland yi St .. 192
1/4/19 Lun fra and lun lieh â^J .. 194
1/4/20 Mao tfa and tnu ... ... 194
1/4/21 Ming and shih f? ... 196
1/4/22 Ming # ... .. 199
1/4/23 Pein ... . .. 199
1/4/24 Sheng № .. 200
1/4/25 Sheng .. 200
1/4/26 Shih^. .. 200
1/4/27 Tai fxf .. 202
1/4/28 Tang % .. 202
1/4/29 Ti ft .. 203
1/4/30 Tsai t£- and ts'un # ... .. 206
1/4/31 Tz'u № ... . .. 207
1/4/32 Wei % .. 209
1/4/33 Wu % .. 210
1/4/34 WuS., ft ... . .. 210
1/4/35 Wuyim&L ... . .. 212
1/4/36 XyuMY t'ung (shuo) m(t&) ... .. . 212
1/4/37 Yi M and hsiang #1 ... ... ... 213
viii Contents
1/4/38 Yi% 214

1/4/39 Yin H 214
1/4/40 Yujwushuo^l^Wi 216
1/5 The stock examples 217
1/5/1 Introduction 217
1/5/2 Ox and horse 217
1/5/3 Dog 218
1/5/4 Crane 219
1/5/5 Miludeer 220
1/5/6 Louse 221
1/5/7 Stone 222
1/5/8 Wall 223
1/5/9 Pillar 224
1/5/10 The circle and the square 225
1/5/11 Illness ..' 226
1/5/12 Tsang M and Huo St ('Jack and JUT) 226
1/5/13 North and South 227
1/5/14 The road 227
1/6 The organisation of the Canons and Explanations 229

1/6/1 The five divisions of the definitions and the propositions 229
1/6/2 The lost definitions of words in the 10 theses of Mo-tzu 235
PART II. TEXT AND TRANSLATION

2/1 Editorial conventions 239
2/2 The ten theses of Mo-tzu 241
2/3 The fragments of Expounding the canons 243
2/4 The Canons and Explanations 261

2/4/1 The definitions (A 1-87) 261
2/4/1/1 Description (A 1-6) 262
2/4/1/2 Action (A 7-39) 269
2/4/1/2/1 The virtues (A 7-14) 270
2/4/1/2/2 Leader and follower (A 15-21) 277
2/4/1/2/3 'Nurture of life' concepts (A 22-28) 280
Contents IX
2/4/1/2/4 Saying (A 29-33) 284

2/4/1/2/5 Ruler and subject (A 34-39) 289
2/4/1/3 Knowledge and change (A 40-51) 293
2/4/1/4 The sciences (geometry) (A 52-69) 301
2/4/1/5 Disputation (A 70-75) 316
2/4/1 /6 Appendix: words with more than one usage (A 76-87) 323
2/4/2 The propositions (A 88-B 82) 336
2/4/2/1 The procedures of description (A 88-B 12) 336
2/4/2/2 Knowledge and change (B 13-16) 364
2/4/2/3 Problems in the sciences (B 17-31) 369
2/4/2/3/1 Optics (B 17-24) 372
2/4/2/3/2 Mechanics (B 25a-29) 385
2/4/2/3/3 Economics (B 30, 31) 397
2/4/2/4 Problems in disputation (B 32-82) 398
2/4/2/5 Appendix: Kung-sun Lung's Essay on pointing things
out interpreted in the light of B 32-82. 457
2/5 Names and objects 469
PART III. APPENDICES
3/1 Abbreviations 497
3/2 The Taoist Patrology text of Mo-tzu ch. 40-45 499
3/3 The arrangement of the Canons at Stage 2 527
3/4 Bibliography 531

3/4/1 Western languages 531
3/4/2 Chinese and Japanese 535
3/4/2/1 Classical texts 535
3/4/2/2 Sinological works 535
3/5 Indexes 547
3/5/1 Chinese glossary 547
3/5/2 Proper names 571
3/5/3 Subjects 577
PREFACE
T H E thought of the later Mohists is the great unknown quantity in
ancient Chinese philosophy. Nearly all that survives of Chinese logic and
some of the most interesting of Chinese science is contained in six chapters
of Mo-tzu so mutilated and corrupt that many scholars prefer to ignore
them as unusable. We know however that in the 4th and 3rd centuries B.C.
the art of pien Wt 'disputation' exerted an important and still unmeasured
influence, that the little we have of the sophists H u i Shih and Kung-sun
Lung gives a wholly inadequate idea of it, and that there is a whole dimension
in the thought of Chuang-tzu, for many the greatest of Chinese philoso
phers, which is inaccessible without the aid of the terminology and proce
dures of disputation, preserved only in the Canons of the later Mohists.
Although this gap in our knowledge is the immediate business only of
sinologists, it is relevant to anyone who finds himself asking certain general
questions about China and about the history of civilization in general. Is
the Greek ideal of rationality a discovery made once only in history, or
does it have parallels in India and China ? Are there episodes in Oriental,
as in Greek and Mediaeval science, which anticipate in part the Scientific
Revolution in the 17th century? T o what extent does language structure
determine thinking, and is the Chinese language an obstacle to analytic
thinking ?
Scholars in China have been working intensively on the Canons ever
since the appearance of Sun Yi-jang's great commentary on Mo-tzu in
1894. Up to the publication of T'an Chieh-fu's Mo-pien fa-wet in 1958
they were making continuous progress in identifying systematic corruptions
and aberrant graphs, and in interpreting the less problematic sections. But
over a large part of the text one editor's guess still looks as good as another's,
and the desperate indulgence in conjectural emendation in Kao Heng's
Mo-ching chiao-ch'uan (also published in 1958) warns us that the traditional
approaches of Chinese textual criticism, adequate for the main body of
Mo-tzu, have come to a dead end. They offer no prospect of the overall
view of the documents as wholes without which even the easily intelligible
parts cannot be seen in perspective. Since Alfred Forke's premature
German version of 1922 no Westerner has dared to include the Canons in
translations from Mo-tzu. Joseph Needham, who could not afford to
xii Preface
dispense with the Mohist testimony in his Science and civilization in China
is almost the only Western scholar who has tried to find a footing in these
quicksands.
But it does not follow that we must abandon the Canons; there are
stricter methods of textual criticism which have not yet been tried. With
rare exceptions such as Luan T'iao-fu, scholars have tended to study the
Canons piecemeal, focusing their attention on each section in turn as though
it had no context, changing its characters and shifting its boundaries until
it seems to make sense, and then moving on to do the same to the next.
Valuable as much of this work has been, it is time to stand back from the
text and ask general questions about the Mohist writings, not only about
graphic and textual problems, but about the grammar, the technical
vocabulary, the recurring stock examples, the overall organisation of the
parts, and answer as many as of these preliminary questions as possible
before committing ourselves to the interpretation of any one Canon. When
this is done certain principles emerge which drastically reduce the oppor
tunities for unverifiable conjecture:
(1) Graphic emendation should not be the first but the last resort,
except when required by parallelism or comparison between Canon and
Explanation, or when the Harvard-Yenching concordance establishes that
corruption is systematic. One should mark the divisions between sections
on purely textual grounds, ignoring the problems of interpretation which
tempt us to shift them to superficially more suitable positions. Indeed in
nearly all cases they can be established definitively. The head characters
which Liang Ch'i-ch'ao first noticed as marking the beginnings of
Explanations seem originally to have been written in the margin, and were
often incorporated below in stead of above the first character of the text;
otherwise they are almost invariably present and rightly placed, and serve
as a reliable guide to dividing the sections. Transposition should also be on
textual grounds. If we are to assume the right to transfer a passage merely
because it makes better sense in another position, it must be within confined
areas where we perceive that fragmentation is general, such as the
Explanations of A 22-39 and the whole of the 'Bigger pick* (a collection of
mutilated scraps from which two documents can be reconstructed,
Expounding the canons, and half of Names and objects, the rest of which
in its companion the 'Smaller pick'). We are also entitled to move illustra
tions beginning with jo 3af 'Like . . a few places forward or backward,
since they are often parenthetic and look like glosses.
(2) The oddity of much of the language, which tempts us to smooth
it away by emendation, is that of logician's language generally. Its syntax
Preface xiii
is extraordinarily clear and strict, confining each particle to a single gram

matical function, and sacrificing ordinary idiom for syntactic consistency.
The vocabulary is also that of a logician, generally simple and regular
(a fact obscured by the frequency of rare graphic forms), but complicated
by a special technical terminology. T o appeal to the rare words in diction
aries or the idioms of other Chinese styles is mistaken in principle. A n
analysis of the language is an indispensable tool for interpreting with
confidence, and sometimes also confirms emendations which had seemed
to be merely conjectural.
(3) Many of the words commonly emended as unintelligible are
either technical usages or stock examples (pillar, wall, milu deer) with a
constant significance; these can generally be elucidated by collating
examples within and sometimes outside the later Mohist corpus. The
technical terms were especially liable to corruption, and graphs confusable
with theirs are often observed in obscure contexts; in stead of emending
them away we should search the text systematically for the places to restore
them.
(4) The idea that Canons which do not fall into easily recognised
sequences are strung at random, so that the possibilities of interpretation
are unrestricted by context, is completely mistaken. The Canons and the
older Expounding the canons constitute an organised summa based on a
fourfold classification of knowledge which also underlies Hsiin-tzu's Right
use of names. In this summa nothing is ever repeated, but nearly everything
connects with something elsewhere. Except in B 32-82, which does seem
to be a miscellaneous collection of problems in disputation, no section can
be profitably read out of context.
I have been intrigued by the Canons almost since I began to work on
Chinese philosophy, both for their content and for the intellectual fascina
tion of playing with riddles. Whenever, following other interests, I have
learned something of Chinese textual criticism, grammar, science, the
taboo system, I have returned to take another look at the Canons. The task
has something in common with breaking a code; at a certain point the
possibilities narrow, each answer generates new answers which interrelate
to support each other, and a coherent message begins to emerge. Many
obscurities remain, and my explanations will not always convince
others as easily as myself. But I believe that the summa of the later Mohists
is now a usable source, a text in which the reader can have some confidence
of finding his way about, and that the time has come to attempt an integral
translation.
xiv Preface
The present edition passes over many older suggestions without

comment. So much of previous work consists of speculative emendations
that to do it full justice I should have to put a note to every second word
explaining why I have left the text as it is. This is not a matter of ingratitude
to my predecessors; not only has their work been indispensable, much of
it reflects the intensive lifelong familiarity with the script and the whole of
the early literature which is beyond the reach of a Western sinologist. But
any reader who accepts my answers to the preliminary questions will
discard a large proportion of older opinions as obsolete, and it would be
pointless to distract him with them; in the Canons it is difficult enough as
it is to see the wood for the trees. Obviously there are places where the
critical reader will prefer rival solutions, some of which I may have failed
to appreciate and to report; but I assume that he will continue to consult
T'an Chieh-fu and more recent editors such as L i u Ts'un-yan and L i
Yu-shu.
Although the book is intended primarily for sinologists, it is to be
hoped that the lay reader curious to find out how much or little ancient
China achieved in logic and science will be able to make use of the transla
tions, the general introduction, and as many pages as are not too crowded
with Chinese characters. He should perhaps be warned, in case he does not
know it already, that no concept of another civilization will have an exact
equivalent in our own, so that no English term can be more than a conven
tional equivalent employed on the assumption that readers will be on the
lookout for the differences as well as the similarities of the Chinese and the
English word. As a general rule, I have avoided English words with too
precise a technical sense in philosophy, since they will carry with them
their whole context in Western thought. But there is one spectacular
,
exception, the use of 'a priori for the Later Mohist hsien ite. It is the only
Western term which resembles hsien in the technical sense which I attribute
to it; but that it is very far from being identical will be obvious from such
,
expressions in the translations as "refer to 'a priori ", even "desire
'a priori* Much of the value of Oriental philosophies for a Western
investigator lies in the opportunity to explore concepts which are to a
greater or lesser degree unlike his own, yet compose a coherent world-
picture; and in the case of the Later Mohists, who are meticulous in defin
ing their terms, this coherence, in ethics especially, is a logical inter
dependence without parallel in Chinese thought.
During the many years in which I have dipped into, played with,
despaired of, and at last set my whole mind on the Canons, I have discussed
them with more people than I can remember. But I should like to acknow-
Preface xv
ledge my especial gratitude to Nathan Sivin, who collaborated with me on

a translation of the optical sections published in 1973, and to D . C. Lau,
Joseph Needham, K . Enoki, L i u Ts'un-yan, my father-in-law Y . P. Chang,
and my former Cornell graduate student Richard Bodman.
I should like also to express my gratitude to the Society for the
Humanities, Cornell University, where I enjoyed a year of leisure in 1972-3
indispensable for the completion of the book: to The Chinese University
of Hong Kong, for undertaking its publication; and to the School of
Oriental and African Studies for assistance towards its publication.
Part I
INTRODUCTION
1/1
THE MOHIST PHILOSOPHY
1/1/1 T H E HISTORICAL BACKGROUND
1/1/1/1 Early Mohism

A M O N G the schools which competed in China during the age of the
philosophers the first to challenge the heritage of Confucius (551-479
B.C.) were the Mohists. We meet them throughout the 4th and 3rd centuries
B.C., as an organised community distributed over the various states,
committed to 10 specific doctrines and recognising a single chief called the
chu-tzu IE-?; by the 3rd century it had split into three branches which
called each other 'heretical Mohists', and there is no evidence of its survival
after the suppression of the philosophers by the Ch'in H I dynasty in
221 B.C. We have however the book Mo-tzu Si-?, the core of which
expounds the 10 doctrines, each in three versions which are presumably
1
the records of a common oral tradition preserved by the three sects. It
tells us surprisingly little about the founder of the school, M o T i 116
(Mo-tzu H"?), who is not even associated with any particular state. The
core chapters are based on sayings ascribed to him; there are collections of
his conversations in the manner of the Analects of Confucius, which situate
him in the second half of the 5th century B.C., but have the look of later
2
dramatisations of issues arising within the school; but there is only one
story about him, in which he saves the state of Sung 3c by his skill as a
3
military engineer.
At first sight it is not easy to find one's bearings among the doctrines
of the Mohists. T o a modern reader the titles of the 10 triads of chapters
seem to cover a curiously heterogeneous variety of subjects, religion
1
Mo-tzu ch. 8-37. That the three versions come from the three sects was first
proposed by Yii Yueh $J$S (Preface to Sun, 1/4-7).
2
Ut sup. ch. 7, 46-49.
3
Ut sup. ch. 50. Attempts (as in Sun 428, 438-447) to identify his native state
as Sung ^ or Lu or to construct a year-table of his life (Sun's from 468 to 376
B.C., Ch'ien Mu's H | g from 479 to 381 B.C., Fang Shou-ch'u's from 490
to 403 B.C.) seem to me to reflect too much faith in the accuracy of the sparse and
confusing data.
4 The Mohist Philosophy
('The will of Heaven', 'Elucidating the spirits'), philosophy ('Rejecting

destiny', 'Universal love'), politics ('Elevating worth', 'Conformity with
superiors', 'Rejecting aggression'), and miscellaneous questions of morals
('Economy in funerals', 'Economy in expenditure', 'Rejecting music'). One
is tempted to think of the community as a church rather than a philosophical
school, for it preached a stringent morality based on the principle of
universal love, condemned luxury and unprofitable amusements such as
music, insisted that Heaven and the spirits reward the good and punish
the wicked, denounced the Confucians for their scepticism and fatalism.
Yet the Mohists are also vigorous independent thinkers who submit all
traditional morality to the test of social utility, explicitly defend innovation,
and support the new kind of centralised state, with merit rather than birth
as the grounds for preferment, which to the regret of Confucians was
emerging from the ruins of the feudal order. As challengers of traditional
values they are the first Chinese thinkers to defend their principles by
rational debate; and by 300 B.C., the period of the documents on ethics,
logic and the sciences which we shall examine, the religious aspect of
Mohism has almost disappeared in the most highly rationalised system
that ancient China ever achieved.
The Mohists, who puzzle us by being at once the most religious and
the most logical of the ancient thinkers, also contrive to be both the most
pacifistic and the most martial. One of their 10 doctrines was the condem
nation of all military aggression, a corollary of their principle of universal
love. But they did not merely allow defensive war, they specialised in it;
they were a military as much as a religious and philosophical community.
The last 20 chapters of the Mohist corpus are devoted to war and the en
4
gineering of defence works. We read in the Lu-shih ch'un-ch'iu §ft#$JC,
a syncretistic philosophical encyclopedia (c. 240 B.C.), that Meng Sheng
j£S§, supreme head of the community, contracted with the Lord of
Yang-ch'eng P§ J$ in the state of Ch'u ^ to undertake the defence of his
5
city. In 381 B.C. the Lord of Yang-ch'eng was condemned to death, and
Meng Sheng was faced with the choice between breaking his contract and
sacrificing all the Mohists of Ch'u in a hopeless fight. When reproached
for being ready to "let Mohism die out from the world" he replied: "I am not
only the Lord of Yang-ch'eng's teacher but his friend, not only his friend
4
The argument of Chu Hsi-tsu ^fc^jj jfi (op. cit.) that the military chapters
are a Han forgery seems to be no longer tenable. It is criticised by Sun Tz'u-chou
J&Sfo;^ (KSP 6), who however still rejects the last two chapters, and by Watanabe
Takashi (1957).
8
Lu-shih ch'un-ch'iu ch. 19/3 (Hsu 894/5-896/8).
i/i 5
but his vassal. If I do not die for him, certainly no one will ever again look
for a reverend teacher, worthy friend or loyal vassal among the Mohists.
It is by dying for him that I shall do the duty of a Mohist and pass on our
tradition". After sending two emissaries to convey the succession to a
Mohist in the state of Sung he fought to the death with his 83 disciples.
The two emissaries, against the orders of the new chief, returned from
Sung to die with them.
If such apparent incongruities surprise us, it is partly because of the
extraordinary impersonality of Mohist writing both early and late, which
allows us no opportunity to experience the Mohist view of the world from
inside. Even the collections of Mo-tzu's conversations consist entirely of
reasoned judgments on specific issues placed in meagre narrative settings,
and tell us more about the doctrinal problems, social status and internal
organisation of the school than about Mo-tzu himself. We have none of
the lively impression of the man and of his relations with his disciples
which we derive from the Analects in the case of Confucius. The Mohists
seem uninterested either in remembering what their founder was like or
in crystallising a legend about him. A few anecdotes in other sources, such
as the story of Meng Sheng's suicidal loyalty, and another of a chief of
the Mohists in Ch'in who insisted on the execution of his only son for
6
murder in spite of the King's willingness to pardon him, give us a much
more vivid idea of what it was like to be a Mohist than almost anything in
7
their own book.
• Lii-shih ch'un-ch'iu ch. 1/5 (HsU 95/1-96/6).

7
Watanabe Takashi makes a very interesting attempt to explain such incon
gruities by the hypothesis that the 10 doctrines emerged at successive stages in the
history of the school. He conceives the Mohists as a radical movement of urban
craftsmen at first dedicated to 'Universal love' and 'Rejecting aggression', later
corrupted by the patronage of King Hui-wen of Ch'in (337-311 B.C.) and
converted to 'Conformity with superiors' and reverence for Heaven and the sp rits.
(Watanabe (1961, 1962)) The trouble with this argument (in which Watanabe
makes many important observations) is that it relegates several of the core chapters
of Mo-tzu in all three versions to thefinalphase of Mohism in the 3rd century B.C.
But in relation to the dialectical chapters, and even to the collections of Mo-tzu's
supposed sayings (ch. 46-49), all 10 triads of chapters share a crudity of thought
and style which surely belongs to an earlier and less sophisticated period. (One of
the interesting things in pre-Han, as in early Greek philosophy, is that we can watch
over several centuries a people learning how to think.) I consider the inconsistencies
only apparent and find it impossible to date any of the 10 triads (at any rate in its
ancestral form) later than 350 B.C. Since they are transparently records of oral
tradition (with crucial passages introduced by 'The Master Mo-tzu said . . .') we
cannot be sure that the substance of all of them goes back to Mo-tzu himself and
there may be signs of late revision: but the school must have committed itself to
the 10 doctrines very early in its history.
But there is a more interesting reason for the peculiar impression made
by the Mohists. Most early Chinese thinkers seem as far as we can tell to
stand rather high up in the social hierarchy; but there is strong evidence
that the Mohist movement was rooted in the trades and crafts of the towns,
8
among people otherwise inarticulate in ancient China. The Mohists called
themselves mo che S ^ f or mo as the Confucians called themselves ju № y
another name of obscure origin; unless it was a surname (and no other

pre-Han school is called after its founder's surname) it would seem that
'the mo T i ' was a man of the lower orders known only by his personal
9
name. On one occasion he is said to have been refused an interview with
King H u i M of Ch'u (488-432 B.C.) on the grounds that His Majesty would
10
not listen to a man of base origins ( B I A ) . In the only episode in Mo-tzu
11
in which he figures as anything but a preacher or debater Mo-tzu hears
that the engineer Kung-shu Pan <2r&3K has invented a scaling ladder with
which the army of Ch'u will conquer the capital of Sung, hurries to Ch'u,
demonstrates to Kung-shu Pan that he can counter all his nine possibilities
of attack, and when threatened with death tells him that it will be useless
to kill him since 300 of his followers are already operating his engines of
defence on the walls of Sung. Passing through Sung on his return journey
from Ch'u he is refused shelter from the rain by the warden of one of the
8
Fang Shou-ch'u in 1937 was the first to assemble the evidence that Mo-tzii
was a craftsman, probably a wheelwright (op. cit. 15-17). That the Mohist move
ment was based on the crafts is now widely accepted, in China and also in Japan
(Watanabe (1961) 1221-1223, Takata Iffffl^ (1967) 26-30), although there is some
disagreement as to whether Mo-tzii was himself a craftsman or a knight (shih ±)
like Confucius, in some way connected with the crafts. For Communist scholars
this issue is tied up with the doctrinal question of whether Mo-tzu's historical role
was progressive or reactionary. Chang Tai-nien 55 (°P- for whom Mo-tzu
is reactionary, treats him as a knight of possible artisan origin. Fung Yu-lan M'iCM
(1964) 141-142 holds that he was a craftsman who rose to the rank of knight, and
that ''Mo-tzu. and the school he founded reflect the demands and interests of the
small-owner producers of the time, especially in the handicrafts". But the data are
perhaps better regarded as evidence for the social composition of the community
than of stages in its founder's career.
For the place of the artisan in Chinese civilization, cf. Needham vol. 4/2,10-50.
9
That Mo is not a surname was first argued in detail by Chiang Ch'uan tCB|
in 1917 (op. cit. 129-151). He took mo in the sense of 'dark', as referring to the
sun-blackened faces of the work-hardened Mohists (cf. Mo-tzu ch. 56, Sun 335/-1,
and a description of Mo-tzu himself as black in ch. 47, Sun 280/7). Since mo
'branding' is the lightest of the Five Punishments, Ch'ien Mu suggested that it
distinguished the Mohists as men who worked like convicts (Ch'ien (1931) 1-7).
Watanabe (1967) 68-72 takes mo in its ordinary sense of 'ink', referring to the
inked cord of the carpenter. The traditional view that Mo is a surname however
still has defenders (cf. Fang Shou-ch'u, B 1-15).
10
Mo-tzu ch. 47 (Sun 276/6).
11
Mo-tzu ch. 50.
1/1 7
city gates. The chapters about military engineering in Mo-tzu, whether

or not actually written by artisans, obviously reflect their specialised
knowledge. We may notice also that when Kung-shu Pan makes a wooden
kite which flies for three days, Mo-tzu tells him, objecting to its useless-
ness: "The kite you made is not worth the linchpin of a wheel made by me;
with a piece of wood three inches long, which it takes only a moment to
12
cut out, one supports a load of 50 shih". This saying, which does not seem
to be intended metaphorically, surely reflects a tradition that Mo-tzu was
a wheelwright. Later legend, which remembers Mo-tzii and Kung-shu Pan
as the two master-craftsmen of their age, credits Mo-tzu himself with the
13
making of the kite.
The evidence of low social status is sometimes more marked in
non-Mohist sources. In one version of the story of the defence of Sung,
Mo-tzu introduces himself to the King of Ch'u as "a commoner from the
1 4
North" ( d k # £ a $ A ) . When offered a fief by the King of Yueh ©
Mo-tzu says that even if he expected the King to carry out his policies he
would be content to "be ranked with the clients and the populace (Slffil),
I would not presume to seek office" (The Mo-tzu parallel has " . . . rank
15
myself among his ministers ( f i f £ ) ; why should I want a fief?"). Mencius
when approached by a Mohist will debate with him only through an
16
intermediary. Hsun-tzu l o ^ derides the idea that a king should himself
do the work he delegates as "the Way of a menial, the opinion of Mo-tzu"
Throughout Mo-tzu one notices that chiin-tzu St"?* gentleman' and

hsiao jen / h A 'small man* are not, as they are for Confucians, moral as
well as social terms. A Mohist, like a Confucian, aspires to be benevolent,
righteous, a sage . . . but not a gentleman. The notoriously graceless style
of early Mohist writing, humourless, ponderous, repetitious, is also signif
icant in this connexion. Other pre-Han thinkers, even when they are not
12
Mo-tzu ch. 49 (Sun 302/-4). Sun's emendation of Ti U to chiang |£
'craftsman* has no authority except that it is the reading in the passage as quoted
in a 10th century encyclopedia, the editor of which no doubt shared Sun's implicit
assumption that Mo-tzu could not have worked with his own hands. The version
in Han Fei tzu ch. 32 (Ch'en 625/1-3), in which Mo-tzu makes the kite himself but
it breaks on the first day (for further versions, cf. Sun 468), ends with the comment:
"Mo-tzii had the truest kind of skill; he was skilful in making linchpins, clumsy in
making kites" (He was skilful at useful but not at useless things).
1 3
Cf. the examples collected in Sun 468.
14
Lu-shih cKun-cKiu ch. 21/5 (Hsu 1018/7).
15
Lu-shih ch'un-ch'iu ch. 19/2 (Hsu 883/5). Cf. Mo-tzu ch. 49 (Sun 298/-1).
16
Mencius 3A/5.
17
Hsun-tzu ch. 11 (Liang 147/1).
such artists as Lao-tzu ^zrf* and Chuang-tzu ffi"?, Mencius %rt and
Han Fei tzu always have some literary sense. The early Mohist
gives the impression of the solemn self-educated man who writes with
difficulty and only for practical purposes, and has no occasion to polish his
style as an adviser of princes. The style of the later Mohist writings in this
volume is no longer clumsy, on the contrary it is neat and functional, but
it is equally remote from literary concerns. T o the extent that the later
writings reveal their social background it still seems to be that of the trades
and crafts. The argument in the Canons (B 30, 31) that the proper price of
a commodity is fixed by supply and demand must always, in China as in
Europe, have appealed more to sellers than to buyers whether noble or
base. The series of Canons on ruler and subject (A 34-39) is brief and
perfunctory, but among the sections on that most abstract of topics, logic,
one notices among the concrete examples recurrent references to coins
(A 14, 75), buying and selling (A 85, B 3), prices (A 88, B 6), carpentering
(B 10), rich merchants (B 53). The sections on optics and mechanics surely
reflect social conditions comparatively rare in history until the 16th century
in Western Europe, where the Scientific Revolution soon followed, the
explosive situation when men with speculative minds are in close contact
with men who work with their hands.
We still know too little of the social history of China during the age
of the philosophers, which coincides with the later stages of the greatest
social upheaval in Chinese history before the 19th century. It was the
period in which the fiefs of the Chou M Emperors, with their serfs and
rigid hierarchies of aristocratic ranks, developed into independent central
ised states with bureaucracies in which social origin became less and less
important, until in 221 B.C. the state of Ch'in completed the conquest of
the rest and united China as a centralised and bureaucratised Empire with
peasant proprietorship. One current in the multiple social changes of the
time was the growth of commerce and the crafts, stimulated by the belated
arrival of iron, which ended in China as elsewhere the near monopoly in
metal weapons and utensils of the Bronze Age aristocracies. In the last
century of the period chronicled in the Tso-chuan £ 4 $ (721-463 B.C.) one
has the impression that some of the smaller states are coming to look like
Greek cities early in their political development, although one knows in
retrospect that conquest by the great states will soon put an end to their
18
evolution. In 562 B.C. Tzu-k'ung regent of the powerless Earl of
1 8
Cf. Rubin (op. cit.), who goes so far as to claim Tzu-ch'an of Cheng as a
democrat. For the examples which follow I have also drawn heavily on Watanabe
(1961), 1230-1231.
1/1 9
Cheng ffl, tries to force the noble clans of Cheng to accept an agreement
submitting them to his authority. "The grandees, ministers and sons of
great families would not accept it." He is persuaded to retract in order to
"appease the crowd" ( S ^ ) , "so he burned the document outside the
Ts'ang gate, and only then did the crowd calm down". In 553 B.C. "the
government of Tzu-k'ung of Cheng was autocratic, and the people of the
capital ( 9 A ) objected to it. . . . Tzu-chan - ? S and Tzu-hsi -?1S led
the people of the capital to attack him, killed him and divided up his
possessions". The victors form a triumvirate with Tzu-ch'an , the
man who finally will come out on top. In 542 B.C. "the Earl of Cheng made
a covenant with his grandees in the Great Shrine, and bound the people
of the capital by a covenant outside the Shih-chih-liang gate", in a common
alliance against another contender, Po Y u ffiW. After Po Yu's defeat
Tzu-ch'an becomes sole regent for 20 years (542-522 B.C.). In 535 B.C. he
makes public the first code of laws in China, in spite of aristocratic objec
tions that now the people will dispute the interpretation of laws instead
19
of doing as they are told.
In the history of Cheng it is not clear how widely or narrowly we are
to interpret those interesting expressions, 'the crowd' and 'the people of
the capital'. But in three other small states, Ch'en pJl, Wei and what was
left of the domain of Chou, we have direct evidence that the struggles of
noble factions involved not only the artisans attached to palaces but the
labourers on the corvee. In 549 B.C., when Ch'ing H u 8£Jt& and Ch'ing
Yin K K rebelled against the Marquis of Ch'en and seized his capital,
"the men of Ch'en were building up the city wall. Someone was executed
because a plank fell; the workmen ($£A) conspired and each group killed
2 0
its overseer; then they killed Ch'ing H u and Ch'ing Y i n " . In 519 B.C.
Prince Ch'ao of Chou "started a rebellion with the support of those
among the former officials and the Hundred Crafts ( " S ^ ) who had lost
21
their jobs, together with the descendants of Kings Ling H and Ching jp:".
In 477 B.C. in Wei "the Duke made the artisans (E) work too long. The
Duke wished to expel Shih P'u S B , but before he could do so trouble
broke out, and on the day Hsin-ssu Shih P'u attacked the Duke with the
support of the artisans". In 469 B.C. another Duke of Wei "made the
Threefold Artisans (HIS) work too long", and also offended a certain
Ch'uan M i 4£5№f, with the result that five of his nobles "started a revolt
19
Tso-chuan Dukes Hsiang 10, 19, 30. Chao 6 (Legge 444/10-12, 481/10,
553/-3, 607/3).
20
Ut sup. Duke Hsiang 23 (Legge 497/3).
21
Ut sup. Duke Chao 22 (Legge 691/9).
2 2
with the support of the Threefold Artisans and of Ch'uan M i " . The
detailed record of the Tso-chuan unfortunately breaks off at 463 B.C., but
it is enough to suggest that in the 5th century B.C., at the time of the rise
of Mohism, there might be craftsmen who would be forming ideas of their
own about what kind of prince deserved their support, for the benefit of
themselves and of the people at large.
There is no evidence of any popular movement seeking power for its
own leaders. Certainly the Mohists themselves are not a revolutionary
league, at any rate in the 4th and 3rd centuries; if there were anything
seditious about them their enemies would be sure to say so. Like the
Confucians, they seek audiences with princes and hope to be appointed to
high office. It would seem however that they expect a Mohist in office to
contribute to the funds of the organisation, which can request his resigna
23
tion if he betrays its principles. We may see them as social outsiders who
have formed the conception of a state which will not benefit only the
privileged, who with the dissolution of hereditary barriers can earn
promotion by their technical skills, and who appreciate the problem that
those who do get promotion may forget where they came from.
The remoteness of the Mohist movement from the central Chinese
tradition, its seeming incongruities and its abrupt disappearance from
history, become intelligible if we think of it as a confluence of merchants,
craftsmen and déclassé nobles, briefly emerging as a power in the cities as
the feudal order disintegrates, but soon to be thrust back by the new
bureaucratised Empire into the station which it has pleased Heaven to
decree for them. Confucius, a teacher of young nobles in L u H - , the state
most faithful to the Chou tradition, has devoted his life to defining and
refining the failing values embodied in the manners and culture of the
Chou aristocracy, behind which he perceives an ideal of the true gentleman
guided by such implicit principles as that the benevolent "love others"
24
and " D o not do to others what you would not wish them to do to you".
But to the man from the lower orders nothing said by Confucians can be
relevant except the occasional ethical generalisation which prepares the
way for the Mohist principle of universal love. He has never shared in the
culture of Chou, cannot afford to complicate his life with elaborate etiquette,
has no leisure to fulfil such ritual obligations as the duty to mourn his father
for three years. He is a needy, or if no longer needy a thrifty man, whose
touchstone of value must always be practical utility. From his worm's eye
22
Tso-chuan Duke Ai 17, 25 (Legge 849/8, 856/6,7).
23
Mo-tzû ch. 46, 49 (Sun 267/-1—268/7, 301/4-8).
24
Analects 12/22, 15/24.
1/1 11
view the cultural refinements of nobles are less obvious than their extrava
gance, which wastes the resources on which he and his like depend. Music,
for Confucius the most civilizing of the arts, suggests to his mind only the
expensiveness of the great court orchestras, an especially insolent exhibition
of conspicuous consumption since they are audible in the distance to those
not privileged to see them; he is in the position of the peasants in Mencius
who, "when they hear the sound of the King's bells and drums, the notes
of his pipes and flutes, all with aching heads and knitted brows tell each
2 5
other 'What else but our king's love of music has brought us to this pass ?' " .
These thrifty, utilitarian attitudes are reflected in the doctrines of 'Economy
in expenditure', 'Economy in funerals' and 'Condemnation of music', given
the same status among the 10 doctrines as others which outside such a
social context would seem to be on a quite different level of importance.
The Mohist's only tradition is his craft, and in this period of swift
social and technological change, innovation is part of the tradition of the
crafts. The artisan attracts attention by having something new to offer—the
most urgent attention if (like the Mohists and Kung-shu Pan) he invents
new engines of war. T o the Confucian precept "The gentleman follows and
does not originate" the Mohist answers: " I n ancient times Y i originated the
bow, Chu armour, Hsi Chung the carriage, Ch'iao Ch'ui the boat. Does it
follow that the armourers and wheelwrights of today are all gentlemen, and
the four inventors were all vulgar men ? Moreover anything that you follow
must have been originated by someone, so that on your own showing you
26
follow in everything the way of vulgar men". This innovatory attitude
does not discourage early Mohists from citing ancient authority. The three
chapters on fatalism all begin by laying down three tests of sound doctrine,
27
for which we quote the first version: " O n what is it based ? Look upwards
to base it on the deeds of the ancient sage kings. By what is it measured ?
Look downwards to measure it by scrutinising what the eyes and ears of
ordinary people confirm to be real. On what is it put to use ? Apply it in
administration, and observe whether it benefits the mass of the people in
the civilized states". This is notable as the first Chinese attempt to formulate
principles of argumentation, but it still gives the first place to ancient
authority. The early writings are full of quotations from the ancients, used
however in a way quite different from a Confucian's. Confucius treasures
and explores the significance of all that survives of the songs, rituals and
music of Chou. He is less interested in more ancient authorities whose
25
Mencius IB/1.
26
Mo-tzu ch. 39 (Sun 186/-4—187/6).
27
Ut sup. ch. 35 (Sun 170/5-7).
traditions have broken off, and even says explicitly that he prefers the Chou
28
to earlier dynasties. The Mohist on the other hand judges doctrines
primarily by his third test, their practical consequences, and combs ancient
documents (most of them professedly earlier than Chou) for supporting
quotations however dubious, like a Protestant sectary using and misusing
Scripture.
The Mohist's deepest objection to the aristocratic moral code is that
it is divisive, that it requires the gentleman to put his duties to his family
and his lord before the interests of anyone else. The result is that each
family and state is entitled, and indeed obliged, to prefer itself to others
and be drawn into war with others, a war in which whoever wins or loses
the common people are always the sufferers. The Mohist sees that a
morality which will not sacrifice him to his social superiors has to be one
unified by a single principle applying to all. It is the principle of "loving
others as you love yourself . . . having as much regard for father and elder
brother and for lord as for yourself . . . having as much regard for younger
brother and son and for vassal as for yourself . . . having as much regard
for others' families as for your own . . . having as much regard for others'
29
states as for your own". He calls it chien ai M which we can hardly
avoid translating as 'loving everyone' or 'universal love', although this may
give the false impression that he is interested in the warmth of the senti
ment rather than the equality of the concern. He uses ai as we use 'love'
when talking of 'self-love', which is concern for oneself; chien ai is being
as much concerned for one person as for another. Although the imper
sonality of their writing makes it hard to imagine ourselves into their
attitudes, one has the impression that Mohists were not people with warm
sympathies towards everyone, but people whose personal affections are
disciplined by a stern sense of justice. (Witness that Mohist in Ch'in who
30
had his son executed.) A corollary of chien ai is 'Condemning aggression';
by 'aggression' (kung$t) is understood one state attacking another simply in
order to benefit at its expense, which is seen as a crime no different from
the private robberies and murders which a ruler punishes inside his state.
The Mohist has nothing against war as such, and even repudiates the
aristocratic code of chivalry which Confucians were trying to keep alive.
31
If your cause is righteous, why give the enemy a second chance ? In any
28
Analects 3/14.
29
Mo-tzu ch. 14 (Sun 66/4-10).
3 0
Cf. p. 5 above.
81
Mo-tzu ch. 39 (Sun 187/6-188/3).
1/1 13
case it is an irony of his situation that what political leverage he has derives
from his reputation as a military engineer.
To understand the Mohist's viewpoint in politics, we may imagine
the traders and artisans as like their counterparts in 15th century Spain,
France and England, welcoming absolute monarchy because centralisation
and bureaucratisation defend them against local magnates and give them
new opportunities of rising in the world. By the principle of 'Promotion
of worth', the sage king "appoints anyone who has ability, even if he is
from the peasants or from the craftsmen or traders", so that "no one in
office is unchangeably a noble, no one of the people is irrevocably among
32
the base". The Mohist believes that government originated in the need
to unify the 'different moralities' H ) of individuals competing in the
primitive war of all against all, and that its function is to "unify and
3 3
assimilate morality throughout the Empire" (— N X T i l l t ) . It is
interesting that he treats anarchy as a conflict not of interests but of
'moralities', by which he means not moral codes but the conflicting family
or state loyalties within the traditional code. The Mohist community can
only preach the unifying ethic of universal love, but an Empire governed
on Mohist lines will effectively submerge local loyalties in wider and wider
loyalties. At each level, village, district, state, Empire, the administrator
will unify the standards of all below him by imposing the standards at the
level immediately above him. At the top of the pyramid the Emperor will
impose the standards of the only power above him, which is Heaven. This
is the doctrine of "Conforming to the level above, not leaguing together
below" ( J l IRI jfn /fv ~~f ,fch). It is to be noticed that at each level the adminis
trator demands conformity to his superior, not to himself; and one Chou
doctrine which the Mohists do not question is that if the Emperor disobeys
34
Heaven his subjects' ultimate allegiance must be to Heaven, so that they
agree with Confucians in admitting the right to revolution. The authori
tarian doctrine of 'Conforming to superiors' is not incompatible with
'Universal love' (that is, equal concern for all), on the contrary it is its
political realisation.
Aristocratic codes such as Confucianism are sanctioned by shame
rather than guilt. The conduct the noble owes to his self-respect does not
have to be backed by promises or threats from above, may even conflict
with them. (Catholic nobilities in Europe observed their codes of duelling
and courtly love at the risk of their souls.) Confucians are content to pay
32
Mo-tzu ch. 8 (Sun 28/-1, 29/5).
83
Ut sup. ch. 11-13 (Sun 47/1, 49/2 and passim).
34
Ut sup. ch. 11 (Sun 49/2-9).
Heaven and the spirits their customary respects without much caring
whether Heaven is a personal being or the spirits of the dead actually
exist. More and more, during the period of the philosophers, they come
to regard Heaven merely as an impersonal power responsible for everything
outside human control, including the good and ill fortune which is to be
accepted as our destiny. The Mohist however comes from a less sophis
ticated class in which the perennial folk religion of China is still alive, and
does not like his masters to forget that they too are subjects of still higher
beings. "The Confucians think that Heaven is insensible and the spirits are
not divine ( f i l X ^ ^ M , £i$Lffi^fiffi). Heaven and the spirits are
35
displeased; this is enough to bring ruin on the Empire." The Confucian
Kung-meng H "says that the spirits do not exist, and also that the
gentleman must learn to perform sacrifices", which is like "learning the
etiquette for guests without having any guests, making a fishnet although
36
there are no fish". The expositions of the doctrines of 'The will of Heaven',
'Elucidating the spirits' and 'Condemning fatalism' insist fiercely that men
will act morally only if they cease to regard changes of fortune as their
destiny, and come to recognise that Heaven and the spirits reward the
good and punish the wicked.
It is plain that more is involved here than a simple fidelity to popular
pieties. Because the Mohist advocates a new universal morality he requires
another sanction than the respect of his peers, for his peers are still satisfied
with custom and its divisive loyalties; he is driven in the same direction as
the great Middle Eastern religions, with their universal moralities ordained
by a personal God who will judge the mighty as they deserve. However
there is nothing to suggest any spiritual dimension deeper than a guilty
fear of ghosts. Even early Mohism is not on closer inspection a religious
movement; it is as man-centred as Confucianism, and insists on the power
and benevolence of the spirits only as a buttress for human morality. The
awe and resignation with which Confucius accepts what Heaven decrees
for him has much more of the sense of the holy in it than anything jn
Mo-tzu.
At first sight one might suppose that on this issue it is the Confucians
who are the rationalists. But the this-worldliness of Confucians has nothing
to do with ratiocination. Confucius offers himself to his disciples simply as
a maturer man with finer perceptions in a shared scheme of values. His
school does not enter into rational debate until it begins to be challenged by
other schools, first of all by the Mohists. The early Mohists are ignorant
35
Mo-tzu ch. 48 (Sun 287/5).
36
Ut sup. ch. 48 (Sun 286/3).
1/1 15
men, excluded from the best culture of their time, but compelled to give
reasons for their tenets, because they are new. Each of the 10 triads of
chapters defending their 10 doctrines is a laboriously assembled collection
of arguments to convince doubters. Some of the argumentation is very
crude, as in the surviving chapter on 'Elucidating the spirits', which refutes
sceptics who deny the existence of spirits by applying the three tests,
ancient authority, common observation, and practical consequences: (1) In
every village there are people who have seen and heard spirits. (Some lively
ghost stories are quoted from the annals of the states.) (2) If you doubt the
witness of humble people, you cannot deny the witness of the ancient sage
kings, who are proved to have believed in spirits by quotations from the
Songs and Documents. (3) People behave better if they know the spirits are
37
watching them. Nevertheless, this is the start of rational discourse in
China. Within a century or so the Mohists will have developed into the
most sophisticated of all the ancient Chinese thinkers.
1/1/1/2 The metaphysical crisis in the 4th century

About 350 B.C. a new competitor enters the scene, the individualist
Yang Chu $1 ^fe. We know little about the teaching of Yang Chu himself
38
except on the authority of his enemies, who call it egoism (wet wo f§ ft)
and say that it was his principle not to sacrifice as much as a hair to benefit
39
the whole Empire. But we do have later individualist writings, certain
40
chapters of the encyclopedia Lu-shih ch'un-ch'iu (c. 240 B . C . ) , and from
these we can get some idea of how his philosophy looked from the inside.
Confucians and Mohists had never been much interested in private
life. For Confucians life is fulfilled in public service, and the only salve
for the misfortune of being out of office is a serene acceptance of the decree
of Heaven. But in the 4th century many people were no longer finding it
obvious that it is really so much better to battle for power and possessions
than to stay at home and do as one pleases. By 300 B.C. the Taoist Chuang-
tzu will have provided a mystical philosophy for all who temporarily or
permanently retire from the world, but the earlier individualists are not
in the least mystical. Their first consideration is simply the danger to life
and limb of plunging into the murderous political struggles of the time.
87
Mo-tzu ch. 31.
88
Menctus 7A/26. For a Mohist example of the phrase, cf. EC 1 below.
89
Menctus 7A/26. Lieh-tzu ch. 7 (Yang 146/2-5).
4 0
Pen sheng Chung chi Kuei sheng Ch'ing yii Shen
wet
External possessions after all are dispensable and replaceable, the life and
health of the body are not. In a typical individualist story the legendary
Emperor Yao ^§ offers his throne to one Tzu-chou Chih-fu ^jWlfc&y
who replies: " I have no objection to being made Emperor. However at the
moment I have an ailment which worries me. I am going to get it put right,
and haven't the time just now to put the Empire right". The narrator
comments: "Nothing is weightier than the Empire; if he would not risk
41
harm to his life even for that, how much less for any other thing?". But
a detachment from even the greatest of material possessions is from another
point of view the irresponsible shirking of an opportunity to benefit the
world by good government. In a story of probable Mohist origin which
42
survives in Lieh-tzu (c. A . D . 300), Mo-tzu's chief disciple Ch'in
Ku-li embarrasses Yang Chu by asking directly: " I f you could help
the whole world at the cost of one hair of your body, would you do it?".
A disciple of the latter recovers the offensive by asking: " I f you could gain
a state by cutting off one of your limbs, would you do it ?". Neither quibbles
over the point that what for one is gaining a state for the other is getting
the opportunity to help it.
Little as we know directly about Yang Chu, it seems that his interven
tion provoked a metaphysical crisis which threatened the basic assumptions
of Confucianism and Mohism and set them on new courses. His truly
explosive contribution is not his individualism, which merely developed as
a third current of thought side by side with the others, but the concept of
43
hsing 'nature' with which he supported it. The word hsing derives
from sheng ^ 'be born, live', and is distinguished graphically by a
radical which is supplied only irregularly in pre-Han texts. The Chinese
concept is a dynamic one, not perfectly represented by our own static
concept of a thing's 'nature'. The hsing is the spontaneous tendency of the
living organism throughout its lifespan; if a man 'keeps his nature intact'
(J* tt) he lasts out his full term in good health, but to do so he must
'nourish his nature' ( J l tt), avoid 'interfering with his nature' ( S for
unlike what we understand by nature in the West the hsing of a living thing
is very vulnerable, as we see in the cultivation of plants. Desires belong to
the spontaneous tendencies of man's nature, but they must be regulated
carefully i n order to strike a mean between the excess which endangers
health and the deprivation which starves vital potentialities, disturbs vital
4 1
Lu-shih ch'un-ch'iu ch. 2/2 (Hsu 106/10-107/6).
42
Lieh-tzu ch. 7 (Yang 146/2-11). For the evidence for its Mohist origin,
cf. G(4).
4 8
This claim is argued in detail in G(9).
17
momentum. The individualism expounded in the Lii-shih ch'un-ch'iu is a

sort of epicureanism, a moderate indulgence of the desires after careful
consideration of benefit and harm, very different from the reckless
u
hedonism ascribed to Yang Chu 500 years later in Lieh-tzu. It agrees with
the description of Yang Chu's teaching inHuai-nan-tzu (c. 130 B.C.),
"keeping one's nature intact and protecting one's genuineness, and not
involving the body in trouble for the sake of external things" ( J k t t ^ c i K ,
Since the way of life which fulfils our nature is independent of our
wills, it belongs to the realm of Heaven and not of man. A simple conse
quence follows: we obey Heaven, not as Confucians and Mohists suppose
by behaving morally, but by nurturing and harmonising the vital tendencies
and spontaneous inclinations which Heaven instilled in us when we were
born. The first of the individualist chapters of the Lu-shih ch'un-ch'iu steals
the metaphysical basis of Confucianism and Mohism to lay its own
foundations:
Ch. 1/2 (Hsu 6/8, 9)
"That which first engenders it is Heaven, that which by nurturing fulfils

it is man. Someone who is able to nurture what Heaven engenders and not
infringe on it is called the Son of Heaven."
It had never been questioned that Heaven, the power responsible for
all things being as they are, for the uncontrollable accidents of fortune and
misfortune, for whatever in man is innate and independent of his will, has
also ordained the principles by which we should live. How can the supreme
power not be on the side of all that by human standards is the best ? Now
for the first time a metaphysical doubt enters Chinese thought, and a rift
opens between Heaven and man, between what is and what should be.
If Heaven is on the side of Yang Chu, on what is the morality of Confucians
and Mohists to rest ?
With the raising of this question Chinese philosophy comes to maturity.
If with the Analects of Confucius in mind one turns to Mencius (c. 371-c.
289 B.C.), defender of Confucianism against Mo-tzu and Yang Chu, we
find ourselves lifted to a new and higher level of philosophical concern. The
Confucians, who had seemed incapable of posing any question more
momentous than " D i d Kuan Chung If № understand the rites?", are now
obsessed with the problem of human nature. Three theories are already
44
Lieh-tzu ch. 7.
46
Huai-nan-tzu ch. 13 (Liu 13, 9B/10).
current, that it is a mixture of good and bad, or neutral, or good in some

men and bad in others. Mencius proposes a fourth theory, the first which
takes the full measure of the challenge of Yang Chu. The debate has turned
his attention inward, from external custom or abstracted moral principle
to man's inner promptings towards good or evil action. He finds in moral
development something comparable with the spontaneous processes of the
body, altruistic inclinations which are not imposed from outside but which
grow or are starved or injured and cannot be forced, like the vitality of the
rice in the fields or the trees on the denuded Ox Mountain. He concludes
that moral inclination belongs to the spontaneous tendencies of our nature
and shares their vulnerability, that at bottom human nature is good, and
that only the sage who fulfils the moral potentialities which distinguish us
from the beasts and birds has fully realised himself as a man. This explana
tion, the only one which can close the gap between Heaven and man from
the Confucian standpoint, took a long time to win its final ascendancy. In
the 3rd century Hsiin-tzu concedes the split between the Way of Heaven
and the Way of Man, and declares that human nature is evil and that
morality is a human invention to harmonise the conflicting desires implanted
in us by Heaven. For the next millennium and a half the problem of human
nature continues to be debated even in periods when other philosophical
issues seem moribund. Only in the Sung dynasty (A.D. 960-1279), when
the old problems were rethought in relation to new concepts influenced by
Buddhism, is the Mencian solution at last restated in the form which
became Neo-Confucian orthodoxy.
For the individualists too the separation of Heaven from man quickly
reveals more radical implications which undermine the prudent and
comfortable epicureanism from which they started. The step from Yang
Chu to the Taoist Chuang-tzu (c. 369-c. 286 B.C.) is as far as from Confucius
to Mencius, or from the early to the later Mohists. If it is by following my
nature that I accord with Heaven, then I deviate from Heaven as soon as
I detach myself from spontaneous process as a thinking self, and manipulate
my Heaven-engendered inclinations according to principles of benefit and
harm which are as man-made as Confucian right and wrong. Chuang-tzu
recommends us to resolve the split between Heaven and man by renouncing
all judgments of right and wrong, benefit and harm, suspending the
assertion of self which is man's initial separation of himself from Heaven,
and surrendering to the pure spontaneity in which we cease to distinguish
ourselves from the unending processes of generation and decay within and
outside us. We accord with the Way as soon as we cease to be diverted from
it by selfish desires or moral demands.
1/1 19
With the posing of deeper questions and the multiplication of compet

ing answers there is a sudden improvement in the quality of intellectual
debate. The philosophical dialogues in Chuang-tzu, Mencius and Mo-tzii
ch. 46-49 are often harder to follow than the ponderous arguments of the
early Mohists; but the farther one gets in pinning down their concepts,
identifying the implicit questions, filling in the invisible bridges across
apparent gaps which debaters with shared assumptions do not need to
46
supply, the more intelligent these debates seem. Late in the 4th century
the 'school of names' (tning chia ) appears, sophists such as H u i Shih
MM and Kung-sun Lung <2^tt who resemble the Mohists in defending
universal love and condemnation of aggressive war, but also practise
disputation (pien Wt) for its own sake. They attracted attention primarily
by their paradoxes, one of which, Kung-sun Lung's ' A white horse is not
a horse', became a standing joke; the sophist was said to have used it to
get his white horse across the frontier in spite of the ban on exporting
47
horses. We know them only by the sophisms listed without explanation
8
in Chuang-tzii and Hsiin-tzu* and by a couple of authentic essays preserved
in the Kung-sun Lung tzu (forged between A.D. 300 and 600), the notorious
White horse, the Pointing things out, and possibly the 'Left and righ
49
dialogue. But all that we have of them has been selected by others, and
there is no reason to suppose that their investigations of names and objects
( £ f f ) , the dimensionless (M ff) and the limitless (M H ) , sameness and
difference ( K M ) and the mutually pervasive (§E & ) , served no other
50
purpose than to astonish people with paradoxes, any more than that Yang
Chu actually preached what Ch'in K u - l i and Mencius saw as the practical
implication of his doctrine, that he would not give a hair to save the world.
Just as Yang Chu's individualism can be understood only in the light of
later documents written from the individualist point of view, so the least
misleading approach to Chinese disputation is through the thinkers who
4 6
A generation ago Waley could say of Mencius that "as a controversialist he
is nugatory. The whole discussion (Book VI) about whether Goodness and Duty
are internal or external is a mass of irrelevant analogies, most of which could equally
well be used to disprove what they are intended to prove" (Waley (1939) 194), but
by now the argumentation of Mencius is much better understood. Cf. the analyses
of his reasoning in Lau (1970) 235-263 and G(9) 231-250.
4 7
Versions of this tale are assembled in Hu Tao-ching ft^M® (1934), 16-18.
Its currency may explain why the 'White horse* essay was thought worth preserving.
48
Chuang-tzu ch. 33 (Kuo 1102-1106). Hsun-tzu ch. 3 (Liang 34/7-35/3).
4 9
For the evidence for rejecting Kung-sun Lung tzu as a forgery, cf. G(2),
6 0
For references to these themes of the sophists, cf. the quotations in § 1 /4/3
below.
actually describe and operate the apparatus of disputation, the later

Mohists.
A kind of disputation with "five wins and three arrivals" ( E S & H I i ) ,
which unfortunately are nowhere specified in the surviving sources, seems
to have had wide currency among the schools. Various forms of its pro
gramme appear in a story about the 4th century thinker Tsou Yen $P #5
(traditionally associated with the Five Elements school), in the Confucian
Han shih wai chuan ^ f f ^ r - W , and in the Teng Hsi tzu fW?-?, a late
51
collection of fragments Legalist i n complexion. Tsou Yen's version is
said to have been his answer to the 'White horse* argument on a visit to
Kung-sun L u n g :
SL » » M m m > m^m · · *mnmk ° ·
"The disputation recognised throughout the world has 'five wins and three
52
arrivals', of which correctness i n phrasing is the least. The disputant
distinguishes separate kinds of thing so that they do not interfere with each
53
other, arranges in sequence different starting-points so that they do not
5 1
Liu Hsiang (79-8 B . C . ) , Pie lu g l j ^ ap. Shih-chi chi-chieh ^dMffl
ch. 76 (Shih-chi 2370 n. 2), for the Tsou Yen story: Han-shih wai-chuan SPTK
6/3B/1-4A/3: Teng Hsi tzu SPTK 4A/4-6. The passage is also echoed in the account
of the School of Names in Shih-chi (ch. 130) 3291/-2.
5 2
Cf. the account of Kung-sun Lung's defeat in disputation by K'ung Ch'uan
^L^p in K'ung ts'ung-tzu JLli-? SPTK, A 72A-76B (3rd century A.D., but revised
from early sources, G(6) 139 n. 18), which ends with the admonition "Don't
dispute matters any more with K'ung Ch'uan; in him reason prevails over phrasing,
in you phrasing prevails over reason" ( ^ A S K i ^ S ^ . ^ S f S I J ^ M ) . It may be
noticed that the programme of disputation concludes with a denunciation of
sophistry as "harmful to the Great Way" (i5A8i) ^ "harmful to being a gentle
an<
man"; K'ung Ch'uan, who as the descendant of Confucius is presented as the

greatest gentleman of his age, has been invited to refute Kung-sun Lung with the
appeal "This man by trivial disputation slanders the Great Way; why not go and
put him right?" (ut sup. 72 A/4 jlfcA/hPffiSb^M)- There is a similar reference to
reason and phrasing in disputation mjen-wu chih A f c S (°h. 4) SPTK, A 18B/1-4,
f
but it is secondary, also echoing K'ung ts ung-tzu,ut sup. 76 A/1. Sheng 'win' might
also be taken as the ordinary word for victory in disputation (cf. § 1/4/24); chih
'arrive' has more than one technical usage (cf. A 32 n. 9, NO 9 n. 4).
5 3
The tuan ^ 'starting-points' are different senses of the word from which
the argument starts; if they are confused the argument will be confused. Hsun-tzu,
arguing that the thesis "To be insulted is not disgraceful" confuses two uses of
'disgraceful', says: "This has two starting-points. . . . There is moral disgrace and
there is social disgrace" (ch. 18, Liang 249/-3 · · · ^ i № i h ^ f t M ? * )·
The Canons have a whole sequence on confusable meanings of words (A 76-87),
but do not use tuan in this sense.
i/i 21
confuse each other, dredges his ideas and makes his meanings intelligible,
and clarifies what he has to say; he shares his knowledge with others and
does not busy himself with misleading them. In this way the winner does
not fail to make his point and the loser finds what he is seeking. When it
comes to elaborating style in order to put up a pretence, adorning phrases
in order to make nonsense of the other's case, using subtle comparisons to
make him shift his ground, stretching what he literally says so that he
cannot get back to his own idea, to behave like this is harmful to the Great
Way. Engaging in tangled debates and competing to keep talking the
longest cannot but be harmful to being a gentleman."
The new art of disputation affected in different ways all the schools
of the 3rd century. Chuang-tzu was a friend of H u i Shih, debated with him
and made fun of his logic. Living at the time when reason first became
self-aware in China, he is the first conscious anti-rationalist. Disputation
is the technique forjudging between alternatives; but according to Chuang-
tzu it is precisely when we distinguish alternatives, the right and the wrong,
the beneficial and the harmful, self and other, that we cut ourselves off from
the world we objectify, and lose the capacity of the angler, the carpenter
and the swimmer to heed his total situation with undivided attention and
respond with the immediacy of a shadow to a shape and an echo to a sound.
A second tendency is represented by the Confucian Hsiin-tzu and the
Legalist Han Fei tzu (died 233 B.C.), who builds his political theory like
Machiavelli on the actual statecraft of princes, and justifies its ruthless-
ness by the amorality of Heaven and its Way. Both make free use of the
resources of disputation, notably in Hsiin-tzu's Right use of names and the
Interpretation of Lao-tzu ascribed to Han Fei tzu. Certain lessons of
disputation, such as the uselessness of arguing without defining your terms,
had been learned by all the major thinkers of the 3rd century. But Con
fucians and Legalists, however much they may have borrowed from the
sophists, despise them for devoting themselves wholly to disputation; they
find it absurd that grown men should let logical puzzles divert them from
serious ethical and political issues. A third viewpoint is represented by the
Mohists. They commit themselves fully to disputation, not because they
have forgotten about serious problems but because (like their contempor
aries in Greece, of whom they know nothing) they think that only logic can
solve these problems definitively. In any case their opportunities of
political power are receding and they have plenty of time on their hands.
By 300 B.C. it was evident to all but Confucians that the authority of
past sages could no longer be a guide to the changed world of the present.
For Chuang-tzu, who in any case denies that the sages could put their
insights into words, any more than the wheelwright can find words to teach
his skill of hand, "the men of old, with their untransmittable message, are
54
dead", and their books are only the dregs which their teaching left behind.
In spite of the persisting convention of expressing ' A sage king would . . .'
by 'The sage kings did . . .', no one but the Confucians still gives real
weight to past authority. In the 3rd century Han Fei tzu appreciates both
the effects of historical change and the unreliability of historical testimony.
" I f today we wish to inquire into the way of Yao and Shun 3,000 years ago,
can there really be any certainty about it ? Being certain about it without
evidence is foolishness, depending on it though unable to be certain of it
55
is error". He thinks that the Confucian ideal of government was suitable
to the small communities of the past, but has become obsolete with the
growth of population. The Mohists too are aware of changing conditions
and the obsolescence of old authority. The relation between knowledge and
temporal change is one of the questions which lead them into the study of
disputation. They discover in disputation a certainty (pi invulnerable
to time, the logical necessity which is eternal.
1/1/1/3 Later Mohism

The Mohists, like the other schools, were forced by the metaphysical
crisis to rethink their doctrines from the foundations. The last chapter of
Chuang-tzu gives a glimpse of them arguing over their 'canons', which were
56
presumably the 10 basic doctrines of the school.
Chuang-tzu ch. 33 (Kuo 1079/1,2)
"The followers of Hsiang-li Ch'in's disciple W u Hou, and the Mohists of

the South such as K ' u Huo, Chi Ch'ih arid Teng-ling-tzu, all recited the
Mohist canons but diverged in different directions, they called each other
heretical Mohists, chided each other in disputation about the as-hard-to-
white and the same-or-different, answered each other in propositions at
odds and evens which did not match."
54
Chuang-tzu ch. 13 (Kuo 4911-2).
55
Han Fei-tzu ch. 50 (Ch'en 1080/-3).
5 6
This interpretation of ching 'canon' was proposed by Hu Shih j^8§j ((1919)
185) The reference is not to the dialectical chapters called the Canons, since dis
k
agreements over questions of logic and science would hardly have led to accusations
of heresy. We shall see in commenting on EC 1 that even in Expounding the canons,
written earlier than the Canons themselves, the word ching refers to the 10 theses
ofMo-tzu.
1/1 23
Han Fei tzu similarly describes the Mohists as split into three branches,
with leaders two of whose names agree with those in Chuang-tzu (Hsiang-li
5 7
* S M K , Hsiang-fu * B * R and Teng-ling § P ^ f t ) . We do not know
which of these sects participated in the great enterprise on which the
Mohists now set out, the organisation of a comprehensive summa of
Mohist disputation, consisting of short 'canons' {ching ) designed no
doubt to be learned by heart, together with explanations (shuo Ift) which
may at first have been oral and fluid but in due course were written down
as a separate collection. It is convenient to date the enterprise about 300
B.C., but it may well have begun considerably earlier and continued until
late in the 3rd century. Since all the writings on disputation can be dated in
58
relation to the Canons and Explanations, it is possible to distinguish three
stages in the development of the corpus:
(1) The oldest surviving document is the fragmentary Expounding
the canons (Yii ching f& $8). Judging by the opening sentences, the title
refers to the problem of 'expounding as canonical' the Will of Heaven and
the other doctrines over which according to Chuang-tzu the Mohists were
arguing and calling each other heretics. The opening section admits the
revolutionary consequences, for Mohists as for Confucians, of the concept
of nature introduced by Yang Chu. If a man thinks that his evil inclinations
come from his nature, and therefore from Heaven, it is useless to urge him
to obey the Will of Heaven. " T o expound Heaven's will as the right one
of the two, and his nature as criminal, is to 'sing' Heaven's will as the
wrong one. . . . The criminal will think that egoism is Heaven's will and
that what men condemn is the right one of the two, and his nature will be
incorrigible." From this point the dialectical writings never again mention
either the Will of Heaven or human nature. Instead of trying like Mencius
to prove the goodness of human nature, the Mohists set out to rationalise
the practical utilitarianism of their tradition. Expounding the canons lays the
foundations of an ethical system independent of the authority of Heaven,
built on the actual benefit and harm, desires and dislikes of individuals.
At the same period the Mohists started the work of assembling a list
of accepted definitions, by defining all words in the 10 theses of the school.
69
This document is unfortunately lost. There is evidence of another lost
document, on geometrised astronomy, the nucleus of their later work on
60
physics. But we cannot be sure that this was of Mohist origin.
57
Han Fei-tzu ch. 50 (Ch'en 1080/3).
68
Cf. § 2/3 Introduction, § 2/5 Introduction.
69
Cf. § 1/6/2.
6 0
Cf. § 2/4/2/3 Introduction.
(2) The Canons and Explanations fall into two parts, the definitions
(A 1-87) and the propositions (A 88-B 82). They cover the entire range
of Mohist learning, the procedures of description, ethics, the sciences and
logic, except that there are no propositions on ethics (these would have
been covered by Expounding the canons). A l l propositions from B 1 onwards
end with the formula shuo tsai X 'Explained by : X \ The formula confirms
that the Explanations already existed at least in oral form, and in any case
some Canons do no more than refer to the Explanations (cf. B 10 "Doubt.
Explained by: . . ."). The closest parallel to this organisation in other texts
is to be found in the six Chu shuo # Ift chapters of Han Fei tzu, each with
half a dozen numbered chingffi'canons' (with the summing-up formula
S?t&6: X "Its explanation is in X " ) , followed by correspondingly numbered
shuo 'explanations'.
It may be guessed that the authors intended to complete their summa
by working in the earlier definitions and inserting Expounding the canons
before B 13. But this would have involved considerable rewriting. As part
of the enterprise of defining their terms and systematising their principles
the Mohists had introduced new stylistic conventions, not yet observed in
Expounding the canons. The régularisation of syntax and strict use of only
a single particle for each grammatical function is so consistent that it can
61
only be the result of deliberate decision.
(3) After the completion of the structure someone perceived the
difference, overlooked in the Canons, between the proposition, which makes
an assertion, and the name, which does not. Names and objects (Ming shih
t
%x S) a mutilated but consecutive treatise, probably written by a single
y
hand, develops a new theory of description.

A general characteristic which distinguishes these writings from all
others of the pre-Han philosophers which survive as wholes is their
deliberately theoretical nature. They do not propagate Mohism, they codify
the techniques for defending it in debate. Other thinkers, from Confucius
down to Han Fei tzù, fail or refuse to detach philosophising from moralising
and practical persuasion. But the later Mohist summa never preaches;
everything it has to say about morals is pure ethics. The propositions
chosen for criticism in B 32-82 can often be recognised as theses of other
schools, but they are considered on their merits without naming the
thinkers who proposed them. This impersonality is unusual in pre-Han
philosophy, where the most interesting examples of disputation tend to be
(or to have been dramatised as) actual face-to-face debates between
Cf. § 1/3/1.
1/1 25
1
Mencius and Kao-tzii Hfi* , H u i Shih and Chuang-tzu. Earlier Mohist
disputation too is dramatised in the conversations of Mo-tzu. Once however
in the older dialogues the point of an episode is precisely that the validity
of an idea is independent of its proponent:
"Mo-tzu, when disputing with Ch'eng-tzu S ^ , cited something from
Confucius. Ch'eng-tzu said: 'You are no Confucian, why do you cite
Confucius' ? Mo-tzii said: 'This is something of his which is dead right and
for which there is no substitute . . .' " (Mo-tzii ch. 48, Sun 288/3, 4 ) .
1/1/2 L A T E R M O H I S T L O G I C , ETHICS A N D SCIENCE
1/1/2/1 Logic and the Chinese language

The danger of imposing his own categories on the thought of other
cultures is one which no investigator can finally escape. The most funda
mental of such categories are those implicit in the structure of the language
in which we think, and before coming to grips with the later Mohists we
must try to identify those most likely to mislead us.
1/1/2/1/1. The distinction between word and sentence, term and pro
position, is much more obvious in inflected languages than in a purely analy
tic language such as Chinese. In Chinese any verb (hsing 'go', ta 'be-
big') can stand by itself as a sentence ('He goes', 'It is big'), and the particle
yeh ife which turns ma M 'horse' into ma yeh 'It is a horse' is much less
obtrusive than the conjugated verb 'is' with subject 'It' which has to be added
in English. Throughout the Canons and Explanations a sentence is assumed
to be simply a name or string of names, and "call it 'ox' " is used
where we should choose the wording "say it is an ox" (B 3 5 ) ; the difference
was first appreciated by the author of Names and objects. One might say that
this distinction, which Western logic could take for granted from the
beginning, was the last and most difficult of the Mohist discoveries.
1/1/2/1/2. The distinctions between different senses of the Indo-

European verb 'to be' are immediately visible in Chinese, as in most other
languages outside the Indo-European family. The Western tradition had
great difficulty in clearly distinguishing 'there is X ' from 'is X ' , succeeded
in the Middle Ages in finally fixing the distinction only with the aid of a
62
terminology borrowed from thinkers in a Semitic language, Arabic, and
6 2
I have argued this claim in 'Being' in philosophy and linguistics (Verhaa
Part 5, 225-233). For the relation between Arabic ontology and Semitic language-
structure, cf. Fadlou Shehadi, Arabic and 'to be' (Verhaar Part 4, 112-125).
even then continued to see it as a distinction between existence and essence

within a single concept of Being. Where we use 'to be' Chinese uses
(A) Yu M 'there-is/have', wu № 'there-is-not/lack\ In Chinese
terms, there is a horse, and in it there are properties such as shape and
colour; or, if one prefers to translate^ by 'have\ the world has the horse,
which has shape and colour (or the horse exists and shape and colour
exist in it).
(B) Noun with particle yeh (HI -tfe 'It is a horse'), negated by the
verb fei $r 'is-not' ffiMfe 'It is not a horse'). A whole sentence may also
be followed by yeh ('It is the case that . . .') or negated by fei before the
verb ('It is not that . . .').
(C) The intransitive verbs translatable by Indo-European adjectives,
negatable like other verbs bypu ^ ( A 'is-big', ^ A 'not is-big/is not big),
and requiring no copula.
The corresponding concepts, each at least partially overlapping with
Being, are represented by nominalised (A) Yujwu 'there-is/there-is-not'
(which however embraces the having of properties); (B) Shih jk/fei
'is-this/is-not (a horse, the case that . . .); (C) Jan 'is-so' (embracing
63
all verbal concepts).
1/1/2/1/3. In Classical Chinese there is a radical difference between

the nominal sentence with final yeh negated by fei and the verbal sentence
negated by pu. This difference by no means coincides with that between
sentences with or without copulative 'is' in English, since we use the copula
with predicative adjectives, which in Chinese are translatable by verbs. In
the Mohist logic it is reflected in the distinction between being this (shih),
a horse, a man, a stone, and something being so (jan) of it, that it is big
or white, that one loves men or rides horses.
In Indo-European languages the nominal sentence, even in languages
such as Greek where it is common, is easily interpreted as a verbal sentence
with the verb 'to be' understood. What an object is, its essence, laid down
in its definition, appears in the Aristotelian tradition as part of its 'being',
as objective as its existence even when clearly distinguished from it. In
Chinese however the word shih 'this', being a demonstrative, simply picks
out one thing from others as the one in question; there is no 'essence',
merely the existence (yu) of the thing with all its properties. The world
has a certain object, and if the object has, for example, four legs, teeth,
a tail, colour (white, black), no horns, it fits the name 'horse' (We take these
6 3
' The relation between the Chinese language and the concepts which overlap
Being' is discussed in G(3) (for sinologists) and G(8) (for the general reader).
1/1 27
miscellaneous properties from B 12, 66. N O 14, 18). There are certain
terms difficult to handle in English without the help of copulative 'be',
such as ku #C 'the thing as it inherently is', ch'ing ! § ( = № ) 'what is
essential to being X ' . But the ku is the object as the world has it, with all
that it has in itself, irrespective of how we name it; its ch'ing is everything
which must be in it if it is to fit the name.
We may connect this with the profound gulf which in Chinese disputa
tion separates names from objects. There are no essences bridging the
realms of words and of things, inclining one to think of a general concept
as a picture, representation, point-by-point equivalent of something
transcending the world of particulars. It is not that the Mohist has altogether
escaped the naive tendency to think of names as evoking mental pictures
(yi M 'ideas', hsiang ffi(= $S) 'images'), but he thinks of ideas merely as
one kind of standard (A 70), with which we compare particular objects
and judge whether they 'fit' (tang Hi), in the same way that we measure
their lengths against a ruler (B 70).
1/1/2/1/4. In Classical Chinese there is some inconvenience in

distinguishing a property from the thing which possesses it. Pat che fi #
or nominalised pat ordinarily refer to a white thing, and can be pinned
down as referring to the colour white only by parallelism or expanded
constructions. Although Kung-sun Lung in his 'White horse' says explicitly
that "'White' is that by which one names the colour" .(fi
he uses pat and pai che throughout for the white thing which the object is
64
in addition to being a horse. The Mohist has progressed beyond this
stage, treats being white as something which is 'so' of a stone (NO 1),
and hardness and whiteness as 'being in' (tsai %£) the stone (B 37). But
the opponent criticised in B 38 thinks of the stone as being 'this', a white
thing, in addition to being "what one refers to 'a priori'" (@f3fcl?), a
hard thing.
1/1/2/1/5. In Chinese there is no presumption that the subject

presents the topic except in the nominal sentence ('X Y yeK). In a verbal
sentence with both subject and object, ' X chien H . Y ' ('X sees Y ' ) , it is a
matter of indifference whether X or Y is the topic; if one is to be marked
as such, special constructions are used, ' X tse JM chien Y ' ('As for X , he
saw Y ' ) or ' Y , X chien chih M* £ . ' ('As for Y , X saw it'). Of course much
the same is true in practice of Indo-European languages. However in
Indo-European grammar the verb takes its person and number from the
subject, which consequently cannot be freely omitted as in Chinese, and
6 4
Cf. G(6) 141-144.
when omitted has to be understood. Moreover we have a verbal copula:

and the ' X is Y* sentence, in which the subject really does present the topic,
tends to be taken as the paradigm for all sentences. The Western tradition
has conceived the sentence in grammar and the proposition in logic as
composed of a subject and a predicate, reflecting a topic/comment dis
tinction which we have supposed to be basic to all thought.
It may therefore seem strange to us that the Mohist appears to conceive
what we call the predicate as the 'root' (ken IB) from which the sentence
grows by the addition of the subject and other dispensable elements
( N O 6, 10). It may also surprise us that when the Mohist talks about
generalisation he is more often than not thinking about the object (for
example "He loves men" and "He rides horses" in N O 14,17). The Mohist
does not recognise the syllogism as a specific logical form, but if one tries
the experiment of extracting syllogisms from quantifying arguments in the
corpus we find rather the type with quantification of the object:
B 74 ftPpÂ . . . «83£0TFP1 o
"He asks about all men. I love all whom he asks about. (Therefore I love
all men.)"
N O 15 ÂA-tfe . . . ^ g S f c o
A 96 f ^ » A °
"Robbers are men. He does not love robbers. (Therefore) he does not love
some men." (The form of an argument attacked in N O 15.)
N O 14
17 » 1 °
" A white horse is a horse. He rides a white horse. (Therefore) he rides
some horse."
1/1/2/1/6. Indo-European languages indicate time and number by

inflection at every occurrence of a verb or noun, Chinese by particles
introduced only when time or number is relevant. Inflections have the
advantage of showing up the structure of the sentence, almost invisible in
Classical Chinese until approached with the resources of modern linguistics.
But they have little practical value; gender is frankly irrational, it is rarely
useful to know whether there is one of a thing or between two and infinity,
we need to be told whether an event is past, present or future no more
often than is indicated by the temporal particles of Chinese. The idea that
there are confusions in early Chinese thought due to the absence of tense
65
and singular or plural seems to me quite untenable. At some places one
6 5
Cf. G(5) 39-48.
1/1 29
has trouble rendering into English without committing oneself to tense or

number (cf. B 49, 51 and the end of N O 18), but this is merely a translator's
problem.
There are however ways in which tense or number may affect the
schemas within which we think. The number terminations of nouns
encourage us to conceive enumeration primarily as counting of things with
the same name, similar to each other and countable only in one way. When
we think of parts in a whole, interconnected by relations other than
similarity, and divisible and countable in various ways, it is as a special
and more complicated case. But the Mohist's starting-point is the counting
of parts divisible and subdivisible according to convenience. A unit is a
t'i f t , primarily a 'member' (of the body), a 'division in a total' (B 2
^M&H), the total itself being countable as one at the next level of enumera
tion. T'i may be either parts in a whole or members in a class, the latter
being simply t'i which are 'of a kind' (lei S§), 'have respects in which they
are the same' (A 86 ^Mffl).
1/1/2/1/7. In Chinese one can place a verb in a nominal position and

treat it as a concept (B 76 t , Sife "Benevolence is love"), even refer to
love or benefit as wu % 'things' (EC 2). But this kind of nominalisation is
almost confined to nominal sentences; in verbal sentences a verb in the
subject or object position tends to refer to the agent (tSjJHSfs 'promote
the worthy and employ the able'). There is none of the overpowering
tendency towards hypostatisation of thinking in Indo-European languages,
where inflection converts verbs and adjectives into abstract nouns, in
varying degrees assimilated to ordinary nouns by number, gender and the
articles, so that it is difficult to deal with any concept except on the analogy
of a nominal concept. Among words which partly correspond to different
senses of 'true', tang 'fit the fact' and k'o 'be admissible' may stand in
nominal positions (A 74 'Sfib " T o win in disputation is to fit the
fact", B 71 8*, ^ "life " T o be self-contradictory is to be inadmissible"),
but they cannot assume the substantiality of 'truths', 'the truth', 'Truth',
'the naked truth', 'serving truth'. In finding one's way about the Mohist's
world one is struck, not merely by the consistency of his nominalism, but
by the inconceivability of any conceptualist or realist alternative. The
nearest to an unobservable entity is perhaps the chih £0 'intelligence' by
which we know, but even this seems to be conceived not as a thing inside
the man but after the analogy of the eyesight by which we see (A 3).
A common objection to nominalism in the West is that it implies one
universal, similarity. But our assumption that 'similarity' is a common
name at all is bound up with the linguistic fact that the adjective in ' X is
similar to Y ' can be converted, not merely into a noun, but into a noun
with number like 'horse'. This enables us to object that although the
nominalist analysis may account for the use of 'horse', we require some
other criterion than similarity to extend the name 'similarity' from one to
all similarities. (Number termination, as we noticed in §1/1/2/1/6, tempts
us to assimilate even the counting of similarities to the counting of horses.)
A thinker in ancient Chinese would certainly see this problem from quite
a different angle. We can imagine the Mohist admitting that jo 'like'
does not fit into any of his three types of name (A 78), and having difficulty
in establishing a fourth to contain it, but he would hardly see that as an
objection to his account of common names.
1/1/2/2 The four objects of knowledge

A modern inquirer into later Mohist thought cannot deny himself the
convenience of categorising it in terms of Western disciplines, as logic,
ethics, geometry, optics, mechanics. But if we take such categories too
seriously we shall of course be blinding ourselves to its true content.
According to A 80 there are three sources of knowledge, report, explanation,
and observation, and four objects of knowledge, names, objects, how to
relate them (ho and how to act (wet j § ) . We shall argue in § 1/6/1 that
the Canons are designed as a survey of all the four branches of knowledge
by explanation (shuo t&):
DEFINITIONS PROPOSITIONS
Explaining how to relate
names to objects A 1-6 A 88-B 12
Explaining how to act
(ethics) A 7-39 (EXPOUNDING THE CANONS)
(Knowledge and change) A 40-51 B 13-16
Explaining objects
(the sciences) A 52-69 B 17-31
Explaining names A 70-75 B 32-82
(Appendix: 12
ambiguous words) A 76-87
Expounding the canons had already provided the propositions for the
second division. After the completion of the summa, the first of the arts
(the relating of names to objects) was re-examined in Names and objects.
Of the two bridging sequences on knowledge and change, the sequence
of definitions ends with chih j h 'staying' (A 50), the temporary validity of
31
names fitting transient phenomena in the first two arts, and/)/ *j& 'necessary'
(A 51), the unending certainty of judgments in the last two.
We do not know what the Mohist called his first three disciplines, but
the last is surely the pien M 'disputation' defined in A 74 and defended
in B 35. We should probably think of these Canons as promulgating a
narrowed conception of disputation proper after the discovery of the
difference between necessary and unnecessary reasoning. (The Mohist
does not however adopt a new word for argumentation in general, and
Names and objects reverts to the practice of treating the other disciplines as
branches of disputation (NO 6).) When we speak of Mohist 'logic' we are
in some danger of confusing disputation proper, which claims the strict
necessity we ascribe to logic, with the discipline expounded in A 88-B 12
and Names and objects, which lays down procedures for consistently
describing changing phenomena. If we have correctly identified the
underlying principles of classification, the last of the four arts is no closer
to the first than to either of the others. Each has an element which is known
'a priori' (hsien 5fe) by explaining the logical implications of the definitions
of its names, and another which is "unknowable 'a priori'" (wet k'o chih
^^T£P). The former belongs to disputation, the latter to the particular
art. If we look at our table vertically instead of horizontally, the whole
series of definitions (A 1-87) belongs to disputation, only the propositions
distinguish what is special to the five divisions.
In order to find our way around the summa, we must put ourselves in
the position of the Mohist who is consulting it as a manual. T o become a
fully educated Mohist I must learn how to apply names consistently, how
to choose between courses of action, how to investigate the causes of physical
phenomena, how to deduce ' a priori' from the definitions of names. How do
I learn how to fit names consistently to objects ? I consult the propositions
of Part 1, where I find the procedures of description laid down in A 88-B 12.
I note that the circle is the typical example of something known 'a priori'
(A 93); since it is a geometrical concept I look it up in the definitions of
Part 4 (A 58), then the words used to define it (A 53, 54). Where do I look
for the procedures of considered choice ? Not in Part 2 of the Canons, be
cause they have already been formulated in Expounding the canons (EC 7-9)
Having consulted Expounding the canons I see that the moral concepts are
established by their definitions as 'desired "a priori"' (EC 2). I therefore
refer to the ethical definitions of Part 2 (A 7-39) and trace back the words
used to define them to the words 'desire' and 'dislike', which being undefined
are noted in the appendix to the definitions (A 84). I am now equipped to
approach the specific problems of disputation in Part 5 (B 32-82).
However the modem reader, who is not using the summa as a practical
handbook, requires a rather different approach. The most convenient may
be to start with the concepts which underlie all four disciplines (knowing,
names and objects, change and necessity) and then examine in turn the
first three, including in each the 'a priori' component which belongs to
disputation.
1/1/2/3 Knowing, names and objects, change and necessity

A 3-6 lays down a series of parallels between knowing and seeing.
The intelligence (chih #P) by which we know is like eyesight; thinking,
which is using the intelligence to seek something, not necessarily with
success, is like peering; knowing (chih (£Q)*I?) is like seeing; wisdom
(chih 3K) is like clear-sightedness. But knowing differs fundamentally from
seeing in that perception ends with the departure of the object, while
knowing persists; we are still able to describe the object (A 5). This confirms
that the intelligence is a special faculty additional to the five senses (B 46).
It remains in us throughout life, even in sleep, when it is quiescent
(A 22-25).
The sources of knowledge are report, explanation, and observation
(A 80-82. Cf. B 9, 70); the Mohist summa is concerned only with explana
tion (shuo f!£, defined A 72). Observation has the limitation that it may be
of some cases but not of all (A 82). Knowledge is of names, of objects, of
how to relate them, and of how to act (A 80).
A name is that by which one calls an object, an object is that which
is so called (A 80). According to A 79 'calling' (wei f l ) is of three kinds,
transferring (yi apparently the defining of one name in terms of others),
referring (chii picking out an object from others by means of its name),
and 'applying' (chia Sfl, addressing, praising, blaming. Cf. A 29, 30).
Like Hsiin-tzu in his Right use of names, the only other pre-Han text
which discusses the problem of common names, the Mohist has a radically
nominalistic approach to naming. We name a particular object 'horse' and
apply the name to all objects which are like it (jo S ) , of a kind with it
(lei !8) (A 78). Judging by a fragment in A 31, the nominalist analysis is
extended to the name itself; we refer to the similar objects by sounds which
are like the initial sound. Referring by name, we present not the object but
an analogue of it (A 32). B 70 disposes of a difficulty in nominalism, that
it implies that through the medium of language we can never know X ,
merely know that it is like something we already know. The Mohist deals
with it by taking the example of 'white'; if we know that something is
similar in colour to a white thing we do know that it is white. Common
1/1 33
names are distinguished from two other kinds of name: the unrestricted
name 'thing', applicable to any object irrespective of similarity: and the
proper name confined to one object (A 78). The nominalist position is not
argued but taken for granted; although Fung Yu-lan and others have tried
tp identify the chih fa (literally 'pointings') of Kung-sun Lung as univer
sal, there is no firm evidence that there were any realists in pre-Han
66
philosophy.
For the Mohist, the deepest and most troubling of problems is the
relation between knowledge and temporal change. As we noticed at the
end of § 1/1/1/2, he lives in an age of rapid social transformation in which
ancient authority is no longer an adequate guide to conduct. He has
developed the moral teaching of Mo-tzu into an elaborate ethical system
justified not by authority but by the procedures of disputation; he believes
that, alone among the sages, Mo-tzu taught principles which are necessary
(pi $&) and therefore invulnerable to time. "The judgments of the sages,
employ but do not treat as necessary. The 'necessary', accept and do not
doubt" (A 83). This is almost the only issue which can provoke a tone of
rhetoric in the Mohist's cool voice. "Among Heaven's constants its presence
is prolonged with man" (B 41). "Even if there were no men at all in the
world, what our master Mo-tzu said would still stand" (EC 2).
This concern with time is already apparent in the Explanation of the
definition of 'knowing' in A 5. What distinguishes knowing from perception
is that "having passed (kuo the thing one is able to describe it". If we
knew only by means of the five senses, "knowing as it endures would not
fit the fact" (B 46). But the object of knowledge may come to anend(jyi B ) ,
so that we have to ask ourselves the question "Is it knowing? Or is it
supposing the already ended to be so ?", which is the kind of doubt arising
from kuo 'having passed' (B 10). Of a past event we can still say that it
'has been so' (ch'ang jan H B 61), but we must be careful to locate it
in the time of its being so, not in the 'not yet so' (wet jan ^ $J, B 16).
The temporal particles yi ('having ended/already', analysed A 76) and
ch'ieh IL 'about to' (defined A 33) are words used to refer to the same events
from earlier and later viewpoints (A 33). T o use either of them of an event
necessarily implies that it has happened (B 61) or will happen (B 51). But
it does not follow that future events are destined; the necessity of " I f it is
about to be so it is necessary that it be so" is simply the logical necessity
of the inference, belongs to the realm of names and not of objects, and has
no fatalistic implications, as can be seen by constructing the more complex
6 8
Cf. G(l) 283-285.
"That which cannot end unless you are about to exert effort necessarily
does not end unless you do exert effort" (B 51).
Of the two bridging sequences on knowledge and change, the defini
tions (A 40-51) begin with the words 'space* and 'duration', while the
propositions (B 13-16) consider the relation between the two concepts. In
ordinary pre-Han usage the words closest to our own 'space* and 'time*
are yu ¥ and chou %; but these tend to suggest rather the 'cosmos as it
extends' and the 'cosmos as it endures', from which the concepts of space
and time are imperfectly abstracted. The use of them involves some
danger of thinking of the two as ultimately identical, so that principles
accepted at a certain time as valid throughout the cosmos in its spatial
aspect might seem to be valid also for its temporal aspect (that is, for all
time). The Mohist retains 'space', but substitutes for the noun chou the
nominalised verb chiu {K 'duration', and his definitions clearly establish
the concepts as abstract: " 'Duration' is pervasion of different times" (A 40),
" 'Space(/extension)' is pervasion of different places" (A 41). In B 13-16 he
shows that space and duration are not mutually pervasive like the hardness
and whiteness of a stone; the whole of space is present at any one moment,
it moves on, and its movement has duration. When the sage Yao ruled the
world his principles of government were appropriate to his times. But when
we say that Yao was a good ruler we are talking about the past from our
present viewpoint; a man of his time talking about the present would have
to call him incapable of ruling (B 16. Cf. also B 53).
The corresponding sequence of definitions proceeds from 'commence
ment', 'transformation', 'reduction', 'circling round', 'rotating', 'moving',
to conclude with the logical words chih lb 'stay' and pi ift 'necessary'
(A 50, 51). " T o 'stay' is to endure as such" ( i t , J ^ V t i i ) , to be ox and non-
horse from the commencement of the object's existence throughout its
duration. "The 'necessary' is the unending" (*&, /pB-tfe). Necessity
/
belongs to one kind of relation (ho & ) ; " i f necessarily we do not have one
without the other, the relation is 'necessary'" (A 83), for example elder
brother and younger brother (A 51).
The two disciplines which precede the bridging sequences concerned
the fitting of names to transitory objects and actions, which has only the
temporary validity of 'staying' (cf. N O 1 ^ I t ^ i f t * ^ "Names and
objects are not related necessarily"). But the two which follow concern the
causal relations between objects in the sciences and the logical relations
between names, both of which are necessary.
i/i 35
1/1/2/4 Description
1/1/2/4/1. The art of description in the Canons

We live in a constantly changing world of concrete and particular
objects (shin W, literally 'solids'), which we describe whether individually
or generally by names or strings of names, such as JfiSifeS "This stone
is white" (NO 1) or "Oxen and horses have four legs" (B 12).
Although in A 78 the names cited as typical are all nouns ('thing', 'horse',
'Jack'), ming & 'name' covers every kind of word, not only 'stone' but
'white' and 'big' (NO 1), even chu ft 'all' and to & 'much' (B 3), or huo &
'some' (B 33). The Mohist does not think in terms of a realm of universals
in which each name can have its own point-by-point counterpart; he thinks
of many names as fitting one mutable object, the name of what it is ('stone'),
and names of what is so of it, either throughout its duration ('white') or
temporarily ('big' until the stone is broken up) (NO 1). Even if a name
fits, it may not be applicable when combined with some other name (B 3.
Fighting men are 'two' but not 'both two'). Ch'ieh JL 'about to' and yi B
'already' have no counterpart at all in the object, being used in saying the
same thing before and after the event (A 33). The authors of the Canons
are well aware that names change their meanings in combination (B 3), and
in A 76-87 list the different senses of 12 words common in disputation.
But in spite of these subtleties they have no conception of the sentence, as
something other than a string of names; for that we must wait for Names
and objects.
The first of the Mohist disciplines lays down procedures for relating
names to objects. The norms presented in A 88-B 12 (and later in Names
and objects) are rules, not for inferring from known to unknown facts, but
for describing known facts consistently. The system of concepts starts from
two phonetically and semantically cognate words, y o / * N j A K 5=f 'like' and
/fl»/*NJAN $t 'so'. It follows from the nominalist analysis in A 78 that
I cannot decide that something is so of an object, for example that it is
'circular' or 'square', without committing myself to extending the same
name to all objects which are like it in this respect. T o use the name
consistently I require a fa $c 'standard', defined as "that in being like which
something is so" (A 70), with which I compare the mao IS, JS 'character
istics' of the object. Thus as the standard for a circle I may appeal to the
idea of it, to another circle or to the compasses (A 70). I may explain 'white'
by saying "It is like the colour of this" (B 70), presumably some white
thing such as snow, or 'tiger' by drawing a picture and saying " I f the
characteristics are like the picture it is a tiger" (A 32). Objects sharing the
same name are not necessarily alike except in the respects covered by the
standard, for example pieces of stone and of wood both of which fit the
standard for 'square* (B 65 "Anything of which they complete the character
istics, as in being square, is so of both the things"). If only parts of an object
fit the standard, as in the case of a man described as 'black* (A 96), we also
require aj>w H 'criterion*, defined as "where it is so** (A 71, on which the
Explanation notes: "Being 'so* is the characteristics being like the standard**).
Although the Mohist allows us a wide choice of standards, the kind
to which he regularly appeals in practice is the verbal definition. When he
mentions the 'idea* (yi M) as his first choice as a standard for the circle,
he must be referring to the definition of a circle in A 58, for the idea could
serve as a standard in public debate only if formulated in words. In any
case there is no term in the Mohist*s technical vocabulary by which he can
be referring to the definitions which are so important to him unless it is
'standard*, for the word yi & 'transferring* of A 79 (the type of 'calling*
which we identify as defining) is never put to use anywhere in the corpus.
Definitions would be the one kind of standard which perfectly fulfils the
ideal of delimiting the respects in which the objects called ' X ' are required
to be alike, the ch'ing ft (= fn) of X , what is essential to being X (cf.
§ 1/4/6).
'Circle* is a name which 'stays* (chih ih) unequivocally in circular
objects (A 93). But many other names require rules for fixing them on
objects, 'making them stay* (chih used causatively). The Mohist lays down
the procedures of consistent description in A 88-B 12. In the first place
we may call Y 'elder* or 'more* in relation to X but 'younger* or 'less* in
relation to Z , but it either is or is not 'black* or 'white*, 'this* or 'not this*
(A 88, cf. B 80). T o justify a description of either sort we follow a procedure
in four stages:
(1) Chih %k, the 'commitment*. T o justify a description I first explain
how I am committed to using the name.
(2) Fa, the 'standard*. If my commitment is questioned, I provide
"the standard for the commitment** (A 94).
(3) Yin, 'the criterion*. The matching of a circle with its standard
is exact (cheng J E , A 98); but in the case of 'black man' we have to decide
whether the parts which are black or those which are not are 'appropriate*
(yi 3H) as the criterion (A 96), which is "fixing the criterion** (A 97).
(4) Lei, the 'kind*. Being of a kind is one of the four types of sameness,
"being the same in some respect** (A 86); objects which are of a kind share
the same name, for example 'horse* (A 78). Having decided how to describe
one object, I am obliged to describe all similar objects in the same way;
1/1 37
the name which I have made to 'stay* in X 'proceeds' (hsingfi) to all objects
of a kind with X (B 1). It is important however to "fix the kind" (B 1),
establish the proper level of classification. Describing an animal with
cavities under its eyes as having four eyes commits me to calling all the
other milu deer four-eyed, describing it as having four legs commits me
to saying that all animals have four legs, but not all living things, since
birds do not (B 2).
Whether a name fits has to be reconsidered as soon as another name
is added. Adding 'both' to 'They fight' and 'They are two', we find that
they both fight but are not both two. It is not that the objects have ceased
to be two, "inherently they are what we called them"; it is simply that
names operate differently in different combinations (B 3). This example
leads the Mohist to distinguish what we should call the distributive and
the collective use of words. Although we cannot say of two objects that
"They are both two" (ft ^) we can say that in some respect "They are
both one" (ft —), pronounce them to be of a kind, and call both by a
common name, 'ox', 'horse'. We may then count oxen and horses as two
things which are themselves one in having four feet, and so combine groups
in wider groups, each of which counts as one on its own level (B 11,12, 59).
Within the theory of description, as in ethics and the sciences, the
analysis of names belongs to the sphere of pien 'disputation'. Disputation
proper is concerned, not with describing what is temporarily so of transitory
objects, but with deciding whether something 'is-this' or 'is not' (shih fei),
is ox or non-ox, and its judgments follow by strict necessity from the
definitions of names. (In this case we have positive evidence that the
Mohist's 'standards' are primarily definitions, in Hsun-tzvi's digest of the
four disciplines in Right use of names (cf. § 1/6/1); to refute a denial of
common sense in the final discipline one "tests it by the convention for
the name ( £ $ J ) , and uses what one accepts to show that what one rejects
is fallacious".)
" O f a thing so in one case, a thing not so in one case, that it stays so
is unnecessary, that it is this or is not is necessary" (A 51). If I imagine an
object out of sight beyond a wall (for this image, cf. § 1/5/8), how much do
I know about it simply by knowing its name? If its name is 'stone', I may
refer 'a priori' (hsien chii 5feS, B 38) to its hardness, but shall not know
whether or not it is white until I cross the wall. I do not know whether a
particular pillar will be round, but do know 'a priori' (hsien chih 5fe^l, B 57)
the idea of a pillar, which the Mohist seems to conceive as a schematic
mental picture with a vacant place to be filled by roundness or squareness.
In the case of a hammer, which is recognised not by its shape but by its
function, the idea is not knowable 'a priori' (wei k'o chih 7fc«J£Q, B 58);
I can form a mental picture of a particular hammer only after seeing what
it looks like.
Among pairs of names we sometimes find that X and Y 'follow from
each other' (hsiang ts'ung E C 3, A 93), or 'dismiss each other'
(hsiang ch'ii ffii, A 93), or that one of them cannot 'be dismissed without
the other' (p'ien ch'ii H № , B 3, 4, 7), for example 'seeing' and 'appearing',
'length' and 'breadth' (B 4). The authors of the Canons, who have no
conception of the proposition, think of all logical implication in terms of
pairs of names of which at least one is the 'complement' (ti T£f) of the other,
after the analogy of 'elder-brother' and 'younger-brother', a pair in which
"both are complements" (A 88). Wherever implication is two-way either
member of the pair may be inferred from the other as its converse (fan ifc
defined in A 73 as " I f inadmissible then on both sides inadmissible"), for
example that if one class of objects is called 'oxen' all other objects are
non-oxen.
As long as we are describing the realm of changing objects we are
obliged only to follow consistent procedures, extending the name we apply
to one object to every similar object. But when we confine ourselves to the
realm of names, these consistencies reveal themselves as complementary
relations which are necessary. If we put the quotation device yeh che i f e ^
after 'like the object' we can say that "For 'like the object' one necessarily
uses this name" (A 78 T J i & l ^ g - i f e . Cf. A 31). Two obscure
and corrupt passages (A 39, 85) suggest that the Mohist conceives being
deemed (wei S ) an X as the complement of being similar to the standard;
the deeming is 'engendered' (sheng £fe) by the object as we perceive it, but
becomes necessary only when 'matured' (shu £ft = fh) by comparison with
the standard. The word 'necessary' is "said of cases where a complement
is matured" (A 51 ffi*l«Sîii).
When in justifying descriptions or in the sciences we offer a 'reason'
(ku $t), we are offering one of complements, which may be either one-way
or two-way. In the case of a 'minor reason' (a necessary condition) it is
one-way: "having this, it will not necessarily be so: lacking this, it neces
sarily will not be so". The 'major reason' (necessary and sufficient condition)
implies a two-way relation: "having this, it will necessarily be so: lacking
this, necessarily it will not be so" (A 1).
In disputation we are not, for example, deciding whether to say of a
certain man who loves some men that "He loves men", but judging some
such question as whether or not the love of some men is the love of men.
A l l issues are between shih 'is-this' and fei 'is-not' (typically, 'ox' and
1/1 39
'non-ox'), alternatives of which to affirm one of certain things is to affirm

the converse (fan) of all the rest (A 72). Disputation is defined as "contend
ing over converse claims" (A 74), not, as a Westerner would have been
inclined to expect, contending over contradictory statements about the
same things. The Mohist has a clear idea of what in Western logic is the
u
law of the excluded middle; they do not both fit the fact, and if they do
not both fit necessarily one of them does not fit" (A 74). It is not true
disputation to argue over whether something is a whelp or a dog (it may
be both), or an ox or a horse (it may be neither); " i n 'disputation', one man
says it is this and the other that it is not, and the one who fits the fact is the
winner". Consequently, "to say that there is no winner in disputation
necessarily does not fit the fact" (B 35).
The Mohist does not use a single term corresponding to English 'true'.
A name or complex of names applied to an object either 'fits' (tang Hf) or
'errs' (kuo №); detached from the object and analysed in disputation it is
judged 'admissible' (k'o "J), 'inadmissible' (pu k'o "J), or 'fallacious,
self-contradictory' (pet l#). How far does he go in divorcing the realms of
fact and of logic? He regards the logically inadmissible as necessarily
mistaken in fact (B 71 " I f what this man says is inadmissible, to suppose
that it fits the fact is necessarily ill-considered"), and uses disputation to
show that a claim does not fit the fact (B 35, 40) although not that it does.
Of one name, 'thing', it is said that "any object necessarily requires this
name" (A 78). A crucial question is whether the thing proved in disputation
to be necessarily an ox is conceived to be the thing as described or the
actual object in front of our eyes. There can be little doubt that the latter
is the right answer. The Mohist raises no epistemological questions, he has
no doubts that whether by observation, report or explanation we can know,
and that when we do "we necessarily do know" (A 3). Presumably the
relation between the animal I see and my knowledge of it is necessary in
the sense that if it is not really there I do not know, merely suppose (yi wet
&s U cf. A 24). In addition to knowing the object by observation I know
by explanation (if it fits my standard for 'ox') that it necessarily is an ox.
But what if it happens to fit someone else's standard for 'horse' ? It may
be assumed that the Mohist recognises naming as conventional, although
he nowhere in the text as we have it says so explicitly; Hsiin-tzu in his
Right use of names (which contains very little not in the Mohist summa)
says that "a name has no inherent appropriateness, we name by establishing
a convention, and when the convention is settled and the usage becomes
customary we call it appropriate". In Seeing things as equal Chuang-
tzu had argued that since my 'that' is your 'this', and we can all
name things as we please, I have only to transfer the name 'horse' to

something else to prove that a horse is not a horse. The Mohist replies in
B 68 that it does not matter which of two things is called 'this' as long as
the other is called 'that*; the point is that they must be called differently.
He has progressed a long way ahead of Chuang-tzu. In disputation as he
explains it, as soon as Chuang-tzu presents his new definition of 'horse*
it becomes plain that there is no disagreement about what the object is;
it is necessarily a horse by the commonly accepted definition, necessarily
not a horse by Chuang-tzu's.
1/1/2/4/2. The art of description in Names and objects

Names and objects is a consecutive treatise on the procedures of
consistent description, badly mutilated throughout the first half. It makes
several innovations in the theory of naming, distinguishing "naming on the
basis of shape and characteristics" (NO 1 'mountain', 'house', N O 7
'sword'), "naming by reference to measure or number" (NO 1, 'big'),
"naming on the basis of residence or migration" (NO 2, the 'Ch'in' of a
'Ch'in horse'). But its most important innovation is the introduction of the
tz'u f$ 'sentence/proposition', for the first time distinguished from the
name. The distinction, grammatically less marked in Chinese than in
Indo-European languages, seems to have attracted attention only after it
was noticed that "knowing is different from having a pictorial idea"
(NO 3). The name 'pillar' evokes a mental picture which might be inter
preted either as a pillar or as a piece of wood, but knowing that it is a pillar
is knowing that it is the former, not knowing that it is the latter (cf. N O 3).
With this discovery the Mohist's attention shifts to the similarities
and differences, not between objects or names, but between the propositions
by which we describe. The Canons had distinguished four types of sameness
(A 86), which Names and objects lists as 'sameness with the same name', and
supplements with four types of 'sameness with the same root' (IRHSÎ^
N O 6). The most important of these are 'sameness in being this' (-jk^M)
and 'sameness in being so' ($*;£N); evidently the 'root' of the sentence
is the part substitutable by 'this' (the complement of a nominal sentence)
or 'so' (the main verb of a verbal sentence). It corresponds closely to the
predicate of Western logic, but is conceived, not as predicated of a subject,
but as the root from which the proposition grows (cf. § 1/1/2/1/5).
According to the theory of description in the Canons, having applied a
name to one object we are committed to applying it to all similar objects;
we 'proceed' (hsing) from what is so of the object to what is so of the kind
(lei). Names and objects restates this principle in terms of the proposition:
1/1 41
N O 10 » » a s s » & m n % & ·
"The proposition is something which is engendered in accordance with a
fact, becomes full-grown according to a pattern, and 'proceeds' according
to the kind."
The part first engendered we take to be the 'root', which is judged to
fit the fact by comparison with the fact; examples would be:
('This') ASffc "It is a horse"
('So') "It has four feet"
('So') St "He rides . . ."
The sentence grows to its full length according to a pattern of pauses
(tuan and sequence-positions (tz'u ^C), terms borrowed from the
geometrical Canons A 61, 69 ( N O 9):
N O 14 S H / B H i l " A white horse is a horse"
X/E3JS. " X has four feet" (cf. B 2)
N O 14 3 ! S S I "He rides a white horse"
Finally the sentence 'proceeds' by the substitution for ' X ' of the name
of 'the thing it is' (shih), for example of 'horse' for 'white horse'. In the
system of the Canons (B 1) one proceeded from "being so in the instance
here" ( j i y y f c i b ) to "being so of the thing it is" ( J i j y M b ) ; in Names
and objects ( N O 13) the second stage is called "being so if the instanced
is this thing" (TjjS-Mffe). We might propose as an example:
X H J £ " X has four feet", "Horses have four feet".
However Names and objects considers only cases in which the word
substituted is the object of the verb:
N O 14 f R £ J § "He rides a white horse", S I S "He rides horses".
We have noticed that for the Mohist it does not matter whether the
word generalised is subject or object; although in Classical Chinese one
expects the subject of a nominal sentence to present the topic, there is no
presumption that it is the subject rather than the object which presents the
topic of a verbal sentence (§ 1/1/2/1/5). The Mohist is especially concerned
with the generalisation of the object because it presents difficulties which
obscure problems arising from the doctrine of universal love. The substitu
tion of the wider for the narrower name does not in certain idioms generalise
at all (cf. an example in N O 15: "She serves her parents", • A
"She serves a man", her husband). Even when it does generalise one is not
always obliged to 'proceed' in the same way as with the subject, yet there
are also cases in which one is. If I describe a man as riding a white horse
I am committed to saying that he rides (that is, can ride) any similar object:
"He rides horses". If I describe him as loving 'Jack' or 'Jill', the bondsman
and bondswoman who are the stock examples of people who cannot be
loved for any reason except that they are human, then I must say that he
loves men (NO 14). However even in these cases "He rides horses" and
"He loves men" generalise in different ways; he loves all men but he has
not ridden all horses (NO 17).
The problem of consistent description is therefore obscured not only
by logical complications but by the semantic shifts of idiom, which are
equally relevant to the Mohist art of description; changes of meaning of
words in combination had been noticed in В 3, and twelve ambiguous
words listed in A 76-87. The Mohist's solution is to divide propositions
into three classes:
(1) "So if the instanced is this thing":
N O 14 " A white horse is a horse, to ride a white horse is to ride horses".
(2) "Not so though the instanced is this thing":
N O 15 "Her younger brother is a handsome man, but loving her younger
brother is not loving handsome men".
(3) "So though the instanced is not this thing":
N O 16 "Reading a book is not a book, but to like reading books is to like
books".
The essential point for the Mohist is that the similar must be described
similarly, the different differently, and that to do this consistently one must
discern the underlying differences between formally similar types of
proposition. For example, suppose we have a man who loves even Jack and
Jill and who also executes robbers; since I say that he 'loves people', must
I also say that he 'kills people' ? This raises one of the elusive problems of
idiom which the Mohist is trying to pin down; if the phrase is understood
neutrally I obviously must, but 'killing people' (ШК) normally carries an
implication of murder or wanton slaughter. The Mohist's decision is that
executing robbers is not killing people, because it belongs with eight
parallel propositions in the second class:
(1) "Jill's parents are jen Л (people), but Jill's serving her parents
is not serving^/* (a husband).
(2) Her younger brother is a handsome man, but loving her younger
brother is not loving handsome men.
(3) A carriage is wood, but riding a carriage is not riding wood.
(4) A boat is wood, but entering a boat is not entering (piercing,
soaking into) wood.
(5) Robbers are people, but abounding in robbers is not abounding
in people.
43
(6) Being without robbers is not being without people.

(7) Loving robbers is not loving people.
(8) Not loving robbers is not not loving people.
(9) Killing robbers is not killing people.
The Mohist is no doubt aware of the possibility of further subdivisions,
but he assumes that all propositions in any of his three classes are true
parallels. In fact of course they arc heterogeneous; it is primarily on the
strength of Nos. 2, 7 and 8 that we are led to agree that we cannot consis
tently describe executing robbers as 'killing people' unless we empty the
phrase of its pejorative meaning, and understand it in the neutral sense
which would allow us to shift the proposition to the first class. The effect
of the parallels, as in Chinese style generally, is to narrow the possible
range of meaning. Nos. 6 and 7 seem at first sight to belong with the
'white horse' of the first class, since being without white horses is not being
without horses. However, a white horse cannot cease to be white as a man
can cease to be a robber; the parallelism requires us to interpret No. 6
as "Being without robbers is not being without the men who are robbers".
Previous studies of the argumentation of Names and objects (including
previous publications of my own) have missed the point that the Mohist is
not proposing forms for inferring from the known to the unknown, but for
describing similar facts similarly and different facts differently. He is not
fumbling in the direction of logically rigorous forms; he has a perfectly
clear idea of logical rigour, but it belongs to another discipline, disputation
proper. Once the point is grasped it can be seen that it was fully understood
by Hsiin-tzu, who rejected "Killing robbers is not killing people" but
placed it correctly among the three types of fallacious proposition dis
tinguished in his Right use of names (cf. § 1/6/1). He does not class it with
Kung-sun Lung's " A white horse is not a horse" among propositions which
"disorder objects by confusion in the use of names" (iSS^ffi^S DMSLJf),
are factually wrong because wrongly deduced from the definitions of names.
To refute these latter you "test them by the convention for the name, and
use what one accepts to show that what one rejects is fallacious" (!&,£, ^5
,
JilK#?:§ №^ft§?i8?), which is precisely what the Mohist understands by
disputation. The propositions among which he classes "Killing robbers is
not killing people" are those which "disorder names by confusion in the
use of names" ( i S K f f i ^ j J ^ S L ^ i ) , which fail as consistent description.
To refute them one "tests them by the purpose of having names and
observes which alternative 'proceeds' " (SfeilBFr [ El ] ,
If we consult Hsiin-tzu's account of the "purpose of having names" we
see that the objection to such descriptions is that they fail to "discriminate
,,
between the same and the different (ISIRIII), SO that "we are sure to
have trouble over the intention not being conveyed'' (rfe ift # % A).
Names and objects makes no new contribution to disputation proper;
it does not take what a Westerner might suppose to be the imminent next
step, the organising of the one-way and two-way implications of comple
mentary names into patterns like the syllogism. (The classification of
parallel propositions of course has nothing to do with syllogistic reasoning;
it is much more like the argumentation of Wittgenstein's Philosophical
investigations and Gilbert Ryle's Concept of mind.) The Mohist seems
assume that the operations of his first two disciplines (deciding how to
relate names to objects and how to act) require formal procedures, but that
in the last two we have only to employ the logical relations between names
and the causal relations between objects to solve particular problems. The
propositions of the sequences on the sciences (B 17-31) and disputation
(B 32-82) are generally answers to specific questions. One might wonder
how the Mohist succeeds in combining the sense of logical rigour exhibited
in many of the solutions of problems (B 73 provides an especially striking
example) with his lack of interest in establishing logical forms. But it
happens that the combination of logical strictness with contempt for logic
as a discipline characterised Western science and philosophy from the
Renaissance to the 19th century. Modern Western rationality has been
nourished by Euclid more than by Aristotle, and in Greece geometrical
proofs preceded syllogisms. T o suppose that the road to rationality inevi
tably passes through the syllogism would be to offend the author of Names
and objects by assuming too close a parallelism between what we know of
the Western tradition and what we expect of the Chinese.
1/1/2/5 Ethics
If we fit the sequences of the Canons to the fourfold classification of
knowledge in A 80, the second corresponds to knowing how to act (level-tone
wei %). This is the one class which is not defined in terms of name and
object. Having fitted the proper combinations of names to the objects with
which we deal, we require in addition to know how to deal with them.
It is not quite clear whether the Mohist thinks of ethical terms as referring
(chii 9) to actions or merely as praising or blaming (cf. A 29-31, 79). He
makes the point that a man's conduct must be 'named neutrally' ( ^ # £ )
before it can be judged (A 10); on the other hand, as we shall see shortly,
he defines his ethical terms by neutral names such as 'love' and 'benefit'.
1/1 45
Since the Mohists were moralists before everything else, it is not

surprising that this is the one branch of knowledge which they fully system-
atised. The ethical theory is in fact beautifully simple, complete and
consistent, an achievement quite without parallel in Chinese philosophy:
Confucianism throughout the whole of its history never attempted, or
indeed wanted, a comparable systématisation. We have its définitions in
A 7-39 and its propositions in Expounding the canons. The latter is badly
mutilated, and so are many of the Explanations of A 7-39, but the only loss
which leaves a perceptible gap in our knowledge is the disappearance of
the definitions of words in the 10 theses of Mo-tzû (cf. § 1/6/2), which
include two words significant in ethics, ai X 'love* and chih 'intent'.
Deciding how to act, like deciding how to describe, requires consistent
procedures, without the strict necessity of the proofs in disputation, but
with the temporary validity of chih i t 'fixing, causing to stay' (cf. A 75
lh0T$K "fix which you prefer"). In justifying the chosen course we appeal
in the case of description to ku #C 'reasons', but in the case of actions to
what they are 'for, on behalf o f (falling-tone wet f l ) . The word wet, used
both adverbially ('for the sake o f ) and verbally ('be for') is the term used
where we should speak of ends, although 'end' is only a partial equivalent
of so wet 'what it is for'. Ultimately the reference seems always to be
to persons ; one chooses this course rather than that for the sake of oneself
( ^ S ) , of others ( ^ À ) , of one's family ( ^ ^ S ) , of the world ( S ^ T ) .
One can choose with the end of earning praise ( ^ # ) , but that is for one's
own sake, and when one's end is to make oneself useful (ISffiB) it is for
the sake of others.
The things one is 'for' are parts within wholes, members within
classes, in the Mohist terminology t'i ft 'units' within the chien É£ 'total'
(cf. A 2). The principle behind the procedures of choice in E C 7-9 seems,
although this is not directly stated in the surviving fragments, to be the
sacrifice of the part for the whole, the finger for the arm, the arm to save
one's life, the man for the sake of the world. As in the four stages of
justifying a description in A 93-B 2 the direction of the procedure is
backward, retracing one's steps from the action to the decisions behind it.
(1) In choosing a course of action one is in the first place 'seeking'
(ch'iu jfc) something beneficial to others or oneself:
E C 7 "Weighing light and heavy among things to be done in practice is
what is meant by 'seeking'."
Whether the man who seeks actually 'gets, succeeds' (te %> A 4,26-28)
depends on whether the chosen course is in fact beneficial. For example
music is commonly supposed to be beneficial, but according to Mohism

is not. " T o suppose that music will benefit one's son and desire it for his
sake is to love one's son. T o suppose that music will benefit one's son and
seek it for his sake is not to benefit one's son" (EC 7).
(2) The next question to arise is whether the benefit to a 'unit'
(part, individual) is beneficial or harmful to the total. We must therefore
'weigh' (ch'uanfil)the relative benefits or harms:
E C 8 "Weighing light and heavy among the things treated as units (so t'i
Wfi) is what is meant by 'weighing'."
Just as in the art of description an object may fit a standard either
'exactly' (cheng IE) or in the appropriate respect, so in the art of considered
choice a preference may be either 'direct, immediate' (cheng) or the result
of weighing (EC 8, A 84). Given the choice between losing an arm fighting
robbers, and losing one's life by refusing to fight, the former alternative is
preferred as the less harmful, the more beneficial (EC 8).
(3) The final question is whether we are benefiting the persons
whom it is our duty to benefit. The benefit of society as a whole depends
on each person performing the duties which are his 'portion' (fen ^);
these require him to give special care to certain categories of persons,
creditors, rulers, superiors, the aged, his elders, his kin. In benefiting
persons we must therefore 'arrange according to grade' (tun lieh ^ M):
E C 9 "Doing more for those for whom duty requires more, less for those for
whom duty requires less, is what is meant by 'arranging according to
grade'."
The procedure for weighing relative benefits and harms in relation to
what one is 'for' may well have originated with the school of Yang Chu,
since it is assumed throughout the individualist chapters of the Lu-shih
ch'un-ch'iu. Mencius too speaks of the weighing of circumstances which
justifies a breach of propriety as ch'iian, asks a king "How can Your
Majesty be choosing for the sake of this?" (HijS^ft'nfe), and says that
of the members of the body "one's love is for the total" (i£№S) and that
since "there are nobler and baser units, major and minor" (fSÛffi,
flf A ' h ) one fosters the major in preference to the minor. A n interesting
67
feature of the scheme is that even if the individualists invented it they have
to stop half way in applying it. The egoist can give a reason for preferring
an arm to a finger, that the choice is for the sake of the whole man. But
there is nothing for the sake of which he prefers himself to others; he can
do so only by an arbitrary choice. If he admits that there can be reasons
67
Mencius 1A/7, 4A/17, 6A/14, 15.
1/1 47
for preferring one man to another the scheme forces him to proceed as
before, choosing between the members or individuals for the sake of the
total, the world. The Mohist does not make this point explicitly in the
texts as we have them (although he may have done so in the mutilated
exposition of E C 10: "That not being for oneself alone is learnable"), but
the whole procedure for offering reasons for conduct implies that action
towards others must be moral if it is rationally justifiable at all.
Disputation, which in the art of description and in the sciences
establishes what is "known 'a priori' " (listen chih), in ethics determines
what "is desired or disliked 'a priori' for the sake of men" (EC 2 5 f e S A
'$> M). This is the most important function of disputation, the one
mentioned first in Names and objects (NO 6 "The purpose of disputation
is, by clarifying the divisions of 'is-this' and 'is-not', to establish the
principles behind order and misrule"). What is desirable 'a priori' follows
of necessity (pi !&) from the ch'ing lit of moral concepts, the essentials laid
down in their definitions:
EC 2 "In the case of all things that the sage desires or dislikes 'a priori'
for the sake of men, men necessarily learn from him by considering their
essentials; but in the case of desires and dislikes born from the conditions
they encounter, they do not necessarily learn from him by considering their
essentials."
If we collate the definitions of 'benefit', 'harm', 'being for', 'bene
volence', 'righteousness', 'filial piety', 'achievement', 'loyalty', which are
scattered over the Canons, we find that all ultimately derive from the
undefined terms 'desire' and 'dislike'. The two basic terms do not them
selves require definition (since the whole purpose of the system is to
establish benefit, harm and the moral concepts as desired or disliked
'a priori'), but they appear among the ambiguous words uses of which are
listed in A 76-87:
A 84 "Yu $C (desire). Directly. Weighing the benefit. T o be about to."
(The last is an irrelevant sense of yu distinguished from the others.)
"Wu № (dislike). Directly. Weighing the harm."
Benefit (It M) and harm (hat 18) are defined in terms of desire and
dislike (although, because of the inconvenience that desire unlike dislike
necessarily precedes achievement, the word used in the former case has
to be hsi I f 'be pleased'). The Mohist said in E C 7 that what we 'seek'
(ch'iu) is not necessarily beneficial; the test of benefit is that we are pleased
when we 'get' (te) it:
A 26 " 'Benefit/ is what one is pleased to get."

A 27 " 'Harm* is what one dislikes getting."
T o be 'for* is defined in terms of desire and the hsiian № ('letting the
scales settle', 'giving the most weight') which is the act of judgment in
chiian 'weighing':
A 75 " T o be 'for' something is to give it the most weight in relation to the
desires, having taken account of all that one knows." (Since wet 'to be for',
to have as an end, is the basic logical term in ethics, this is included among
the terms in disputation defined in A 70-75.)
The moral concepts are derived in turn from the concepts of benefit
and of love. The loss of the definition of love in the missing document on
the 10 theses is not as serious as might be feared, since any reference to
loving men can always be reduced to already defined terms, "desiring
benefit and disliking harm to them, for their own sakes" (Love as an
emotion was not recommended by the Mohists, who distrusted all the
passions. Cf. Mo-tzu ch. 37, Sun 277/8 *t§*lB,*Xifn
ffi t H . "Be sure to get rid of rejoicing, anger, joy, sadness, love, and
employ benevolence and righteousness.")
Love of man (aijen X A ) is the love of persons for their own sakes.
It is most clearly exhibited, on the one hand in self-love, on the other in
the love of Tsang W and Huo 3ff ('Jack' and 'Jill'), the stock examples of
persons too humble to be loved except for themselves alone. Love of self
and of others are strictly parallel; I do not love others if I benefit them
solely for my own sake, for example for the sake of praise or blame (EC 11);
similarly "Jack's love of himself is not for the sake of the men who love
him" (EC 10), and "Love of oneself is not for the sake of making use of
oneself. It is not like loving a horse" (A 7). We cannot love men for their
own sakes without loving all men (chien at ^ $t 'love of the total/of
everyone'). Love of man is a concept unlike, for example, the riding of
horses in that I can be said to ride horses however few I have ridden, but
cannot be said to love men unless I love them all (NO 17). The love of
man, being simply the desire to benefit any man for his own sake, does not
vary quantitatively with the number of its objects; there is no more love
in me when I love all men than when I love one (EC 13. Cf. E C 5). Several
Canons answer objections, based on misunderstandings of the Mohist
position, that one cannot love men without knowing their location, their
number, whether their number is infinite (B 73-75).
Although the love of men for their own sakes is equal, a man may
deserve additional love "for the sake of the man he is" (a phrase found in
E C 11 M R A i f e ) . Since benefit to those most beneficial to others is itself
1/1 49
beneficial to all, we shall desire their benefit more, and therefore love them
more. " T o love Y u B more for the sake of the world is to love for the sake
of the man Y i i was (H R ^ A ) " (EC 5). The sage on his side will love us
less than we love him, because we are less beneficial to the world. "The
great man loves the small man less than he is loved by the small man,
but benefits the small man more than he is benefited by the small man"
(EC 7).
Except for unequal moral worth, no grounds for loving unequally are
recognised. Love is equal for kin and stranger, self and other, men past or
present or not yet born (EC 4, 5, 12). It is laid down explicitly that we owe
no more love to our own parents than to the parents of others (EC 12).
However within society each man has his 'portion' (fen ft, ??), which
prescribes special care (hou №) but not greater love (hou at % S.) for
certain categories of people. Besides kinsmen these include creditors, the
ruler, superiors, the aged, one's elders (EC 9. The list does not include
friends, as it would for Confucians). Here again the Mohist is careful to
show the exact parallelism of the duties owed to oneself and to others.
One of the persons to whom special attention is owed is oneself; and just
as the special care due to parents is independent of their moral conduct
(EC 9), so "giving special attention to oneself is not on account of one's
worth" (EC 10).
The two basic moral virtues, jen t 'benevolence' and yi H 'righteous
ness', are defined in terms of loving and benefiting:
A 7 " T o be 'benevolent' is to love individually."
A 8 " T o be 'righteous' is to benefit."
Filial piety is subordinated to righteousness:
A 13 " T o be 'filial' is to benefit one's parents."
At first sight one is surprised that benevolence should be identified as
love of individuals, not the 'love of everyone' (chien at Ik S ) which is the
characteristic ethical doctrine of early Mohism. But 'love of everyone' is
the principle which requires us to extend the love of one to all, it is not
itself a moral virtue. It is "the love of Jack" and "the love of J i l l " which
has moral value (cf. E C 2). Benevolence as such is not necessarily practically
effective, and how much or little the agent desires to benefit others is
irrelevant to the judgment of his actions (EC 7 "The conduct of the
slightly and of the greatly benevolent are of equal worth"). The love which
effectively benefits is born from pondering the needs of persons, which
are not the same for Jack as for Jill, for the people of yesterday as for the
people of today (EC 2).
Confucians would not have objected to the Mohist's definition of

benevolence, although they did not share his reservations about its value.
But for them conduct is righteous if it is appropriate (yi '£.) to one's
relationships as ruler or minister, father or son, to what the Mohist calls
one's 'portions' (fen). The bald statement that to be righteous is to benefit
is therefore the most challenging of the definitions. It is also the vaguest,
and might seem to invite us to judge a man by the amount of benefit he
actually achieves (which will vary with his opportunities) rather than by
whether he is the kind of man who benefits others. But the Explanation
clears up this point, using the distinction between chih S 'intent' (a word
in the 10 theses of Mo-tzu, and therefore undefined in the Canons) and
kung $J 'achievement' (defined in A 35: " 'Achievement' is benefit to the
people."):
A 8 "(Righteousness). As for his intent, he takes the whole world as
portion; as for ability, he is able to benefit it. He is not necessarily
employed."
Cf. A 13 "(Filial piety). As for his intent, he takes his parents as portion;
as for ability, he is able to benefit them. He does not necessarily succeed."
Thus a man is judged not by his achievement, nor by the mere desire
to benefit (which would be benevolence), but by whether he has both the
intent and the ability. Similarly we are told in Expounding the canons (EC 3)
that " i n the matter of righteousness being beneficial and unrighteousness
harmful, one must distinguish between intent and achievement", and that
"no external condition can make me more beneficial". Consequently
righteousness is independent of opportunity and social position: "One
exalted to the throne of the Empire is not more beneficial to man than any
ordinary fellow".
On the traditional assumption that righteousness is the appropriateness
of actions to certain fixed social relationships it had been plausible to
suppose that righteousness is external to man although benevolence is
internal. This position was attacked by Mencius, for whom both virtues
68
belong to human nature. The Mohist cares nothing for human nature or
the inwardness of man (cf. § 1/1/2/7), but he wants the same status for
both virtues. He deduces a refutation immediately from his definitions:
B 76 "Benevolence is loving, righteousness is benefiting. Loving and
benefiting are on this side, the loved and the benefited on that side."
A Westerner will perhaps expect the edifice to be crowned by a
definition of 'good'. But the Mohist is uninterested in defining a word of
68
Mencius 6A/4.
i/i 51
such generality unless (like 'benefit') it contributes to the structure as a

whole. Oddly enough however Mencius has a definition which fits perfectly
into the Mohist system, "The desirable is what is meant by the 'good* **
69
( " I ^ ^ l l i #). He may even have taken it from the Mohists, much as
he detested them. In a period of vigorous philosophical controversy
Confucians or Taoists can hardly have escaped picking up the less contro
versial formulae of the school most interested in the art of definition.
A remarkable innovation of the later Mohist ethic is that it conceives
morality in terms, not of fixed social relationships between father and son,
ruler and subject, but of individuals benefiting themselves, each other, and
the world (t'ien-hsia X ~ F ) . This may be seen as another aspect of the
Mohist reaction to the challenge of egoism. Yang Chu has forced the
Mohists to pose the central problem of ethics in his own terms: shall I
benefit myself alone, or other individuals as well as myself? We are
expressly warned against loving men only in the abstract (EC 2); the 'world*
is the totality of individuals, each of whom, however humble his social
position, is to be estimated according to his will and capacity to benefit the
rest. It is recognised that men must be allotted for their mutual benefit
special duties to superiors, elders, parents, but these 'portions' (fen) are
merely instrumental to the good of the whole (EC 9 " B y the standard laid
down by the sages you forget your parents when they die, for the sake of
the world. Doing more for parents than for others is your portion, and ends
,,
with the rites of death and farewell. ). We should benefit a man 'for the
sake of the man he is*; and although the degree of love owed to him depends
on his moral worth, even the worthless individual is to be loved simply as
a man (EC 2 "Even if to get rid of the love of them would benefit the world,
we cannot get rid of it."). At the same time there is a tone of intellectual
ruthlessness about the Mohists (with their utterly dispassionate use of the
word 'love*) which warns one against saying too easily that they conceive
men as Kantian 'ends in themselves*. Moral worth is independent of
external conditions (EC 3), but nothing in the system forbids us to sacrifice
an individual for reasons external to himself. The crucial passage on this
question (EC 6) is badly mutilated, but is explicit that " i f a death and a
life are equally beneficial there is nothing to choose between them**. Such
a sacrifice however would be made to benefit others as individuals. The
entire system is based on the desires and dislikes of individuals, and all
morality reduces itself finally to the benefiting of 'Jack and Jill* (Tsang and
Huo, the most plebeian proper names). Although its underlying principles
69
Mencius 7B/25.
are necessary and eternal, and would be valid even if the human race were
extinct, they can be applied only after pondering the actual benefit and
harm of individuals in changing situations (EC 2).
It will be noticed that the Mohist holds both that we should desire to
benefit all men equally and that the benefit of society as a whole depends
on the unequal 'portions' of its members. This position is not inherently
self-contradictory, but there must always have been some tension between
the egalitarian principle and the inegalitarian practice, as in Christian and
liberal democratic theories of society. The doctrine of Hsii Hsing ff ff
in the 4th century, that the ruler should farm with his own hands like his
subjects, and that commodities equal in quantity should be equally priced,
70
was very probably an offshoot of Mohism. Mencius attacks the Mohist
Y i Chih M £ for holding that "Love is without degrees but the application
starts from one's kin", objecting that conduct requires a single underlying
71
principle (—* but Mohism has two. Hsiin-tzu ignores the inegalitarian
side of Mohism, and even credits Mo-tzu with believing like Hsii Hsing
7 2
that the ruler should "do it himself" ( @ f e £ ) . He accuses Mo-tzu of
"confusing degrees, being incapable even of allowing room for distinctions
and differences, or giving their relative weight to ruler and subject", and
says that he "had an eye for the equal but not for the unequal; if we have
73
equality without inequality, the decrees of government will not be applied".
The Mohist is aware of one radical objection to his whole beautiful
system. He must insist that the desirable is the one unquestionable value;
but there were individualists who treated length of life as an end in itself,
and disapproved of both desire and dislike as injurious to health. In B 44,
45 he replies that an ethic of longevity implicitly assumes desire and
dislike. The modern objection to confusing value with the psychological
fact of desire (as in G . E . Moore's criticism of the naturalistic fallacy) is
outside his scope.
7 0
The identification of Hsu Hsing with Hsii Fan jfÊ. disciple of Mo-tzu's
disciple Ch'in Ku-li (proposed by Ch'ien Mu, KSP 4/300-301), is however very
much open to question (cf. Sun Tz'u-chou, KSP 6/189-190). In Mo-tzU a certain
Wu Lu ÎjK preaches that it is one's duty to make one's own pots and plough with
one's own hands; Mo-tzu refutes him with much the same arguments as those used
by Mencius against Hsu Hsing (Mo-tzu ch. 49, Sun 297/9-298/9, Mencius 3A/4).
Mo-tzu however takes the problem more seriously than Mencius, and his answer
is much more sympathetic. He goes to visit Wu LU personally, while Mencius
merely receives a disciple of Hsii Hsing.
71
Mencius 3A/5.
72
Hsiin-tzu ch. 11 (Liang 147/1).
73
Ut sup. ch. 6, 17 (Liang 61/5-10, 231/-2).
1/1 53
1/1/2/6 The sciences

It is at first sight surprising to find in the Mohist summa a sequence
of geometrical definitions (A 52-69) and a parallel sequence of propositions
in optics (B 17-24), mechanics (B 25-29) and economics (B 30, 31), as well
as traces of the loss of an older document on geometrised astronomy
(§ 2/4/2/3). But once we appreciate the design of the Canons it can be seen
that the fourfold classification of knowledge requires that disputation, the
explaining of names, should have as its counterpart a discipline which
explains objects. The sciences form an integral part of the whole scheme:
the alignment of gnomons in astronomy is used not only in a geometrical
definition (A 57) but to illustrate how one can point out a common property
in two objects (B 38); two geometrical terms are applied to propositions in
Names and objects, to designate what seem to be the pause and the phrase-
position (NO 9); illustrations of 'a priori* demonstration are taken from
the geometry of the square and the circle (A 80, 93, 98); several times in
sequences on ethics, change, and disputation one notices other verbal or
material connexions with sections on optics (A 47, 48) and mechanics
(A 21, B52, 62). As we shall see in § 1/1/2/7, the Mohist has no metaphysic,
but he does have a world-picture which unites his four disciplines; it is of
a cosmos of concrete and particular objects, each with its mutually pervasive
properties, located in space and changing through time, interconnected by
necessary relations like the logical relations between their names.
In China as elsewhere astronomy was the first science to develop high
standards of observation and a mathematisation only partially vitiated by
numerology. In other sciences (and in astronomy as well, to the extent that
they succeeded in incorporating it in a unified cosmology), the Chinese
explained natural phenomena by the interactions of active Yang B§ and
passive Y i n , the sequences of the Five Elements, and the numerology
of the Book of Changes, concepts of the same order as the four elements,
four humours and Pythagorean number of the West. Their hypotheses in
medicine, alchemy, geomancy and the other proto-sciences and pseudo-
sciences were of the same untestable kind as prevailed in all civilizations
until the Scientific Revolution in 17th century Europe. But the Mohist
researches are a very remarkable exception, and deserve to be classed with
similar brief episodes in Greece and Mediaeval Europe among the move
ments which in retrospect look to us like abortive efforts in the direction
74
of modern science.
7 4
Needham in his Science and civilization in China, which provides the indis
pensable background for work in Mohist or any other Chinese science, is among the
few Westerners to appreciate the importance of the later Mohists and the first since
The Mohists do not formulate their solutions in physics in mathe

matical terms, nor do they lay down any programme of controlled experi
ment. But they do confine themselves to solutions which are geometrically
visualised and experimentally testable. In one case, the 'criss-cross* {wu
of light converging at the centre of curvature of a concave mirror and
opening out (B 19, 23), they conceive a geometrical figure not visible to the
eye. Instructors in the optical and mechanical Canons would certainly have
had to draw for their pupils the kind of diagram which a modern editor
tries to supply. The crucial respect in which the Mohist falls short of
Archimedes is in his failure to mathematise his conclusions—a difference
which attracts attention especially when we find him expounding the
principle of the lever (B 25b). He has not taken the decisive step from
measuring and calculating in astronomy, to estimate the cycles of the planets
or the height of the sky (cf. § 2/4/2/3), to doing the same in physics without
interest in the particular results, purely in order to formulate a law in its
most rigorous form. Here there is a striking contrast between the science
of the Canons and the technology of the military chapters, which are full
of detailed measurements.
The advantage of the Mohists over the Yin-Yang and Five Elements
schools (which laid the foundations of traditional Chinese science) is that
their logic gives them a clear conception of what they find acceptable as an
adequate explanation. "One uses explanations to bring out reasons"
(NO 10), and ku #t 'reasons* are of two kinds, the necessary conditions
without which something "necessarily will not be so**, and the necessary
and sufficient conditions having which it "necessarily will be so** (A 1).
In the sciences a ku is a cause, for example a wound or dampness as the
cause of illness; a wound is "why the thing is so** (B 9 ^ £ § r 2 l $ t ) ;
"dampness is a cause: it is necessarily required that what it does comes
about'* (A 77). The necessity of causal relations accounts for the placing
of the sequences on the sciences with those on disputation after the bridging
sequences on knowledge and change, the definitions of which end with pi
'necessary* (A 51). The word pi appears twice in the scientific Canons
(B 24, 29) and six times in their Explanations (B 22-25). In all but two of
Forke to risk a translation of the scientific sections (Needham vol. 4/1, 17-27,
81-87). But I think that when we find answers to the preliminary questions (the
textual and syntactic problems, the organisation of the summa as a whole), it turns
out that we can form a more coherent picture of Mohist science than Needham's,
and make bolder claims for it; the Mohists were the first and perhaps the last
thinkers in traditional China to break out of the conceptual framework of Mediaeval
science.
1/1 55
the latter the clause with pi is followed by an explanatory clause ending in

yeh 'It's that. . .', 'It is because . . . ' :
B 23 . . . iffi&jE · te^t ' l*IEffn*XiStil ·
" . . . and it (the shadow in the concave mirror) is necessarily upright. It is
because the light opens out from the centre, skirts the upright object and
prolongs its straight course." (Also later in B 23 and twice in B 25b.)
In the other two cases the next clause has ku 'therefore' (which is ku
'reason/cause' used as an initial particle):
B 22 . . . ' ttl^jK °
" . . . and necessarily recede beyond the mirror plane, therefore they share
the same place." (Also B 24.)
The Mohist never mentions the Y i n and Yang or the system of the
Changes,™ but he once criticises the theory that the Five Elements follow
a regular sequence:
B 43 "The Five Elements do not have constant ascendancies. Explained
by: whichever is appropriate."
The Explanation notes that the element fire does not invariably
conquer metal; it does so only if there is sufficient fuel. The word yi !iC
'appropriate' in the summing-up of the Canon refers to the classification
of three kinds of relation (ho ' o ) in A 83, the exact, the appropriate and
the necessary. The Mohist's fundamental objection to the Five Elements
type of explanatory principle is that it lacks the necessity he finds in causal
explanation. But he does not try to replace it by anything resembling the
strictly testable laws of nature in modern science. He looks for causal
relations only in specific phenomena, just as in the final sequence, of
propositions in disputation (B 32-82), he explores logical relations in
specific problems, without looking for syllogistic or other logical forms.
(It is human operations, consistent description and considered choice,
which he sees as requiring codification.) His problems arise from the
manipulation of mirrors, balances, ladders or masonry. Why is the image
in a concave mirror upside down ? And why only if the object reflected is
outside the centre of curvature ? (B 23). The problems, although suggested
by practical situations, are purely theoretical; one point not mentioned is
the practical function of the concavity, to make a burning-mirror. General
principles emerge only incidentally from the specific explanations. The
explanation of the inversion of the image implies two, that light travels on
7 6
There are two references to Yin and Yang as influences behind the weather
in the older parts of Mo-tzu (ch. 6, 27, Sun 22/8, 9: 129/8). The Five Elements
are mentioned only in a quotation from the Book of documents (ch. 31, Sun 154/-4).
a straight line and enters a curve along its axes; but if he presents them
explicitly it is by combining them in a simile, in the controversial sentence
which we translate "The light's entry into the curve is like the shooting of
arrows from a bow" (B 19). The most carefully formulated principle is in
B 27, introduced by the generalising fan J\x 'In all cases': "Whenever a
weight is not pulled up from above or received from below or forced from
the side, it descends vertically". This underlies all the causal explanations
in terms of 'pulling sideways', 'being vertical', 'supporting from below' and
'pulling from above' in B 26-29. In B 29 the Mohist discerns that the chu &
'supporting from below' by which he is explaining why a wall does not fall
is the same as the shou 4fe 'receiving from below' which arrests the vertical
descent of a weight: "Without any alteration except the substitution of a
name, it is shou."
Although causal explanations are necessary, causes are unknowable if
circumstances coincide (yu S ) , so that unlike disputation the sciences
admit of doubt (B 10). The point is illustrated by the stock example of a
happening with multiple causes, sickness (§ 1/5/11); a fighter's collapse
may be ascribed either to drunkenness or to the heat of the sun. We can
now understand why the Mohist, in spite of many incidental references to
the causes and cure of illness (A 76, 77, 85; B 9,10,45), selects his problems
not from established disciplines such as the ve lerable Chinese science of
medicine but from optics and mechanics, which never had the status of
organised sciences at all. The explaining of objects which parallels the
explaining of names in disputation requires phenomena with causes which
are easily isolated and clearly demonstrable.
We have noticed that in each of the disciplines the 'a priori' element
which belongs to disputation is contained in the definitions of A 1-75
(§ 1/1/2/2). In the third discipline this is especially obvious, since the terms
defined are not optical or mechanical but geometrical (A 52-69). The circle
is the Mohist's favourite example of something known 'a priori' (hsien
chih) ; "by the things which follow from each other or exclude each other
we may know 'a priori' what it is" (A 93). Of a thing perfectly matching
its standard we are told that "when there is explanation, you assent to more
than that they match (for example, to a circle being nowhere straight)"
(A 98). 'Explanation' is one of the three sources of knowledge distinguished
in A 80, where the example is "Something square will not rotate" ('Square'
was defined side by side with 'circle' in A 59). It is remarkable that the
Mohist seems to have the idea of geometrical proof, for the absence of strict
proofs in geometry is perhaps the most obvious weakness in Chinese as
compared with Greek science. One would dearly like to see that lost
1/1 57
document on geometry and astronomy which we postulate; without it one

cannot know whether he appreciated the value of trying to fill in all the
intermediate steps between the definition of the circle and the conclusion
that it is nowhere straight. One of the problems in disputation (B 62)
suggests that he did not; to show the possibility of a figure which cannot
be made to leave the perpendicular he is content to say "because it is
spherical".
We saw in § 1/1/2/5 that the moral concepts are established as desired
,
'a priori by a system of definitions built up from the undefined term
'desire*. The circle is proved to be "known 'a priori* ** by the same method,
by deriving it from the undefined terms jo Jff 'like* and jan 'so*, on which
as we saw in § 1/1/2/4/1 the whole of the nominalist theory of knowledge
by explanation is based. (/aw/*NIAN does not require to be defined in
#
terms of / o / N I A K since etymologically it is * N J A K + - N 'like it*, as
yu/*lAN is yul*\0 W: + - N 'in it*, and throughout Chinese history
76
has been defined by phrases equivalent to 'like this*.) In § 1/1/2/4/1 we
noticed that standard and criterion are defined in terms of them:
A 70 "The 'standard* is that in being like which something is so.**
A 71 "The 'criterion* is where it is so.**
The same is the case with the quantifiers, which are defined by adding
negatives before jan:
A 43 " 'Exhausting* (chin ffe, adverbially 'all*) is of none not so.**
NO 5 " 'Some* is not all.**
A measurement which is 'so', like its standard, is equal to it, and
hsiang jo ffi T£ 'like each other* is in fact the Mohist*s ordinary term for
'equal* (A 54, B 25b, 26, E C Appendix 3, 5, 6, 7, 13). He sees that to
proceed to the definition of the circle he has only to add a definition of
'straight* (chih i t , of a path, but the more general cheng JE is used, for
77
example, of a straight edge). Like the Greeks he is unable to define it
except in terms of visual alignment:
A 57 " 'Straight* (chih) is in alignment**.
The rest follow on directly:
A 53 " 'The same in length* is exhausting each other when laid straight
(cheng)."
A 54 "The 'centre* is <the place from which(?)> they are the same in
length.** (The Explanation resumes the stages from jo 'like*: "Outward
from this they are equal/like each other*'.)
7 6
#nitfc» &fl;Jik- See the examples assembled in Juan, 230.
7 7
Cf. A 57, comment.
A 58 " 'Circular* is having the same lengths from a single centre."

Except for the concealed dependence of the second definition on visual
experience, all are derived from purely logical concepts; in principle we
could learn from them what a circle is without any empirical experience,
just as we could know that benevolence and righteousness are to be "desired
'a priori* for the sake of men" from the ethical definitions, "even if there
were no men at all in the world*'.
The sophist H u i Shih had raised the problems of the point (wu hou
M№ 'the dimensionless*) and the infinite (wu ch'iung 'the limitless*),
which are the smallest and the greatest units countable as one (hsiao yi
-
'h— ta yi A * ) . These concepts involve such paradoxes as that points
both do and do not accumulate (cf. A 55 n.) and that space is both finite
and infinite. (Although we do not have H u i Shih*s explanations of his
paradoxes, both would follow from the assumption that finite quantities
are multiples of the 'smallest one* and divisions of the 'greatest one*.)
Some of the Mohist's definitions seem to be designed to avoid these
paradoxes. He does not treat the infinite as countable as one; for him, to
speak of infinity is simply to say that there is always room for another
measurement (A 42). "That which being demarcated cannot be referred
to without referring to that from which it is demarcated is 'space* ", so
that "advancing in space we never get nearer** (B 63). Divisions of a
measured length always allow further divisions (B 60), so that we never
arrive at any points which are not the starting-points of divisions. Through
out the Canons and Explanations the Mohist recognises no points which are
not the starting-points (tuan 388) of measurements (A 61), no moments
(wu chiu MfK 'durationless*) which are not the commencements (shih #n)
of periods (A 44). It is measured lengths, not points, which can be 'doubled*
(pei № A 60).
The Mohist*s analysis does not escape another paradox of Hui Shih,
that "the sun is simultaneously at noon and declining, a thing simultane
ously lives and dies** (Chuang-tzii ch. 33, Kuo 1102/3). If two measure
ments can share one starting-point (A 60, 67), it would seem by analogy
that successive periods will share one moment, at which what is ceasing
and what is beginning will be simultaneous. In certain rather obscure
passages the Mohist does appear to say that at the moment of death a thing
is both 'horse* and 'non-horse* (A 50), a man is both alive and dead (A 88),
and also that at the moment of commencement something is both about to
be and 'just now so* (fang jan J5$t A 33).
t
1/1 59
1/1/2/7 The absence of a Mohist metaphysic

In most early Chinese thought (Confucianism, early Mohism, Taoism,
even Legalism) ethics is rooted in metaphysics, in the idea of a Heaven
which man should obey or a Way with which he should accord. We have
noticed the metaphysical crisis in the 4th century B . C . , provoked by the
realisation that morality conflicts with human nature (§ 1/1/1/2). Since man
owes his nature to Heaven it seemed to follow that the wicked can justify
their own natural inclinations as the will of Heaven, a difficulty admitted
at the beginning of the first of the later Mohist documents, Expounding the
canons (EC 1). The Mohists reacted to this problem by deriving a ration
alised ethic from the actual desires and dislikes of men, putting aside the
problem of its relation to the will of Heaven. We know from casual references
in illustrations to arguments that they still paid at least lip-service to the
power of Heaven and the spirits (EC 1, 7, 11, N O 18), and the lost docu
ment on words in the ten theses of Mo-tzu (§ 1/6/2) must have offered
definitions of t'ien X 'Heaven', kuei % 'spirits' and ming # 'destiny'. But
there is no evidence that they tried to organise these concepts in a new
metaphysic. They show great interest in the objective study of the natural
world, in optics, mechanics and probably astronomy, but none in the
mystery of man's relation to this world.
Although the metaphysical crisis of the 4th century affected all
philosophical schools, the Mohist school is unique in simply ignoring
metaphysics. Even the Legalist Han Fei-tzu makes at least a formal
profession of deriving his amoral principles from a Taoist conception of
man's place in the world. One is again reminded that the Mohists were men
of a kind very unusual in traditional China although familiar in Western
civilization for the last two or three centuries—men who care only for the
moral, the useful and the rational, and exhibit no sign of having any inner
life. They never use, except in non-philosophical senses, such words as
tao M ('the Way') or te W (the personal quality of the man who has the
Way). Not only are they silent about the problem of hsing tt. 'human
nature', they define sheng ^£ 'life' (A 22) in a way which seems to deny
that there is such a thing as human nature (a solution which abolishes the
conflict between morality and Heaven's will and puts aside the problem of
equating them).
A n especially remarkable omission is the word hsin & 'heart, mind',
a word which (since the heart was regarded as the seat of thought) was used
rather as we use 'brains' in English. This is very important in the thought
of Mencius, Chuang-tzii, Hsun-tzu, and some of the writers of Kuan-tzu;
but although there are a few references to it in Names and objects (NO 9,
15, 16) and elsewhere in Mo-tzu, the Canons and Explanations seem
deliberately to avoid it in favour of chih #P 'the intelligence', defined in
A 3 as the ts'ai 'resources, capability' of knowing, and compared to
eyesight which is the capability of seeing. The living man is described as
having intelligence as he has configuration, the two at least partially
pervading each other (A 22).
The few psychological terms employed by the Mohist all confirm his
wholly extroverted view of life. If we ask what he conceives as being
present in man even when not exhibited externally, the answer is chih y
the 'intelligence' by which we know (present throughout life although

latent in sleep, A 23); chih the 'intent' to act, and neng fs, the 'ability'
to act (both latent in adverse circumstances, A 8, 13); and ch'i IR 'energy,
zeal', which like the intent is liable to be disguised (A 11). We know by
means of the intelligence as we see by means of the eyesight (not the eye.
A 3, 5). It is not altogether clear whether the intelligence is conceived
merely as one of several capabilities (the gift of reason) or as embracing the
others, in which case it would approximate to a concept of consciousness
or mind fully abstracted from the physical organ, the heart. We use the
intelligence to think, but it does not itself think (A 4); on the other hand
it is once implied that it can have desires and dislikes (A 25). Hsiin-tzu
gives 'intelligence' and 'ability' exactly parallel definitions (quoted A 3-6 n.),
both agreeing with Mohist usage, which favours the former of the two
interpretations. What is quite clear is that the Mohist is uninterested in
anything inside a man which does not express itself in action as a social
being.
It was the breakdown of the authority of Heaven which started the
Mohists on the great enterprise of building a fourfold system of knowledge.
But the summa docs contain one reference to Heaven which is not peri
pheral. This is the claim in B 41 that "among Heaven's constants its
presence is prolonged with man", which we connect with E C 2, "Even if
there were no men in the world what our master Mo-tzu said would still
be present". The Will of Heaven is never mentioned after E C 1, but here
the eternal norms which disputation puts in its place are called Vien ch'ang
'the constants of Heaven'. In any other text we might ignore the
phrase as formulaic, but in the Mohist summa, with its peculiarly terse and
compressed style, words in combination regularly keep their full weight
(§ 1/3/14). We cannot base too much on a single reference, but it may be
that the Mohist saw himself as discovering Heaven's design for the cosmos
and for man in the causes of physical events and the implications of ethical
concepts.
1/1 61
1/1/2/8 Relations with other schools

The later Mohist summa never mentions rival philosophers by name,
but the final series on disputation refutes a number of positions held by
Confucians (B 34, 36), Taoists (B 44, 49, 69, 77, 81), sophists (B 47, 54),
Kao-tzu (B 76), the Five Elements school (B 43). No less than seven
Canons defend disputation against anti-rationalist theses to be found in
Chuang-tzu (B 35, 48, 68, 71, 72, 79, 82). The ethical definitions contain
a series on sheng ^ 'life* (A 22-28), apparently intended to clarify the
Mohist attitude to the 'nurture of life* taught by the individualists (who
are the target also in E C 1, B 44, 45). There are no references to the
Legalist doctrines of the 3rd century, although we should certainly expect
the Mohists to defend their ethics against the ruthless amorality of Shang-
tzu #5·? and Han Fei-tzii if they knew of them. Presumably Legalism
either did not yet exist or was an arcanum of government, reserved for
those whose political interests it served.
The loss of almost all the writings of the sophists makes it impossible
to judge how much the Mohists owed to the founders of disputation in the
4th century. Chuang-tzu ch. 33 lists two series of sophisms, the first
ascribed to H u i Shih, the second to the sophists in general. The Mohists
mention half-a-dozen sophisms of the second list, and seem to know some
of the others; they accept " A shadow does not shift" (B 17), "The eye does
not see" (B 46) and a variant of " A stick one foot long, if every day you
take away half of it, will not be exhausted for a myriad generations" (B 60),
but reject " A whelp is not a dog" (B 54), "Fire is not hot" (B 47) and a
variant of " A white dog is black" (NO 18). Except for a casual reference
to "The South has no limit" (B 73) they ignore H u i Shih's paradoxes,
although some of the geometrical definitions may be designed to avoid
them (cf. § 1/1/2/6), and B 63, 64 attack some kind of spatial and temporal
paradox. As for Kung-sun Lung, we find only one incidental reference to
"rejecting the 'white horse* " (EC 1); Names and objects includes ' A white
horse is a horse* * among its examples as though no one had ever questioned
it (NO 14). The Kung-sun Lung tzu has no parallels with the Canons and
Explanations except in the three forged chapters (ch. 4-6), which are
decorated throughout with grossly misunderstood borrowings from the
78
Mohists.
However the paradoxes of the sophists, recorded without their
explanations except in the 'White horse* and 'Pointing things out* of the
7 8
For evidence that parallel passages in Kung-sun Lung tzu are borrowed and
misunderstood, cf. G(2) 156-164.
Kung-sun Lung tzu, are much less important than their methods of argu
ment, of which we know very little. How much of the apparatus of Mohist
disputation is inherited from H u i Shih and Kung-sun Lung ? One possible
approach is to search the summa for key terms which are undefined,
concepts which are taken for granted. For the historian of Chinese
philosophy two of the most remarkable items in the terminology are hsien
'a priori' and pi 'necessary', neither of which as far as I have observed have
the same strict logical application anywhere else in the philosophical
literature. Since the Mohists pay close attention to necessity and define pi
at a crucial place in the organisation of the summa (A 51), it is very probably
their own discovery. But they never define hsien, which reveals its signi
ficance only when we collate the examples, note the contrast of hsien chih
"know 'a priori' " and wet k'o chih "not knowable 'a priori' ", and connect
both with various neighbouring references to walls (cf. § 1/4/13, 1/5/8).
We may guess that the practice of deducing what can be known about an
object hidden behind a wall from the implications of the definition of its
name was already a commonplace of disputation. Probably it goes back to
Hui Shih, who is described in Chuang-tzii ch. 33 as "tabulating the ideas
of things" (M % A). Yi 'idea' is also undefined, and would belong to the
same constellation of terms; so would the undefined ch'ing (the 'essentials'
as formulated in the definition) and mao (the perceptible characteristics),
both of which we shall notice later in a dialogue between H u i Shih and
Chuang-tzu (§ 1/4/6). If we run through the rest of the terms analysed in
§ 1 / 4 looking for items undefined in the Canons we find that nearly all
belong to the common vocabulary of pre-Han philosophy.
T o what extent did the Mohist logic affect the thinking of other
schools ? Its influence can certainly be seen in the writings of the followers
of Chuang-tzu, although there is some difficulty in distinguishing it from
the influence of H u i Shih on Chuang-tzu himself. The Keng-sang Ch'u
iPtH/§ chapter (some of which may be as late as the 2nd century B . C . ) con
tains an interesting attempt to organise the basic concepts of Taoism in a
string of definitions like those from which the Mohists derived their 'a
priori' concepts in ethics and geometry. The items on knowing have verbal
parallels with the Explanation of A 4 and the Canon of A 5; and it is
especially instructive since unlike the Mohist, whose definitions have to be
reassembled from their places in the summa, the Taoist lists them in a
single series and says explicitly what he is doing ("The names are opposed
but the objects take their courses from each other"). His view of life is of
course radically different from the Mohist's. He believes that one should
abandon conscious choice, perceive things as they are without discursive
1/1 63
thought, and respond to them as spontaneously as a shadow or an echo.

The concept at the basis of his system is, therefore, not the 'desire* from
which the Mohist starts, but pu te yi ^Ftlf B 'having no choice*. He uses
it to define the te the sage's unreflecting grasp of the way to deal with
things and people.
Chuang-tzu ch. 23 (Kuo 810/4-6)
>« £ & O i M ° ' ° • £±W-fe ° > IB
# ± s f ^ » i » M i l ! * o u ^ t f B ± S B r s j ° » « f # a ^ i B r © J ° ^ffi
Riffi mare*<>
"The 'Way' is the layout of the te. 'Life* is the radiance of the te. Some
thing's 'nature* is its resources for living. What is prompted by its
nature is called 'doing'. Doing which is contrived is called 'misdoing'.
To 'know' is to be in touch with something, 'knowledge' is a representa
tion of it. As for what knowledge does not know, it is as when we are peer
ing in one direction.
That which prompts on the course which is inevitable is what is meant by
'te'. The promptings being from nowhere but oneself is what is meant by
'ordered'. The names are opposed, but the objects take their courses from
each other." (Of these the Mohist defines 'living', 'knowing', 'knowledge',
'ordering', A 22, 5, 6, 28.)
The Confucian Hsiin-tzu's Right use of names begins with a much
longer series of interrelated definitions starting from sheng ^ 'living'. This
too has parallels with the Canons on knowing, which will be considered in
the commentary on A 3-6. (A point in common between both schemes and
the Mohist's is that all three fail to connect the definitions of knowing with
the rest.) The Right use of names is indeed so closely related to the summa
and to Names and objects that one is strongly tempted to see it as a digest
of the techniques of Mohist disputation adapted to Confucian purposes.
Hsun-tzu may of course share with the Mohists common sources in the lost
literature of the School of Names. But the overall arrangement of his essay
seems to follow the Mohist fourfold classification of knowledge (§ 1/6/1).
Nearly all its significant ideas are found in the Mohist dialectical chapters,
the nominalist theory of common names (A 78), the differentiation of wider
and narrower kinds of thing (B 2), the weighing of desires (EC 8, A 75),
the distinction between knowledge and perception and between chih £fl 'the
intelligence' and chih I? 'knowing' (A 3-6), even the idea of the tz'u i $
'proposition' which first appears in Names and objects. It quotes as a soph
ism "Killing robbers is not killing people" (NO 15). Hsun-tzu presents
his summary of the techniques of disputation as the essential minimum
without which a Confucian cannot compete in these troubled times when

people are no longer content to follow the teachings of the sages. He is not
likely to have contributed much to it himself; he thinks there is too much
disputation in the world already. Although he is not the man to admit a
debt to the enemies of Confucianism, we can hardly doubt that he owes
most of his observations about names and objects to thinkers such as the
Mohists who were interested in disputation for its own sake. It is especially
unlikely that he discovered the proposition for himself, if as we have argued it
was a very difficult discovery to make in the Chinese language (§1/1/2/1/1).
It is notable that he makes no specific reference to the concept of logical
necessity; probably he did not see its importance. Hsiin-tzu is a strong
and clear-headed reasoner, but like other Confucians and Legalists he
thinks it frivolous to be too much interested in logical puzzles.
1/1/2/9 The loss and rediscovery of the Mohist logic

The later Mohist summa is a masterpiece with an unlucky fate. In the
3rd century B . C . it must have served as a powerful weapon in debate,
enforcing new standards of logical rigour on all rival schools which deigned
to compete with the Mohists. We have noticed evidence of this in the
Keng-sang Ch'u chapter of Chuang-tzU and in Hsun-tzu, who in spite of
his limited commitment to disputation is the most rationalistic of all the
Confucians (§ 1/1/2/8). But the centuries of political division and social
change in which philosophy flourished were almost over. In 221 B . C . the
Ch'in dynasty completed the reunification of China and suppressed the
philosophical schools, quickly to be succeeded by the stabilised order of
the Han 91 (206 B . C . - A . D . 220) with its Confucian ideology. We know
nothing about the end of the Mohist school except that after 221 B . C . we
79
no longer hear about it. The simplest explanation of why the Mohists
79
During the first century of the Han, before the victory of Confucianism
under Wu-ti ( 1 4 0 - 8 7 B . C . ) , some of the pre-Han schools temporarily revived;
the predominant tendency was a Taoist-Legalist syncretism which freely absorbed
elements from Confucianism, Mohism, the sophists and the Yin-Yang school. Since
Taoists and Legalists had always seen the two moralistic schools as barely distinguish
able, the vague references in 2nd century sources to moralists as Ju-Mo fflH§
'Confucians and/or Mohists* do not imply the revival of the Mohists as an organised
school (Fukui 3-5). But Fukui f g ^ M J S argues persuasively that the common
adversity of the two moralistic schools threw them together, that the Mohist
remnant was absorbed into Confucianism and influenced it. His most striking
evidence is a Mohist-sounding definition of jen 'benevolence' in a memorial
o n e
presented in 134 B . C . by Kung-sun Hung ^J^*}A> °f the leading Confucians
elevated by Wu-ti: "To promote benefit and get rid of harm, and love everyone
without selfishness, is called 'benevolence' " (Han shu 9tHf ch. 58, 2616/-5 gCÎ]
i/i 65
disappeared although the Confucians survived is sociological; craftsmen

and traders could aspire to influence forms of government in the small and
unstable states of the 5th and 4th centuries B . C . , they could no longer hope
80
to do so in the reunited Empire.
Readers of the summa must always have depended on a teacher to
explain its peculiar terminology and the principles of its arrangement as a
manual. When the tradition of interpretation lapsed with the extinction of
the Mohist school it must have impressed readers as a very difficult book.
During the last century B . C . it suffered a bibliographical disaster which
effectively eliminated it from the usable literature of China for 2,000 years.
A complete text of Mo-tzu was assembled for the Han Imperial Library,
81
consisting according to the Han bibliography of 71 p'ien M. Since this
is the same number as in the extant text it would seem that the Canons
(p'ien 40, 41) were already grouped separately from the Explanations
(p'ien 42, 43), and Expounding the canons and half of Names and obj
already reduced to the fragments collected as the 'Bigger pick' (p'ien 44).
Assuming that, as in the text as we now have it, the five divisions of the
definitions and the propositions were unmarked, and that both Canons and
Explanations were written consecutively, with the head characters which
distinguish the Explanations incorporated into the text, a Han scholar must
have been in much the same position as a modern reader who looks at a
plain text of Mo-tzii ch. 40-45 without the aid of a critical edition; except
for the 'Smaller pick' (p'ien 45) it would be only intermittently intelligible.
|&TS» 5ffe36M?A» I I / i t - quoted Fukui 7). However Kung-sun Hung may be

echoing Hsiin-tzu's "benefit universally and love everyone" (Hsiin-tzu ch. 25 (Liang
347/-4) ©fljjjfcSE quoted Fukui 16), itself Mohist in inspiration. It did not require
a rapprochement between the schools for Confucians to borrow from Mohists; their
opponent's concepts were so much better organised that it would be impossible to
debate with them without being influenced by them (much as Neo-Confucians
could not avoid the influence of Buddhism).
80
Wang Ming 3 E ^ (op- cit.) has pointed out some affinities between Mohism
and the next popular ideology to emerge in China, the revolutionary Taoism of the
2nd century A.D. reflected in the T'ai-p'ing ching ^C^pfS- But hefindsno evidence
of direct influence of Mo-tzu on the T'ai-p'ing ching. Mo-tzu himself did, perhaps
at this period, win a modest place among the Taoist immortals; by A.D. 300 at latest
alchemical treatises were being forged in his name (cf. Sun 469-471). No doubt
we owe to this canonisation the preservation of the book in the Taoist collections
which became the Taoist Patrology; that includes a variety of other unorthodox
philosophers such as Han Fei tzu and Kung-sun Lung tzu but none perhaps which
t
is as blatantly un-Taoist as Mo-tzu.

81
Han shu (ch. 30), 1738/5. The copy in the Imperial library of the Sui
dynasty (A.D. 589-617) still contained the report of its editor, Liu Hsiang or his son
Liu Hsin SyHt (died A.D. 23) (Sui shu Pai-na "gffft edition, 34, 6A/-2).
Since the Chinese civilization which stabilised during the Han has
never been much interested in logic, it might be supposed that the Mohist
summa would have left no lasting mark in any case. But as a matter of fact,
in spite of the desperate condition of the text, it attracted a surprising
amount of attention as long as the complete Mo-tzu continued to circulate.
When independent philosophising revived in the 3rd and 4th centuries A . D .
the Neo-Taoists were fascinated by the little that survived of ancient
82
disputation. Commentators on Chuang-tzu, from Ssu-ma Piao
(died A . D . 306) to Ch'eng Hsiian-ying J $ £ 3 S (fl. 631-650), quote the
83
Canons to explain the sophisms of ch. 33. Three of the four passages
from Mo-tzu in Lieh-tzu (c. A . D . 300) are taken from the Canons and
8
Explanations. * Although most of the Mohist terminology would have been
unintelligible, the Neo-Taoists seem to have noticed the recurrent phrases
p'ien ch'i MM> p'ien ch'il p'ien chii ; Wang Pi EEK (A.D. 226-249)
uses the phrase ^\*IM9^ "cannot be referred to without the other" (B 63)
in commenting on Lao-tzu 2, and p'ien chii is also used by Ssu-ma Piao in
85
explaining a sophism.
About A . D . 300 a certain L u Sheng H # wrote a commentary on the
Canons, which he supposed to be the work of Mo-tzu himself. The book
was soon lost but the preface survives in his biography in the Chin
86
History:
"Names are the means of arranging the similar and the different and
clarifying the right and the wrong; they are the gate to the Way and to
righteousness, and the waterlevel and carpenter's line of perfect govern
ment. Confucius said 'Surely, the right use of names! If names are not
right nothing is successfully performed'. Mo-tzu in the book that he wrote
instituted canons of disputation in order to establish the basis of naming.
H u i Shih and Kung-sun Lung continued his tradition and became famous
8 2
For Neo-Taoist references to Hui Shih and Kung-sun Lung, cf. Hou ^
2 5 0
Chao and T u £fcfflj?, / » 51. Hsu Fu-kuan 63, 64. Among
the pseudonymous writings of the period, Yin Wen tzu ^$1^ (c. A.D. 200) begins
with a discussion of names and objects, Lieh tzu (c. A.D. 300) introduces a disciple
of Kung-sun Lung defending a string of sophistries (ch. 4, Yang 86/2-89/7), and the
present Kung-sun Lung tzu is itself a forgery of between 300 and 600. Even the
Confucian K'ung ts'ung-tzu contains refutations of Kung-sun Lung's 'A white
horse is not a horse' and 'Jack has three ears' (SPTK, A 72A-76B).
8 3
The direct quotations are listed on p. 76 below.
8 4
Cf. the list on p. 76 below.
85 1 1 0 8 n 9
Chuang-tzu ch. 33 (Kuo IRKtit - >«
86
Chin shu H H (Pai-na 94, 6A/6-6B/3). The book does not appear in the Sui
or any subsequent bibliography.
67
in the world by correcting forms and names. Mencius condemned Mo-tzu,

but in his argumentation and demonstrations is the same as the Mohists.
Hsun-tzu and Chuang-tzu both reviled the School of Names, but could not
find a substitute for their methods. . . .
From Teng Hsi down to the time of Ch'in, from generation to genera
tion there were writings of the School of Names, of which little can be
known; none of the later scholars continued their transmission and study.
By now, more than 300 years later, they have completely disappeared. O f
the Mohist disputation we have the upper and lower Canons, with an
Explanation for each Canon, four p'ien in all. Since they are incorporated
with the rest of the p'ien of the book, they alone have survived.
Now I have extracted the Explanations and attached them to the
Canons, each in its own section, leaving doubtful passages without com
ment, and have also composed a 'Forms and names' in two p'ien (JPJ^Zl^)
selected from miscellaneous collections. I have given a rough explanation
of their general drift, and hope that some gentleman who elucidates
obscurities and restores lost traditions will take an interest in them."
As a modern reader of the Canons knows only too well, the task of
fitting the Explanations to the Canons is only a preliminary to understand
ing them. It is unlikely that L u Sheng recognised the head characters
which connect the Explanations with their Canons, since in all passages
quoted in Lieh-tzii and Kung-sun Lung tzu the head characters are treated
87
as part of the text. The forger of Kung-sun Lung tzu, who modelled his
nonsensical 'Ox and sheep* argument on the Explanations of B 66, 67, must
88
have read them as a consecutive passage. Very probably he was reading
them in L u Sheng's arrangement (he was certainly later, since L u Sheng
knows nothing of a book ascribed to Kung-sun Lung); this would explain
why he borrows nothing from the 'Bigger pick* and the 'Smaller pick',
which L u Sheng did not include. He may also have found his authentic
essays of Kung-sun Lung (the 'White horse* and 'Pointing things out') in
89
the two p'ien of 'Forms and names'.
87
Lieh-tzii ch. 5 (Yang 107/3-5), quoted under B 52 below; Kung-sun Lung tzu
ch. 5 (Ch'en 167, 177, 188), quoted p. 176 below.
88
Kung-sun Lung tzu ch. 4 (Ch'en 126-139). The relation to B 66, 67 is
discussed in G(2) 161-164.
89
This speculation is supported by a probable reference to Lu Sheng's book
in Pao-p'u-tzu (c. A.D. 300), B ch. 42 (WYWK 743/9, 10). Writers of
irresponsible verbiage are described as "having what resembles the book(s) about
hard and white and length and breadth, and Kung-sun's essays about forms and
T h i s l o o k s l i k e a d i r e c t
names" ( ^ © g & ( 4 ! r f № J f l I £ )· reference
to the last phrase in B 4 (JfK*H8Ill=3 )· The phrase is imitated in Kung-sun Lung tzu
At some time before the end of the Sui Pf dynasty (A.D. 589-618) the
dialectical chapters suffered a second and almost fatal bibliographical
disaster. A text of Mo-tzu consisting of only the first thiee chuan &
(ch. 1-13) appeared and gradually drove the complete text out of circula
tion. This version is first mentioned explicitly by bibliographers of the
Sung dynasty, some of whom describe it as having a commentary by Yiieh
9 0 1
T ' a i ^ S , or as combined in one volume with the Ho-kuan-tzii^M^.*
But since Yiieh T'ai appears in the Sui and T'ang bibliographies, as the
92
author of a commentary on Kuei-ku-tzu J&£r*?, it cannot be later than
the beginning of the 7th century. That this was the version commonly
available during the T'ang )S dynasty (A.D. 618-907) is confirmed by the
essay of Han Y u ft* (A.D. 768-824) " O n reading Mo-tzu", which is so
favourable to the book that the great Confucian cannot possibly have read
more than the relatively innocuous first 13 chapters. Moreover his only
other essay on a heretical philosopher is his " O n reading Ho-kuan-tzu";
evidently he was using the book which combined the abbreviated Mo-tzu
9Z
with Ho-kuan-tzu.
The shortened Mo-tzu survives only in the unpublished Y i i № man
uscript, now in the National Peking Library, collated by W u Yu-chiang
94
and described in detail by Luan T'iao-fu *f lilff. It lacks the
commentary of Yiieh T*'ai, but has one textual note which may derive
95
from it; Luan infers that the rest of the commentary has been excised.
Since there is no obvious purpose in circulating only the first 13 chapters
of Mo-tzu we may guess that Yiieh T'ai happened to abandon his com
mentary at this point, and that most readers lacked the stamina to grapple
ch. 5 (Ch'en 177, quoted B 4 n. 284 below), but I know no other pre-T'ang reference
to hard and white which introduces length and breadth (cf. the quotations assembled
in § 1/4/3 below). It seems likely that "the book about hard and white and length
and breadth" is Lu Sheng's edition of the Canons, and "Kung-sun's essays about
forms and names" are its supplementary " 'Forms and names' in two p'ien".
9 0
Cheng Ch'iao ftf^ (A.D. 1104-1162), Tung chih MS; ch. 68 (WYWK
797A/-4). Chiao Hung j&tfc (A.D. 1541-1620), Kuo-shih ching-chi Vao H&|£J№9
ch. 4B (TSCC 177).
91
Chun-chai tu-shu chih IflUfUftfe. Taipei 1967, (ch. 10), 707.
"Sui shu 34, 6B/-2. Chiu Tang shu (Chih 27) 6B/3. Hsin
Tang shu 0f)jfi} 59, 9A/9 (Pai-na editions). The name T'ai H is probably a
corruption of Yi — (Yen Ling-feng (1969), 2).
93
Han CKang-li hsien-sheng ch'uan-chi ft,i,^3fc£ik^ ch. 11.
9 4
Wu Yu-chiang, Appendix 1/1B. Luan (1957) 147-158. I have consulted the
microfilm of it in the Library of the School of Oriental and African Studies,
University of London.
95
Luan (1957) 149.
1/1 69
with the remaining chapters without his aid. This accident had the unfor
tunate result that, although the full text survived in the Imperial library
at least until the Sung T£ ( A . D . 960-1279), for nearly a thousand years
scarcely anyone seems to have read it. Shen K u a Yfcffi ( A . D . 1031-1095)
discusses the inversion of the image in a concave mirror without mentioning
96
the Mohist optics. The Neo-Confucian Ch'eng Y i (A.D. 1033-1107),
asked his opinion of Han Yii's praise of Mo-tzu, betrays his ignorance of
more than the opening chapters when he remarks: "Moreover Mencius
said that Mo-tzu loved his elder brother's son in the same way as a
neighbour's son. In Mo-tzu's book, where does he ever say anything like
97
that?". Among the Sung bibliophiles, Ch'ao Kung-wu JS&lft in the
12th century owned a complete Mo-tzu as well as a Ho-kuan-tzu combined
with the first 13 chapters; the full text was also known to L i T'ao
( A . D . 1115-1184), who mentions its corruptness and says that he tried to
98
correct it. But late in the 12th century the account of Mo-tzu in the
Tzu liieh of Kao Ssu -sun "ffiffi^ exhibits knowledge only of the
first 13 chapters; and in the 13th Ch'en Chen-sun №$5·^ possessed only
the shorter version and supposed the rest of the chapters to be lost. Huang
Chen K M (graduated 1256) and Sung Lien %9k ( A . D . 1310-1381) describe
only the text in three chiian, and the excerpts from Mo-tzu in the Shuo-fu
ftff of T'ao Tsung-yi (ft. A . D . 1360-1368) are all from the first 13
99
chapters.
Fortunately the complete Mo-tzu had survived in the Taoist Patrology,
which was printed under the Sung and again in 1445. Eventually someone
noticed it among the 1,476 titles in the later edition. The complete text
returned to general circulation with the L u S edition of 1552 and the
T'ang St edition of 1553, the latter based on the Ming Taoist Patrology
100
and the former descended from the same Sung exemplar. The dialect-
96
Meng-ch'ipi-Van ^ggÊgfe (Hu Tao-ching ^ ^ 1 ^ (1956) item 44), translated
Graham and Sivin 144-146.
97
Ho-nan Ch'eng-shih yi-shu fflffilfitgflf BSS (ch. 18), 254/9-14.
98
Chun-chai tu-shu chih (as n. 91) (ch. 10), 707; (ch. 11), 722. L i T'ao, ap.
Wen-hsien t'ung-k'ao îDcffi^ ch. 212 (WYWK, p. 1740 C/12-16).
99
Tzu liieh SPPY 3/3A. Chih-chai shu-lu chieh-t'i ]g]|g£3$#PJfi TSCC, 285.
Huang-shih jih-ch'ao Sfffi B t £ (Ssu-k'u ch'uan-shu chen-pen 5
23B-25A. Chu-tzu pien fg-j-Jfc, edited Ku Chieh-kang ffifjgM, Peking 1928, 22.
Shuo-fu (Taipei 1963) 46, 11B-12B.
The Tzu liieh account of Mo-tzu does not give bibliographical information,
nor does it directly quote; but the passages echoed verbally are from ch. 7,12 and 13
(Sun 23/-2, 56/12-57/1, 61/-2, also Mencius 3A/5, Han shu (ch. 30) 1738).
100
Cf. p. 74 below.
ical and also the military chapters were at first considered unintelligible;
most of the late Ming collections of philosophers leave these chapters
101
out, like modern translations of Mo-tzil into Western languages. How
ever the first part of Mo-tzu to be annotated was one of the dialectical
chapters, in the Mo-tzil Ta-cWii p'ien shih fi^^flfclSW of Fu Shan
102
( A . D . 1607-1684). For the first time in their textual history the Canons
were now favoured by a stroke of good fortune; they had become
available just when China, simultaneously with and independently of
Europe, was developing the kind of scientific philology required to
understand them. P i Yuan Wdt ( A . D . 1730-1797), the first modern
commentator on Mo-tzu (or rather his underling Sun Hsing-yen 3ÎL#}
( A . D . 1753-1818), whose handwritten draft on which Pi Yuan's commentary
is based is preserved in the National Central Library at Taipei), repeated
some of the work L u Sheng had done more than a thousand years earlier,
by attaching the Explanations of Part A to their Canons. Another almost
contemporary commentary by Wang Chung t£$* ( A . D . 1745-1794) has
not survived; it appears from the preface reproduced by Sun Yi-jang
( A . D . 1848-1908) that Wang Chung was the first to recognise the
103
relatively late date of the Canons. A little later Chang H u i -yen c=t
( A . D . 1761-1802) wrote the first special study on the Canons. His manu
script, completed in 1792, was not printed until 1907; the first to publish a
reconstruction of Part B was W u Ju-lun J U S * ! (A.D. 1840-1903) in 1887.
At last, in 1894, the magnificent commentary on Mo-tzil of Sun Yi-jang
threw open the sanctum of the Canons to all comers. It also made it plain
that Mo-tzu is not a book by Mo-tzu, that the Mohist school had a history,
and that the Canons belong to the age of the sophists.
Returning for a moment to L u Sheng's preface to his lost work, one
is struck by its perceptiveness; L u Sheng appreciated, not only the value
of the Mohist science of names and objects, but the influence of Mohist
disputation on rival schools ("Mencius condemned Mo-tzu, but in his
argumentation and demonstrations is the same as the Mohists. Hsiin-tzu
and Chuang-tzu both reviled the School of Names, but could not find a
,
substitute for their methods. '). The first modern reader of the Canons to
1 0 1
Cf. the collections listed in Yen Ling-feng (1969) 7-15.
102
Shuang hung k'an cht MUMM ch. 35.
1 0 3
Sun 417/-2—418/2. Sun Yi-jang himself did not see this commentary
(Sun 419/-6 n.). The earliest Japanese study of Mo-tzu, the Bokushikai H^PP of
Hatta Ryükeiy\HHg| (A.D. 1692-1755), also seems to have been lost. I am indebted
to Prof. K. Knoki, Prof. Abe Ryüichi, and Prof. Hayashi Shüichi, for their
unsuccessful attempts to trace this book for me.
i/i 71
see them as something more than documents of a long dead heresy is

perhaps Tsou Po-ch'i S M S ^ (A.D. 1819-1869), a mathematician acquainted
with the Western sciences. He did not write on Mo-tzii, but his friend
Ch'en L i Eiffi (A.D. 1810-1882) records several of his observations:
"Tsou Po-ch'i said that the Canons and Explanations in Mo-tzu
contain both Chinese and Western mathematics, so I took and read them.
For example Canons A says ' "Level" is of the same height' (A 52); this is
what the Hai-tao suan-ching ^JftlP/ffi refers to as 'the equal height of a
pair of gnomons' (Four more parallels in Chinese mathematical works
follow). . . . It also says: ' " A tuan" is what precedes everything else in a
body without hsu' (A 61 18, 9 £ №f¥ fffi ft M # -ft); the Explanations
says: 'Tuan: it has nothing the same' (A 60 dl^Hifa). What is here
called a tuan is what Western mathematics calls a point, and 'a body
1
without hsu is what it calls a line. Hsu is the word as in 'East hsii and
'West hsii\ it is like saying 'two sides'. The Elements of geometry (Euclid
says: ' A line has length without breadth', which is what is here meant by
'without hsii'; that is, it does not have two sides. . . . Po-ch'i also said that
in Canons B 'If you stand looking into the mirror the shadow turns over'
(B 22) refers to a concave mirror. I observe that Explanations B says
'The legs cover the light from below and therefore form a shadow above;
the head covers the light from above and therefore forms a shadow below'
(B 19); this explains the reason why a man's shadow showing up in a
concave mirror is upside down. It also says: 'If the man looking at himself
is near the centre, everything mirrored is larger and the shadow too is
larger; if he is far from the centre, everything mirrored is smaller and the
shadow too is smaller' (B 23). In this case it is referring to a convex mirror.
The skill of Western makers of mirrors today does not go beyond the two
types concave and convex, and Mo-tzu already knew about them. It is a
pity that his writings are so mutilated, corrupt and difficult to interpret.
(It also says in Explanations B 'There is effort in lifting from above, no
effort in pulling' (B 26 S^^Jife, Slfefrtfe). I suspect that this is the
mechanics of the Westerners. It is a pity Po-ch'i has already passed away;
104
he would surely have been able to interpret i t " . )
Sun Yi-jang wrote in a letter to Liang Ch'i-ch'ao:
"In the six chapters Canons and Explanations A and B and the Bigger
pick and Smaller pick, not only is the meaning terribly obscure, even the
punctuation is garbled. Former scholars all dismissed them as impossible
to read. . . . It was two pioneering scholars of your own district, Ch'en L i
1 0 4
Ch'en Li 12, 13A-14A.
and Tsou Po-ch'i, who were the first to use the sciences of astronomy,
105
mathematics, optics and mechanics to discover their significance."
It will be seen that from the middle of the 19th century the recognition
of parallels with Western thought quickens interest in the Canons. After
the appearance of Sun Yi-jang's commentary their problems became
inseparable from the general problem of Westernisation, and involved two
of the leaders of political and cultural modernisation, Liang Ch'i-ch'ao
m&m ( A . D . 1873-1929) and H u Shih (A.D. 1891-1962). In order to learn
from the West without ceasing to be Chinese it was important, as H u Shih
explains in the introduction to his Chung-kuo che-hsueh shih ta-kang 4
U H A ^ S i (1919), to break with Confucian preconceptions but also to
find the alternatives within the Chinese tradition which point in the new
direction. H u Shih explored the p re-Han thinkers for "what I consider to
be the most essential part in every history of philosophy, the development
106
of logical method", which he perceived as culminating in the later
Mohists. Liang Ch'i-ch'ao in his Mo-tzu hsiieh-an (1921) and
Mo-ching chiao-shih BffiftS (1922) presents Mo-tzu and his school
(with its logic, science, utilitarianism, and quasi-Christian morality and
religion) as the ancient model most relevant to modern China.
It is a commonplace that nothing is more vulnerable to time and
chance than an important book. We know from our own history that a
text can fall from sight for a millennium and still be rediscovered by a
civilization which has not yet caught up with it. But bibliographical
accidents eliminated the Canons from the usable resources of Chinese
civilization throughout almost the whole of its history. We are not suggest
ing of course that the Canons alone would have transformed China;
whatever one's theory of history, one hopes to find deeper causes than the
loss or recovery of a book. But in such a period of technological and
economic advance as the Sung, highly creative in philosophy, science and
mathematics, the Canons would certainly have found appreciative readers
if they had been available and intelligible. If it pleases us to play for a
moment with the idea of Europe continuing its downward course during
the 14th century, and missing the Renaissance and the Scientific Revolu
tion, the potential consequences of Sun Yi-jang's commentary for China
and the world would have been incalculable. However by the time that they
rediscovered the Canons the Chinese were already learning from the West
everything that the Mohists could have taught them.
The letter is quoted in Ch'en Chu (1934) 183.

Prefatory note to Hu Shih (1922).
1/2
T E X T U A L PROBLEMS
1/2/1 PROBLEMS COMMON TO T H E SIX CHAPTERS
1/2/1/1 External evidence

A reader approaching the Canons for the first time will probably see
it as an irreparably corrupt text which editors emend and interpret more
or less as they please. Fortunately, as he soon discovers, much of the
corruption is systematic, and errors can be corrected without hesitation
because they recur throughout the Canons, sometimes throughout the
whole of Mo-tzu. But the fact remains that for many, perhaps most, of the
181 sections, interpretation has also depended on graphic emendations,
transpositions and section divisions which are largely conjectural. Nor is it
at all obvious that the textual criticism of the Canons has made much
progress during the last half century. From P i Yuan in 1783 to Liang
Ch'i-ch'ao in 1922 the progress is unquestionable; but when one looks at
the text edited by that excellent critic Kao Heng fli^ in 1958, which is
not so much emended as rewritten at his pleasure, one can hardly avoid
the suspicion that research has already come to a dead end, leaving us
with no honest choice but to discard at least half the document as
unintelligible.
There is no point in attempting a translation of the Canons unless the
text of most sections can be established on a relatively firm basis. Although
this has not yet been done we need not despair of the possibility, for the
strict methods which the Canons require have not yet been tried. Most
editors have been content to emend passages in the process of trying to
interpret them, without standing back from the text to get an overall
picture of its condition. Since Liang Ch'i-ch'ao's important discovery that
each Explanation begins with the first word of its Canon repeated as a
heading only Luan T'iao-fu has taken an interest in the textual problems
of the document as a whole. But the interpretation of a large proportion
of the Canons will remain a matter of guesswork until we can establish
criteria by which to settle most textual problems before, not in course of,
exploring the meaning.
74 Textual Problems
The Mo-tzu of the Han bibliography had 71 p'ien, of the Sui biblio
1
graphy 15 chiian, figures which recur regularly in the later lists. The
extant text preserves the numbering of the p'ien and of the chiian in which
they are grouped, although 18 of the 71 p'ien are missing. The Ming
Taoist Patrology {Tao-tsang MM) contains the earliest surviving edition,
in the part printed in 1445. Traces of older taboo characters point to a
Sung exemplar, presumably the lost Sung Taoist Patrology, although some
of the taboos are later than the printing of the main body of the collection
2
in the Cheng-ho JE&I period ( A . D . 1 1 1 1 - 1 1 1 7 ) .
It is fortunate that Mo-tzil is one of the early texts which has been
published with a full critical apparatus. Wu Yu-chiang's edition, which
appeared in 1944, notes the variants of all the older manuscripts and
editions preserved in China and in Japan, so that we no longer have any
excuse for mistaking misprints and conjectural emendations inherited from
Pi Yiian's edition of 1783 for genuine alternative readings. In a valuable
appendix W u describes the textual authorities. The T'ang edition, printed
by T'ang Yao-ch'en Jif^lE in 1553 and reprinted in facsimile in the
Ssii-pu ts'ung-k'an E9p£Hf(l, derives directly from the Taoist Patrology
and has no independent weight. The following are the important manu
scripts and editions which include the dialectical chapters:
(1) The Taoist Patrology (Tao-tsang) edition of 1445.
(2) The W u manuscript, the unpublished manuscript written by Wu
K'uan ^ % (died 1506), cognate with the preceding but with interesting
variants.
(3) The L u edition of 1552, edited by L u Wen BIH. It was long
confused by scholars with the T'ang edition published a year later (which
has a preface by L u Wen).
(4) The Mao edition of 1581, edited by Mao K ' u n ^ t t , cognate
with the L u edition.
(5) The Mien-miao-ko 8 № I edition of 1602, in the Hsien Ch'in
: /
chu-tzu ho-pien 3 f e ^ ^ f o i S .
(6) The Japanese Horyaku S № edition of 1757, edited by Akiyama
Tadashi $C[i4#i under the title Bokushi zensho i i ^ i i i i i . Akiyama based
it on a collation of the Mao edition with manuscripts preserved in Japan,
other readings of which it also records.
W u Yli-chiang himself follows the traditional practice of simply
listing the authorities for each reading without attempting to weigh them,
1
The references are assembled in Sun 399-403.
2
Cf. § 1/2/1/4/3.
1/2 75
but he provides the materials for a more sophisticated approach. For

example in A 68, where the Taoist Patrology and the cognate W u manu
script read all other early authorities (Lu, Mao, Mien-miao,
Horyaku) read i l for the second character. W u Yu-chiang however prefers
a third reading @, in favour of which he lists three more editions. But
when we look at his own descriptions of these authorities we see that this
reading can have originated only by the obliteration of a stroke in the @
of the Mao edition:
(7) The T'ang-ts'e-hsien edition of the T'ien-ch'i
period (A.D. 1621-1627), based on the Mao edition and disfigured by
conjectural emendations.
(8) The ' K u collation with the L i copy' (JKR$*), a list of 55
variants from an edition which W u shows to have been No. 7.
(9) The Ssu-k'u ch'uan-shu E9Wlk# copy, also directly based on
No. 7.
Of the six major authorities I have myself had access only to the
Taoist Patrology, Mao and Horyaku editions, and for the rest depend on
Wu's collation. Readings superior to those of the Taoist Patrology are rare
in any case. A notable general difference between textual traditions is a
recurrent tendency, not limited to the dialectical chapters, for the L u and
Mao editions to have the phonetic ^ in graphs which have the phonetic $
in the Taoist Patrology and the W u manuscript. Apparently the two were
written indistinguishably at some stage of the transmission and must be
treated as inextricably confused, although in the dialectical chapters it
seems always to be the Lu/Mao reading which is the better (A 81, B 20,
57,62).
The textual tradition of the L u and Mao editions is sufficiently
distinctive to show that they do not derive immediately from the Ming
Taoist Patrology. But there can be no doubt that all these editions descend
from a single Sung exemplar, since they agree in avoiding certain Sung
3 4
taboos. This was presumably, as Luan T'iao-fu notices, the lost Sung
Taoist Patrology, since no other Sung edition of the complete Mo-tzu is
attested. There is no possibility of independent readings surviving except
perhaps in the Horyaku edition (since we know nothing of the Japanese
manuscripts which it utilised). The main value of the collation is therefore
simply as a means of eliminating post-Sung misprints peculiar to the Ming
Taoist Patrology (cf. A 53, B 56).
3
C f . §1/2/1/4/2.
4
Luan (1957) 51.
76 Textual Problems
For external evidence before the Ming we have only the occasional
parallel or quotation. Although sparse these are sometimes very valuable.
I have noticed the following, none of them later than the T'ang (when, as
we noticed in § 1/1/2/9, the complete text of Mo-tzu passed out of general
circulation):
1
Chuang-tzti, jffi-T (c. 300 B . C . ) , Kung-sun Lung tzU ch.
SPTK 4-6 (between A . D . 300 and 600),
Ch. 14 (5,48B/1) . . N O 15 Taoist Patrology
23 (8,17B/1) . . A 4,5 B, 3B/2-4A/6 . B 66,67
4A/10-4B/3 B 12
Hsun-tzu €^ (c. 250 B . C . ) , S P T K C, 1A/2 . B 15
Ch. 22 (16,7B/4) . . N O 15 IB/3,4 . A 66
2A/6-2B/3 B 4
Huai-nan-tzu MlM^f* (c. 130 B . C . ) , 3A/2,3 . B 37
SPTK 3B/6 . B 70
Ch. 9 (9,13B/4) . . N O 11 4A/3,4 . B 5
16 (16,5B/5) . . EC 8 4A/5,6 . B 46
16 (16,15A/3,4) . N O 18 5A/7,8 . B 70
5B/2-4 B 72
Ssu-ma Piao (died A . D . 306), 6A/1-3 B 68
ap. Ching-tien shih-wen iffiftf? 6A/4,5 . B 33
X , SPTK
28,32A/5 . . . . B 17 Ch'eng H s i i a n - y i n g ^ C ^ ^ (ft. A . D .
28,32A/11 . . . . N O 18 631-650), Nan-hua chen-ching
chu-su Taoist
Lieh-tzu (c. A . D . 300), S P T K Patrology
1,3 A/4 . . . . A 45 35,45B/3,4 . . . B 54
4,7B/2,3 . . . . B 17
5,5B/1,2 . . . . B 52 Yi lin '&№ (latest preface A . D . 787),
SPTK
Chang Chan 3ftS (c. A . D . 370), on l,25B/6 . . . . EC 8
Lieh-tzu, S P T K
4,7B/3 . . . . B 17
1/2/1/2 Graphic conventions

1/2/1/2/1. The Mohist Canons are notorious both for the abundance
of otherwise unknown graphs and for the extraordinary variations in the
writing of single words. As random examples of the latter we may point to
sui 'although* (88, ff£, * ) , cheng 'correct, exact' (JE\ fi), yi 'slant' (№, flk,
i&, M),piao 'tip' ( S , JS), ch'ui 'hammer' ( 1 6 , I S ) , wu 'match' (ft, E ) ,
1/2 77
chieh 'suspend' (IP, yuan 'circular' ( J i , IS), t'uan 'spherical, cylindrical'

( » , H), fen 'share' ( # , 3?), tAiot 'mirror' ( g , * ) . The back of this
problem was already broken by 1893 in the great commentary of Sun
Yi-jang, and the identification of aberrant characters has been one of the
great successes of scholarship in this difficult field. Graphic corruption is
only part of the explanation of this confusion, although it is easy to see
that the radical Tfc, for example, has tended to degenerate into ^t", and the
latter in its turn into ± . The main reason, as is generally recognised, is
that the graphic standardisation undergone by all ancient texts was in
this case carried out hesitantly and inconsistently by editors who could
not follow the meaning.
1/2/1/2/2. In some cases we can still see that the Mohists were using
different radicals to mark distinctions which later standardisation has
obscured. It seems clear that originally the text consistently distinguished
1
chih £0 ' the intelligence', the faculty by which we know, from chih H
'knowing' and chih & 'knowledge', all three being key terms which are
defined in A 3, 5, 6. The last word is used in A 75 as the opposite of yii jS
'foolishness', so that although its graph is otherwise unknown we can be
sure that it is merely the ordinary falling-tone chih 'wisdom'. In standard
later usage it is the second of the graphs (W) which is used for this word.
A scribe therefore eliminated the radical from chih W 'know' throughout
the Canons and the first part of the Explanations until his patience ran out
at the beginning of Part B. The result is that chih 'the intelligence' and
chih 'know' sometimes stand close together in one sentence without any
longer being graphically distinguished (A 23 ftlftlftl-tfe "Sleep is the
intelligence being without awareness", cf. also A 3, 5, 6). But from B 9
he kept the radical in chih 'know' almost consistently. At four places it
survives in the head character of the Explanation although corrected out
of the Canon (B 34,40,46,48). The scribe did not recognise the unorthodox
graph 58 and therefore did not correct it. It tended however to be mistaken
for the familiar shu $S, which has replaced it twice in the Taoist Patrology
(A 6, 88) and nearly always in other editions, as can be seen in W u Y i i -
chiang's collation.
1/2/1/2/3. The sorting out of this graphic confusion is helped by a

very important fact which has not been sufficiently appreciated, that the
writers of the Canons seldom use rare words. Their vocabulary consists
of the ordinary words current in the 3rd century B . C . , and gives us trouble
only because they are often disguised by aberrant graphs or adapted to
technical uses. For this reason dictionaries, even the Shuo wen i&3$C itself,
78 Textual Problems
can sometimes be misleading rather than helpful in the tracing of Mohist

graphs. A more fruitful approach is to take the common word written with
the phonetic of the graph and then ask what sense of it is distinguished by
the addition of the radical. For example, one is struck by the frequency of
rare graphs with the 'man* radical, which certainly do not represent the
rare words written with these characters and provided with definitions of
varying intelligibility in the Shuo wen:
t'ung m erh № (ft*) hsuan It
p'i #b (SHfe) yu ffi ( « - & ) ch'ien U (tftft)
When examined in context, it can be seen that each represents a quite
common word used in a sense connected in some way with human action,
and therefore distinguished by the 'man* radical:
1/2/1/2/4. T'ung M 'agree, conform to* (A 39). The radical has been
corrected out of the Canon but survives in the head character. This verbal
use of t'ung 'same* would be important to Mohists because it occurs in one
of the 10 theses of Mo-tzvi, shang t'ung fnlffl 'conforming to those above*.
1/2/1/2/5. Pi 'lay side by side* (A 68). This graph too survives
only as the head character; in the Canon it is corrupted to 4H. The radical
distinguishes/)/ in this sense from/)/ til 'compare* (A 88, N O 9, 11), itself
distinguished by a radical in B 6 (№).
1/2/1/2/6. Erh m represents two separate words, both of which are

defined, the former as action on one's own initiative (A 15: what looks
like the same word is written as I B in A 17), the latter as the feature by
which something is judged to accord with a standard (A 71). The latter
word seems to be the yin 'criterion* elsewhere written S ; presumably it
was originally written AS, and corrupted in the single place where the
radical had not been eliminated (§ 1/4/39).
1/2/1/2/7. Ch'u U 'divide off from others, group separately* (B 63).

Ch'u without radical is used of a demarcated space (A 48, 63 E / C 'bounded
hollow*, B 22 S@k 'reduced area*). It is interesting that the other examples
of the verbal ch'u 'divide off* are also disguised by strange radicals (B 12 Bfc,
A 73 S . Cf. A 49, where the latter graph does represent the ordinary shu
'pivot*). It may be suspected that these have replaced an original 'man*
radical, since in B 12 ch'u as head character is corrupted to {Mr. Cf. § 1/4/7
below.
1/2/1/2/8. Fan f5 'converse* (B 30,72), corrupted to f&, ife (A 73,74),
R , ft (A 83), n (B 3). Cf. § 1/4/9 below.
1/2 79
1/2/1/2/9. Hsiian fit 'to circle, circumambulate' (A 47).
1/2/1/2/10. #| 'to let the scales settle' (A 75). Although Sun

is undoubtedly right in identifying this word as hsiian IS 'hang', it may not
be a coincidence that this is the one place where hsiian is used of settling
for one course rather than another, a metaphor derived from the sinking
of the steelyard on one side or the other (cf. A 75 n. 194 below). Since the
weighing (Milan fll) of relative benefits and harms is so important for the
Mohist ethic it would have been worthwhile to distinguish this usage by
a radical.
1/2/1/2/11. Hsien U (=%L) 'present, at hand' (EC 2). Elsewhere

the present text does not distinguish Men 'see' and hsien JL 'appear',
even when they stand in the same clause: B 4 "Seeing and appear
ing are apart" (cf. also A 1). But in B 4 the second occurrence is corrupted
in the Canon to ft, confirming that it was originally distinguished by the
'man' radical.
1/2/1/2/12. It will be noticed that the radicals of Mohist invented
characters tended to be eliminated in the later attempts to standardise the
text, so that in cases where a familiar character lacks its radical we cannot
be sure that it was originally written without one (for example M ^ = 8
'evening': t'ai t6 = fil 'posture', once written lf& in the military chapters,
Sun 313/9: hsiang *B=& 'image': hao S = S§, U 'crane'). The most
extraordinary example of missing radicals is the use of yeh ife, not only for
*'o *E, flb 'other' (A 17, B 3, N O 15, 16, as well as Sun 325/9, 328/2), but
also, as Kao Heng showed in B 39, for she 12, fife 'snake' (B 39, 50,
E C Appendix 13). The former is attested elsewhere ( M . 171 def. 14), but
the latter is quite unintelligible except on the assumption that some
graphic distinction has been eliminated by standardisation.
There are several other cases where textual corruption preserves traces
of original graphic differentiation:
1/2/1/2/13. Chih ii as pronoun and as particle. In the Explanations

the graph % (A 31, 75, 78, B 3 twice, 50 twice, 74, 75, 82 eight) is always
a graphic error for Mh, except for one case in B 50 where it does seem to
be wen 'pattern'. This corruption is found only in the Explanations, never
in the rest of Mo-tzii including the Canons themselves. It is remarkable
that the corrupted graph always represents the pronoun, never the particle.
Twice the particle is written with the correct graph in the same sentence:
A 75 S ^ I K tX-fe "the consideration which the intelligence gives to it",
B 75 W f t S ^ l a f i t S l i b "know that love of the people is for all of them".
80 Textual Problems
There are also examples of chih corrupted to yeh ifa (EC 8, Appendix 2,
B 26, N O 9, also Sun 116/9, 280/2, 3); in all of these chih is the particle.
We shall see in Part 1 /3 that the Mohists were consciously concerned with
the syntactic lucidity which is indispensable to the study of logic, so that
it is inherently plausible that they would use some graphic device to dis
tinguish the two words commonly written HL.
1/2/1/2/14. Chii 9 'lift' and 'mention': yii jft 'together with' and
interrogative particle:
(1) Chii IS, IS 'lift'. In this usage it lacks the radical twice in the
Taoist Patrology (A 21, B 5) and once more in the Horyaku edition (B 5),
has it in all editions only once (B 10).
(2) Chii 9 'mention, pick out by name', a recurrent word in logical
contexts, always has its radical.
(3) Yii IS 'together with' is twice corrupted to ft (A 83, B 41) and
twice more to ftf (A 67, Sun 288/—3). The latter graph, to judge by its
confusion with 5fe in B 57, 73, was originally written It is therefore
likely that the conjunction yii was originally written without its radical,
as M or This would explain how readers were able to disentangle such
a sentence as B 2 W H ^ J f t H t e f t H ^ / h - t f a " 'Is it an animal? Or a
living thing or a bird ?' is 'all of the thing* and 'the wider and the narrower* ".
(4) Yii IS, interrogative form of yeh -tfe (B 2, 10).
(5) Yii IS 'give to* is never found; in stead we find yii ¥ (B 69,
N O 10, 11).
These differentiations would hardly make sense unless there was a
further untraced differentiation between Nos. 1 and 4.
1/2/1/2/15. Mao B is used both nominally (A 48, B 22, 65, N O 2, 7
'features*) and verbally (A 5 'describe*). The unknown graph in A 95
(preserved in the Taoist Patrology though corrupted in other editions to IS)
may be identified as representing mao 'describe* and distinguishing it by
the 'word' radical.
1/2/1/2/16. Chung 4* 'hit on'. If we are right in identifying the word

represented by the corrupted graph •¥· (A 85, 88) as this (§ 1/2/1/3/22),
it was probably graphically distinguished from the commoner chung
'middle'.
1/2/1/2/17. Fortunately graphic confusion is limited to the radicals
and we can depend on the constancy of phonetics. Interchanged graphs
5
Luan (1957) 154.
1/2 81
with the same readings are generally ones that we meet everywhere in
pre-Han literature, such as yu ^ ( = X ) 'again', hou Jn 'after', yuan
j l , HI 'circular' and the phonetic « « / E / ^ ; the only unquestionable one
which is at all rare is chou j\\ (=JS) 'all round' (B 82). A t various places
scholars have proposed to emend a graph to another with a similar reading,
such as S to £ in A 32 (Sun, generally followed), to & in A 49 (Chang
Hui-yen), & to $L in A 83 (T'an Chieh-fu), but to my mind only the
second of these is plausible. In general the suggestion that one graph is a
sound loan for another deserves suspicion unless backed by other examples.
1/2/1/3 Graphic corruption

1/2/1/3/1. There is much less graphic corruption in the dialectical
chapters than is commonly supposed. The number of seemingly necessary
graphic emendations steadily diminishes as one learns to identify the
textual, linguistic and terminological problems of the corpus as a whole
and ceases to be content with piecemeal solutions. The demoralising habit
of using conjectural emendation as a panacea has been the greatest of all
the obstacles to progress in the understanding of the later Mohist writings.
1/2/1/3/2. It is important that the grounds for emendation should

as far as possible be detached from problems of interpretation. The mere
fact that one does not understand a passage is not a sufficient reason for
correcting the text. But there are a large number of emendations, most of
them already made by Sun Yi-jang, which we need not hesitate to accept,
because the corruption is systematic. The Harvard-Yenching concordance
to Mo-tzti (based by the way on the text as edited by Sun, so that the
apparent variants at the bottom of the page are often the Taoist Patrology
readings which he was emending) is a useful tool for checking whether
Sun made the same correction elsewhere. The following are cases where we
find the same graphic forms confused, sometimes recurring in more than
one character:
1/2/1/3/3. £ , ± (§ 1/2/1/2/13).
1/2/1/3/4. a , JL (A 40, B 14, 50), ft (A 43, 60, 67). Tan (B. 'only'
does not occur anywhere in Mo-tzii.
1/2/1/3/5. 41, IK (B 4 twice, 64. Sun 24/3,189/4). In B 64 the correct

reading survives in the Explanation. For BK!§ in B 4 the parallel in Kung-
sun Lung tzU ch. 5 (Ch'en 181) has fftf 'length and breadth'.
1/2/1/3/6. 0 , S (B 10, 20, 47. Sun 123/7).

82 Textual Problems
1/2/1/3/7. * , S5 (A 6 twice, 88) » , *P (A 11. Sun 184/5, 315/2).

The Confucian moral term shu $8» 'forbearance* never appears in Mo-tzu.
The later Mohist syntax, which avoids synonyms among particles, employs
jo 5g 'if, like*, never ju #P (§ 1/3/1).
1/2/1/3/8. A , A (A 98, B 19 thrice, 71. N O 2, 15. Sun 87/14,331/7,

367/14).
1/2/1/3/9. ft, fl& (A 49, B 13, 17). The correct reading № survives
in the Explanations of B 13, 14 and in a quotation from B 17 by Ssu-ma
Piao (Chuang-tzu ch. 33, Kuo 1109/16).
1/2/1/3/10. l h , (A 96, 97, B 26 twice, 32. Sun 361/9, 370/2).

Hsin 'fr 'mind* is consistently avoided in the Canons and Explanations for
chih% 'the intelligence* (A 3). Cf. also l h , I E (Sun 379/8, 383/6, 381/14);
i t , & ( A 5 1 ) ; <t\ J E ( A 53).
1/2/1/3/11. 0 , §# (B 34, 79). The former character is not found
elsewhere in Mo-tzu, the latter occurs in B 71, 73.
1/2/1/3/12. 5 , ^ (A 78, B 63).

s
1/2/1/3/13. £c (A 50. Sun 321/-3, 369/13, 380/12).

1/2/1/3/14. J», FP (A 61. Sun 95/4, 101/14, 326/2). The word hsii¥f
does not appear in Mo-tzu ; Sun finds the graph corrupt even in the single
place where he emends otherwise (Sun 98/8).
1/2/1/3/15. № (graphic variant S ) , M ( A 69. Sun 130/2 n., 387/5).
1/2/1/3/16. ft, • JrC (A 63. Sun 313/10 and ch. 62 passim).
1/2/1/3/17. H , H ( A 62, 63,80).

1/2/1/3/18. JS (B 5. N O 13 twice. Sun 94/7).
By consulting the Li pien of K u Ai-chi it will be seen
that most of the less similar of the confused graphs have closely similar
forms in the li-shu MUr script of the Han inscriptions ( f t , ftfe,
±, st-: f¥, I * : ^ , J¥, I f : ft, K: JS). The Mo-tew of the Han
Imperial library would have been written in this script, and Luan T*iao-fu
has pointed out a number of li-shu forms surviving in the Y i i manuscript
6
of Mo-tzu ch. 1-13, which represents an independent textual tradition
7
going back at least to the 7th century. We may infer that a large proportion
6
Luan (1957) 152-154.
7
Cf. p. 68 above.
1/2 83
of the systematic corruption originated in misunderstandings of the graphic

forms in the Han text, and that li-shu forms have a special claim on our
attention.
1/2/1/3/19. These systematic confusions have been recognised by all

20th century editors. The peculiar consistency of later Mohist vocabulary
gives reason to hope that further additions can be made to the list. In stead
of emending piecemeal, paying attention only to the immediate context, it
is more profitable to start from two general observations about the text as
8
a whole, that as we saw in the last section the technical coinages were
especially liable to corruption, and that certain problematic graphs tend
to recur, to give trouble wherever we find them, and sometimes to resemble
each other in form. Investigation of the technical terms must start from the
definitions of A 1-75; we can expect to find all but the geometrical terms
in regular use in Mo-tzu, the moral terms in the body of the book, the
logical in the dialectical chapters themselves. Of the logical definitions
(A 1-6, 70-75) two are of corrupt terms, A 71 73 $C. Both have the
'man* radical common in Mohist coinages; all the other 10 in the two
series are frequently attested in the dialectical chapters; where else are we
to find them, if not in the practical exercises of Mohist disputation ? We
shall later give the evidence for identifying them as yin 'criterion' and
fan ffi 'converse', and also for recognising mao 'characteristics' in certain
9
puzzling examples of the graph JSi. Attacking from the other direction,
we can collate certain graphs or families of similar graphs which are
puzzling at every recurrence.
1/2/1/3/20. № (A 48, B 19 only), M (A 80 only; the li-shu form of

10
the phonetic is 3JL). Sun saw the affinity of the graphs and emended the
former to the latter. But W u Yu-chiang has since shown that the word in
11
A 48 is certainly yiin M 'rotate'. It stands in a series of common words
for different kinds of change: . . . A 45 'Transformation', 46 'Reduction',
47 'Circling round', 48 'Rotation', 49 'Movement'. Since in bronze
inscriptions the graphs 3^ and both appear written indistinguishably in
1 2
both the forms $ , the word is simply yiin written without its radical.
Wu Yu-chiang did not apply the same explanation in B 19 and A 80, but
it fits at both places.
8
Cf. § 1/2/1/2
9
Cf. § 1/4/9, 20, 39.
1 0
Ku Ai-chi 2/29.
1 1
Wu Yu-chiang 10, 9A/2.
12
Tuan Wei-yi gftllS! 247, 649.
84 Textual Problems
1/2/1/3/21. =ft (A 83, B 41 only). I take this is a mistake for H written

without its radical, which there is some reason to think was the form used
to d i s t i n g u i s h ^ 'together with/ (§ 1/2/1/2/14).
1/2/1/3/22. -¥· (A 83, 88 only). This graph is used for tsao 'early'
only once in Mo-tzu (Sun 181/4), and even there the W u manuscript has
the ® used everywhere else. In the single other occurrence, in the military
chapters (ch. 52, Sun 332/6), the phrase -¥-11 was recognised by Sun
Yi-jang as an error for the ^ i H "bore out in the middle" of the same
chapter (Sun 316/-3). The contexts in A 83, 88 are both obscure, but in
1
both the word is intelligible as chung 4 'hit on'. I suspect that in the
dialectical (although not the military) chapters the corrupted graph,
whatever it may have been, distinguished the verbal use of chung
(§ 1/2/1/2/16).
1/2/1/3/23. * (A 51, 85 only: graphic variant K ) , S (B 5 only:

graphic variant HI). P i Yuan recognised the graphic affinity, and emended
13
the latter to the former in A 51. I prefer to recognise all as a technical
term ti TSf 'complement' which appears with a radical in A 89 and in a
vulgar form in A 22 (§ 1/4/29).
1/2/1/3/24. ft (A 73, 74 only), ft (A 83 only), which I take to be
corruptions of fan ifc 'converse' (§ 1/4/9).
Where we observe systematic corruption we are entitled to pick
whichever of the confused graphs fits better into the context, almost as
freely as we choose between interchangeable graphs or even dictionary
meanings of a word. The graphs A , A for example must at some stage
of the text have been almost indistinguishable, so that to choose the latter
where the context requires a verb barely counts as an emendation at all.
We can also emend with confidence in the rare cases where there is a
parallel with a better reading in Huai-nan-tzU or the forged chapters of
Kung-sun Lung tzfi, and the very frequent cases where the graph is written
correctly elsewhere in the same section. A fortunate consequence of the
fact that the Canons and Explanations were collected in separate chapters
of Mo-tzu is that there has been no temptation to assimilate their readings
to each other, so that we can very often correct Canon from Explanation
or Explanation from Canon.
Finally we have the pure conjectural emendation, the choice of a
graphically or phonetically similar character simply because it makes
better sense. With a text as obscure as the Canons it is easy to forget that
1 3
Pi Yuan 10, 7A/8.
1/2 85
conjectural emendation depends on having a relatively firm context in

which it can be seen that a reading does or does not fit. Sometimes the
context may be so firm that the true reading may be virtually certain, for
example in parallel sentences in which T 'below' contrasts with i t ,
obviously a corruption of -h 'above' (B 19). But unless one has already
found one's bearings in a passage emendation is useless. The great vice of
editors of the Canons has been to offer interpretations which depend on
conjectural emendation at every level, even for establishing the punctuation.
1/2/1/4 The Sung taboos

1/2/1/4/1. The practice of tabooing the graphs of the personal names
of Emperors continued throughout the history of Imperial China. During
all dynasties down to the T'ang ( A . D . 618-907) it involved the substitution
of synonymous words, such as kuo S3 for pang % and ch'ang % for heng f3
under the Han and tat ft for shih 1ft and jen A for min S under the
T'ang, even in the copying of older books; the consequences for the
transmission of texts and the names of offices, persons and places are
examined in the Shih hui chii li jfefl№#l of Ch'en Yuan W*!. The taboo
system of the Sung (A.D. 960-1279), although in other respects the most
stringent in Chinese history, covering the names of all previous Emperors
of the dynasty and all their homophones, allowed the printing of tabooed
graphs provided that the final stroke was omitted, so that at first sight there
would seem to be no danger that it affected the text of Mo-tzu. Its regula
tions, including full lists of homophones and instructions for the treatment
of graphs with different readings, may be studied in the appendix of a
rhyme dictionary of the Shao-ting $3Sr period ( A . D . 1228-1233) Fu
shih-wen hu-chu li-pu yun-liieh F № 3 C £ t t $ I § P 8 i l B & .
1/2/1/4/2. Unfortunately Luan T'iao-fu has uncovered evidence that

the Sung text of Mo-tzu did make substitutions for tabooed graphs and
that the original readings were not fully restored in the Ming editions. He
ascribes this unexpected complication to a revival of the practice of
14
substitution after an edict of 1162. In Mo-tzu ch. 49 (Sun 296/-4) a
character has been replaced by the note: ^ C f l f i S J I ^ - h ^ "the first
character of T'ai-tsu's taboo name". The reference is to K'uang-yin HUH,
personal name of the founder of the Sung. Since this note is shared by the
15
Ming Taoist Patrology, L u , Mao and Horyaku editions, we can be sure
that all extant editions of the complete Mo-tzu derive from a single Sung
1 4
Luan (1957) 100.
16
Wu Yu-chiang 13, 5A/3.
86 Textual Problems
exemplar, probably that of the Sung Taoist Patrology. The homophone

k'uang St however did appear in the Sung text, since in its eight occurrences
in ch. 53 the Ming printers of the Taoist Patrology still omitted the
bottom stroke.
1/2/1/4/3. The single authority independent of the Sung textual
1
tradition is the abridged Mo-tzu in three chiian * (containing only ch. 1-13)
which survives in the Y u ifa manuscript (A.D. 1506), the readings of which
may be consulted in W u Yii-chiang's collation. This reads wan % three
times for the ch'iian ^ of the standard text and cWeng M six times for
17
shih J ? . These are certainly the original readings, since both are confirmed
18
wherever there are pre-Sung quotations. It seems clear that the Sung
editors eliminated wan as an officially listed homophone of Huan ffl, the
name of Ch'in-tsung i f e ^ (A.D. 1126-7), and ch'eng as an element in
Kuei-ch'eng ft IS, former name of Li-tsung (A.D. 1225-64). Since
the avoidance of taboos was so often inconsistent in practice we need not
ask such questions as why they eliminated a homophone of Huan yet were
content to omit a stroke in the case of a homophone of the first element in
K'uang-yin.
1/2/1/4/4. To what extent can Luan T'iao-fu's results be applied to

the dialectical chapters, for which we have no independent textual author
ity ? In the Canons we find shih U as the odd-man-out in a series of other
wise familiar moral terms ( A l l ) , which is puzzling until we recognise that
this must be another instance of shih replacing the tabooed ch'eng. There
are two cases of graphs which seem to be corruptions of Sung homophones:
A 53, 59 graph with phonetic E corrupted to ffi, ft.
B 62 A corrupted to
1/2/1/4/5. In the first case at least it looks as though the original
graph had been written without the final stroke. Other homophones have
survived uncorrupted:
K B 52 (confirmed by a parallel, Lieh-tzu ch. 5, 5B/1), 27, 29, 62
* A 45
« A 75
91 A 45 (confirmed by a parallel, Lieh-tzu ch. 1, 3A/4)
if N O 10.
1 6
Cf. p. 68 above.
1 7
Luan (1957) 100, 155.
1 8
Luan (1957) 100.
1/2 87
1/2/1/4/6. Apart from the corrupt yin IS of B 55, there are two
actual tabooed graphs:
hsü BS B 19 (homophone 8«J A 47), name of Che-tsung S S (A.D.
1086-1100).
jang Ä B 36, element in Yün-jang A Ä , name of father of Ying-tsung
^ ( A . D . 1064-1067).
These deserve close attention since although the Sung text may have
dealt inconsistently with homophones the presumption must be that any
actual tabooed character in Mo-tzü was restored after the end of the
dynasty, perhaps incorrectly. In the Canons and Explanations, unintelligible
until the study of Mo-tzü revived in the 18th century, the chances of
correct restoration would not be high. The jang of B 36 seems acceptable,
but hsü has baffled commentators in both A 47 and B 19. Evidently what
ever graph stood in the Sung text was later mistaken for the taboo substi
tute. The only case that I have observed of a Sung substitution for hsü
is in the Meng-ch'i pi-Van (§ 29) of Shen Kua, where a certain L i u Hsü
S|9pJ is more than once called L i u Chii SBI^J. It is typical of the complexities
of the Sung system that this graph too, when given the reading kou, would
require omission of a stroke as a homophone of Kou $t, name of Kao-tsung
c
A S (A.D. 1127-62). In both A 47 and B 19 kou curve' is the word
suitable to the context; A 47 concerns circular movement, B 19 the
inversion of the shadow in either the camera obscura or the concave mirror.
1/2/2 T H E CANONS A N D EXPLANATIONS
1/2/2/1 The order of the Canons

1/2/2/1/1. The Canons are collected in ch. 40, 41 of Mo-tzu, the
Explanations separately collected in ch. 42, 43; the two series are arranged
in different sequences, and there is nothing either to link an Explanation
to its Canon or to distinguish one section from the next. Of the two
arrangements there is no doubt that it is the order of the Explanations
which is original, since they are grouped by topic, while the Canons have
been systematically redistributed according to what at first sight seems to
be an arbitrary mathematical principle:
A 1, 50, 2, 51, . . . 48, 96, 49, 97, five characters («]fc«3&f? "Read
this book horizontally"), 98.
B 1, 42, 2, 43, . . . 40, 80, 41, 81, 82.
The reason for this rearrangement was already discerned in the 18th
century by Pi Yuan. There must have been an earlier stage at which the
88 Textual Problems
Canons, although written vertically, were laid out in two horizontal rows:
A 1 2 . . . 48 49 "Read this book horizontally''
50 51 . . . 96 97 98.
The scribe who wrote them consecutively overlooked the instruction
to read horizontally, and read straight down the successive columns
according to the usual practice.
1/2/2/1/2. There is however one place where even when the two
rows are laid out the order of the Canons is not that of the Explanations:
B 12 13 14a (22-24a) (14b-21) 24b 25
52 53 54,55 56-63 64 65
Evidently the two items we bracket have been transposed. Both B 14 and
B 24 have been broken in two by the dislocation, at points first located by
Luan T'iao-fu; even more recent scholars such as T'an Chieh-fu have
sometimes failed to disentangle themselves from the assumption that the
blocks transposed must have consisted of whole Canons. Another observa
tion of Luan which has not always been sufficiently appreciated is that
since the transposition is confined to the upper row of Canons there must
have been a still earlier stage at which they were written consecutively in
19
the same order as the Explanations. Similarly A 89-92, as will be shown
20
later, appear to be an interpolation confined to the lower row of the
Canons, and therefore can only have entered the text at the stage postulated
by Luan.
1/2/2/1/3. We must therefore distinguish three stages in the history

of the Canons :
Stage 1: the order of the Explanations (A 1, 2, 3, 4, . . . )
2: the two horizontal rows.
3: the order of the printed Mo-tzu (A 1, 50, 2, 51, . . . ).
Luan suggests that at Stage 1 the Canons were written on wooden
strips, that Stage 2 dates from their transference from strips to paper in the
3rd century A . D . or later, and that it was the impossibility of fitting the
horizontal rows into the shorter columns of early printed books which
made it necessary to print them consecutively at Stage 3. The most ques
tionable of these suggestions is the dating of Stage 2. There is no obvious
reason to connect the horizontal layout with a change of writing materials.
The most likely motive would be convenience of study, which suggests that
it was done before the extinction of the Mohist school and therefore before
1 9
Luan (1957) 52.
2 0
Cf. p. 109 below.
1/2 89
wood was superseded by paper. Luan's reason for denying that strips were
used at Stage 2 is that the scribe would write on separate strips before bind
ing them and therefore could not write the top before the bottom row. The
discovery of the Yi li $kWt strips, which were bound before writing, has
21
disproved this argument. There is a very strong counter-argument; there
22
is one copy of the Yi li with as many as 100-110 characters to a strip,
but the columns of paper manuscripts are hardly longer than those of
printed editions. At Stage 2 the 23 characters of B 9 were followed by the
9 characters of B 49, which implies columns allowing room for at least 32
characters. Luan, who had made this calculation, admitted that he could
not find columns of as many as 30 characters in any of the Tun-huang
23
manuscripts known to h i m . It is therefore likely that Stage 2 belongs to
the 3rd century B . C . and is not much later than Stage 1. Stage 3 on the
other hand is the work of a much later scribe or printer who was not trying
to understand the text, merely fitting it to the shorter columns of the
manuscript or block.
1/2/2/1/4. The Explanations were written at Stage 1. This is evident,

since they preserve the original order of the dislocated Canons of B
14b-24a.
1/2/2/2 The Canons at Stage 1

1/2/2/2/1. The transposition of B 22-24a and 14b-21 is presumably
the result of a displacement of writing strips. Luan T'iao-fu recognised
that from this and several other transpositions in Mo-tzu it should be
possible to identify the number of characters on a strip at Stage 1. The
figures for exchanged passages in Mo-tzu ch. 1-39 as edited by W u Y u -
chiang are as follows:
Mo-tzu ch. 5 (Wu, ch. 1, 14B/4) 40 characters
10 (Wu, ch. 2, 15B/5) 37 (with preceding lacuna), 41
12 (Wu, ch. 3, 7A/3) 37, 43, 39
15 (Wu, ch. 4, 3B/9) 40,38
36 (Wu, ch. 9, 8A/2) 40,40
The copy of the Mu t'ien-tzu chuan WR^tif- found in A . D . 281 in
a tomb of the 3rd century B . C . was written, according to the preface of
Hsiin Hsu K J U (died A . D . 289), with 40 characters to a strip. Evidently
21
Wu-wei Han chien 58. The early Han silk manuscripts found at Ma-wang-tui
JH3y$L in 1973 also have sufficiently long columns.
n
Ut sup. 63.
2 3
Luan (1957) 53.
90 Textual Problems
the same was true of Mo-tzil, and also, as I have argued elsewhere, of the
24
White Horse essay of Kung-sun L u n g .
1/2/2/2/2. However, the figures for the two transposed blocks of
Canons are 66 and 35, and although Luan is content to infer that the
25
number of characters to a strip varied between 35 and 40 it does not seem
reasonable to force the Canons into the same scheme. We shall return to
this question when examining the Ta-ch'u fragments, and conclude
that the copy of the dialectical chapters which suffered dislocation was
26
written on strips of 33/34 characters.
1/2/2/2/3. D i d the editor who rearranged the Canons at Stage 2 ever

mistake the divisions between them ? The consequences of such an error
would be serious, since at Stage 3 a Canon from the other row would be
interposed between the misplaced fragment and the Canon to which it
belongs. A clear example can be found in B 2, 3:
B 2. . . r * / h J (B 43). . . B 3 r J. n% . . .
Compare the Explanations:
B2... rJ%r%>hJ-fe...B3(m&)....
Two points immediately attract attention:
(1) The two head characters of the Explanation of B 3 should repeat
the first two characters of the Canon; but if so the phrase felfe does not
belong to B 3.
(2) The phrase joins up with the end of the Canon of B 2, as is
confirmed by the Explanation.
We may conclude that the editor at Stage 2 made the division between
the Canons two places too soon.
1/2/2/2/4. The number of Canons at Stage 2 can be confidently
estimated as 98 in Part A and 82 in Part B. Since we cannot have the
same confidence about the Canons of Stage 1, we shall number them as
they stood at Stage 2 (except for the transposed Canons of B 14b-24a).
At several points we may suspect differences between the two stages.
1/2/2/2/5. The two definitions in A 74 may have been different
Canons at Stage 1. But this is doubtful; they seem to share a single Explana
tion and a single head character.
2 4
Cf. G(6) 134-136.
2 5
Luan (1957) 53.
2 6
Cf. p. 107 below.
1/2 91
1/2/2/2/6. Two Canons, on desire and on dislike, may be suspected

at A 84. Since there is no Explanation we cannot check this possibility.
1/2/2/2/7. The fragment in front of A 47 may be from a lost

Canon on 'increase' pairing with the one on 'reduction* (A 46), as suspected
by Sun. But again there is no Explanation.
1/2/2/2/8. The sequence A 89-92 is a displaced fragment, probably

27
from Names and objects.
1/2/2/2/9. B 25 seems to combine the body of one Canon with the
28
summing-up of another; both have Explanations. The two Canons, which
we number B 25a and 25b, would have been fragmented by the same
accident which dislocated B 14b-24a.
1/2/2/3 The Canons at Stage 2

The layout of the Canons at Stage 2 can be recovered by writing the
Canons of the standard text so that they alternate in two rows, and using
the order of the Explanations to check which belong to the top and which
to the bottom row. We reconstruct it in § 3/3 below. Since editors continue
to ignore Luan T'iao-fu's discovery of Stage 1 the reconstructions printed
in modern editions are vitiated by attempts to recover the primitive text
without abandoning the division into two rows. T'an Chieh-fu prints
reconstructions of as many as three supposed stages, the last of which
29
however approximates to our Stage 2.
Parts A and B of the standard text both end with a pair of Canons
from the bottom row; there are also three other examples of consecutive
Canons from the same row without an interposed Canon from the other
row, A 40, 41 : B 3, 4 : 66, 67. Such exceptions demand explanation, since
we cannot suppose that arbitrary gaps were left in the columns. In none
of the three cases is there any possibility that the paired Canons could have
been mistaken for single Canons by the Stage 2 editor. There appear to
be only two possible explanations:
(1) A Canon has been lost at the corresponding position in the other
row. A n interesting implication is that since nine out of ten Canons have
30
Explanations, and the Explanations as we have seen date from Stage 1,
27
Cf. p. 109 below.
2 8
Cf. B 25 Textual note.
2 9
T a n (1958) 33-49.
3 0
Cf. § 1/2/2/1/4.
92 Textual Problems
we should expect to find the Explanation of the missing Canon surviving

at its proper place. In one of the three cases we do in fact find this:
B 26 (27) 28
66 67 68.
T'an Chieh-fu already recognised that the passage following the
Explanation of B 26 is the Explanation of a lost B 27. It is a remarkable
example of the general indifference to the overall arrangement of the Canons
that no editor since seems to have appreciated the force of his argument.
More recent editors attach the passage either to B 26 (Kao, Liu) or to
B 28, by ignoring the head character of the latter or transposing (Wu, Li).
But for anyone who has considered the layout of the Canons as a whole
T'an's suggestion is not merely one of several possible solutions; we are
bound to accept it as the only solution yet proposed which explains why
at Stage 3 the Canon of B 67 follows straight on to B 66.
(2) After the top row was written it might turn out that in some
columns there was not enough room for the Canon which belongs at the
bottom, one or both Canons being too long; it would then be written in
the next column or in both, pushing the two rows out of alignment. In
both Parts the bottom row does in fact continue one place beyond the
top row.
Now since a column had room for at least 32 characters (B 9 plus
B 49) this suggestion will not serve as an alternative explanation in the
case just discussed. B 26 (7 characters) has ample room below it for B 66
(10 characters), and in any case the gap is in the top row. But it does
account for our still unsolved problems. Take first the absence of a bottom-
row Canon under B 3:
B 2 3 4
(12 characters) (31 characters) (17 characters)
43 — 44
(8 characters) (11 characters)
The columns had room for at least 32 characters. But after the 31
characters of B 3 had been written there was no room to write the 11
characters of B 44 underneath; the latter had to be pushed along to the
31
next column. Once again the solution is an achievement of Luan T'iao-fu,
and once again he has been almost ignored by more recent scholars, who
have failed to see the textual problem which he was attacking. Apart from
Luan, most editors break up B 3, by attaching part to B 4 (T'an) or by
dividing it into two (Wu, Kao, L i u , Li) or even three Canons (Sun, Liang).
3 1
Luan (1957) 53.
1/2 93
The effect of splitting B 3 is of course to increase the number of apparent

gaps in the bottom row; yet no Explanation survives of even a single lost
Canon between B 43 and 44. The main reason for the general reluctance
to accept B 3 as a unity is probably the impossibility of reading it as one
sentence if we take the phrase as its beginning. But we have already
concluded on other grounds that at Stage 1 this phrase belonged to B 2,
32
so that the difficulty disappears.
It may be noticed that except for the mutilated B 42 every Canon in
Part B has a summing-up in the form 'Shuo tsai t&fi: X* and that B 3 has
only one summing-up. The textual evidence is strongly in favour of the
unity of B 3, and we must deny ourselves the liberty to break it into bits
which we think will make better sense than the whole. We may draw the
further conclusion that a column did not have room for 42 characters
(B 3 plus B 44).
Turning to the problem of A 40 and 41, we find under A 39 a mutilated
Canon of 7 characters, A 88. This ends with the first of a long series of
examples taken up in the Explanation, all of which (to judge by the analogy
of A 76-87 and all similar series) should be listed at the end of the Canon.
If so, it was exceptionally long, about 32 characters:
A 39 40 41
(8 characters) (5 characters) (5 characters)
88 — 89
(32 characters) (5 characters)
If as we have calculated a column did not have room for as many as
42 characters (and not necessarily for more than 32), we can understand
why A 88 could not be written in full in its own column and would overflow
into the next, usurping the space under A 40.
We can now explain why both Part A and Part B of the standard text
end with two items from the bottom row:
A 1 39 40 41 49 —
5 0 . . . . 88a (88b) 89 97 98
B
1 2 3a 4 41 —
42 43 (3b) 44 81 82
In both Parts the unavailability of one space has pushed the rest of
the bottom row one place forward.
There is one instance of a supposed transposition from the bottom to
the top row, first proposed by Liang Ch'i-ch'ao. A Canon on similarity
(A 39) stands immediately above the last of three on similarity and differ-
3 2
Cf. § 1/2/2/2/3.
94 Textual Problems
ence (A 86-88), and at first sight it is tempting to suppose that it belongs

to the same group. But the proposal is only another illustration of the
reluctance of scholars to see the organisation of the book as a whole. Luan
33
T'iao-fu already saw that the transposition is quite impossible. The com
mon topic can only be chance, since A 39 is a definition ending my eh *ife like
all Canons of A 1-75, while A 86 and 87 belong to the lists of meanings
of ambiguous words (without final yeh) collected in A 76-87. Moreover it
is difficult either to conceive a purpose for transferring A 39 to the top row
or to understand how it could happen by accident. How is it that the bottom
is still the longer row after the loss of one Canon to the top ? We should
have to suppose that all Canons after A 39 in the top row and before A 88
in the bottom row have moved one place forward, and that one has been
transferred from the end of the former to the beginning of the latter.
Finally, we should have to suppose that the order of the Explanations was
also readjusted, which did not happen in the case of the known transposition
in Part B.
1/2/2/4 The two parts of the Canons and Explanations
There is no natural break between Part A of the Canons in Mo-tzu

ch. 40 and Part B in ch. 41. The series of propositions in the art of
description runs consecutively from A 88 to B 12, and inside it A 97 pairs
with B 1. The number of columns at Stage 2 is suggestive; in Part A , the
longer, they amounted to exactly 50. The rearrangement of the Canons in
two rows at Stage 2 would, if we assume that the number of characters to
a strip remained as before, more than double the number of strips required.
We may guess that at Stage 1 there was room for all Canons on a single
scroll, and that they were divided at Stage 2 into two scrolls, the first and
longer containing 50 strips.
There is however an important difference between the two Parts.
From B 1 each Canon ends with a phrase in the form Shuo tsai Iftffi X
("The Explanation is in X " ) , which we call the summing-up. The connec
tion with the Explanation is not always easy to establish, which suggests
that the summings-up were originally reminders of oral expositions and
are older than the written Explanations (which themselves belong to Stage 1,
since they preserve the correct order of Canons dislocated at this stage).
That the summings-up belong to Stage 1 is in any case clear from the
counting of their characters into the 66 and 35 of the misplaced strips of
33/34 characters (B 14b-21, 22-24a).
3 3
Luan (1957) 9.
1/2 95
Why is it then that they begin exactly at B 1, although they would be

as appropriate in A 88-98 (although not in the definitions and analyses of
ambiguous words in A 1-87) as they are in B 1-12? A plausible answer is
that originally the summings-up did begin at A 88, and that someone
excised them from the first scroll but for some reason left the second scroll
intact. Their deletion is understandable since they would have lost most
of their value with the writing down of the Explanations and the ease of
referring to the Canons after they were laid out in two rows at Stage 2.
Since the Han bibliography records a Mo-tzu in 71 chapters, the same
number as in the extant text, it would seem that in both Canons and
Explanations Parts A (ch. 40, 42) and B (ch. 41, 43) were already divided.
If we are right in placing the division at Stage 2, this confirms that the
arrangement of the Canons had arrived at Stage 2 by the last century B . C .
1/2/2/5 The head characters of the Explanations
1/2/2/5/1 Sun Yi-jang already recognised that the first character of

an Explanation is very often its heading. Later Liang Ch'i-ch'ao in his
Mo-ching chiao-shih made the very important observation that
the heading is always the first character or characters of the Canon itself,
and that even Explanations without obvious headings generally begin with
the first character of the Canon. He inferred that every Explanation took
its heading from the first character or characters of the Canon irrespective
of meaning, and therefore that the first character is not to be read as part
of the Explanation even when it is a particle such as yi Ek, yii W: or pu ^
This last suggestion was greeted sceptically by H u Shih (in the postscript
he wrote for the Mo-ching chiao-shih), but has since won general acceptance.
Liang's discovery provided the first objective test for locating the start of
an Explanation. However, modern editors still assume that the head
characters are often missing or transposed, and that they are at liberty to
divide the Explanations on other grounds. This assumption allows much
latitude for disagreement over the placing of divisions. I wish to show that
the head characters obey quite strict rules, which enable us to decide
confidently between proposed divisions in nearly every case.
1/2/2/5/2. (1) The head character sometimes stands not before but
after the first character of the Explanation. Editors frequently advise us to
reverse the first two characters—T'an Chieh-fu, for example, as many as
ten times (T'an A 32, 39, 54, 93. B 5, 13, 63, 66, 78, 79)—but without
noticing this as a recurrent feature which requires explanation. Why does
the head character often appear one place too late, but never apparently
96 Textual Problems
one place too early ? The interest of this question is that it enables us to
guess how the head characters were distinguished on the strips, where they
could not have fulfilled their function if they were already submerged in
the surrounding text. The head character must have been written at the
side of the first character, so that it could from time to time enter the text
below in stead of above it; the Explanations must have been written
consecutively with their beginnings marked only by the head character in
the margin, since the mistake would have been impossible if they had been
separated by intervals. Consequently we no longer need to treat the
transposition of the head character as a conjectural emendation. When
editing the text we can reserve the right to bracket either the first or the
second character of the Explanation as head character.
1/2/2/5/3. (2) The Explanations of the transposed Canons in B 14b-

24a all stand in their original order and with their head characters. The
head characters, like the Explanations themselves, therefore belong to
Stage 1.
1/2/2/5/4. (3) There are few if any exceptions to the rule that an
Explanation has a head character at first or second place. Editors have
continued to suppose that head characters have shifted or disappeared,
without noticing that their own work has undermined this assumption; if
one compares the various modern editions it is quickly evident that almost
every head character in a total of 180 has been identified already by one
scholar or another, and that it is in its right place. Even among the few
apparent exceptions we find that there is a head character which has
escaped notice because it is at the second place and is a negative or numeral
which fits easily into its context (B 2, 4, 67, 75, 82).
The head character is normally present even when the Explanation
happens to begin with the same character (A 4-6, B 8, 24, 40, 54, 76).
There are however three cases of one of the repeated characters being
dropped (B 2, 64, 69). The single instance of a head character being
displaced (B 46) is quite exceptional, due to entanglement of the head
34
character in the margin with a marginal gloss. With these exceptions all
divisions between sections in the present edition are made on the assump
tion that there is a head character at first or second place. The presence
and position of a head character is one of the most stable and dependable
features of the text, and in most cases should be taken as the decisive test
in the controversies over the placing of divisions which still harry students
of the Canons.
3 4
Cf. § 1/2/2/6/2.
1/2 97
1/2/2/5/5. (4) There are occasionally two head characters, but it is

unlikely that there are ever more. In a few cases Canon and Explanation
share more than two words at the beginning (cf. A 88, B 61), but on exami
nation the surplus words turn out to belong to the Explanation itself.
Generally there is only'one word in common, even when successive Canons
begin with the same word (cf. B 17-21, five Explanations with the same
head character). But in a few cases there are two-word headings:
A 3 *P*t (cf. A 4 » ) .
A 62 ^T«|ffl (cf. A 63 • H ) .
A 88 mm (cf. A 86 IRI, 8 7 %
B 3
B 38
I have failed to find a clear principle behind these exceptions. The
scribe perhaps began with the idea of avoiding the same single head
character in successive Explanations, but abandoned it after A 3. The
other phrases all stand out very sharply as syntactic units in their Canons.
(In A 62 'Without interval* is the geometrical term being defined, and in
A 88, B 3, 38 'Being the same or different', 'Having the same name* and
'Pointing something out* are all preliminary phrases detached from the
rest of the sentence.) We may picture the scribe as mechanically copying
in the first word or pair of words, but rarely perceiving words as paired
because he is only intermittently aware of the sense of the text.
1/2/2/5/6. (5) When a Canon has no Explanation there is some

danger that its first character may occur fortuitously in the previous
Explanation: we should then be tempted to mistake it for the head character
and the succeeding passage for the missing Explanation. This danger is
especially serious in the definitions of A 1-75, since each Canon begins
with a technical term which may occur quite frequently in neighbouring
Explanations, the sections being grouped by topic.
A n innovation of the present edition is that it twice rejects head
characters which have been commonly accepted since Liang Ch'i-ch'ao
first proposed them:
(1) Liang found the head character of A 32 (B*) embedded in the
phrase 45cff-tfa#, from which it cannot be detached without drastic
35
emendation.
(2) He found the head character of A 61 ($&) in the middle of a
sentence in A 60, forcing him to split one intelligible Explanation into two
unintelligible ones. But this belongs to a series of geometrical definitions
8 6
Cf. A 32 n. 104.
98 Textual Problems
(A 52-69) of which three have no Explanations (A 52, 56, 57) and three
more have Explanations which use the word tuan (A 63, 67, 68). The
presence of tuan is therefore probably accidental.
1/2/2/5/7. In view of the complicated textual history of the dialectical
chapters the almost perfect preservation of the head characters is a very
36
remarkable fact. We shall see later that it has an explanation which will
partly qualify our satisfaction with the result. The head characters, at any
rate as we have them, were added after the text had already been mutilated.
However, the area of serious fragmentation is limited to A 22-39. In any
case it remains true that the identification of head characters and dividing
of the Explanations are problems to be solved on textual grounds, not on
subjective considerations of which division makes the best apparent
sense.
1/2/2/6 Glosses
1/2/2/6/1. A recurring feature of the Explanations is the illustration
introduced by jo 'Like, for example . . .', sometimes an integral part of
the sentence, more often itself a detached sentence standing at the end of
the Explanation or one of its parts. In some cases it is clearly parenthetic;
thus the phrase ^MftSlfi "For example, a circle is nowhere straight" in
A 98 interrupts the whole organisation of the sentence and invites confusion
with the concluding phrase 3? "it is as though it were so of itself".
T'an Chieh-fu noticed that in B 10 the illustration Wfflft "Like stone
and feathers" stands six places too late. (Cf. also B 11). There is a very
interesting displacement in B 45, 46, immediately preceding what I believe
37
to be the only displaced head character in the whole document:
1/2/2/6/2. B 45 a U T f i f f n a S o
B wxm^& °
46 » # ( ^ S B p i ± [ ± ] K 8 H b ) ( « ) B i a M M a
"(B 45) . . . Moreover there are cases where we have increase only after
reduction. (For example, the malarial crisis in the case of malaria.)
(B 46) The knower sees with the eye and the eye by means of the fire but
the fire does not see. . . . "
Clearly the illustration stands two places too late; but how did it come
to push down the head character of B 46 in front of it ? We have noticed
evidence that head characters were originally written in the margin opposite
3 6
Cf. § 1/2/2/7/3.
37
Cf. § 1/2/2/5/4.
1/2 99
the first character of the Explanation, and that there was no interval
38
between one Explanation and the next. Presumably in this case the
illustration was written in the margin immediately above the head character
of fi 46. The copyist who incorporated them into the text assumed that
the head character belonged to the illustration; this left him without any
indication of a new Explanation, and he wrote in the whole two places
too late.
1/2/2/6/3. We may conclude therefore that some and perhaps all of
the detached illustrations are marginal glosses, and that there is always
some risk of misplacement. I was at one time convinced that some illustra
tions are so far from their proper places that they must have been written
39
in from a separate document. Sun Yi-jang had already proposed that the
illustration to B 53 belongs to B 36, and it seemed to me obvious that the
final illustration of A 1 has been transposed from A 6. But I have since
come to appreciate that both are after all intelligible in their contexts, and
that I had been guilty of the self-indulgence of emending to suit my own
convenience which I criticise so freely in others.
1/2/2/6/4. There are also a few examples of detached clauses begin
4
ning withjw M It is as with . . (A 32, B 65, 66). These are unlike the jo
illustrations in that they refer to examples mentioned earlier in the
same context. But it is likely that they too are glosses, since two are
parenthetic:
A 32 f i m , ( » £ ) , Sc-lb ·
"To say calling by name (as in the case of the stone) is to convey.''
B65 m$L,
"If they have all the characteristics (as in the case of being square), the
things are all so."
1/2/2/6/5. There are other parenthetic expressions which may be
glosses, such as the possibly misplaced A^S$ "that is, not self-refuting"
of B 71. However, the problem of glosses is important only in a text with
a long tradition of study. Any glosses in the Explanations probably belong,
like the text itself, to the last century of the Mohist school, and we need
not hesitate to accept their testimony. It is sufficient to mark suspected
interpolations by brackets, in order to detach them from their contexts and
call attention to the possibility of misplacement.
8 8
Cf. § 1/2/2/5/2.
8 9
Cf. G(5) 24 n. 1, G(12) 164.
100 Textual Problems
1/2/2/7 Fragmentation
1/2/2/7/1. The Ta-ch'ii and Hsiao-ch'ii /WR, the former of
which is palpably no more than an assemblage of mutilated scraps, will be
examined in the next section. The Canons show little evidence of fragmenta
tion, except for damage to B 14, 24 and 25b which may be connected with
the transposition of B 14b-21 and 22-24a at Stage 1, and the loss of the
summing-up of B 42, which at Stage 2 stood at an exposed position at the
bottom of the extreme right-hand column of Part B.
1/2/2/7/2. The organisation of the Explanations raises two initial

questions. Why do some Canons lack Explanations? When they are missing,
why is the head character sometimes written in and sometimes not ? In the
first place we may notice that the lacunae are almost confined to the
definitions of A 1-87, some of which may have seemed to require no
comment. But the propositions of A 88-B 82 make no claims to be self-
explanatory, and from B 1 have the summing-up formula referring to an
oral or written explanation. There is in fact only one clear case of a
proposition without its Explanation (B 11), although another is commonly
but on the present arrangement wrongly assumed in B 75. There can be
no doubt that the Explanation of B 11 is either lost or transposed. We
identify two fragments of it in the scraps packed in at the beginning and
end of B 43.
1/2/2/7/3. Turning now to the definitions, we find:
(1) At 10 places there is neither Explanation nor head character
(A 52, 56, 57, 61, 72, 84, 89-92). The longest sequence is A 89-92, which
do not appear to be Canons at ail, but four consecutive sentences misplaced
4 0
from Names and objects at Stage l . In A 84 every word requiring comment
has already been explained in E C 8 or defined in A 26, 27, 33. The defini
tions of A 52, 56, 57, 61, 72 are easily accepted as self-explanatory.
(2) The head characters of A 23-25 are written consecutively, with
the first two Explanations missing and only two words surviving from the
last. The head characters of A 31, 32 "8") seem at first sight to be
missing, but can be recognised squeezed together as # , at the head of
what looks like the mutilated Explanation of A 32. The head characters of
A 36, 37 are also written consecutively, but this time A 36 does have an
Explanation transposed to the end of A 37 (But the Explanation of A 36,
as also of A 38, is identical with the Canon). On closer inspection nearly
all Explanations from A 22 to 39 (and also A 12,14) turn out to be mutilated,
4 0
Cf. p. 109 below.
1/2 101
lost or transposed; yet the complete sequence of head characters is intact

and in its right order. There can be little doubt that the head characters
as we have them were written in after the dislocation of this part of the
text. Probably the Explanations were at first marked off from each other
by some device but without any indication of their Canons, and it was this
textual disaster which impressed their next transcriber with the urgency
of providing them with headings. Fortunately there is no evidence of
mutilation in this scale outside A 22-39.
1/2/3 T H ERECONSTRUCTION OF 'EXPOUNDING T H ECANONS'A N D
' N A M E S A N D OBJECTS' F R O M T H E T A - C H ' U A N D HSIAO-CH'U
1/2/3/1. The last two of the dialectical chapters have contrasted

titles, Ta-ch'ii and Hsiao-ch'u, 'The bigger pick' and 'The smaller pick*.
The Hsiao-ch'u is a consecutive treatise on the art of description, the most
easily intelligible of all the later Mohist documents. The Ta-ch'u on the
other hand is a random collection of fragments, some concerned with
logical problems in the manner of the Hsiao-ch'u, others with ethical
questions raised by the doctrine of universal love. O n both topics the
argumentation is very sophisticated; but even the longer fragments break
off before we can catch the drift of the argument, with the result that
almost the whole chapter has remained virtually unusable. Of the many
modern editors of the Canons and Explanations only T'an Chieh-fu (in his
admirable Mo-pien fa-wei MHIlM) and Chang Ch'i-huang 36ftfii have
ventured to attack the problems of the Ta-ch'u, and commentators on the
whole text of Mo-tzu have offered only minor contributions.
41
Two of the fragments start with phrases which look like titles or
headings (№'&, the former already recognised as a heading by
42
Sun Yi-jang. It is therefore likely that the Ta-ch'u contains the mutilated
remains of more than one document. The similarities of content and
contrasted titles of the Ta-ch'u and Hsiao-ch'u also suggest the possibility
43
that they are 'The bigger pick' and 'The smaller pick' from a single
4 1
T C 2B/9, 4A/2.
4 3
Sun 255/6.
4 3
For the meaning of ta cKii and hsiao ch'ti, cf. Kuan-tzu J- (ch. 38) 2/70/2
l
jfcSH$B&. /jNfti*H/hfMB. ^ * S f i J ^ # ® , mffZMJiTM- "This is called
the Way. If you take a little from it you win a little good fortune, if you take much
from it you win much good fortune, when you practice the whole of it the Empire
submits". I previously followed Pi Yuan in supposing that the chapters "were
evidently misnamed by someone who read no further than a discussion of choosing
corpus, 'The smaller pick* consisting of more homogeneous material

sufficiently well preserved to be recognised as an entity. If so there is
danger that bits of the Hsiao-ch'u source may have slipped into the rag-bag
of the Ta-ch'u, and the lacunae stuffed with material which might be from
any of the sources. Fortunately the Hsiao-ch'u is an exceptionally closely
knit treatise, but there is some disjointedness near the beginning, as well
as a sentence broken in two by the intrusion of a small fragment followed
44
by 20 characters repeated from earlier in the chapter. The proposition
" A man's ghost is not a man, but your elder brother's ghost is your elder
brother" appears both in the Hsiao-ch% where the context is suitable, and
45
in the Ta-ch'u, where it is not.
The manner in which fragments sometimes begin and end in mid-
sentence and are arbitrarily juxtaposed may be illustrated by the following
four passages, in which we mark the proposed divisions by a stroke:
A B
T C 4A/2 » B A 9 f 3 f e * A » / * K « f f ^ « .
C D
4B/5-7 5 f c » / J B # A # « K » # g » № ^
E F
5B/2-3 **^®^7K*/i^4Sc*aaftR«ft*«.
G
5B/5-6^»K*ff#*.
It will be seen that A can be joined to D to make a sentence (in which
the graph i& has been corrupted to 3a) parallel with the next sentence in D ;
and that similarly F follows on to C to make a sentence with an intact
parallel in G . Clearly the wooden or bamboo strips on which the text was
written have been displaced and mutilated, and the fragments copied in an
arbitrary sequence. Commentators have made various tentative suggestions
46
for linking disconnected scraps in this way. But when we attempt the
total re-ordering without which the fragments will remain unusable, we
the greater benefit and the lesser harm early in the Ta-ch'u" (G(5)l cf. Pi 11, 1A/3).
But in that passage the order of words is reversed: ch'ii ta "choosing the greater"
ch'uhsiao "choosing the lesser" (EC 8), and the resemblance to the title is probably
accidental.
4 4
NO 13.
4 5
EC 5, NO 18.
4 6
Cf. Sun 254/5, 8, 255/1, 5, -5, -2, 256/3, 6, 7, 257/-4, -2, 258/-2, 259/7.
Chang Ch'i-huang (op. cit.) has proposed a rearrangement of the Ta-ch'u as radical
as the one offered here, but vitiated by failure to find any guiding principle.
1/2 103
soon find ourselves involved in a problem complicated by too many

imponderables:
(1) Much of the text has certainly disappeared; this can be seen from
every other instance where parallelism offers guidance. There is no phrase
in the Ta-ch'ii which will fill, for example, the gap at the beginning of
T C 3A/9 . . . M r t t o f 8 « t » . Moreover, in spite of occa
sional broken phrases of this kind, it would seem that the compiler tended
to discard the remains of lost sentences at the beginning or end of a
fragment. The overall impression made by the Ta-ch'u is of an aggregate
of intact sentences in no discernible order.
(2) There is considerable difficulty in distinguishing the effects of
fragmentation from the change of topic inside a single fragment which in
modern typography would be marked by paragraphing:
T C 4B/2-3 -B7!/£fFn^r.B^
If this is a consecutive (although no doubt corrupt) passage we must

4
take the three words ssu yiieh ch'iang as one clause, 'The fourth is called
'strong' ". But if the confusion is due to fragmentation there is no reason
to locate the break at an apparent syntactic pause; we cannot be sure that
the first fragment extends farther than the last yiieh "is called ".
(3) However casually the fragments were assembled it is likely that
the compiler was to some extent guided by their content. Thus there are
three definitions parallel in form ( T C 1A/9-10, IB/7-8,2A/6-7), separated
by almost equal intervals, and followed by explanatory material which
however in each case seems to break off before we reach the next definition.
We may suspect that the compiler juxtaposed fragments in which he
observed parallel definitions, but did not necessarily get them in the right
order. This awakens the suspicion that apparent continuities may be the
result of the compiler juxtaposing fragments which quite fortuitously have
a word or phrase in common:
T C 2B/5 &h±Mfol¥№h£&&.
Syntactically this makes a quite acceptable sentence: "It is that he
desires the benefit of others, it is not that he dislikes the harm of others".
Yet if any idea emerges clearly from the Ta-ch'ii fragments it is that desire
for benefit and dislike of harm are equivalent. The doubt arises that the
two phrases belong to different fragments, juxtaposed because of an
accidental resemblance. Desire for benefit and dislike of harm are concepts
so common in Mo-tzu that it would not be in the least surprising if the
former appeared at the end of one fragment and the latter at the beginning
of another.
(4) If the displaced strips were intact we should have a good chance
of identifying the number of characters to a strip, which would establish
a firm basis for rearrangement. We have noticed that there are half-a-dozen
transpositions of intact strips in Mo-tzu, which show that at some stage in
47
its transmission the number of characters to a strip was about 40. But
unfortunately many Ta-ch'u fragments are quite short, implying that the
strips were not only disarranged but broken. A sufficient proportion of
broken strips would destroy any uniformity in the number of characters
to a fragment; even when a series of strips survived intact, one broken strip
at the start or end would make the total diverge from a multiple of the
number (which may not have been 40 in all documents which found their
way into Mo-tzu). The number of characters to a fragment varies (on the
present estimate) from 201, 179, 119, 102, down to 14, 12, 10, 9, 7, 5.
A problem with so many imponderables might well be quite insoluble.
But it happens by a fortunate accident that the Ta-ch'u contains an
unnoticed key to its reconstruction. This is a list of 13 propositions on
ethical questions, each with the concluding formula lei tsai $S & X "the
analogy is in X " which survives intact at the very end of the chapter.
The formula resembles the shuo tsai tfe & X "the explanation is in X " at
the end of each Canon in Mo-tzii ch. 41, referring to the Explanations
assembled in ch. 43. Commentators have therefore speculated that origin
48
ally the 13 propositions must also have had written explanations. Now
there are many close similarities in content and phrasing between the
propositions and fragments earlier in the chapter, nine of which were
49
already noticed by Sun Yi-jang. Is it possible that the fragments con
cerned with ethics are the remains of the missing explanations ? A simple
way to test this possibility is to write out the propositions in sequence on
different pages of an exercise book, cut up a xerox copy of the Ta-ch'u at
the points where fragmentation is most evident, and group the pieces under
the relevant propositions, making further cuts and rearrangements as the
experiment progresses. It must be admitted that the fragments seldom if
ever connect with the analogies provided in the lei tsai X formula, as the
Explanation of a Canon connects with its shuo tsai X formula; this con
sideration indeed deterred me from trying the experiment for many years
after the possibility first occurred to me. But if we decide that it is worth
while to test the possibility in spite of this difficulty (to which we shall
4 7
Cf. § 1/2/2/2/1.
4 8
Cf. Sun 259/-5, Wu Yii-chiang 11, 9B/-3.
4 9
Sun 259/13-260/8.
1/2 105
50
return later), we find that all the fragments concerned with ethics quickly
begin to sort themselves out like pieces of a jigsaw puzzle, revealing them
selves to be a series of expositions of the 13 propositions, some apparently
almost complete, others badly mutilated, and in the same sequence as in the
list of analogies (evidence of mutilation increases towards the end, and the
exposition of No. 13 seems to be entirely lost). Such a reconstruction must
of course be uncertain in detail, and a student of pre-Han philosophy with
a taste for this kind of game might be advised to put aside the book at
this point and try the experiment for himself; an independent try without a
knowledge of my results might well hit on an arrangement at some points
demonstrably superior. But several considerations confirm that the general
principle is sound and that the success of the experiment is not accidental:
(1) The consideration which seems decisive is that in five cases
expositions follow on to each other inside a single fragment, and that in
each case the order is that of the list of the 13 analogies:
Propositions 3, 4 T C 5A/2-6 66 characters
6, 7 IB/4-8 63
7, 8 1A/3-1B/4 201
8,9 2A/3-9 102
9, 10 2B/7-9 35
These five fragments (the fourth of which actually quotes Proposition
9 verbatim) provide a firm framework for the reconstruction.
(2) The three parallel definitions scattered near the beginning of the
Ta-ch'u, which tantalise us by implying some principle of organization
51
which we cannot grasp, as well as a sentence elsewhere which refers to
52
one of them, turn out to be the introductions to successive expositions
(Nos. 7-10). Each stands in the middle of a fragment:
Proposition 7 ( T C IB/7), in T C IB/4-8,
8 ( T C 1A/9), in T C 1A/3-1B/4,
9 ( T C 2A/6), in T C 2A/3-9,
as well as the sentence which refers to the last definition:
Proposition 10 ( T C 2B/8), in T C 2B/7-9.
It can now be seen that the compiler recognized the parallelism of the
three definitions and therefore juxtaposed the fragments (we shall consider
53
shortly the single interruption, T C 1B/9-2A/3), but in the wrong order.
5 0
Cf. § 2/3, EC (Appendix) below.
5 1
Cf. p. 103 above.
5 2
EC 10, based on the definition of lun lieh in E C 9.
5 3
Cf. p. 107 below.
(3) The exposition of the first proposition turns out to begin with
one of the two embedded headings which we have already noticed, Yii
5 4
ching f § S . Sun and T'an took this as an adjunct/head construction
("The canon of sayings"), but we can now see that the 18th century
commentator Pi Yuan was right in analysing it as verb/object ("Expounding
the canons"). We do in fact find the verb yii 'expound' in the exposition
of the first proposition. So Yii ching is the title of the whole document.
(4) Having established the starting-points of Nos. 1, 4, 7-10 and
the conclusions of Nos. 3, 6-9 one finds that all remaining fragments
concerning ethics (but not those concerning logic) can be fitted comfortably
into the gaps. They do not however fill them; much of the text is irretriev
ably lost. Indeed if it were complete we should expect all the 12 transitions
between expositions to be inside single fragments, apart from any cases in
which transitions happened to coincide with the ends or breaks of writing
strips. The evidence for placing the fragments inside the framework will
of course be of variable weight, and the critical reader, if he accepts the
argument up to this point, may still make other choices in dividing and
distributing the fragments. But even when the content of a fragment does
not at first sight connect with any of the 13 propositions it is reasonable to
postulate that it belongs somewhere in the document. In general the
connexions in matter and phrasing are as close as between the Canons and
their Explanations; this point being granted, the fact that all ethical
fragments can fit into a context in one or other of the 13 sections becomes
evidence that they do. If we tried to distribute all the ethical fragments
among any sequence of 13 in the Canons we might have a few fortuitous
successes, but we would not find plausible positions for all of them.
Having secured a foothold in the apparent chaos of the Ta-ch'ii, let
us reconsider the question of whether at the time of the dislocation the
number of characters to a strip was 40, as elsewhere in Mo-tzu at some
stage in its transmission. There are grounds for thinking that this question
is answerable, although for reasons which will soon emerge we shall take
no further account of it in the present reconstruction. When listing the five
55
framework fragments above we noted the number of characters to each,
56
66, 63, 201, 102, 35. It is remarkable that all these figures are within or
very close to the multiples of 33/34. We may add the evidence of the two
54
Cf. p. 101 above.
5 5
Cf. p. 105 above.
56
As emended in the present edition the numbers become 67, 66, 200, 102, 35;
all except the last are within the range of multiples of 33 and 34. But there may of
course have been corruption before the transposition of the strips.
1/2 107
transposed blocks of Canons, and of the four Canons which seem to be from
57
another source :
A 89-92 32 characters
B 14b-21, 22-24a 66, 35
It would seem then that in a certain damaged copy of the dialectical
chapters the number to a strip was 33/34, amounting to 100 characters to
every three strips. But unfortunately there is reason to suspect a further
dislocation at the stage when Mo-tzti, was written on 40-character strips.
There are two fragments with approximate multiples of 40 characters, and
these are suspiciously close together in the text as we reconstruct it:
E C 1 Fragment A (26) characters T C 2B/9-10
B 80 1B/9-2A/3
EC 2 C (19) 2B/4-5
D (12) 4A/1-2
E 119 4B/5-5A/2
F (9) 2B/10-3A/1
One explanation of the dislocation could be that C and D are scraps of
a broken strip and that B C D E were a sequence of six 40-character strips
dispersed by the snapping of the bands at this point, leaving nothing
between A and F . As a matter of fact F does follow immediately on to A
in the present text of the Ta-cWii ( T C 2B/10). Moreover if we look at the
present position of Fragment B in the Ta-ch'u we find that this suggestion
accounts for the anomaly at the beginning of the chapter which we have
58
already noticed. The compiler started off with the very longest fragment
at his disposal, followed by the two which are related to it by parallel
definitions. In between the latter he found a place for two scraps related in
subject matter. But in the text as we have it there is also a longer interrup
tion which we cannot account for on grounds of content, the 80 characters
of Fragment B :
T C 1A/3-1B/4 201 characters
IB/4-8 63
> 276 characters
IB/8 5
IB/8-9 7
(1B/9-2A/3) (80)
2A/3-9 102
It would seem that after the chapter was transferred to 40-character
strips Fragment B (two strips) intruded after the first seven strips of the
chapter (276 characters).
5 7
Cf. p. 109 below.
5 8
Cf. p. 105 above.
But if at one stage there were transpositions of 33/34 character and at

another of 40-character strips, not to mention the fragmentation of strips
and the compiler's apparent reluctance to copy incomplete phrases, the
practical application of these hypotheses becomes extremely complicated.
We can no longer take it for granted even that sequences of short scraps
must have resulted from the assembling of broken strips; the same effect
might follow from dislocation at the 33/34 and 40 character stages happen
ing to overlap (although under very limited conditions, the fulfilment of
which would be significant in itself). The present reconstruction is based
on the content of fragments, and ignores the problem of recovering their
position at different stages in the history of the text. But the problem may
not be insoluble for an inquirer stronger in mathematics than myself; if
successful, he would be able to reduce the possibilities of arbitrary choice
indulged in the present reconstruction.
There remain the fragments which discuss logical problems in the
manner of the Hsiao-ch'u. Some of these can be arranged in a sequence
with the aid of a resumptive passage, "In the first case we say that it is this
thing and is so, in the second that it is this thing but is not so, in the third
69
that it has changed in place, in the fourth . . , " . At the beginning of this
sequence one is gratified to find the other of the two embedded titles, Ming
shih "Names and objects". We have then the opening passage of a
second document. Examining the remaining fragments we find our suspi
cions increasing that the compiler could not distinguish between the Names
and objects and fragments detached from the Hsiao-ch'u source. Two
neatly dovetailing fragments about the tz'u l№ 'sentence/proposition* seem
to belong to the discussion of propositions in the Hsiao-ch'u, while an
apparently random series of definitions in the Hsiao-ch 'u turns out to have
60
a definite sequence when related to a passage in Names and objects. At
first one does not doubt that the Hsiao-ch'u source was a separate docu
ment, since the long passage on disputation with which it begins has the
look of a formal introduction. But this impression vanishes when one finds
reason to break up the passage by inserting material from the Ta-ch'u and
also, as we shall see in due course, from the Canons. Finally one is driven
to the conclusion that both chapters contain the remains of only two
treatises, Expounding the canons and Names and objects. 'The bigger pi
is simply a scrap-heap of fragments which the compiler despaired of
reducing to order (including the whole of Expounding the canons). 'The
smaller pick' consists of the pieces which the compiler succeeded in fitting
6 9
NO 4.
6 0
T C 4B/4-5 and 5B/2-6 (NO 9, 10); H C 6B/3-5 (NO 5).
1/2 109
together to make a consecutive chapter. It consists of 732 characters of

Names and objects (the longest fragment at his disposal), preceded by bits
of the same document joined together to make an introduction.
At the very end of the Ta-ch'u there is a disturbing example of the
confusion caused by juxtaposition of fragments which seemed to the
compiler to have something in common. The chapter concludes with the
list of analogies (lei S§) to the 13 propositions which is the key to the
reconstruction, but introduced by the word ku #C "therefore" and immedi
ately following a section from Names and objects on the analogies (lei) of
sentences. The first fragment presumably broke off at ku. The fortuitous
sharing of the word lei misled the compiler into establishing an artificial
61
continuity.
The fragments from the Canons which may come from Names and
objects are Canons A 89-92 and two characters in A 94 (10/2B/1). There
are strong reasons for thinking that A 89-92 come from another source:
(1) They have no Explanations; the sequence of the Explanations in
ch. 42 runs straight on from A 88 to A 93.
(2) They interrupt a consecutive series on the theme of similarity
and difference (A 88-96. Cf. also A 86, 87).
(3) Unlike any other Canons in A 76-98 they end in yeh , reverting
in this respect to the form of the definitions in A 1-75. Yet they do not look
like definitions, and the yen 'saying* of A 91 ("8", PiflJtfe "Saying is
fluency with the mouth") has already been given a true definition in A 32
("B", tBW-tfe " 'Saying* is uttering references"). Indeed they do not look
like Canons at all, but like four successive clauses in consecutive discourse.
(4) A 89-92 have 32 characters, approximating to the 33/34 of a
62
single intact strip.
(5) We twice find the word hsin *b 'mind*. It cannot be an accident
that such an important philosophical term appears nowhere else in the
Canons and Explanations and is absent from the 75 terms defined in the
first series of Canons. In its place we find chih *R\ 'intelligence/consciousness',
(defined in A 3), used just where we might expect hsin (cf. A 23
£fl$l$niil " 'Sleep* is the consciousness not being conscious of anything").
6Z
Hsin does however appear elsewhere in Mo-tzU, including the Hsiao-ch'u.
It seems clear that A 89-92 have strayed into the Canons by the
displacement of a single strip. The use of hsin shows that the fragment did
6 1
T C 5B/2-6 (NO 10) 5B/6-6A/7 (EC Appendix).
6 2
Cf. p. 107 above.
6 8
HC 7B/1, 8A/7 (translated NO 15, 16).
not come from elsewhere in the Canons and Explanations, and that if it
belongs to the dialectical chapters at all its place must be among the
disjecta membra assembled in the Ta-ch'u and Hsiao-ch'ii.
A little further on in the Canons, in A 94, we find after the unknown
graph W& (assumed by Sun to be IS na) two characters which in the Taoist
Patrology and all other early editions are printed small in the form of a
gloss, B ^ I I . One might take them as a phonetic gloss ("Pronounced ft");
but the reading li seems not to fit the graph, and there are no other notes
printed as glosses in the dialectical chapters. Sun and T'an emend one
graph and try to fit them into the text (the latter placing them in A 46).
We may suspect that they are a scrap from the same source as A 89-92
(in which we have just noticed a curiously similar phrase, n £ $ j ) ,
mistaken for a phonetic gloss and therefore written in after the next obscure
graph. Since the 32 characters of A 89-92 are one or two short of the
average for a strip (33/34), they could even be from the broken off end of
the same strip.
Evidence of fragmentation in the Hsiao-ch'ii is confined to the intro
ductory passages; the remaining three-quarters of the chapter is unques
tionably homogeneous and consecutive, constituting about half of what
remains of Names and objects. For the first half, we have the resumptive
passage which provides the key to the opening sequence; otherwise we can
only shuffle and re-shuffle the pieces until they assume a plausible order,
and any arrangement must be largely conjectural.
1/3
GRAMMAR
1/3/1 INTRODUCTION
P R E - H A N literature reflects the historical stages and the diverging

local dialects of a rapidly developing language, so that nearly every text
is to some extent peculiar in grammar as well as in vocabulary. Inside the
Mohist corpus, which accumulated over a long period (c. 4 0 0 - 2 0 0 B.C.),
we can discern pronounced general differences between the ancient core
(Mo-tzu ch. 8-37) and the dialectical and military chapters, less obvious
differences even inside the groups of cognate chapters. In the dialectical
chapters, as we shall see, the grammar of Expounding the canons differs from
that of the Canons, Explanations and Names and objects. We need to kn
as much as possible about the grammar of these writings since we can be
sure that an ordinary knowledge of Archaic Chinese usage, however wide the
reading experience on which it is based, will often be inadequate and even
treacherous when applied to texts dealing with such unfamiliar subjects as
logic, optics and mechanics. Obscure Chinese texts tempt us to force
meaning out of the apparently meaningless on the assumption that their
grammar is exceptionally loose. But this is a short-sighted indulgence;
only if their grammar turns out to be strict can we hope to read them with
any confidence.
The later Mohists, more than any other Chinese philosophical school,
were concerned with logical precision, to which grammatical precision is
an essential means. We should therefore expect in principle a grammar of
exceptional strictness, the study of which might be relevant to the whole
question of the possibilities and limitations of the Chinese language as an
instrument of analytic thought. But this general consideration hardly
prepares us for the precision and consistency of the grammatical system
as we actually find it. We may note in the first place the almost complete
absence of rhetorical usages. For example there are no exclamatory sentence
patterns; the exclamatory final particle tsai Sfe is absent, the final hu ^ is
purely interrogative, we find ch'i % as the possessive pronoun ('his, her,
its, their') but not as a pre-verbal modal particle. Indeed there are no final
112 Grammar
particles at all except for yeh ife, hu ^P, yii $i (= yeh hu -ti№) and
occasionally the perfective yi The only rhetorical questions are a
couple in Names and objects (3£^$IPP . . . "Why shouldn't I too . . .",
o*S . . . "How can I . . . ?"); otherwise all questions expect answers.
The manner in which particles serve only the logical structure of the
sentence can be seen in the use of the pre-verbal particles yi <fc yu% 'also,
again', which in many pre-Han texts contribute to idiomatic constructions
which are very difficult to analyse. In the dialectical chapters their functions
stand out nakedly and are sharply differentiated, with yi referring back
ward, generally to the subject, and yu (always written with the graph 'ft")
referring forward to the verb or object:
B 23 Bfftfc, « : * A o
"What is mirrored is big, and the shadow too is big."
B 33 ttSîfc-fc, ^ r » J & ± ^ 9 E № i b <>
"Knowing that the one in question is not this, knowing too that it is not
here "
The Canons and Explanations are perhaps unique among pre-Han
texts in having no pronouns or particles which are logical synonyms:
yii j& 'together with' Not chi R
5
yii H (= yeh hu i k ? ) Not yeh %
tz'u it 'this' Not ssu #r
tzU £J 'from' Not yu £
ho M 'what?' Not hsi %hom
wu §§ 'to, in, from what?' Not an i£ yen M
yi E. 'already' Not chi IS
ch'ieh JL 'about to' Not chiang Jrf
yu S 'still' Not shang fnl
fang ~)J 'just now' Not shih 31
tse № '(if. . . ) then' Not ssu #r chi BP
ku $fc 'therefore' Not shih-yiMM
erh houffD?£'only then' Not jan hou fk'ik
jo 'if, like' Not ju tn
chii ff: 'all' (external Not chieh W
distributive, pre-verbal)
chin ft 'all* (internal distributive) Not hsi Mi
The handful of apparent exceptions are in passages commonly agreed
to be corrupt (A 12 jr$, B 25 #P), or use the word in another sense (5S
'merely'), or allow an alternative interpretation which the uniformity of
usage entitles us to prefer ( A l l ftA(#P)*£n3 'enable others to know
1/3 113
one': A 46, "Like cutting": A 98 W(= fit)* A "enter together").

We have noticed some evidence that the later Mohists also graphically
distinguished some words which in the present text, as in other pre-Han
documents, are written with a single character, such asjyw J5| 'together with'
and yii S i = yeh hu -ifef, and particle and pronoun chih
There is no evidence of these restrictions i n Expounding the canons,
which uses chi t£ 'already' (EC 1, 8) and jan hou 'only after that*
(EC 2). This is one of our reasons for deciding that Expounding the canons
2
is the earliest of the documents. The establishment of a perfectly uniform
system of particles was no doubt a by-product of the great enterprise of
co-ordinating definitions and organising the principles of disputation in
the Canons and Explanations. We shall also notice occasional divergences
in the latest of the documents, Names and objects; but these have the look
3
of refinements and improvements to the system.
1/3/2 SENTENCE STRUCTURE
Writers in Western languages who find themselves in the position of

having to quote from these imperfectly understood documents seem often
to be translating on the assumption that they are dealing with fragmentary
notes rather than consecutive sentences. It may therefore be advisable to
state plainly from the start that the later Mohists wrote true sentences,
not less but more strictly organised than those of other pre-Han writers.
The Canons of one series in Part A (A 76-87) simply list meanings of a
word (A 76 T E J ^C, t " 'Finish': make complete, abolish"), and a few
in Part B are nominalised clauses exposed in front of the summing-up
formula (B 77 S ^ - t & i & f f i U I ^ "That learning is useful: the explana
tion is in the objector"); with these exceptions sentences can always be
identified when the text is intelligible at all. Even headings in the Explana
tions are tied to the succeeding sentences; when nominal they are treated
as subject (B 4 T —— J ^t@<& " 'One and two' do not fill each other")
or as exposed element resumed by a pronoun (A 1 [ /h#t J , ^ t e ^ f ^ ^ t
"In the case of the 'minor cause', when we have it this will not necessarily
be so" cf. also B 25, 39), when verbal they behave as preliminary clauses
(B 26 r ¥ J , ftlt^T "When you 'pull up', the longer and heavier is
below" cf. also B 11).
1
Cf. § 1/2/1/2/13, 14.
2
Cf. § 2/3 Introduction.
8
Cf. § 1/3/4/3, 1/3/11/4-7.
114 Grammar
As in Classical Chinese generally, it is convenient to distinguish

between the verbal sentence (with a main verb precedable by pu 'not')
and the nominal sentence (with a complement precedable by fei Ifc 'is not'),
although we shall see shortly that the nominal sentence presents certain
complications:
Verbal sentence
SUBJECT VERB OBJECT DIRECTIVE
(substitutable (preceded by yii
by chih 1£) substitutable by yen M)
B 18 ( ~ ) * & (—) #
"Two lights flank one light"
B 19 jft *
"It forms a shadow above"
B21 * A
"The shadow is bigger than the piece of wood"
N O 18 m (18) £
"Some horse is white".
Nominal sentence
SUBJECT NEGATIVE COMPLEMENT (followed by yeh ife)
N0 14(S)M Sife
" A white horse is a horse"
B 67 *
"It is not an ox".
Subordinate units (such as those bracketed) precede a sentence
position. The subject itself, although for present purposes it is convenient
to treat it as a sentence position, can be seen as a unit subordinate to the
sentence pattern as a whole.
The later Mohists are notable for the rigour with which they adhere
to the standard word-order of the verbal sentence. With certain exceptions
4
which will be considered later the directive is never exposed at the head of
the sentence, and instrumental phrases with yi &,, however lengthy, precede
the verb as subordinate units. Exposure is governed by a firm rule: the
exposed unit (or its subject if it is a nominalised clause) is in apposition
with a resumptive pronoun inside the sentence structure:
(1) Apposition with object chih :
B 57 © ± * t f e , E± . . .
"In the case of the pillar being round, when you see it . . . " (Cf. also A 1,
B 48, N O 1.)
4
Cf. § 1/3/12/4/6, 1/3/12/6.
1/3 115
In negative sentences chih is omitted and the ordinary negatives

5
replaced by fu and wu :
A 83 m%\ ffljfn^fc o
"The sagely, employ but do not treat as necessary . . ."
(Followed by two more examples):
B 27 J i m , ...
" A l l weight, if you do not pull it up from above . . ."
(Followed by two more examples, cf. also B 39):
We may also note a case of two exposed units resumed by adverbial
'huo . . . huo. . . .':
B53 < ? > , « , ^ B t « » ( = ^ ) A , M * i A «
"With (?) and with soup, in one case we show to others by means of the
name and in one case by means of the object."
(2) Apposition with directive yen:
B 25 •jtafiji . . .
"The cross-bar, if you put a weight on it . . ."
(3) Apposition with the ' X ' of the ' X chih 1£J substitutable for ch'i:
B 25 » , J t a K » K - » . . .
"The steelyard, if you put a weight on one side of it . . ." (Cf. also A 64,
B 39, 57, 74.)
(4) Apposition with complement shih or mou:
NO 2 mmm^, °
"As for those which name according to shape and features, such ones as
'mountain', 'mound', 'house' and 'shrine' are all examples of these."
"As for those which name according to shape and features, it is necessary
to know that this is ' X ' , only then does one know X . "
(5) Apposition with two subjects, the first qualified by shih:
B 53 mz.m& s » t e < £ > i & 4 ,
9 °.
"As for Yao being an example, this vocal sound is born in the present, the
object taken as example resides in the past"
In the absence of a resumptive pronoun, even a long nominalised
clause ending in yeh is not an exposed element; its subject is the subject
of the whole sentence. In the following example the subject is chih W
(generally written £P), 'the intelligence', a concept preferred by the Mohists
to the hsin *t> 'mind' of other schools:
6
Cf. For the extreme consistency of this rule, cf. § 1/3/5/1, 2.
116 Grammar
A 75 ...
" I f the intelligence in its consideration of it overlooks none of the harm
in i t . .
Leaving aside occasional syntactically obscure passages in which one
is tempted to postulate exposure, the only type of unit which is exposed
without resumption is the single word followed by yeh che *tfe# which
stands at the head of the sentence eleven times in the Explanations. This
is always a word quoted from the corresponding Canon, followed by an
explanatory sentence with which it has only a loose relationship of topic
6
and comment, as is especially clear in the four examples in A 3-6. In one
case it actually allows a second exposed element, this one attached to the
sentence structure by a resumptive ch'i:
A 64 T Jffi J i f c t , f B * ± H , °
" 'Empty*: of the interval between two pieces of wood, it refers to that of
it in which there is no wood." (The interval between two pieces may be
partially or wholly filled by another piece.)
Although the verb is consistently passive after k'o "I and tsu J£, in
accordance with ordinary pre-Han usage, passive sentence-patterns are not
admitted. The single case of a passive verb after wei % is in Expounding
the canons (EC 5). The pattern 'passive verb—yii 2ft—agent* is not found
at all.
1/3/3 'XisY'
The later Mohists as logicians are much interested in propositions of

the form ' X is Y \ For these they use the ordinary nominal sentence negated
by fei, and are punctilious in supplying the final yeh even in the negative.
But they also have occasion to negate fei itself. T o do this they have to
treat it as an ordinary verb, negate by pu and drop the final yeh. They also
have two other verbal copulae negated by pu and used without yeh, wei %
'constitutes, counts as, is deemed* and wei 1ft 'is specifically' (ox or horse,
as distinct from being an animal or thing). The sentence with copula thus
resembles a verbal sentence, but with the decisive difference that the
pronoun chih which substitutes for the object cannot substitute for the
complement:
SUBJECT COPULA COMPLEMENT
B67 * *
"The ox is not not an ox"
6
Cf. §1/3/10.
1/3 117
B2 SI
'All count as milu deer'
B3
B 12
B72
In philosophical as in other Chinese 'wet 1§ Y ' is not quite equivalent

to 'Yyeh\ Much as wet ch'en is to fill the role, perform the functions,
of a minister, so wet at jen S S A is to satisfy the conditions for being
deemed love of mankind:
N O 17 T J ^ J S S A , ffnftJSSA °
" 'Love of man* requires that one loves men without exception, only then
is it deemed to be love of man."
This use of wet is also frequent in the White Horse and Pointing
things out essays of Kung-sun Lung:
Kung-sun Lung tzu ch. 2 (Ch'en 65) J g * j 5 i £ S £ S , £ * * J £ j 8 £ °
"Horse not yet combined with white is deemed horse, white not yet
combined with horse is deemed white."
The Mohists as logicians must often discuss what X is or is not
without specifying what it is or is not. But they cannot combine fei or wet
with the relative pronoun so or object chih without rendering them transitive
('what he condemns/does' 'he condemns/does it'). Consequently they use
both verbs freely without complements, producing sentences in which at
first sight wet seems unintelligible and fei can be mistaken for fei 'wrong'.
When the complement is dropped, the Mohist has different usages
according to whether the implicit question is 'Is it X ? ' or 'What is it?':
(1) 'Is it X ? ' : shih 'It is', fei 'It is not'.
A 73 B 35 * i B ± * , * I B ± * °
"One calls it 'ox', the other calls it 'nori-ox'."
"One calls it 'this' (says it is), the other calls it 'not-this' (says it is not)."
N O 18 A ± A * A - t e , . . . jfc71-*ifiH##fe °
"The ghost of a man is not a man, the ghost of your elder brother is your
elder brother. . . . These are examples in which in one case it is and in the
other is not."
A 97 ^ S A * # f f n ^ * ·
"For example, the sage has respects in which he is not, yet he is."
118 Grammar
A 98 J£M1k . . . g i l l o
"The exact nowhere is not. . . . For example, a circle is nowhere straight."
B 8 fi&#teifnfiMg °
"The loan-named could not be loan-named an ' X ' unless it were not."
(Fei without complement also A 73, N O 2, 5, 6.)
(2) 'What is it?': nai shih 'It's this', pu shih 'It isn't this*.
(Cf. Chuang-tzu ch. 2 (Kuo 108/1) ?&^f& ° "Treat the not this
as this, the not so as so".)
B 8 2 * f t , f l № & r & J jgo
"As for the things which are not this, a 'this' is about to be recognised
as 'this' among them."
The other examples concern the problem of whether something so of
the object instanced is universally so of shih 'this' (what it is judged to be):
N O 4, 13 TjMn%& "So if the one instanced is this thing"
4 TbMM^ffe "Not so though the one instanced is this thing"
13 *^fvllffn$t "So though the one instanced is not this thing".
Wei M appears six times without a complement, and wei № once:
A 83 "What Jack is deemed to be" (Cf. E C 12 "Jack as
he is in himself").
A 89 "When a crane is deemed to be something . . .".
B 3 U K "When deemings are strung together . . .".
B 57 S £ ( = BS)J3tf8,*3tefcH!l "The place at which one deems is known
'a priori'."
NO 5 o £ r # : f , 0f E i J S i l & i b °
" A n example is a standard for being deemed something. What it exemplifies
is the standard by which it is deemed."
B 72 nmrnrnvmi °
"The caller has no thing which is specifically an X for the ' X ' it is called."
Both fei ^ and wei ^ may be preceded by hsiang 'mutually':
A 66 J l j f r F f c f i , ·
"Different places do not fill each other. Not being each other is excluding
each other". (Unique example.)
B 30 ummnn °
"Coin and grain are each the price of the other."
Wei ff£ is the pre-classical affirmative copula corresponding to the
7 y
negative copula fei. Its functions were taken over by the ' X Y yeh pattern,
7
Cf. Pulleyblank, op. cit.
1/3 119
but it survived before the subject ('It is X which . . .' narrowing to 'It is
only X which . . .') and the verb ('It is that . . .' narrowing to 'It is only
that . . .'), as well as in front of a complement ('is only'). Historically it is
difficult to determine the point at which it had ceased to function as copula
and had become a particle dependent on subject and verb ('only'). In the
sparse examples in the dialectical chapters (where the graphs 'Hi, °f£ gen
erally represent sut $8 'although') it appears to be a copula, since except in
A 55, B 4 6 clauses containing it do not have either a final y eh or a main
verb (B 11 "It is only a matter of convenience", B 12 f t S "It is
specifically this thing (ox, horse)", B 7 2 'ffi-SH "It is specifically what I
call it"). Syntactically it behaves quite differently from tu S 'only', which
> k
is limited to sentences with a main verb or yéh (cf. B 21 # ® / J < A > ' i È
"It is not only the size", B 38&$iB^ffi& "Be sure to point only at what
I mentioned"). Wei is significant in philosophical argument only in B 1 2
8
and 7 2 ; in both Sun Yi-jang took it as the expression of assent &fi wei 'Yes*
used verbally ('respond'). But the argument of B 7 2 seems to me to make
sense only on the present interpretation. The Mohists use chih i h 'stop' to
deal with a name which 'stops' in, is confined to, one thing (A 78, B 68) ;
wei would be suitable for dealing with a thing confined to one name.
In parallel ' X Y yeK sentences a contrastive not Tb sometimes precedes
the complement of the second ; in every case but one the first sentence is
negative (EC 2. N O 1, 3, 18. Cf. E C 5). Nat may also serve to carry an
equation one step forward, ' X Y y eh, nai Z yeK, ' X is Y , which is Z '
(EC 2).
Throughout the Explanations, when verbal phrases are linked by yeh,
we frequently find the pronoun shih M 'the said, the aforesaid' marking
the start of the complement. But it is not used between nominal units in
the ' X Y yeK sentence ; it is a resumptive pronoun clarifying the organisa
tion, sometimes of short linked phrases, sometimes of long sequences of
clauses.
1/3/4 PRONOUNS
1/3/4/1. The first person pronouns are wu and wo 5£ (the former

once subject in Names and objects but otherwise possessive); in place of
the second person tzù ^ is used; in the third cKi ft is exclusively posses
sive, chih object and yen directive (equivalent to yii chih fôi).
There are examples of chu ^ as a fusion of chih yù (B 16, N O 10)
but not of chih hu ^ ^ P .
8
B 72 has three more examples, discussed on pp. 133, 134 below.
120 Grammar
1/3/4/2. Of greater linguistic interest are the demonstratives. The

near demonstratives tz'u jlfc 'this' (the thing here) and shih Si 'this' (the
said thing) are only vaguely differentiated in most pre-Han writing, but
in the Canons and Explanations (although not in Names and objects)
distinction is unexpectedly sharp. Of crucial importance in the use of shih
is the fact that disputation is concerned with alternatives which are shih S
and fei and that these are not, as generally elsewhere, simply the 'right'
and the 'wrong'. Disputation is conceived primarily as debate as to whether
a particular object is 'the said thing' (shih, typically an ox, cf. A 73, 74)
or 'is not' (fei). The subject of debate may of course be a general concept,
and if it is particular what it is judged to be may also be singular ('North'
or 'South'). But in contexts of disputation shih will refer not to an object
but to what is meant by a name, in the case of class names not to an ox or
horse but to oxen or horses in general. In B 38 'know what it is'
contrasts with chih chih 'know about them'. In B 1, 2, where the
Canons are about lei S3 'kinds of thing', the Explanations use shih of the
kind which a debater rightly or wrongly selects in arguing from particular
instances. Here there is an explicit contrast of shih and tz'ii: the point of
the distinction may be brought out by translating shih as 'what it is judged
to be' :
B 1 tftBljlfcK«!t*lftft*»-i& *
"He, because of its being so of the thing here, explains that it is so of the
thing it is judged to be."
B 2 IQftft&S o
" I f what is so of the thing here is necessarily so of the thing it is judged to
be, all will be the 'milu deer'."
Cf. B 33 %\&±#ik-&, ti(=X)9a&±ttik&, f&ffimjk[mt] ·
(Of North and South, which change positions as we move.) "Knowing
that what this place is judged to be is not this place, knowing too that it is
not in this place, we none the less call this place 'the North' or 'the South'."
1/3/4/3. In Names and objects the distinction between shih and tz'u
has faded, no doubt because the specialised shih has been replaced by a
more satisfactory technical term, mou K ' X ' :
NO 2 MV&t&fc, iiS**o
"With things named according to shape and features one must know that
this is ' X ' , only then does one know X . "
1/3/4/4. Although in ordinary pre-Han discourse shih/fei and also

jan 'so' are used freely in judging proposals as right or wrong, we do
not find these usages in the dialectical chapters. A thing is the 'said thing'
1/3 121
(shih) and something is 'so' of it (jan) ; a. proposition is logically admissible

(k'o Hf) and fits the fact (tang Hi). However, we are once told explicitly
that shihjfei can be said of "discourse, conduct, learning or an object"
( A 88 traffic Jf, ^:#-tei), and they are used of right or wrong courses of
action in one Explanation and several times in Expounding the canons. Th
words when final are followed by yeh, suggesting that they are nominalised,
'the right/wrong one of the two'. Most examples follow wei % 'deem',
9
with shihjfei yeh as an embedded '(X) Y yeK :
B78 » » , № 4 0
"Although he criticises a lot his criticising is the right thing to do."
EC 1 R ^ ± ( = i S ) f t ( d H & ) . . . &A#3§(^-ifa) ·
"Expound Heaven's will as being the right one of the two . . . suppose
that what others condemn is the right one of the two."
EC 8 *
"Re-estimate the wrong as being the right one of the two, or condemn the
wrong as being the wrong one of the two."
1/3/4/5. In the case of pairs of contrasted concepts, shih marks the
one under discussion:
A 26,27 n&wm » mmm ° ° m&wM»m&w& · &m
"If you are pleased to get this one, it is this one which is beneficial; the
one of them which is harmful is not this one. If you dislike getting this one,
it is this one which is harmful; the one of them which is beneficial is not
this one."
y
From this example we may note that ch i ft 'their, of them* often
refers to the implied alternatives:
B 35 'the one'/ fti& 'the other of them*
A 4 6 ft## 'the one of them which remains'
B 5 2 ^ftl&iife^lM " I f we equalise the one of them which snaps, none
will snap".
1/3/4/6. With one near demonstrative a distinction is made between
the independent pronoun and the pronoun as adjunct. Tz'u ]th 'this' (the
thing here) is always the independent pronoun: A 9 6 Ifcjlfcj¥fjfc "choose
this, pick out the other", B 7 6 g$]iitife, J^SJJffiJffc-fe "The loving
and the benefiting are on this side, the loved and the benefited are on the
other side". The corresponding pronoun used as adjunct is chih: A 3 9
9
A pattern discussed on p. 155 below.
122 Grammar
1£r— 'this one thing', A 31, 78 'this name', B 71 £ A 'this man',

N O 18 £ l § 'this horse',~ZJ%*'this ox'. In one passage it can be seen to
contrast with adjunct shih:
A 78 T ^ J *mfo*G*}№*±*£fc°&±fi&\n&° riV* J&*»
&Blft£te ° ^ TMJ » · it***** ·
" 'Thing' is unrestricted: any object necessarily requires this name.
Naming something 'horse' is classifying; for 'like the object' we necessarily
use the name ' X ' . Naming someone 'Jack' is private: the name ' X ' is
confined to the object X . "
Here even the adjunct shih is confined to individuals or kinds, of which
one can say shih/fei 'It is/it is not'. The obstacle to using shih with 'thing'
is presumably that one cannot say 'It is not a thing'.
1/3/4/7. The far demonstrative pi ffe 'the other' is not simply a 'that'
contrasting with tz'u it 'this'. It can form pairs with personal pronouns
as well as demonstratives, in each case marking the other of the pair:
A 96 ftjltjfifc °
"Choose the one here, pick out the other."
A 31
"By means of this name refer to the other object."
A 16 j B & 2 : * * « - f e , ·
"Where doing the right thing interferes with the other he will not do it".
B i »ajikft^teie«*«&* > nszik&yttmM&mm ·
"The other man, because it is so of the one here, argues that it is so of the
thing it is judged to be; I, because it is not so of the one here, doubt that
it is so of the thing it is judged to be."
1/3/4/8. Pi refers to the other of alternatives; when the contrast is

with something indefinite, the word used is t'o "Hi (— ftb) 'another':
N O 15, 16 ftttbSfcjg *
"There is no other reason than this".
B 3sHSHMH-fe»mmmm ·
"If it is said of this thing, it is the case that in itself this thing is beautiful;
if it is said of another thing, it is not the case that this thing is beautiful."
1/3/4/9. When a demonstrative is the subject of a nominaiised phrase

it is often followed by chih in accordance with the ordinary rule that
the subject of a nominaiised phrase is either followed by chih or is the
possessive pronoun ch'i X . When this is found in parallel phrases the
contrast is between verbs or complements:
1/3 123
B 33 #(=X)»Jfe£?%lfcte . . .
"Knowing that what this place is judged to be is not this place, knowing
too that it is not in this place . . ."
B 80 . . . .
"In the case of this being this or not being this . . ."
But we also find a demonstrative followed by ch'i. When there are
parallel phrases it is the subjects which are contrasted:
B i №№tm№m&M:f№ °
"He, because it is so of the one here, argues that it is so of the thing it is
judged to be."
B 70 S M i l f e ^ f t f e o
"The colour of the thing inside the house is like the colour of this."
Cf. also E C 3 # » £ f ? £ - & o
"It is not that the performance of the other one is improved."
The resumptive ch'i ft emphasises the subject in the same way that
shih & and chih *L emphasise the inverted object (Tso chuan Hsiian 12/3
№S&&%. "It is only the enemy that we seek").
1/3/4/10. There remains one more demonstrative, the adjunct fu
This is used to mark a general concept: B 7 0 'names', B 7 2
'the crane', N O 10 5^8? 'the proposition', N O 9 'disputation',
N O 12, 13 ifetf 'things', N O 16 'to read a book'. In each case
the context shows that the reference is to names, cranes, propositions in
general, not to particular instances. There is also one example of jo fu
3gf5fc 'with regard to' (B 27).
1/3/5 NEGATIVES
The nine negative verbs and particles are those general throughout
pre-Han literature, and are very clearly differentiated.
1/3/5/1. (1) pu ^ and (2) fu pre-verbal negatives. The rule-of-
thumb that a verb negated by fu may be translated as though it had an
object chih *L applies without exception to the fourteen examples of fu.
While this is generally true of pre-Han texts of the classical period the
dialectical chapters are very unusual in not only confining fu to this type of
construction but also excluding pu from it, except when there is an inter
vening adverb. In the two cases where an adverb intervenes the negative
is pu:
A 4 "not necessarily find it"
B 3 8 ^ f g ® j t "unable to point at it alone".
124 Grammar
It may be noticed that in both cases the reference of the verb is

specified, by chih"k.'it* or by tu $S 'alone*. In the absence of an adverb
the reference of an objectless transitive verb is indefinite after pu, definite
after fu :
A 13 "not necessarily succeed"
B 16 3<m№ "not able to govern"
B 39 ^MB "not able to point it out"
B 4 4 ftfe "not govern it"
EC 2 "not able to get rid of it".
There is none of the stylistic reluctance to repeat fu in parallel
sentences so common in pre-Han literature:
B27 « S ( T » f t , M ....
"If you do not pull it up from above or gather it in from below or interfere
with it from the side . . ."
1/3/5/2. (3) wu ¥i and (4) wu ty negative imperatives. The four
y
examples are sufficient to confirm that, as in the other pre-Han literature

of the classical period, the relation is the same as between pu and fu :
A 83 %Ji& . . . ... °
"Do not treat it as necessary . . . do not doubt it . . . do not present one
of them without the other.**
B 37 mm^^m °
" D o not mention what I did not mention.**
1/3/5/3. (5) fou S , negative verb equivalent to the immediately

preceding verbal unit negated by pu ( B 1 2 ^ ^ S "in one case re
peated, in the other not**, B 3 4 ^ P ^ . S "know it or not**, B 73 3ftS . . . fi^lS
"fill or not . . . fill it or not**, B 7 8 " I S "admissible or not**). In other
texts it frequently stands alone in dialogue to express denial, and this seems
to be its function in one of the two cases where in the present text it is
written without the radical (A 9 4 ffc^f* "Yes or no**, cf. B 5 2 IS^P "snap
or not**).
1/3/5/4. (6) wei ^ 'not yet*. This twice follows a verb in the manner
of fou (A 8 9 "complete or not yet complete'*, B 3 "one or
other rejected or not yet rejected**), a rare usage for which Chou Fa-kao
10
JS^fSj notes only a single example (Shih chi ifefH ch. 107, 2844/3
B i i M c "Is it already finished or not?**).
1 0
Chou Fa-kao Mffift vol. 3, p. 251 n. Cf. also p. 161 below.
1/3 125
We may note the distinction between pu k'o chih ^ "Jftl 'unknowable*

(B 10) and wet k'o chih 'not yet knowable' (A 75, 94, B 58, 73: opposite of
hsien chih 5te$P 'know beforehand', A 93, B 57), which is observed
throughout Mo-tzu. Wei k'o ^ " J 'not yet admissible* appears only in
B 67; there it refers to positions not yet conceded in the course of the
argument and the conclusion is marked by pu k'o 'inadmissible'.
1/3/5/5. (7) mo ^ 'none'. This has been treated with the distri
11
butives.
1/3/5/6. (8) wu M 'there is not, have not', opposite of yu ^ 'there is,

have'. We have elsewhere considered its use in front of verbs and the
1 2
reason why it is sometimes written with the graph ® .
There are no less than four cases of pu yu (A 65, 83, 87, B 44),
as well as one of wu yu 'not have any' (B 66). I have failed to discover a
single factor accounting for the use of pu yu in stead of wu.
1/3/5/7. (9) fei negative copula 'is not', negating the ' X Y yeh'
pattern. The constant concern of the dialecticians with proving that X is
or is not Y drives them to some astonishing permutations, negating fei
itself by pu, preceding it by hsiangffl'mutually', or by quantifying yu and
13
wu, freely omitting the complement.
The copula fei as main verb, whether or not it has a complement,
regularly takes a final yeh ; this is especially apparent in the many examples
of affirmative and negative ' X Y yeh' sentences throughout Names and
objects. The final yeh is missing only in the pattern ' . . . yeh, fei . . .' (B 3
^ H H i l i . . . J k ^ J i "It is the case that in itself this one is beautiful.
. . . It is not the case that this one is beautiful" Cf. also A 18, B 47). The
absence of yeh may therefore be significant, as in B 60 WtM^r^ "There
is nothing which when combined with it is not half", which cannot be
taken as "What has nothing combined with it is not half".
1/3/5/8. At several places we find the phrases which

seem at first sight to be clauses ('It is not an ox', 'It is not a horse'). But
if so we should expect the yeh, which is used regularly even in front of
conjunctions:
B 8 fl&#*ff58HH ·
"The loan-named necessarily is not the thing in question, otherwise it
would not be loan-named."
1 1
Cf. pp. 127-130 below.
1 2
Cf. pp. 131-135 below.
1 8
Cf. § 1/3/3.
126 Grammar
B72 r t S I J
" 'It is specifically what I call it* is inadmissible if it is not its name."
The phrases are in fact nominalised, 'what is not an ox/horse',
'a non-ox/non-horse':
9
A 73 »flg-feo
" ' A l l oxen* and 'non-oxen separately grouped* are the two sides."
B i 2 m r ^ j r s j . . . A so a r j s i j T ^ S J o
"Fit 'ox* or 'horse* . . . Fit 'horse* or 'non-horse*.'*
We have elsewhere noticed fei niu 'non-ox' in constructions which
require it to be a nominal unit, after wet chih IBÎ 'call it' and huo j&
14
'some'. The latter is found in the curious argument as to whether an ox
and a horse are oxen:
B 67 a ^ j f f i * * * "J » fl9*#*sR*ifn*i& "I... · '
"If it is admissible that with one as a non-ox they are not oxen, it is admis
sible that with one a non-ox and one an ox they are oxen. . . . Without the
ox not being the ox or the horse not being the horse, there is no difficulty
about the ox and the horse being a non-ox and a non-horse."
In the first sentence there is a clear contrast between on the one hand
niu ('ox') and fei niu ('non-ox'), on the other niu yeh ('is an ox') and fei
niu yeh ('is not an ox'). In the last sentence it will be seen that the absence
of yeh is crucial; if it were present we should have the statement " A n ox
and a horse are not oxen and are not horses", which whatever it might
mean does not fit the argument.
There is a striking instance of fei ma 'non-horse' at the beginning of
ch. 1 of Kung-sun Lung tzu:
(ch'en 38/i) m&mn^m^ °
"He said that a white horse is to be deemed a non-horse."
1/3/5/9. The absence of the yeh also serves to distinguish the fei of
the pattern fei . . . pujfujwu . . .', 'If it is not . . . then not . . .':
N O 10 ^ A ^ M M S f f f o
"Now men have nowhere to walk except a road."
This is the only construction in which the difference between fei and
wu is allowed to fade:
A 83 o
"There cannot be one without the other."
1 4
Cf. § 1/3/12/1/1 and p. 128 below.
1/3 127
In the dialectical chapters we do not find any similar construction with

wu in the place of fei, and indeed wu at the head of the sentence is the
main verb and predominant over a later negative (B 36 ^ 5 1 8 "There
is nothing he does not yield", B 60 i S J & l ^ ^ "There is nothing which
combined with it is not half").
1/3/5/10. We have already noticed the rare cases of fei 'wrong',

15
mostly in the pattern S S / ^ i f e 'deemed the right/wrong thing to do'.
There are also cases of the causative fei 'condemn, reject', discussed with
16
other causative usages.
1/3/6 DISTRIBUTION
The dialectical chapters show a theoretical interest in quantification,

partly inspired by the Mohist doctrine of loving not some but all men.
They define 'all' as 'none not' (A 43 [ft J ^fafcifa) and 'some' as 'not all'
(NO 5 fsKj ife^r, ^H-ife), and have several quantifiers of their own, used
adverbially like the regular distributive particles: chou JS 'in all cases',
Hang M 'in both cases', pHen M 'in one case but not the other'. It is there
fore of some interest to see how they deal with the standard distributive
words of Classical Chinese. These are either verbs or pre-verbal particles
(some of which are basically verbs used adverbially), although in the case
of universal quantification the Mohists are able to detach a noun from its
context with the aid of the adjunct fan (A 73 'all oxen').
The basic distinction is between external and internal distribution, the
former relating things to each other (chit $k 'all', as a verb 'associate'),
the latter relating parts or individuals within one thing or collection of
things (chin ft 'all', as a verb 'exhaust, use up'):
EXTERNAL INTERNAL
All chii (verbal sentences, 28 times) chin ft
chieh (nominal sentences, 4 times)
Some huo $i yu M
None mo ML wu M
Of the four cases of chieh, one is in Expounding the canons (EC 9) and
the rest in Names and objects ( N O 2). (The graph also appears in A 98,
but probably for chieh 'together'.) On the other hand chu occurs only
once in Names and objects (NO 11) and 27 times in the Canons and
Explanations. These figures might be interpreted to mean that the Canons
1 5
Cf. § 1/3/4/4.
1 6
Cf. p. 159 below.
128 Grammar
and Explanations use chii alone while the other documents prefer chieh.
But since chii is confined to verbal sentences throughout the concordanced
pre-Han literature it seems safe to discount this distribution as an accidental
consequence of the growing prominence of the ' X Y yeh' sentence after
the Mohist shift of interest from the name to the sentence. It would
appear that they deliberately confined chieh to nominal as well as chii to
verbal sentences.
Distribution is comparatively rare in nominal sentences. We find only
chieh and huo, used when there is more than one subject, and a single case
of chin, referring to the parts:
N O 2 m&,mm^m » ^ r a 4 j r £ j r s j r * j °
"As for those which name according to shape and features, such ones as
'mountain', 'mound', 'house* and 'shrine' are all examples of these."
NO 1 » flteftfi-fe o
"As for those which do not name by referring to measure or number, when
you break the thing up all of it is this."
In both cases with huo there is no final yeh and the whole phrase
depends on a main clause elsewhere; one remains in doubt whether it was
(
legitimate to say huo X yeh', 'One is an X ' :
B 65 m w & r f n f l ' dfcfclfcS *
" A l l have the standard but they are different, some wood and some stone."
B 67 * S # ^ i f n # ^ t & « r » fliJ^^^ffi^-tfaÎ o
"If it is admissible that with one a non-ox they are not oxen, it is admissible
that with one a non-ox and one an ox they are oxen."
We find no sentences of the form basic to Western logic, 'All/some/no
Xs are Ys'. The Mohists write not ' A l l white horses are horses' but simply
'White horses are horses' (NO 14 S M S i f a ) , not 'Some horses are white
horses' but the verbal sentence 'Some horse is white' (NO 18 H s £ S ) .
Had they been interested in sentences of this type they would presumably
have used the internal distributives (cf. Mencius 7A/36 ^ ^ f t A ^ / f ' J f i i ?
"Are not all sons of men?").
In verbal sentences the three external distributives all refer backward,
to the subject or to an exposed element. Unlike the nominal sentence the
verbal sentence with such a distributive can have a single subject referring
to the things associated in the action:
B 22 № S & & o
/ A l l its members are so."
1/3 129
A 74 J i ^ f t * ° «» * &lfcFlB» °
"These do not both fit the object. If they do not both fit, necessarily one
does not fit."
B 48 ^ * f f c f f i £
" I f he is able both to choose and to discard. . . ."
B 67 * PldCF °
"In one case it is admissible, in the other not."
B 80 H f t f e J l °
"Nothing is longer than this."
Occasionally chii is translatable by 'together', but only when association

is implied by the assertion (A 87 "Both live in the house/they
live together in the house"). Even in a case where one might seem to
recognise the pattern ' X yii Y chii . . .' ('X does . . . together with Y ' ) the
yii IS is simply the conjunction and the chii is independent:
A 67 (81) ° R I S R f f c ^ · *8 * » * * « « ! » R H < ·
"('Coinciding') Of measured feet, neither is completely covered; of
starting-points, both are completely covered; of foot and starting-point,
one is completely covered and the other is not."
When phrases with huo are paired the second huo is twice missing:
A 46 jfcfeff "One is removed, the other remains."
B 11 s K U S "In one case repeated, in the other not."
But in both cases the cause is probably textual corruption. In A 74

also the second huo in a pair of phrases is missing in the editions of Pi Yuan
and Sun Yi-jang, but survived in all older editions collated by W u Y i i -
chiang. (Cf. also Mo-tzu ch. 48, Sun 284/6 ^ f c ^ / F f c "One was bene
volent and the other not": the second huo is missing from a parallel in the
previous line.)
In several cases huo and mo seem to refer to place ('somewhere,
nowhere'). This is a curious usage, and editors have generally followed
Sun Yi-jang in taking the graph in such contexts to represent yii
'region'. But this ignores the parallel use of mo:
A 42 r m J * * ( = X ) M ^ S R * °
Canon " A 'limit' is where, if you advance again in some direction, there is
no room for the foot-rule."
Explanation " I f in some direction there is no room for the foot-rule, it is
limited; if in every direction there is room for the foot-rule, it is limitless."
130 Grammar
A 49 r » J * « f e f t »
"To 'move* is to shift in one or other direction." (Cf. also B 13, 33.)
A 65 r S J J G F W t t o
" T o 'fill' is to be nowhere absent."
Proceeding to the internal distributives we find that their function
changes according to whether the verb is transitive or intransitive. When
it is intransitive they may like the external distributives refer back to the
subject, but to parts of a thing or to instances of something inherently
uncountable ("a 'saying'):
N O 1 ^ A S - t f e S » SScftS-tfe » ftJfiiSI^ o
" I f this stone is white, when you break up this stone it is everywhere the
same as the white thing."
A 60 — » R | i i R * ^ i c - f f i »&mmte ·
"When they are two, the measured feet both leave one starting-point,
which is being nowhere the same."
A 99 1EM1¥ . . . 5 g a « i i 6 *
"The exact nowhere is not. . . . For example, a circle is nowhere straight."
A 98 ##JTn^# o
"In some respects is not but is not not."
The last is the only example of yu in this construction. In general the
Mohist assumption is that things either are or are not the same or black
or straight; we can say that they are in all or no respects so, but as soon as
a choice arises attention shifts to the parts or respects 'by means of which*
(yi &J) the things as wholes are to be judged so or not so:
A 97 fitA£W£#W*J!*#teiL ("HAJ°
"By means of what is black or what is not black in a man fixing 'black
man'."
A 86, 87 W R P » r « J P № . . . ^ ^ » r ^ « J - t b o
"Having respects in which they are the same is being of the same kind. . . .
Not having respects in which they are the same is not being of a kind."
(The first three words are more literally "Having something by means of
which they are (judged to be) the same".)
N O 12 * 4 № H H i f 5 W ! f t l R l °
"Things have respects in which they are the same but it does not follow
that in all respects they are the same." (This is the unique instance of
shuai ^ as a distributive.)
A 68 ^\>imm · °
"One part coinciding and the other part not."
1/3 131
It may be noted that the yi is dropped in the negative, but that yu is

negated by pu (A 87) in stead of being replaced by wu.
When there is no reference to the subject yu and wu in front of an
intransitive verb generalise the action of the verb and convey that there
are or are not circumstances in which it occurs:
B 30 nmn ° . . . ^jmm > m#m
"In no circumstances is buying at too high a price. . . . If the royal coin
does not change from year to year but the harvest does change. . . ."
B 35 mmmm °
"Saying that in no circumstances does one win in disputation. . . . "
The technical terms wu hou I f f 'in no circumstances thick* (dimen-
sionless) and wu chiu 'in no circumstances long* (durationless), used
of the point and the moment, presumably imply comparison; there is no
thing than which the wu hou is thicker, no time than which the wu chiu
is longer. This agrees with the definition of hou (A 55 f" J¥ J * WSf^Cife
17
" 'Dimensioned* is having something than which it is bigger*').
When on the other hand yu and wu precede a transitive verb the
reference is forward, to an implicit object:
A 4 'seek something*
A 23 M90 'not know anything*
A 25 IffifrM 'not desire or dislike anything*
A 65 I S 'not fill anything'
B 25 U P 'not put anything on top'
B 38 fffta 'point something out'
B 66 I f f 'not have any'
B 71 ff "J 'allow something'.
We do not find chin 'all' with a similar function, in stead a double
negative with wu:
B 36 M^Wt "There is nothing he will not yield." (Contrast the backward
reference of the external distributive mo: A 43 "None are not so".)
With chin the transitive verb may be followed by an object, with yu
and wu by a directive with the preposition yu:
B 74 « K A · № J « * R № K °
"If he asks about all men, one loves all whom he asks about."
A 97 ffSSÂ · ff^gSÂ o
"Love some men, not love other men."
1 7
Cf. § 1/3/12/3.
132 Grammar
N O 17 mmm... ^mtksn °
"Ride all horses without exception . . . ride some horses."
NO 2 · WWKM-tb o
" T o have some Ch*in horses is to have some horses."
E C 11 M S ( = « ) K A °
"Leave out no men."
A 75 ftsmjytim °
"Overlook none of the harm in it."
The directive 'yu X* may be taken as literally 'among X* ( A 97 "love
some among men"). ' X ' may in fact be a numeral, and to enforce a contrast
the directive may be preposed and resumed by yen in one case and placed
in the main verbal position in the other:
B 37, 38 (Contrasted Canons) WfclS ' ^ f t l S . . . ° Wfit · iffi
"In one thing you know some of it and do not know some of it. . . .
Pointing out something is inescapably from two."
The difference of behaviour between chin and yujwu is not of course
limited to the use of an object with the former and directive with the latter.
Chin is primarily the verb 'to use up, expend* (B 26 ±q£ffl№!fc » S)
"When the leverage and weight of the one above are spent it falls to the
ground"). In more typical Mohist contexts ' X chin Y* implies that X comes
to the end of Y , that Y is completely included in X . This is especially clear
in the geometrical description in A 67 just quoted, but is also relevant to
logical passages; the usages in the following two examples throw light on
each other:
B 65 finm · ftWrSffnJI... ° » °
"(Of something being so of different things.) Things in which the
characteristics of the square are complete all have the standard but are
different. . . . If they have all the characteristics, as in being square, all
the things are so."
B I j t * . . . B 2 * i t r * / h j · r*y£j ...mm&> rwu»> r * j
"Fixing the class. . . . Fixing a wider or a narrower, or the things all

included. . . . If you say it has four feet, is it 'animal* ? Or 'the living', or
'bird'?—'the things all included* and 'a wider or a narrower*.**
When chin is used in connexion with a transitive verb it implies that
the action comes to the end of Y , applies to all of Y . Chin may precede the
verb in the position of the other distributives, as already noticed (ftFÂ
1/3 133
'ask about all men'). But if it is the chin rather than the verb which is
syntactically prominent it will itself become the main verb, with the other
verbal phrase nominalised as its subject. In the following example the
whole clause is in its turn nominalised as object of the verb chih ( = £ 0 )
'know*:
B 74, 75 o ffi#([gfcH]£*»£-fe). . . o 5g£JUfejfD»([S]£*
" I f they are two, we know the number: how do we know that love of
mankind applies to both of them ? . . . Like knowing that love is for all
of them when you do know their number."
In translation a Chinese verb preceded by yu/wu is often conveniently
represented by an English abstract noun (A 23 №%\ 'without consciousness*
A 25 1 ® U § 'without desires and dislikes'). We have deliberately avoided
such equivalents in order not to commit ourselves to the position that the
Chinese verb after yu/wu is nominalised. But certainly there are verbal
units after yu/wu which are nominalised without quantification, and some
are followed by directives. This raises the question whether there are
ambiguities in the forward-referring devices which might hinder thought
on quantification. The following are all the examples I have noticed of
comparable sentences which do not quantify. When collated they reveal
certain very striking features: wu is always written with the graph S , in
the single case where we should expect yu we find huo che the
preposition is not yii but hu . We underline the nominalised verbal unit
after wu:
B 43 Hft&ft# o
"There are no constant ascendancies among the five elements." (Contrast
B 35 i i i l "Saying that in disputation in no case does one win".)
B 60 °
"There is nothing which when combined with it is not half." (This is the
interpretation which fits the context; with the standard graph for wu it
would presumably be 'Without anything combined with it it is not half,
lB
which would however require a final yeh.)
B 72 mmm^&m °
"The speaker has no thing which is specifically an X for the ' X ' that it is
called."
1 8
Cf. § 1/3/5/7.
134 Grammar
В 74 о
"Some are left out of his question." (Cf. the examples of З&ШЙ . . . in
A 75, E C 11, quoted above. W S S £ f t F p № would presumably be "leave
out some that he asked about".)
E C io ттшш > ттш±жт»на °
"Beneficial action in which one does more for some and less for others,
but without the beneficial action which grades according to relationships,
,
has a selfish motive." (Not 'in no case grading according to relationships ;
the motive would be selfish whenever one fails to do so.)
N O 15, 16 °
"The heart has no hollow inside it."
If we proceed to look for other examples of the graph Ш, we find none
except for a single case of the negative imperative wu (B 38). The only
other example of the preposition hu immediately follows the sentence
quoted from В 72, and seems from the parallelism to be a scribal error
( « » [¥] ftffl ШтёЪ'ШЖШ " I f the other still is specifically what
it is called. . . . If the other is not specifically what it is called. . . .")
It would seem then that the Mohist dialecticians deliberately reserved
the pre-verbal yujwu for quantification, and avoided the confusion which
might result from their use in other constructions by choosing other graphs
and particles.
Another possible difficulty in quantification by yu and wu would be
the quantification of yu and wu themselves. Of the four theoretical possi
bilities (yu yu 'have some', yu wu 'lack some', wu yu 'have none', wu wu
'lack none') the second would be ambiguous, since yu and wu might be
taken as co-ordinate, 'having and lacking'. We do in fact find two passages
where yu wu might be expected, and in both it is avoided. In one the yu is
absent from a theoretically possible yu wu yen:
в 49 %мп»ти±мтш °
" I n the case of there not being some of something, there is not only after
there was."
In the other a theoretically possible Ш^Ш 'wish to be without some'
is replaced by a construction with huo che:
В 44 ШШЪШ °
"Wish that you did not have some."
Thus both of the examples of huo che in the corpus (B 44, 74) serve to
avoid difficulties in quantifying with yu.
Of the three unambiguous combinations, there is no example of wu wu
or a substitute. But we find not only wu yu (B 66 '^М'ЙМШ^ "They do
1/3 135
not in one case have and in the other not have any") but an extraordinary
example of yu yu with a directive, assumed by previous editors to be
corrupt:
N O 2 m^mmm^mfe °
"To have some Ch'in horses is to have some horses."
It will be seen that there is no need to emend this sentence. From the
point of view of style, elegance means nothing to the Mohist, syntactic
clarity means everything; and within his system the sentence is unam
biguous.
Among the interrogative pronouns shu IfX and shut ffi 'which* are close
in behaviour to the external and internal distributives respectively. The
former refers back to the subject, of which there may be more than one:
B 6 ?
"Which is longer, a tree or a night?"
We commonly think of shut as referring exclusively to persons ('who ?').
But since a person is an individual picked out by the question there is a
close connexion between words translatable as 'which?' and as 'who?';
in other texts shu often refers to persons and shut sometimes to things. The
following example is from elsewhere in Mo-tzu :
ch. 46 (Sun 264/6) r nmn^»^»»* j . . . r i№K«№ J ·
" 'With either thoroughbreds or sheep to yoke to your carriage, which
would you drive?'. . . . 'I would drive thoroughbreds'."
In the same chapter we find a construction comparable to the
'yu/zuu- verb-yu X ' pattern discussed above:
Ch. 46 (Sun 268/10) ^f6*^jIfc-=lA ?
"Of these two men which will you honour?"
The dialectical chapters contain three examples of shut:
B 41 ^fctfySHHb o
(Canon) "Not know which he refers to", a question rephrased in the
Explanation as MUMfe "What does it refer to?".
B 44 S£X o
"Which person does one love?" (of self and others).
B70»£;SJRte»«iII* ?
"It is just as with, 'it is white or it is black, with which answer does one
win?'"
In the first two shut as object is comparable with shu as subject. In the
last case it is possible to translate simply 'Who wins?' But the symmetry
with the distributives suggests rather that it should be assimilated to the
136 Grammar
wu of B 3 5 ("In disputation in no circumstances does one win"),

'in which circumstances does one win ?' We may note also that with the
addition of shu and shut three of the four pairs of distributives show a
phonological similarity:
huoj * G ' W a K yuj * G I Ü G
moj *MÄK wul*MlWO
shu/ *DJÖK shui/*Dl№R
This suggests that the Mohists conceived the three external distri
butives as the internal distributives modified by a final - K . It may be
added that ko & / * K L A K 'each', which occupies the same pre-verbal
position and has the same final - K , is not used in the dialectical chapters
(nor is met 48 'each').
1/3/7 QUESTIONS
The interrogative particle hu ^, used after verbal sentences, appears

once ( B 41). Yü $3, the interrogative form of final yeh, appears in two
pairs of questions ( B 2, 10). In three cases it follows nominal units; in one
case where it follows a verbal clause belonging to the Type 1 C elsewhere
19
defined ('It is that . . .') both yeh and yü are present:
B 10 « H ? JäBft«H&Ä ?
"Is it knowing? Or is it that one supposes to be so what is already past?"
Since so many questions are about the alternatives of disputation the
commonest interrogative pronouns are shu $K and shut ft 'which?', the
former referring backward and the latter forward. These are more con
20
veniently discussed with the distributives. The other interrogative
pronouns are ho M (object: 'what?') and wu M (directive: 'to, in, from
what?'). In two passages they can be seen to be in correspondence with
object chih and directive yen:
B41 rUHIlib? J « H » riUfeJ№)»£°
" 'What does X refer to ?' If he says ' A n X is a Y ' then one does know it."
B 38 · m n n - m m t ^ M » °
"This would be knowing the very thing one does not know, and in what
can it be supposed that, the reference being to a single thing, there is
something one knows in it and something one does not?"
In two passages the graph S£ appears to represent wu :
1 9
Cf. p. 158 below.
2 0
Cf. p. 135 above.
1/3 137
B 4 2 C. § f # H S < # > ^ ° f e # ? H » # ?
"Where it belongs and what belongs, T n what does it belong ?' and 'Which
belongs ?'."
E. JB ? . . . t№ ?
"In what may it be considered to belong? . . . Which belongs?"
B 57 E t » t t f t » » S»Bl«»Jfe»H& o
"When we deem a pillar round, that wherein we deem is known a priori."
(For another interrogative pronoun with chih 'know', cf. B 41 ^£fl№?t
i1№ "Not know which he means".)
Evidence for this graphic interchange is sparse except for the exclama
21
tory wu Hi . But it is firmly attested in Mo-tzii by a series of questions
appearing in different forms in the three versions of 'Rejecting destiny'
(ch. 35-37):
Sun 170/5 Ink H * £ ? . . . " O n what does one base it? . . ."
174/-1 ?. . . " O n what is the basing of it? . . ."
178/2 M ¥ ^ £ ? . . . "In what does one investigate it? . . ."
1
Wu appears also in the phrase S S (B 74 'Whence does one k n o w . . . ?').
This is conveniently translated 'How does one know . . . ? ' , but there is no
reason to assume that the common pre-Han practice of asking this question
with wu or with an 3c 'where?' implies any weakening of the directive
function (cf. the question 'Whence (an) do you know that the fish are
happy?', which Chuang-tzii derided by the answer 'I know it from up
22
above the Hao' ?**P>^Jl'tfe).
It is remarkable that there are no interrogative adverbs in the Canons
and Explanations. But in Names and objects we find ch'i J S , rhetorica
'How . . . ?' implying the answer 'No', and two adverbial combinations
with hsi : ftJ§№P"Why should it not be so in my case too?"
"By what means shall we make it clear?" (NO 6, 11, 15).
1/3/8 COMPARISON
The standard word to express similarity is the verb jo 3? 'be like',

negatable by pu . Jo provides a striking example of disputation imposing
a consistent use of words regardless of accidents of idiom. In ordinary
pre-Han usage jo is negatable only when degree is implied (pu jo 'not as
much as, not as good as') but in the dialectical chapters negative is com-
2 1
Cf. P'ei Hsueh-hai 253.
22
Chuang-tzu ch. 17 (Kuo 607/4, 5).
138 Grammar
pletely assimilated to affirmative usage (pujo 'not like' A 7, 11, 74. B 3, 81).
We can understand that it would have been very inconvenient for the
Mohist dialecticians to have to switch from jo to some other word such as
ssu ffit 'resemble' (which they never in fact use) whenever they wished to
negate it. It is interesting that the only other pre-Han example I have
noticed uses the animal illustrations and therefore perhaps the language
conventions of the dialecticians:
Lu-shih ch'un-ch'iu ch. 23/5 (Hsu 23, 14A/2) ^Z&^¥^±^^W<>
"The ox's nature is not like the sheep's nor the sheep's like the pig's."
As the introductory word in the formula of the parenthetic illustrations
jo is used very loosely and is often conveniently translated by 'For
example . . .'. On the other hand where degrees are being compared it has
a very precise sense, 'as much as' (B 25 ttMlfH^f "They are equal in
leverage and in weight").
Jo in front of an exposed element at the head of the sentence ('With
regard to . . .') is found only in the combination jo fu (B 27). We do
however find jo in front of the verb ('as though, seems'):
B 22 %W&'P °
"There is more but seems to be less." (Cf. also A 98, B 38.)
Yu is not negatable, and must therefore be classed not as a verb
but as a particle, used in the pattern '(X) yu Y ' , '(With X ) it is as with Y \
It does not like jo imply a general similarity between X and Y , but intro
duces Y as a clearer example to which the specific observation being made
about X will more easily be seen to apply:
B 66 ^ f f i ^ i b » » * № t t J S W S °
"This is picking indiscriminately, as with 'Oxen have teeth but horses have
tails' " (as evidence that an ox is not a horse).
In two cases a phrase with yu seems to be parenthetic:
B 65 stoKSJE » ^W&JJ£№*&& o mm » ( » # - & ) » °
(On things which share one standard all being 'so'.) "Some are wood and
some stone, but it does not affect their agreement in being square. If they
have all the characteristics, as in the case of being square, all of the things
are so." (Not " I f in all characteristics they are like the square. . . .")
A 32 MB » ( » B ) · Scife o
" T o speak calling by name, as in the case of the stone, is to communicate."
(However we understand this obscure sentence, we cannot take it to be
comparing some other object with a stone.)
In three cases the theme illustrated by yu 'as with . . .' is introduced
by chin yeh 'in the present case . . .':
1/3 139
"In the present case, the fact that when you stand up a footrule on flat
ground the weight does not descend is because it has no inclination to the
side; as for the pull of the rope on the trundle, it is as with a pull on the
crossbar from inside a boat."
"It is as with 'White or black, with which does one win? This is like its
colour, and what is like white is necessarily white*. In the present case too
we know its colour is like white, therefore we know it is white."
B 78 ^ f t i B £ № 8 f * « J » * » K S 3 6 S o
" T o say in the present case that it is inadmissible to criticise too much is
as with the case of sorting out the shorter in comparison with the longer."
Shih yu always, as in B 27, 78, introduces a new example (A 75
cf. N O 6). But yu alone generally refers back to an example already
introduced (A 32, B 60, 65, 66). The two exceptions are B 70, just quoted,
and:
B 8 » M R T UJ til °
" A dog being loan-named a crane is as with naming it 'Crane'."
But both may be taken as backward references to:
A 88 » gJR№ . . . . « £ » ttSSMb °
" O n both sides winning absolutely: white or black. . . . In the case of what
'crane* is deemed to be, 'surname or the thing itself'.**
The temporal adverb yu 'still* is probably a derivative of the con
junction ('as before, as it was'). We find it both as a single word (A 75,
B 72, E C 2) and in combinations which appear to be transitional between
conjunction and adverb, yu shih yeh MSffe 'as it was' as complement of
the sentence (B 50) and adverbial yu chih (B 72). The latter is syntac
tically remarkable since it treats yu as verbal. Both combinations are
common in other pre-Han texts:
9
Mo-tx* ch. 47 (Sun 281/4) S » E t № f t E - t b ° « ^ ~ F £ № » *
"This is as with throwing eggs at a stone. After throwing all the eggs in
the world the stone is still as it was; it cannot be broken." (Cf. Mencius
4B/28, 6A/2 for other examples.)
Kuan-tzu (ch. 64) BSS 3/40/-4 ft№*/H6fi4> » feSI
140 Grammar
"Therefore although his land was small and people few, he still became
Emperor. . . . Therefore although their lands were great and people many
they still perished in misery and disgrace.''
1/3/9 NUMERALS
The numeral may stand in a nominal position (B 59 E W — " F i v e

has one in it"). But in the main sentence position it is verbal, without final
yeh and negated and distributed by the pu ^ and chti 1R which precede
the verb, not the fei $r and chieh which precede the complement:
B 3 fficW, ^ f f t H "They both fight, they are not both two"
B 7, 12 "They are both one"
B 12 " A n ox and a horse are two"
B 18 J!:— "The shadows are two/there are two shadows".
Numeral adjunct with head may stand in a nominal position (B 18
~%'$i-~$t "Two lights flank one light"). But in a verbal position it
functions verbally, without the addition of a verb such as yu W 'there is':
A 39 ÂrfnfflcM^Mife "They are two men and both see that this is a
pillar"
A 86 *f? "There are two names but one object"
B 12 ^ J S № "Ox and horse have four feet"
B 18 — " T h a t which has one light"
N O 7 —*Sft# "Things one in appearance"
N O 18 —JUffOaKS "There are two horses and one is white".
1/3/10 QUOTATION
The particle combination yeh che ife^j is equivalent to 'unquote',

implying that what is under discussion is the meaning of the preceding
word or phrase. In Names and objects it marks technical terms introduced
for definition (NO 5 J "'Some' is not all"). In the
Explanations its function seems at first sight to be the same, since in the
earlier examples the word so treated is in fact the term defined in the
Canon (A 3-6, 32). But throughout the first four the sentences lack the
final yeh of definitions; and on closer inspection one sees that every one
of the 11 single words followed by yeh che is a direct quotation from the
corresponding Canon (A 3-6, 32, 46, 64, 71, 83. B 31, 35). Even when
there is a definition it is of the word as used in the context of the Canon
1/3 141
(cf. A 46, 71, B 31). In A 83, in a series of three explanations of terms, the
only yeh che is after the second, the one which is from the Canon:
9
A 8 3 is^f mmto&»r &j » ^ < r f n > ^ m ° *\m · nwm °
"The sagely, employ but do not treat as necessary; the 'necessary', allow
and do not doubt: the converse apply on both sides, not on one without
the other."
The yeh che phrase always stands at the head of the sentence. Much
more common is the quoted word or phrase attached to the end of the
sentence by yeh. This device is used systematically in A 7 6 - 8 9 , B 2, 3, 9 , 1 0 ,
42, 58, but also occurs in isolated sentences elsewhere (B 27?, 29?, 36, 55,
61-64, 66, 70). It accounts for a number of apparently clumsy sentences
in which what might conveniently have been treated as subject stands at
the end:
B63№F«HS»» r * ? J i f e °
"That which when divided ofT from things cannot be presented apart from
them is (what the Canon means by) 'space'."
In three cases yeh che follows a unit of more than one word:
N06 rS»3HJte#- IRH&O I" Jte# » m b ·
" 'This is like saying . . .' implies similarity, 'How can I say . . . ?' implies
difference."
A78*£["JSJ r ^ f Jftf
"When one names it 'horse', it is classifying. For 'like the object' one
necessarily uses this name." (Cf. also A 31.)
In the last case it is tempting to emend, omitting the yeh: "For what
is like the object one necessarily uses this name." But there seems to be no
objection to retaining it, in which case the Mohist is making the much
more sophisticated assertion that a class name is an abbreviation of 'a thing
like the individual X ' .
1/3/11 LOGICAL IMPLICATION
1/3/11/1. The Canons and Explanations have two conjunctions of

implication, tse №\ '(if . . . ) then' indicating that A is a sufficient condition
of B, and erh hou \R№k ( Jn ) 'only then' indicating that A is a necessary
condition of B.
B 70 · mm%\± °
"If you hear that something you do not know is like something you do
know you know them both."
142 Grammar
A l T«CJ » №#ffi«Jfcfe ·
" A 'cause* is what must be got before it will come about.'*
The adverb pi 'certainly, necessarily* appears only in clause B in
the case of tse, only in clause A in the case of erh hou:
B 25 mmmmm > mm<&r °
" I f you put equal weights on both of them the tip is certain to decline.''
(Cf. also B 48, N O 10.)
B 64ft%<&ft&mm*
"The traveller is necessarily at first nearer and only afterwards farther
away." (Cf. also B 8, 51.)
Pi in the second of successive verbal phrases always implies that B is
the consequence of A, whether tse is present or not:
B 25 J q m i & K - ^ & f i °
"If you put a weight on one side of it it is certain to decline."
This allows a very economical way of putting one implication inside
another:
B 2 itm&i&mmum °
" I f it is necessarily so of the thing it is judged to be if it is so of the thing
here, then all will be milu deer."
1/3/11/2. The essential function of tse is to authorise inference from
1
A to B ; the sentence may itself make the inference ( B 73 ffSJH ] "I ft " I f it
is limited it is exhaustible") but more often, as is especially clear in the
passage on mechanics just quoted ( B 25) entitles us to infer B when we
encounter A . The logical relationship is allowed to weaken only in a few
cases of parallel clauses which merely contrast what happens in contrasted
conditions:
B i o m±mm » m ± % m » °
" A case where if he lifts something it is light and if he puts it down it is
23
heavy is not a test of strength." (Cf. also B 35, 49.)
1/3/11/3. The strictness of the implication when tse is used in formal
argument can be seen from this example:
N O i8 zm±g*№· a»j*ii±M*n> ° ±mz^»^m±mx °
" I f this horse's eyes are blind we say that this horse is blind; when this
horse's eyes are big we do not say that this horse is big."
In this pair of sentences (and in two other pairs in the same series)
tse in the first is replaced by erh in the second, because we cannot infer that
23
The syntax of B 35 is discussed on p. 158 below.
1/3 143
a horse with big eyes is not big. In English we can tolerate ' i f in both
sentences, and indeed I used the word in both in a previously published
translation of the Hsiao-ch'u.
1/3/11/4. We find a similar strictness of implication with erh hou:

NO 17 gÀfêiiSÀïïnfâSgÀ o
" 'Love of men' requires loving men without exception, only then is it
deemed love of men."
" 'Riding horses' does not require riding all horses without exception before
being deemed to ride horses."
"It requires that one rides no horse at all, only then does one not ride
horses."
In the second case erh hou is abandoned for jan hou, because the
necessity of the condition is denied. The formulae of N O 17, 18 are there
fore correlated :
p JM q (p a sufficient condition of q)
p ÏT5^ q (p not a sufficient condition of q)
p ffiîîâ q (p a necessary condition of q)
^Fféf p q (p not a necessary condition of q).
The distinction between erh hou and jan hou may be an innovation of
Names and objects, but it is also possible that the absence of even this use
of/aw hou from the Canons and Explanations is merely accidental. Jan hou
also appears in Expounding the canons, written before the grammatical
restrictions were established; but there it is a mere synonym of erh hou
(EC 2 H ^ & J t : , £ * f â £ J ^ £ . . . "Only when the three things are
provided together are they adequate to generate . . .").
1/3/11/5. Another possible innovation of Names and objects is the

rare conjunction^/* St (NO 2,6). This seems at first sight to be synonymous
with erh hou and jan hou, as it is elsewhere in Mo-tzu :
Mo-tzù ch. 14 (Sun 65/1) & » * £ 0 r i J E » M^fëZ ° ^1LZ9iU1& »
"He must know the source of disorder before he will be able to reduce it
to order : if he does not know the source he will not be able to reduce it
to order."
But we may assume that the Mohist would not use yen in addition
to erh hou without a special reason. In The grammar of the Mohist dialectical
chapters I proposed to take the yen of N O 2 as 'only in it', on the strength
144 Grammar
of an example in Mo-tzii ch. 46 (Sun 271/-5). But this does not account
for the yen of N O 6, which I now follow Sun Yi-jang in recognising as a
further example of the conjunction. We may notice that throughout the
dialectical chapters it is always the verbal phrase immediately preceding
erh hou which states the necessary condition (A 1, 68, 69. B 3, 8,45, 49, 51
bis, 64. N O 17 bis). In N O 6 on the other hand the necessary condition
before yen is developed in a whole string of preceding phrases; and in
N O 2 it is expressed not by the preceding phrase but by the main verb,
which as in the example just quoted from Mo-tzii ch. 14 is chih 'know':
NO 2 · » ·
" I n the case of those which name according to shape and characteristics,
it is necessary to know that this thing is ' X ' , only then does one know ' X ' . "
The distinction between erh hou and yen may therefore have served
to clarify confusion arising when there is more than one possibility in
identifying the necessary condition. Since the problem never arises in the
Canons the absence except in Names and objects of yen (as of jan hou) m
be accidental.
1/3/11/6. Expounding the canons, which is unique in never using tse,

has no less than three particles translatable as ' i f , jo 3§ (EC 1), kou ^
(EC 6) and chieh W (EC 4). The last is contra-factual:
E C 4 mm^mxrm · ^mm&nfe °
"Supposing that if Jack died the world would be harmed, my special care
for Jack would be myriadfold."
The Canons and Explanations, in accordance with their policy of
avoiding synonyms among particles, use only jo. I have failed to identify
any device marking a hypothesis as contra-factual; they avoid even the
shih $1 'supposing that' used for example in the two genuine essays of
Kung-sun Lung. In more than one passage a modern reader would be
grateful for some grammatical indicator of the contra-factual (cf. the final
sentences of B 2, 57).
Names and objects replaces jo ' i f by kou (NO 1, 2), perhaps in order
to avoid confusion with jo *like\ Jo is used only in the formula ^f3=f;Sk,
№\ . . . 'if it is like this, . . .' (NO 15,16), itself a reminder of the advantages
of kou.
1/3/11/7. A tendency for Names and objects to elaborate and refine

the terminology of implication is also visible in the use of adverbial yin H
'on these grounds' in addition to kuftiC'therefore*. While ku is primarily
a noun ('reason', on the evidence of its phonetic and graphic relation to
1/3 145
ku 'ancient times' basically 'what is at its origins'), yin is primarily a

verb, 'take as basis for doing something' (cf. E C 1). It appears in the
Canons as a noun, the 'grounds' or 'criterion' of a judgment (A 98, B 3, 15),
clearly distinguished from ku, which is used of reasons in general but in
particular for causes (cf. A 1, 77). Adverbially also it refers to the grounds
of a judgment, not the cause of an event:
N O 17 ^rntikm»SUSHIS °
"If he has ridden some horses, on these grounds he is deemed to ride
horses." (More idiomatically, to bring out the force of the yi, "these are
sufficient grounds for him to be deemed to ride horses".) (Cf. also N O 2.)
1/3/11/8. Besides the conjunctive yii M 'and', used between nominal
units which may be of some length (cf. A 97, B 10), there is a disjunctive
particle jo 3§ 'or'. The two examples (B 10, 70) are discussed elsewhere. 24
The disjunctive jo, very rare in other texts, appears about 50 times in the
military chapters of Mo-tzu, in instructions for defence allowing for
alternative possibilities. It may intervene between either nominal or verbal
units, which may be single words or clauses of considerable length:
Ts'en 105/5
"Morning and evening stand or sit."
56/2 s m ^ & t o
"Put chaff or ashes inside."
109/9 #&ftwmm± · mimm%mmn± » m ·
"If without authority he takes over what is not his allotted office, or
without authority deals with matters which are not his business, he is to be
sentenced."
55/3 m{=№)A®m > mm®±m% · »jiwt±
"The enemy behaving differently, building walls or piling earth i n an

unusual way, or an unusual amount of water at the side, are signs of
tunnelling."
1/3/12 KEYWORDS
1/3/12/1 Wei m
Uses of the verb wei 'call, say o f fall into three patterns:
1/3/12/1/1. Pattern 1: wei with two objects, the first a pronoun and
the second a noun:
1 4
Cf. p. 139 above, 150 below.
146 Grammar
A 74 l £ S I £ r ^ J » * S H ^ r * * J "One calls it W , the other 'non-ox* " 2 5
B 33 fllifc T M15 J "call the place here 'Southward* '*

B 72 IBS r 8 J "call the said thing 'crane' *'.
1/3/12/1/2. Pattern 2: w« before noun and verb:
A 46 "say of what remains of it that it is reduced"
B 35 WMMftfc "say of disputation that there is no such thing as
winning".
1/3/12/1/3. Pattern 3: wei before noun and complement with final
yeh:
B 72 IBtftJtJtife "say of both that and this that they are this".
The last is the sole instance of its pattern, but I take it to be dis
tinguished from Pattern 1 by the final yeh.
1/3/12/1/4. In the last two patterns it is not clear whether the noun
after wei is its object or the subject of the succeeding verb or complement;
our translations assume the former. One is tempted to reduce Pattern 2 to
Pattern 1 by understanding A 46 as "call what remains of it 'reduced'".
However, it seems that in Pattern 2 the noun is not substitutable by the
pronoun chih: cf. B 54 Mk. T $.3: J "call it 'killing a dog' " (not "say of
him that he killed a dog").
1/3/12/1/5. In Pattern 1 the second object may be dropped; the

pronoun object continues to indicate what is referred to, not what it is
called:
B 3 MkMii "only then does one call it ('X')"
I I S . . . Ilife (-#) "call the said thing ('X') . . . call another ('X')"
B 41 K i l l ! * "which thing is so called"
H§№ "What is so called?"
A 80. B 35, 49 №11 "that which is so called" (cf. A 80 0rEtiI, ifrfe.
0?S!, ffife "That by which one calls is the
name; that which is so called is the object").
1/3/12/1/6. In Patterns 2 and 3 either the noun or the verb/comple
ment may be dropped; the latter case is marked by final yeh:
B 2HE№ "say it has four feet"
A 62 (cf. 51, 63, 64) !B3fc£#fe "is said of those which flank it".
1/3/12/1/7. Nominalised wei refers to what an object is called, the
name used of it, on a particular occasion:
25
Yorfei niu 'non-ox', cf. § 1/3/5/8.
1/3 147
B 72 511 "what I call it" (my term for it)

%M "what it is called" (the term for it)
II "the man so calling it" (the man who terms it)
B 38 —II "one of the things it is called" (one term for it)
B 3 MM "it is not so called" (it lacks the term).
1/3/12/1/8. The familiar construction ' X 2M Y ' ('It is X that is

meant by Y ' ) appears only in Expounding the Canons (EC 7-10). But there
are a couple of examples of ' X chih weiyeK ('It is of X that it is said'):
A 21 S £ S B < & > "It is of weight that it is said"
NO 8 ^S^SBffa "It is to longer and shorter distances that they refer".
Outside this pattern with wei there are no cases of the inverted object
with succeeding chih; this accords with the Mohist's general avoidance of
26
inversion.
1/3/12/2 Che%
Except in hsi che m% 'yesterday' (EC 2) and at the end of phrases
containing so, the particle che stands exclusively after nominalised verbal
units. In subjectless phrases the reference appears to be exclusively to the
agent, although there are a few cases where one may hesitate to insist on this
point (cf. A 7 S S ^ ^ H f f l S i b "With the man who loves himself, it is not
for the sake of making himself useful"). In A 83 and N O 5 we find defined
terms ending in yeh che mixed with apparently similar words followed by
che alone, but in each case the difference emerges on closer inspection.
That che can be assumed to mark the agent of a preceding verb and
not merely to nominalise it is significant for the analysis of many obscure
phrases:
A 98 ... 'what is so . . . what is not so'
A 51 — ·— ' a thing in one respect so, a thing in one respect
not so'
B 65 — ' t h i n g s sharing one standard'
B 17 . . . —%3k 'one light . . . what is under one light'
B 22-24 H . . . 5S^f 'the mirror . . . the man looking at himself in
the mirror'
B 27 ft ( = S K ) 'what makes the ladder glide'
B 68 J E £ ^ I 'one who uses names correctly'
A 31 ^ f £ # 'what is like the name'
B 70 'what is like the white'
B 72 IB# 'the user of the term'.
2 6
Cf. p. 114 above.
148 Grammar
The collective pre-nominal particle chu IS (cf. A 32 IS P 'speakers in

general') sometimes precedes long phrases ending in che:
EC 2 H B A J f f f t f t A » S i # o
" A l l that the sage desires or dislikes 'a priori' for the sake of men."
N O 2 mu&m^m°
"Those which name according to shape and characteristics." (Cf. also
E C 1, N O 1.)
1/3/12/3 So fix
In the relative pronoun we do not find the distinction between object
and directive which is made in the 3rd person (chih yen M) and among
interrogative pronouns (ho wu M). In this as in most other sources we
do not find the combination soyii $ ? T h e use of so as object followed by
a transitive verb ('him whom, that which*) is much the more frequent and
presents no difficulty. But so as directive followed by an intransitive verb
shares the ambiguity of all directive units.
There is one case of so ta @f A . Since Ha yu X ' can only be 'big in
relation to/'bigger than X ' , this seems to be unambiguous:
A 55 T * J » # № A t e ° № < * > » T A < >
"Canon 'Dimensioned' is having something than which it is bigger.
Explanation Only <the starting-point> has nothing than which it is
bigger."
This is confirmed by a sentence in a well-known exposition of relativ
2 7
ism in the 'Autumn Waters' chapter of Chuang-tzu : " I f on the basis of
something (in relation to which/) than which it is bigger you treat it as big,
none of the myriad things is not big" (H3t§?Aff5A;£, №I1S#^>P A ) .
We also find two examples of so jan 0f!& (A 71, B 16). We cannot take
the jan as transitive ('what one approves') since this use oijan is unknown
in the dialectical chapters. The so is therefore directive ('where it is so').
In one case its significance is in any case fixed by the verb tsai 'be in'
(used causatively):
B 16 · ·
"Locating it in (the time) in which it is so or in the not yet so."
So jan appears also in the 'Autumn waters' passage. Syntactically it is
similar, although jan is also used transitively in the same sentence. (The
Mohist distinction between jan 'is so' and shihjfei 'is this/is not' is also
disregarded) :
27
Chuang-tzu ch. 17 (Kuo 577/-3).
1/3 149
"If on the basis of a standpoint from which it is so you treat it as so, every
one of the myriad things is so. If on the basis of a standpoint from which
it is not you treat it as not, every one of the myriad things is not."
Cf. Chuang-tzi ch. 27 (Kuo 949/-1, 950/1) # § " № $ 1 . . . »
&... ®wm&...
"There is a source from which something is so. . . . From what is it so?
It is so from something which is so. . . . A thing really has something
from which it is so. . . ."
Han Fei tzu ch. 20 (Ch'en 365/1) JB#X4&207&-& °
"The Way is that from which all the myriad things are so."
1/3/12/4 Yi &
1/3/12/4/1. Yi serves primarily as a preposition before a nominal unit
('by means of, because o f B 48 J^&Ifc 'choose by means of the name') or
as a conjunction between verbal units, marking the first as the means and
the second as the end (A 98 i h H K S O J l " F i x the grounds and thereby! in
order to distinguish courses"). In the latter case it is seldom material
whether one decides that the yi refers back ('thereby') or that it refers
forward ('in order to').
1/3/12/4/2. Yi also appears in the familiar combinations:
§ ? ' t h a t by means of which, the reason why'
'deem*
& X f § Y 'deem X to be Y '
"JK( 'may . . .',fiL\&'is adequate to . . .' (followed by active verbs,
the verb after "RT or £ alone being passive)
X , 7$]^ 'the purpose of X is to . . . \
1/3/12/4/3. We also find the combinations yu yi (A 68, 86,
N O 12), wu yi (A 73, B 34), but not in their ordinary senses 'have/lack
the means'. The yi seems always to refer to what one uses, judges by, in
comparing things or deciding what they are. The yi is dropped when the
phrase is negated:
A 68
"To some extent coinciding, to some extent not."
A 86 ... o
"Having respects in which they are similar. . . . Not having the similarity."
1/3/12/4/4. Yi presents problems only when it stands as the main

verb of the sentence. In two cases it is equivalent to yi wet 'deem':
150 Grammar
B 47 x±mm °
"It is not that one thinks of the heat of the fire as belonging to oneself."
B 71 o
"To suppose that it fits the fact is necessarily ill-considered."
It is possible that in both cases wei has dropped out of the text, but
this usage is attested elsewhere in Mo-tzii (ch. 39, Sun 189/9 ftKIA-fe
" A l l think him an excellent man").
1/3/12/4/5. In other cases the instrument with yi is all that is explicit
of a clause with a dropped verb, and its significance is only plain when
one recognises the type of verb:
(1) Preposition yi as in A 3 X 9& J ' UffiWiHb ° ("As for the
'intelligence', it is the means by which one knows"):
B 46 %w^m№ o
"One knows, but it is not by means of the five senses."

(2) Preposition yi as in A 80 Si&M, £tiS ("That by which one calls
something is the name"):
A 78 r £ * J t t # » & J 2 l J & £ - u l °
"For 'like the object' one necessarily uses this name."
(3) Preposition yi as in B 9 %kBr£№& ("The reason why a thing
is so"):
B 10 H # £ « M & » £i1km%&.* 0 * > S * « J » - « i o
"Whether the fighter's collapse is because of drinking wine or because of
the midday sun cannot be known."
(4) Preposition yi as in B 1 M ^ - t f e t S S X ^ i k . " O n the grounds
that it is so of the instance here, argue that it is so of what it is judged
to be":
B 78 o
"Whether criticisms are admissible or not is not (to be argued) on the
grounds that they are many or few."
(5) Preposition yi, with li M 'link':
B3&M °3gisr5fcj
"When deemings are linked, one cannot be sure either that they will be of
a kind or that they will conflict. . . . For example, someone deemed a fu,
if you use yung (to link with it: yung fu 'bold fellow') is not being deemed
a fu (husband); something deemed chu, if you use mat yi (to link with it:
mat yi chii 'buy coat and shoes') is being deemed chii (shoes)."
1/3 151
(6) Conjunction yi, with a preceding verbal unit dropped. (Cf.

B
Mo-tzu ch. 10 (Sun 40/7) ] l f c « I B A £ « , * J § 1 # * « @ K « ^ - & "This is
to say that the virtue of the sage is long-lasting in its glory, extensiveness,
firmness"):
B 64 ffHHSkX . . . SfTffil» ^JÂ-ifa o
"If one travels for a certain distance one does so for a certain time. . . . If
people travel for a certain distance they necessarily do so for a certain time."
9
A 50 jh °
" T o 'stay* is to be so for a certain time."
1/3/12/4/6. The instrument with yi precedes the verb. With a couple
28
of exceptions to be considered below, this rule applies even when the
verb is objectless:
B 48 "choose by means of the name"
9 1
N O 10 &MM K t S ? "accept according to the kind, present accord
ing to the kind".
Nor is it affected by the length of the instrumental phrase, however
unwieldy:
9
A 96 J ^ ( A ± ^ H # f ^ H f ) ] t H A *
"Using what is black or what is not black in a man tofix'black man'. . . ."
N 0 11 r « J * * ^ ( * № ^ ± ^ * 9 f ^ ) ^ ± i b °
" T o 'infer* is to present to him on the grounds of a similarity between
what he does not accept and what he does accept."
1/3/12/4/7. The elsewhere familiar formula K X J S / ^ / * Y ('give/
tell/show X to Y') is consistently avoided in the dialectical chapters. This
is easy to understand, since when the Mohist discusses names or kinds of
things in abstract terms it is important for him to avoid confusion between
the secondary object and the instrument. When a verb has two objects he
puts the secondary object last:
B 69 ft A ? A S "Cause someone to give someone wine."
By shunning the construction K S ? A he escapes the danger of
ambiguity in N O 10 "present according to the kind" (not "present
the kind").
When a secondary object stands alone he does mark it by yi, but
distinguishes it from the instrument by placing it after the verb:
A 32 £ r & ; £ £ "inform of this name."
(Contrast B 48 £ t £ & "choose by means of the name" B 53 J ^ £ i §
(=7J\) A "show to others by means of the name"):
2 8
Cf. § 1/3/12/4/7.
152 Grammar
B 41 M± "answer him"
№&>!%%i "answer that you do not know it".
In an obscure phrase in B 44 (f£$^Mll) the word-order shows that
it is not to be taken as 'explain by means of. . (cf. B 1 3&£ljlt;R$*"ffe,
t£ . . . "He, on the grounds that it is so of the instance here, explains . . ."
cf. also B 66), but as 'advise to . . (cf. Mencius 5A/7 tft^JJltfeJC "advise
him to attack Hsia").
1/3/12/5 ErhM
The rule that both units linked by the conjunction erh are verbal
applies universally in the Canons and Explanations. The relationship is
either of co-ordination or of subordination of the first unit to the second.
Throughout pre-Han literature the unit after erh is verbal, but inside
subordinate, embedded or auxiliary clauses the first unit is sometimes a
subject or exposed element. In Expounding the canons there are cases of erh
after an exposed directive in a preliminary clause. The purpose in each
case is to contrast directives in parallel clauses.
EC 8 ( < » > 0 f * # i f f i * * ) » · (ftjfrttYifoX*)»m±
"Choosing between things that one does not yet have is choosing the greate
among benefits. Sacrificing one of the things that one already has is choosing
the lesser among harms." (Cf. also the contrasting definitions at the
beginnings of E C 7, 8.)
1/3/12/6 Yii n
The directive preposition is yii, replaced by hu ^ only in a special
29
pattern discussed elsewhere. In the 3rd person there is a directive pronoun
yen My equivalent to yii chih 1£. (to be distinguished from the conjunction
30
yen 'only then', found in Names and objects and elsewhere in Mo-tartf.)
In a few cases the directive unit stands in the main verbal position:
B 37, 38 (Contrasting Canons) » # ^ * 0 * § · . . . # » · fe—ffi
" I n one thing there is something he knows and something he does not
know. . . . As for pointing one of them out, it is inescapably in two things."
(Cf. also the last sentence of B 53.)
The directive unit occupies the last nominal position in the sentence,
a rule of word-order observed with the usual grammatical rigour of the
2 9
Cf. pp. 133, 134 above.
3 0
Cf. § 1/3/11/5.
1/3 153
dialectical chapters. It is a rule with one exception; when the main verb
is wu M followed by so $f (there is no example of yu it is necessary to
detach the directive from the final verb and attach it to wu, and it is there
fore transposed in front of wu:
B 22 mm±*$k » ftSAKBftUE ·
"What is lustrous(?) in the man looking at himself is mirrored in the
mirror without exception." (Cf. also A 65.)
The directive unit may like other units be exposed, but for the sake
of a contrast significant to the understanding of the sentence. In the single
example in the Canons and Explanations, quoted above (B 37), it heads the
sentence and is resumed by yen ; in the three examples in the Expounding
31
the canons it precedes erh in a preliminary clause.
1/3/12/7 Yi &
Except for the interrogative particles and yeh the only final particle
is yi, which marks the point of transition from one state to the next in the
manner of colloquial le ~T. It appears only six times (A 28, 98, B 26,
N O 10,15,17), and in each of them the transition point is easily recognised:
A 28 a-^ffp^ "Once our affairs are at the point of being properly
ordered"; B 26 H M S B J *ih#l "When the leverage and weight of the
two have equalised they come to a stop"; N O 15, 16 ...
"If it is like this (with A , B , C . . .) then there will no longer be any difficulty
(about X , Y , Z ) " . In two cases the transition point has been described as
'awaited* (ffi) in an earlier phrase:
N O 10 . . . &№&^&M'&& . . . ay&B£ °
" I f . . ., he can be expected to get into trouble at any moment. I f . . .,
then he will certainly get into trouble."
N O 17 T * * A J XttXXfLA » « * H * ^ « A * °
" 'He does not love men* does not require him to (literally 'wait for him to')
love no men at all; he does not love men without exception, on these
grounds he is deemed not to love men." (Here we could bring out the effect
of yi by inserting 'it is sufficient that* after the semi-colon.)
1/3/12/8 Yeh *
Since in most texts it is difficult to come to grips with the particle yeh
it deserves close attention in a document in which words are never used
casually. The units which it concludes are of two main types:
9
Type 1: the sentence pattern '(X) Y yeh ( ' X is Y ' ) .
3 1
Quoted § 1/3/12/5.
154 Grammar
Type 2: the nominalised verbal or directive phrase in the form ' X chih
Y yeh* (B 77 "not know that learning is useless", B 44
3=f (S^J&Aifa) "Like the effect of wine on a man"). X is a nominal unit
subordinated to the verb, not necessarily its subject (EC 2 (E3^t;£j£A'tfe)
$r (^^SA-tfe) "Yesterday's love of men is not today's love of men",
( * K ^ * A * ) 75 ( * S I ^ * A - f e ) "The love of man which is love of Jack
is the love of man which is love of Jill"). When subject, ' X chih* may be
replaced by ch'i ft. Chih itself is replaced by ch'i when X is a contrasted
32
demonstrative subject.
The reference of the nominalised verbal ' X chih Y ' phrase is to the
action and not, as in the subjectless phrase nominalised by che ^f, to the
agent. (A 98 t S * (Jfcft^-tfJ) "He refers to what is so and
deems that it is so of the thing here", B 70 ( 5 g f i ^ ) # G ° . . . WS{%&№)
"What is like the white is necessarily white. . . . Therefore I know it is
white"). Exceptions however are found where ch'i stands in the place of
' X chih* (A 26 s № & "The one of them which is harmful is not
this one", cf. B 52).
The ' X chih Y ' phrase is generally although not invariably followed
by yeh, wherever it stands in the sentence.
B 57 ( © ± 1 « ) » M £ » o
"In the case of a pillar being round, when we see it, its place in the idea
is unchanged."
A 20 fr£o. . .
"One gives him the name ('Brave') because of what he does dare. . . . "
B 65 ( — & # £ * B f H & ) « » o
"The agreement of things which have one standard is complete, like the
coincidence of squares."
Apart from this type of phrase, what kinds of element are marked off
by yeh at the beginning of the sentence ? There appear to be no convincing
examples of other types of verbal unit, whether nominalised or not. We
do find nominal units, but confined to the following two types:
Type 3: a temporal word or proper name followed by a contrastive
yeh. The contrastive yeh is frequent after chin ^ 'now', ch'ien M 'before',
shih № 'in the beginning': B 32 (Mtii) ^ ' | f , IS "Previously he was
not afraid, now on the contrary he is afraid". There is one instance with a
proper name (a usage familiar in the Analects):
EC 4 ^(îb)^ffnAT*»(^wmm)M^ °
3 2
Cf. §1/3/4/9.
1/3 155
"Supposing that if of all men Jack were to die the world would be harmed,
my special care for Jack would be ten-thousand-fold."
Type 4: A subject to which shih 'the said* is adjunct is always followed
by yeh: A 78 lhS£;SkJ№ "The said name is confined to the said
object". B 53 <&>»4% "This vocal sound is
born in the present, the object taken as example resides in the past".
N O 1 ^ (diS-tfe) £ " I f the said stone is white", * (=88) A
"Although the said stone is big", N O 3 (:&S№) S S i f e "This half-disc
is this jade". The purpose of this usage, also found in such texts as
Chuang-tzu and Hsiin-tzu, is presumably to avoid confusion with a
resumptive shih at the beginning of the sentence.
Returning to Type 1, since the Mohist logicians are especially
concerned with X being or not being Y the '(X) Y yeh' pattern appears
in a variety of guises. It is frequently embedded in verbal phrases, some
times following verbs after which one would not normally expect it:
E C 7 JX!«S(*ffiife)rfnfIJ± o
"Benefiting Jack on the supposition that he is one's parent."
(Contrast the immediately following K^S$I*-?ffnS*^^±
"Desiring music for one's son on the supposition that it will benefit him".)
B 53 *(3Cttffi-fe)» Ate » SBtKi&A-fe °
"Mentioning a friend as being a rich merchant is showing to others by
means of names. Pointing something out as being soup is showing to others
by means of the object."
A 39 — A f f i f t ·
"They are two men and both see that it is a pillar."
B 8 35HS(*fe)» fftKS-fe o
" 'Loan-naming' a dog as being a crane is like using 'Crane' as a surname."
This usage is important for the understanding of certain difficult
sentences in Names and objects, in which we mark crucial phrases by letters:
N O 3 Mm0M^
A
» c
o *(B}g£A-fe)#«A-tb ' « ( » f e )
"Visualising a pillar is not visualising wood, it is visualising the wood of

the said pillar. Visualising a finger as being the man is not visualising a man;
but visualising as being game on the contrary is visualising birds."
E
N O 7 (°№±AtiL)m(-n±A&)m » A£f*#-«#fe °
"The finger which is the man is different from the head which is the man,
because the man's members are not things identical in appearance."
cf. B 53 { m±m^)^^mm^mmm
F
-
156 Grammar
"Yao's being an example originates in the present yet he resides in the past
and they are different times."
We cannot take B or D as 'the finger's man' or F as 'Yao's example*
because when followed by yeh ' X chih Y ' is always a nominalised verbal
phrase (cf. E , where yeh is absent). We cannot take D as 'pointing out this
man* or C as 'visualising the game* because only units of Types 2, 3, 4
can have a concluding yeh at the beginning of the sentence (cf. A, where
yeh is absent). The strictness with which this rule applies to the '(X) X yeK
sentence can be easily confirmed by running through the long sequences
of such sentences in Names and objects.
We find the ' X Y yeh' pattern also in front of the conjunction erhffn:
A 3, 4 r » J · » B5&» · . . . T * J * * » J^*n##-til ·
"In the case of the 'intelligence* it is the means by which one knows, and
one knows with certainty. . . . In the case of 'thinking*, it is to seek some
thing by means of one*s intelligence, but one does not necessarily find it.**
When X and Y are nominal units the logical relationship is of identity
or class membership. There is of course no counterpart to the Indo-
European use of the copula before predicative adjectives, since words
translatable as adjectives operate verbally and do not occupy the Y position
of this pattern (NO 1 (Jk^rtfe) £ " I f this stone is white"). We may
class the pattern with nominal X and Y as Type 1A, and distinguish from
it Type 1 B, in which X and Y are verbal units, which may be single words
or lengthy clauses.
In this sub-class no chih is added to the subject to nominalise the
phrase, and the logical relationship between X and Y is much loosened.
The nearest English equivalent of yeh in this construction is perhaps the
unstressed 'it's that* of the spoken language, used much more freely than
the standard 'It is that*:
B 43 !km± » o · o
"When the fire melts the metal it*s that there's more fire, when the metal
uses up the charcoal it's that there's more metal."
Here Y indicates the cause, and it is convenient to translate by 'it is
because . . .'. In the sections on optics and mechanics, where yeh is
infrequent, it is generally the marker of an explanatory following a
descriptive clause:
B 25 J P * » * - » ^ S » mmmfa o . . . · mnmfc °
" I f you lay a weight on one of the sides it is certain to decline, because they
are equal in weight and leverage. . . . The tip is certain to fall, because it
has gained in leverage."
1/3 157
Shih M. intervening between X and Y is found only in Type 1 B.

It generally preserves the strict relationship of identity or class membership;
if there is loosening, it is in the direction of allowing Y to be an implication
of X (not its cause or its grounds):
B 71 * ± * A±b^ » ( d & W ) » f l « M n * № o
"If this man's statement is allowable (that is to say, not self-refuting), this
implies that he allows something.''
y
' X shih Y yeh is used very freely, sometimes closely linking very short
phrases (in the bracketed phrase in the sentence above, which however
33
looks like a gloss, the 'shih Y ' is even parenthetical):
A 66 mwtmm. ° °
"Different places do not fill each other. Not being each other is excluding
each other."
B 77 » °
"This is causing to know that learning is useless, which is teaching."
l
An ' X shih Y yeh* clause may itself be a main clause after tse B J. In
A 75 and in B 38 we find four examples of a curiously elaborate sentence
pattern: A , K l B t e f t C - f e .
In each case B affirms or rejects a possibility raised in A : 'That it's
the case that you do/'do not. . .':
A 75 » mZM%*7i » fl!l»*-fiftfil9f*JtBf«-ft ·
"If whether it is beneficial or harmful beyond the wall is not yet knowable,
but if you hurry for it you will get money, then that it's the case that you
do not hurry for it implies that you are determining which you prefer in
terms of what you doubt."
B 38 - f · W ( = X ) ® * ^ t e M »WttMfam&fttiMfc »£-
"If you know what this is, and know too that it is identical with what I
referred to 'a priori', then that it's the case that you do not by knowing
what this is know what I referred to 'a priori' implies that even in one of
the things it is called there is both a known and an unknown."
There remain many ' Y yeK sentences in which we cannot identify the
preceding clause as X . Before considering them we must exclude cases of
final yeh which belong to other types. The nominalised clause of Type 2
(Noun verb ife) is as common at the end as at the beginning of a sentence.
There are also other patterns, which we may group under a Type 5:
3 3
Cf. § 1/2/2/6/5.
158 Grammar
Type 5A: Yeh after a passive verb following k'o "J, as in A 75 ^ "I
#flife "is not yet knowable". This usage with k'o (and also with tsu is
common to most pre-Han texts.
Type 5B: The pattern "It is said of. . . " . 34
There remains our final sub-class of Type 1:

Type 1 C : The affirmation or denial of one of alternatives, expressed
by yeh at the end of a verbal sentence or fei ^ in stead of pu /J* in front of
its verb. In general we may say that a verbal sentence answers a 'What ?'
question, which allows any number of answers, and that the addition of
yeh turns it into the answer to a 'Yes or no?' question. Most Chinese
thinkers use this formula so freely that one gets into the habit of assuming
that any final yeh can be put in this sub-class and then ignored. But the
later Mohists use it more rarely and with discrimination, in contexts where
the alternatives envisaged are generally plain. For example B 3 points out
a difference between the words erh 'two* and mei H 'beautiful*. Whether
we describe as 'two* depends simply on whether or not there are two
objects; if not, there is no saying of 'two*. The alternatives are yujwu
'there is/there is not*:
B 3 #±j№ff5Sefi£»№±nmmm& °
"Only if it's that there are these objects does one refer to them; if it's that
there are not, then it's that there is not the reference."
In the case of 'beautiful' however, the alternatives are that the thing
is either beautiful or the converse of beautiful:
" I f the reference is to this, then it is that this is in itself beautiful; if the
reference is to another, then it's not that this is beautiful; and if it does not
have the reference, it's that the converse applies."
In the next example the alternatives are that names are or are not
synonyms:
B 35 mmm&. s u n t ° mmmz r m j»immz. r * J* · mm
"As for the things so called, if it's not that they are the same then it is that
they are different. When if they are the same it's that one calls it 'whelp'
and the other 'dog' (a synonym), or if they are different it's that one calls
it 'ox' and the other 'horse' (not a synonym), and neither wins, this is
failing to engage in disputation."
3 4
Cf. § 1/3/12/1/6.
1/3 159
1/3/13 CAUSATIVE USES
In pre-Han Chinese certain intransitive verbs are used freely in

causative constructions: the main examples in the dialectical chapters are
'treat as admissible' and pi*j&'treat as necessary':
B 32 :foftR?E£ "Not be sure whether he is dead or alive" (cf. A14, B 3,73),
B 71 W "J "Admit something",
B 67 ^^mZ^M^kU "There are the same grounds for denying
that oxen and horses are not oxen as for allowing it."
The causative use of pu k'o 'disallow' is paralleled elsewhere:
Hsun-tzu ch. 21 (Liang 293/8) * "Taffi «T#3I °
"Reject the Way and admit what is not the Way."
The paired terms chih j h 'stay' (in an object) and hsing f} 'proceed'
(from object to kind) are both used causatively:
A 96 Jh T H A J . . . Jh r § A J "Fix 'black man' " . . . "fix 'love of man' "
(cf. also A 75, 97. B 1, 2).
B 1 ft A "Make the man proceed."
The latter is a unique example, but may be paralleled from Hsun-tzu:
Ch. 9 (Liang 108/10) J2URfr*l > E i — f l B °
"By means of kinds make the miscellaneous proceed, by means of the one
make the ten thousand proceed."
Except for such special words the ordinary practice in pre-Han
Chinese, which in the dialectical chapters as usual becomes a meticulously
observed rule, is that an intransitive verb is used causatively only with
certain words which mark it as transitive, the pronoun object chih or
fu wu ty, chu ftf, k'o RT, tzu @, so @f:
A 83 ffiM^® "Apply it on both sides, not on one without the other",
B 16 ffistTfi"... Ê^/T' "Locate it in the past.... Locate it in the present,"
B 78 «J№ "rejcctable",
B 79 "not reject it",
B 42 "may be treated as belonging",
N O 15, 16 "condemn themselves",
"condemn them",
(It may be noticed that fei is not one of the verbs which are freely causative,
although we do find it once in the combination A2£ 'others' disapproval'
(EC 1). Cf. B 3 A ^ : 'others' approval'.)
B 53 B r i l (=IK) "What one takes as an example",
B 81 0fW (=fn) "What one takes as the ch'ing (what the thing essentially
is)",
E C 8 0fW "What one treats as the t'i (units/parts)".
160 Grammar
Some pre-Han philosophers were fond of another causative construc

tion, with repetition of the same verb, noun or pronoun. The repeated word
could be either affirmative 'treat things as things') or negative
("Hpf* 'admit the inadmissible'). The affirmative repetition is common in
the dialectical chapters:
E C 8 2 № "reject the wrong",
B 79 2£f# "reject rejection",
B 82 AM "treat this as this".
The second example is remarkable since (as the argument of B 79
confirms), fei 'criticise' is used as fei 'wrong' rendered both causative and
syntactically mobile by the added radical.
More exotic examples are the iifcjlfc 'this-ify this' and ?$jit 'that-ify
this' of B 68. Examples of this sort in philosophical writing are seldom
confusing when a simple rule is grasped; the pattern is 'causative verb
object' unless the word on repetition is marked as a verb by particles, in
which case it is 'subject passive verb':
Chuang-tzu ch. 6 (Kuo 253/1) $l£M^$L °
"The engenderer of the engendered is unengendered."
Ch. 20 (Kuo 668/5) mWtt^Wto °
"Objectify objects but not be objectified by objects."
Ch. 22 (Kuo 757/6) °
"The unformed which forms forms."
In B 68 and also in B 82 the pronoun is marked as verbal and passive
by yi $\ ch'ieh EL and erhTO.There is however an extraordinary example
in the Canon of B 68 which is not elucidated by the rule:
A B C D E F
« j l f c « ] l f c J S « j ! f c l R l "
35
The sentence pattern signifies ' A B C D is required by E F ' . A B C D
should be verbal, since it is co-ordinate with E F (That-ifying this'). Since
C D repeats A B they cannot be co-ordinate, and A B is therefore subordinate.
We may take A B as 'There and here', on the analogy of a similar discussion
in Chuang-tzii:
Ch. 2 (Kuo 66/4) tt*—» jifc#—ft^ ·
"From there you have one 'is it' and 'isn't', from here you have another
'is it' and 'isn't'."
If so, C and D will be co-ordinates like A and B, and transitive since
as the core of A B C D they are co-ordinate with the verb in E F , which is
transitive:
3 6
Cf. § 1/4/36.
1/3 161
"That-ifying and this-ifying both there and here is required by

that-ifying this." ("Using both 'that* for this and 'this* for that is required
by using 'that* for this.")
There is one type of causative pattern common in Chuang-tzu which
risks confusion with a co-ordinate construction: " T ^ " J 'admit the
inadmissible', also interprétable as 'the admissible and the inadmissible'
(as in B 78). The Mohist consistently avoids it. We have noticed his
tendency to join verbs with fou ï 5 and even wei as in B 78 " I S 'admis
36
sible or not' ; his motive is no doubt to avoid this confusion.
1/3/14 STYLE
The style of Chinese dialectical writing is marked by syntactic

regularity, sparsity of idiom, and a vocabulary which is generally simple
but complicated by the presence of technical terms which can be under
stood only by collating examples and examining them in context. This
style, to which philosophical writing tends in all languages, is apparent in
the propositions of the sophists, in the White Horse and Pointing things ou
of Kung-sun Lung (although not in the forged chapters of the Kung-sun
Lung-tzu) and, within the dialectical chapters of Mo-tzu, in the Hsiao-ch'u
and all other parts which are immediately intelligible. If it is not so obvious
over a large proportion of the Canons this is partly because of the frequency
of obsolete graphs, which disguises the simplicity of most of the vocabulary
although such graphs when identified have generally proved to represent
quite ordinary words, and partly because of the piecemeal approach of
editors, who will give a word the rarest or least firmly attested of meanings
as a short-cut to understanding its immediate context. A high proportion
of proposed interpretations may be rejected out of hand on grounds of
style. The Mohists are often obscure to us because we have no independent
information about problems which they assume to be known to their
readers, but it is not their style which is obscure; they do not write
elliptically or paradoxically like Taoists, do not hunt out rare words, do
not show off their literary education.
The peculiarity of the later Mohist style can be seen in the behaviour
of certain key words such as fei 'is not', jo 'like', tsai 'is in'. These often
appear in unexpected contexts, tempting us to look for some special idiom
attested elsewhere in early texts. But the point is that the Mohists have a
much greater interest in the logical permutations of X not being Y or being
like Y or being in Y than any other pre-Han thinkers; they consequently
8 8
Cf. 1/3/5/3, 4.
162 Grammar
give such words much of the syntactic mobility of ordinary verbs, in

defiance of ordinary usage. We have already noticed the extraordinary
37
mobility of such words as fei 'is not* and wet 'constitute, be deemed'.
Two more such words which deserve attention are/o 'like* and tsai 'be in',
both of which may serve to illustrate a point of some importance. In most
kinds of Chinese writing one expects a word to have a variety of meanings,
distinguishable in theory by looking up a dictionary, distinguishable in
practice only when we have become familiar with the sort of context to
which each is confined. But the key words of Mohist dialectic are shorn of
all but their basic meanings; and this semantic restriction, together with
the precision of the syntax, frees them from their ordinary contextual
limitations.
Jo 'like' often appears in contexts which have tempted editors to resort
to some rare use of the word (such as M . 30796 def. 14 'this' in A 31, 78
below, or def. 20 'follow' in A 70, 71 below). But in fact it is used with
perfect consistency although with the syntactic mobility of ordinary verbs,
as a member of a family of phonetically cognate words, / o / * N J A K 'like',
/tfw/*NJAN 'so' and also nuo/*NAK IS 'assent' (A 93, 98).
A 70 m ' Bpeffa&te °
"The 'standard' is that in being like which something is so."
A7i r ^ J t e # » * a ^ t e o
"Being 'so' is the characteristics being like the standard."
A78 r^tJifef
"For 'like the object' one necessarily uses this name."
A31 r S S J t t f » I X f * t i f c «
"For 'like the stone' one necessarily uses what is like the name."
B 70 7g&m&& ° ^ - t M S K f e f c e s - t b » ^ K e - t b °
"What is like something white is necessarily white. In the present case we
know that in colour it is like something white, therefore we know that it is
white."
Jo frequently has an implication of degree (as in ffiS 'as much as
each other', 'equal', B 25, 26. E C 3, 5-7, 12, 13). In ordinary p re-Han
idiom jo is never negated without implying degree (pujo 'not as much as,
not as good as'). But for the Mohists the syntactic mobility of a key word
is more important than idiom; they therefore assimilate the negative to the
38
affirmative :
37
Cf. p. 118 above.
3 8
Cf. pp. 137, 138 above.
1/3 163
A 20 "Like committing a robbery",

A 7 ^%StM "Not like loving a horse",
(/>w> 'not like' also A 11, 74. B 3, 81).
Tsaiffi'is in* is used nominally ('presence') and causatively ('locate'):
B41 'ffi*J&£Aft o
"Among Heaven's constants its presence is prolonged with man."
B 14 mtfc*R » m(=X)&M °
(On the mutual pervasiveness of space and time.) "North and South are
in the morning, and again are in the evening."
B 17 o
"For example when being in it is finished the past ceases."
B 16 £ s f ft§ff£ ' ° "Locating it when it was so or in its future."
@ ^ j £ t f "frtfe ° "From the present locating it in the past."
H l K E ^ ^ ° "From the past locating it in the present."
If we found the tsai of B 16 in any other text we should probably give
it the rare sense of 'scrutinise' ( M . 4881 def. 4). But in the Mohist literature
common words which have not been adapted as technical terms are likely to
keep their common meanings; it is the abstractness and complexity of Mohist
thinking which causes them to appear in unexpected syntactic positions.
It will be seen from these examples that an ordinary sense of Chinese
idiom is sometimes a misleading guide to the reading of the later Mohists.
Thus no modern Chinese editor accepts the text of the extraordinary
sentence W W f e ^ ^ S , ^ ^ j K S i b . (NO 2). But when we understand that
the Mohist was especially interested in the quantification of the object,
39
regularly used the formula 'yu—verb—yu—object' and cared more
for syntactic consistency than for idiomatic smoothness, we see that this
was the natural way for him to say " T o have some Ch'in horses is to
have some horses". There are even cases where a beginner in Chinese,
looking up the words in the dictionary and applying the rules of syntax
learned in class, might do better than an experienced reader habituated to
pre-Han idiom. For example previous editors have either emended the
Canon of B 56 ( f t K H ) or discarded it as unintelligible. They do not
consider the possibility that it means "Causing a mansion to be beautiful"
because they know that in no recognised Chinese style would the idea be
expressed in this way. We should not be surprised to find it expressed by
the two words in a series of parallel phrases which establish it as a
'verb-object' construction. Let us suppose however that the Mohist is not,
like the writers whom one would expect to use such a phrase, talking about
39
Cf. pp. 131, 132 above.
164 Grammar
mansions and their beauty, but is analysing the ambiguity of the phrase
itself, which might mean either 'beautify a mansion' or 'deem a mansion
beautiful' (this does in fact seem to be the theme of the Explanation of the
Canon). If he uses the phrase in isolation it will be taken as an 'adjunct-
head' construction ('beautiful mansion'); to discuss it he therefore has to
expand it with the addition of shin 'cause'.
A further characteristic of later Mohist style is its extraordinary
bareness and economy. The older chapters of Mo-tzu are notoriously
long-winded and repetitive, but the authors of the dialectical chapters
consistently refuse to make the same point twice. It is remarkable that no
definition is ever repeated; the Canons omit those already laid down in
40
Expounding the canons (EC 7-9) and the lost document on the 10 theses,
Names and objects defines 'some' and 'exemplify' (NO 5) without repeating
the definitions of 'all' (A 43) and 'standard' (A 70) on which they depend.
Every lacuna is an irreparable loss, since whatever was said will have been
said nowhere else. A n editor must sometimes sigh for the repetitiveness of
the older Mohists, so useful both for emending and for interpreting the text.
A more fortunate consequence of this close-packed writing is that the six
dialectical chapters are incomparably richer in content than all the rest of
the first 51 chapters taken together. (The 20 military chapters, because of
their specialist nature, are in another category.) The impression that one
has poorer materials for the study of later than of early Mohism very soon
turns out to be the reverse of the truth.
One feature of this bare style is the absence of the pairs of near-
synonyms, verging on compound words, which are so common in pre-Han
writing (cf. Mencius 1A/3 ^WdW:M^^ "carry on their backs and heads
on the highways and roads", translatable simply as "carry loads on the
roads"). Even when we meet such a phrase as liyung ?']ffi 'benefit and use'
(A 93) or t'ien ch'ang %1$ 'Heaven's constants' (B 41) the presumption
must be that the Mohist is giving both words their full weight. In other
texts the word wu ffi, 3 £ 'match point-by-point' is seldom used except in
combination with ts'an # 'align/co-ordinate', or opposite it in parallel
phrases; but the Mohist uses either ts'an alone (A 57, B 38) or wu JE, fip
41
alone (A 98, B 58, 76). There are however certain combinations which
look like reduplicatives:
A 47 n. 133 chii-chih (1),
A 88 n. 237 MM huan-yiian,
A 88 n. 238ffiffiyung-t'ung.
4 0
Cf. § 1/6/2.
4 1
Cf. §1/4/33.
1/3 165
A l l three contexts are appropriate to words descriptive of directions

of motion (of the type of English 'zigzag').
The Mohist does not reduplicate the same sound. In a couple of cases
of what look like words elsewhere reduplicated, he does not do so:
B 10 n. 301 M p'eng for p'eng-p'eng,
B 57 n. 494 yang-jan for yang-yang-jan.
Postscript
The observations on hu ¥~ on pp. 133f, 152 require revision in the light

of my " A post-verbal aspectual particle in Classical Chinese: the supposed
preposition H U " , forthcoming BSOAS, 1978. If hu is a post-verbal con-
tinuative particle the deletion proposed on p. 134 is unnecessary (B 72
"If the other still goes on being specifically. . . . If the other is not speci
fically. . . . "). The other translations also require revision (B 72 "The
speaker never has anything . . . " : B 74 "Some go on being left . . . " : N O
15 "The heart goes on lacking a hollow . . .").
1/4
TECHNICAL TERMINOLOGY
1/4/1 T H E P R O B L E M OF T E C H N I C A L T E R M I N O L O G Y
W E have more than once noticed that the vocabulary of the dialectical
1
chapters is in general very simple, although this simplicity is often dis
guised by graphic confusion due to imperfect standardisation by later
scribes. But there is one important qualification to the claim that the later
Mohists use only ordinary pre-Han words in their ordinary senses; since
logic, optics and mechanics are topics rarely discussed elsewhere in early
literature we must expect to find words appearing in unexpected contexts
and adapted to peculiar technical uses. The Mohists define many of their
special terms (A 1-75, E C 7-10, N O 5), but the usages which seemed to
them to require explanation did not of course include all that puzzle a
modern reader.
It may be useful to formulate certain general principles for dealing
with technical terms:
(1) When we find a common word recurring in unfamiliar contexts
(chih #1, kuo My mao I&) the presumption must be that it has a special
technical use. With this as with other problems presented by the dialectical
2
chapters if we resort immediately to conjectural emendation we may miss
important observations.
(2) In spite of the effects of later graphic régularisation it is still clear
that the Mohists distinguished some of their technical terms by the
addition of radicals. We have noticed elsewhere the differentiation of
chihftl'the consciousness', chih ^ 'know' and chih 25 'knowledge', as well
as special coinages with the 'man' radical some of which are technical
(fan IS 'converse', thing M 'do the same', pi № 'commensurate', ch'iï ffi
'group separately/mark off', hsùan M 'settle on the preferred alternative'). 3
1
Cf. § 1/2/1/2/3, § 1/3/14.
2
Cf. Sun Yi-jang's emendations of these graphs in Sun 205/3, 209/-4, 237/-4,
238/-4, 240/-5, 258/-1.
3
Cf. §1/2/1/2/2-10.
168 Technical Terminology
4
It is unsafe to assume, with most editors, that an unknown graph is simply
an obsolete form of a known one.
(3) Technical terms were especially liable to corruption after their
significance had been forgotten. This is most obvious in the case of specially
coined graphs, our knowledge of which almost always rests on a few surviv
ing examples which have escaped being stripped of their radicals or
5
mistaken for more familiar graphs. Consequently we should use textual
emendation, not to avoid the problem of technical terms, but on the
contrary to recover other instances of terms of which we have found
uncorrupted examples. In this connexion it is interesting that most of the
few corrupt or unidentified terms defined in A 1-75 have the 'man' radical
characteristic of Mohist coinages ( № A 15 and again A 71, ifo A 73, cf. №
A 39, iM A 47, lit A 68). The most promising place to look for such words
is not in dictionaries but among technical terms used by the Mohists but
which lack definitions.
(4) Putting aside its ordinary uses, a word has only one technical use
in one field of discourse and has no synonyms. Among the words examined
below there are, however, a few which have different uses in logical and
y
in scientific contexts (cheng IE, ch uan W , ch'ii B , sheng M ) , or logical an
ethical contexts (wei S , hsing fs, hat Hf). These three fields of discourse
fortunately tend (although one must look out for overlaps) to be compart
mentalised in separate parts of the documents:
Ethics: E C , A7-39.
Logic: A 1-6, 40-51, A 70-B 16, B 32-82, N O .
Science: A 52-69, B 17-31.
(5) A familiar word used technically is likely to assume an unexpected
syntactic mobility. Thus we need not hesitate to interpret chih as not
only verbal ('uphold one alternative') but nominal ('the alternative upheld'),
provided that we can find supporting examples in the corpus. We are
familiar with Vi I t as a noun ('limbs, members') and chien i as a verb
often used adverbially ('collectively'); since the Mohists use them as a pair
of technical terms, which we examine under the definition of Vi in A 2
('unit and total', 'part and whole'), it is necessary for them to assimilate the
two words syntactically, regularly employing chien as a noun and on occa
sion using Vi adverbially ('individually', A 7).
(6) It is important not to equate too easily a Mohist term and a word
from our own philosophical vocabulary. A reader of Chinese does not
4
Cf. Sun 198/7, 199/-5, 207/4, -6.
5
Cf. § 1/2/1/2/2-12.
1/4 169
expect the cardinal virtues jen t and yi IS to correspond more than

approximately to their conventional English equivalents 'benevolence* and
'righteousness', or even yang ¥ to be exactly equivalent to 'sheep' and
yii 3£ to 'jade'. In the case of logical concepts however we are involved in
the very interesting question, revived by the work of Noam Chomsky, of
whether there may not after all be an ultimate identity of the forms of
thought in all civilizations. When the Mohists not only use the current
word pi № 'for sure, certainly' in such a sentence as " I f they do not both
fit the fact, necessarily one does not fit the fact" (A 74 ^flcllf, ),
but also make it clear that for knowledge by explanation this is the only
kind oipi (A 51, 83), it seems pointless to question that what is pi for them
is logical necessity for us. On the other hand it would be a disastrous
mistake to suppose that because 'is an ox' and 'is not an ox' are called
fan ifc there must be an exact equivalence of fan and 'contradictory'. The
point about fan is that if one kind of thing is an ox we can 'reverse' and say
that all others are not oxen; as contradictories only one is true, as fan both
are true (A 73, 83). For us oxen and horses are 'classes', for the Mohists
they are lei SB, but it does not follow that we can translate lei by 'class'.
In English X and Y are or are not members of the class Z, in Chinese X
9
and Y are 'the same in lei or are 'not leV (Sfiî, ^>M). In such cases
it is advisable, even at the cost of making the Mohists look less sophisticated
than they are, to choose a non-philosophical word (lei 'kind', 'of a kind').
(7) One needs to be on the lookout for terms which form pairs, even
if one never meets both together. Examples are hsing ff 'go' and chih i h
'stay', tang IK 'fit' and kuo № 'go beyond'; ku H 'inherently' and ku #C
'the thing as it inherently is'; mao JfiL 'features, characteristics' and ch'ing If
'what is essential to the thing'.
1/4/2 C H E N G I E , -flr
Cheng I E , T& 'exact, direct, straight, upright' (EC 1, 2, 7, 8. A 53, 56, 83,
84, 99. B 21-24, 28, 31, 51, 62, 68, 70. N O 9, 12)
The graph f& for cheng was one of the special graphs promulgated by
6
the Empress W u 1§i (A.D. 684-704). It is remarkable however that in
Mo-tzu it appears only in the dialectical and military chapters, the ones
richest in archaic graphs. It is used fairly regularly in the Canons (with
exceptions in B 28, 51), but only twice in the Explanations (both in B 31)
and never in the other dialectical chapters.
6
Cf. Luan (1957) 99. Tung and Wang, op. cit.
Cheng is one of the terms which have different uses in scientific and
in logical contexts. Its basic sense however is always 'exact, dead on'
(cf. A 56 i & S 'due South'), assuming coincidence with an implicit standard.
In the geometrical and scientific sections a body is cheng (not yi № 'slanting'
or yi J?r 'inverted') if it is upright (B 21-24, 28, 62), an edge or surface is
cheng if it is even, straight (A 53, B 22, 23). It may be noticed that in the
latter usage cheng is not synonymous with chih lit, the closest equivalent
of English 'straight', used of a rope, a path of light or the direction in which
a finger points (A 57, 99. B 23, 26, 27, 38).
In scientific contexts cheng is frequently nominal, whether it refers to
bodies, edges or surfaces: B 23 $i iE 'skirting the upright figure", A 53 DUE
'along a straight edge', B 22, 24 M J E 'going beyond the plane'.
In logical contexts cheng contrasts with yi 3EC 'appropriate', ch'uan W
'weighing' and perhaps fu \% 'compound'. A black man has black parts and
parts which are not black, and to decide whether he is a black man we have
to ask 'which is appropriate?'; but the agreement of a circle with its
standard is cheng 'exact' (A 83, 97, 99 cf. Mo-tzu ch. 31, Sun 146/-1
lE^J "The shape of his face was exactly square"). Similarly there are things
which are desired or disliked only after weighing one against another, but
there are also desires which are cheng 'direct, immediate' (A 84, E C 8). In
one obscure passage (NO 9) names seem to be described as compound or
as cheng, exactly and immediately fitting the object.
p
Cheng is sometimes transitive, 'put right': B 28, 51, E C 1 / F J i E
'cannot be righted' (in the first example the reference is to restoring to the
perpendicular): B 68 J E £ I 'get names right': E C 2 lEft 'hold the limbs
right' (cf. also B 31, 70). It may be noticed that these are the only instances
in which it is not misleading to translate cheng by English 'correct', and
even of these only a couple in Expounding the canons concern the correcting
of ethical behaviour (EC 1, 2). The flavour of moral rectitude which the
word conveys in Confucian vocabulary is completely absent from its usages
in the dialectical chapters.
1/4/3 CHIEN PAI
Chien pai i £ & 'hardness and whiteness' (B 37): 'as-hard-to-white, distinct

but mutually pervasive' (A 66, 67. B 4, 14, 15)
Previous editors have tried to explain the Mohist references to chien

pai in the deceptive light of the Essay on hard and white (Kung-sun Lung tz
ch. 5), which argues that a white stone is two things, the hard stone which
one touches and the white stone which one sees. In 1957 I offered evidence
1/4 171
that this and the other two essays of Kung-sun Lung tzii ch. 3-5 were
forged between A.D. 300 and 600, utilising misunderstood scraps from the
1
Canons and Explanations. But such is the hypnotic effect of familiarity with
the spurious essay that I continued to take it for granted that the many
pre-Han and Han references to Kung-sun Lung's separation of hard from
white imply a specific sophism, against which the Mohists defend the
common-sense position that the two properties are mutually pervasive. It
was not until 1967 that I came to appreciate that references earlier than
A.D. 300 are not to a sophism at all, but to a theme in disputation, the
separation of distinct but mutually pervasive properties in general, for
which Men pai is a technical term (for example, Kung-sun Lung's
8
separation of shape and colour in a white horse).
Liang Ch'i-ch'ao long ago put his finger on the puzzling fact that the
six Mohist references to Men pai seem to have nothing to do with the
9
Essay on hard and white. This led him to suspect that in nearly every case
one or both words is an interpolation. More recent editors dismiss this
observation and propose drastic emendations and transpositions to make
the Mohist say what a reader of Kung-sun Lung tzii would expect him to
be saying. (See, for example, T'an Chieh-fu's reconstructions of A 66,
B 14, 15, in T'an 94, 140). But once we are free of preconceptions imposed
by the forged essay it becomes plain that Men pai is simply a general term
for distinct but mutually pervasive properties, of which hardness and
whiteness are taken as the typical example in B 37. It is defined among the
geometrical terms:
o
A66â,;wHii
"Chien pai is not excluding each other."
This can be understood as a proposition in reply to Kung-sun Lung
("Hardness and whiteness do not exclude each other") only by ignoring
that its position is among the 75 definitions ending in yeh, not among the
propositions of A 88-B 82. Syntactically Men pai operates as a compound
verb:
B 4 №«§g£ °
"Length and breadth are Men pai."
B 14 ? ^ / f M o
"Space and duration are not Men pai."
7
G(2).
8
G(7).
9
Liang (1922) 108, 109.
b15-*ha№?Kfi"
"The durationless is chien pai with space."
The technical use of chien-pai and its curious syntax may be further
illustrated from Han Fei tzu:
Han Fei tzu ch. 34 (Ch'en 753/-1) * A # & A £ 0 f * « * b ° A £ # f g S i
" A weighty man is inevitably a man whom the ruler greatly loves. T o be a
man whom the ruler greatly loves is to be to him as hard to white. Wishing
with the resources of a commoner to separate from the ruler his 'as-hard-
to-white' loved one is advising the right buttock to get rid of the left; the
consequence will be that you will surely lose your life and your advice will
be ineffective."
If the Mohists knew nothing of a paradox of hard and white, what of
the sophists? H u Tao-ching has collected 21 references to 'hard and white'
disputations of Kung-sun Lung and others in texts from Chuang-tzU
(c. 300 B.C.) to the first century A.D. Reading them consecutively in his
10
book one is struck by their curious air of generality. In no less than 14 of
them the phrase chien-pai appears beside t'ung-yi f^H 'the same and the
different' or wu-hou MM 'the dimensionless', themes which provided
11
matter for many sophisms, but never beside a particular sophism. One
has the impression that a 'hard and white' disputation is one kind of
disputation. Fung Yu-lan already saw this point, and concluded that 'the
separation of hard from white' and 'the unity of the same and the different'
were terms used to characterize the doctrines of Kung-sun Lung and
12
Hui Shih respectively. Since Fung Yu-lan accepts the genuineness of the
'Essay on hard and white' he does not doubt that the more general use of
the former term is derived from the name of the sophism. But if we reject
the essay and give chien-pai the technical sense which we have noticed in
the Canons there is no need for us to postulate a lost sophism; a disputation
about the chien-pai ('as hard to white') might be simply an argument about
supposed inseparables, space and time and length and breadth in the
10
Hu Tao-ching (1934), 16, 55, 75-80. All are mentioned in this section except
Shih chi (ch. 76) 2370/3. I have noticed one more: Ch'ien-fu lun yg^fre (SPTK)
3, 13B/1.
11
Chuang-tzu ch. 8, 10, 33 (Kuo 314/4, 359/-4, 1079/2). Hsun-tzu ch. 2, 8, 19
(Liang 20/3, 83/-3, 260/9). Lu-shth ch'un-ch'iu ch. 17/3 (Hsu 755/-3). Han Fei tzu
ch. 41 (Ch'en 899/5). Huai-nan-tzu ch. 11 (Liu 11, 18B/-3). Shih chi (ch. 23, 74)
1172/1, 2349/1. Liu Hsiang, memorial on Hsun-tzu {Hsun-tzu SPTK 20, 35A/8),
also Pie-lu SHU ap. Han shu commentary (ch. 30), 1737 n. 5.
1 2
Fung (1952) v. l,214f.
1/4 173
Canons, shape and colour in Kung-sun Lung's 'Essay on the white horse',
in contrast to disputations about 'the dimensionless' or 'the same and the
different', which would be concerned with paradoxes involved in the
concept of a geometrical point or in the treatment of similarity and
difference as absolutes. The following are some characteristic references to
chien-pai disputation:
Chuang-tzu ch. 17 (Kuo 597/2) £ I R I H » B I S S » gfcf & » R J ^ R I O
(Kung-sun Lung): "I join the same and the different, separate the chien-pai,
treat the not so as so, admit the inadmissible."
Ut sup. ch. 33 (Kuo 1079/2) K S S ^ ^ ^ ^ f f l * » &№№^ft±%t№B o
(The later Mohists) "reviled each other in disputations about the chien-pai
and the same and the different, answered each other with odd and even
propositions which did not match."
Hsiin-tzu ch. 2 (Liang 20/3) * S & K ^ * « f t » ± » ^ P * - & » f&ffiM?
» ih^lffe o
"It is not that inquiries into the chien-pai, the same and the different, the
dimensioned and the dimensionless, are not perspicacious; that the
gentleman nevertheless does not engage in disputation is because they are
outside the limits he sets himself."
When chien-pai is mentioned alone it seems to be a metaphor for
hair-splitting debate in general. Thus Chuang-tzu says that thinkers who
break up the unity of the world by distinguishing 'right' alternatives from
'wrong' (shih fei ;Sk^) 'end up in the obscurities of chien-pai* (££§££;£
13
ft$£) and mocks the sophist H u i Shih with the phrase 'You crow about
14
chien-pai' (^fJÎHSHir). The following example is interesting since it
shows that the image of a hard white stone was still alive in the early Han,
yet it is clearly metaphorical:
Yen t'ieh lun ^.Wim ( S P T K ) 4/12B/9 » titBME 'gift
9
lt°
"Tung-fang Shuo judged himself a clever debater, and had no rival in his
age for melting the hard and dissolving the stone."
What we do not find in pre-Han and Han literature is any suggestion
that there was a specific paradox of hard and white. We miss such a paradox
in the lists of sophisms in Chuang-tzu and Hsiin-tzu and among the three
sophisms of Kung-sun Lung recorded by Hsu Shen ft'K (fl. A.D. 100) and
18
Chuang-tzu ch. 2 (Kuo 75/1).
14
Chuang-tzu ch. 5 (Kuo 222/-1).
15
the seven listed in Lieh-tzu (c. A.D. 300). On the other hand Kung-sun
Lung's famous thesis that ' A white horse is not a horse' is listed by all these
16
authorities except Chuang-tzu, which mentions it elsewhere; and in
nearly all the early references it is unmistakably a particular sophism,
17 18
generally presented in its full form or as ' A horse is not a horse'. The
contrast is especially conspicuous in certain references by Han authors who
seem to know the writings of Kung-sun Lung, no doubt from the original
19
Kung-sun Lung tzil in 14 chapters recorded in the Han bibliography:
Huan T'an © J ? (c. 43 B.C.-A.D. 28), Hsin lun Sfflra, apud T'ai-p'ing yu-lan
X^MW ( S P T K ) 464/5A/11-13 ^ f f i ^ B P S I I H r t b ° g ^ S ^ f i * > 1S#J
"Kung-sun Lung was a sophist of the time of the six kingdoms. He com
posed essays about chien-pai, took comparisons from things borrowed as
illustrations, and said that a white horse is to be deemed a non-horse."
(Huan T'an continues with a summary of the 'white horse' argument.)
Wang Ch'ung S £ (A.D. 27-c. 100), Lun heng m№ ( S P T K ) 29/2A/6
"Kung-sun Lung wrote essays about chien-pai, making hair-splitting

analyses of propositions, busying himself with perverted statements and
unreasonable comparisons, of no use to government."
Huai-nan-tzu ch. 11 (Liu 11, 18B/-3) &Wm№9iX$ > SUNS- * SSMS °
"Kung-sun Lung engaged in unreasonable disputations and shocking
assertions, distinguished the same and different, separated the chien-pai/'
Hsu Shen's comment: < £ № f S A ' ftfttfrm^ZS ° E t S J S § ; W # £ « —
"Kung-sun Lung was a man of Chao who enjoyed hair-splitting and par
adoxical talk. Considering that 'white' and 'horse' cannot be joined as one
thing he separated them and regarded them as two."
Ut sup. ch. 14 (Liu 14, 7B/-2) ^ ^ f f i ^ ^ S ? f f B S ^ °
"Kung-sun Lung bought himself undeserved fame by brilliance with
words."
15
Chuang-tzu ch. 33 (Kuo 1102-1106). Hsun-tzu ch. 3, 22 (Liang 24/6-25/3,
315/11-317/8). Huai-nan-tzu commentary ch. 14 (Liu 14, 7B/-2 n.). Lieh-tzu ch. 4
(Yang 88/4-89/1).
18
Chuang-tzu ch. 2 (Kuo 66/6).
1 7
Cf. the eight pre-Han and Han references collected in Hu Tao-ching (1934),
15-18, three of which we quote here.
18
Chuang-tzu ch. 2 (Kuo 66/6). Hsun-tzu ch. 22 (Liang 316/-3).
19
Han shu (ch. 30), 1736/-5.
1/4 175
Hsu Shen's comment: feMMM. f £H2{Mf J ' T i * ^ * J ' [" # ^ S U f§

m °
"Kung-sun Lung's themes for his essays were ' A white horse is not a
horse', 'Ice is not cold', 'Coals are not hot'."
Plainly these authors do not take 'hard and white' and 'white horse'
as two paradoxes; Hsu Shen even quotes the 'separation' (It $1) of 'white'
and 'horse' as an example of 'separation' (It) of the chien-pai. Nor do they
say that Kung-sun Lung wrote an 'Essay on hard and white' (SSIra);
they say that he wrote essays on chien-pai (llfl^Hira), implying that he
specialized in the separation of supposed inseparables (shape and colour
in the horse, ice and cold, coals and heat) rather than in analyses of the
'dimensionless' and the 'same or different', the topics which especially
concerned Hui Shih.
20
In my earlier paper I assumed that Kung-sun Lung defended a
sophism about hard and white the nature of which was soon forgotten and
that Han explanations in terms of the 'white horse' paradox are merely
guesses. But in the light of the present inquiry we must look with new
respect at Han and even at post-Han writers. About A.D. 300, after the
21
disappearance of the original Kung-sun Lung tzu, we begin to meet with
writers who ask what chien-pai is and disagree about the answers. L u Sheng
(fl. A.D. 291), editor of the Canons, understood that chien-pai was a theme
of disputation like 'the dimensionless' and 'the same and the different', and
that the separation of shape and colour in the white horse was an example
of it:
L u Sheng, preface to the Mohist Canons, apud Chin shu (Pai-na edition)
biographies, 64/6A/9 £>&^W> ' ^f<?£>il£n9Jft » ftfSSii °
"There must be a shape corresponding to a name, and the best way to

examine the shape is to distinguish the colour from it; therefore there is
disputation about chien-pai. There must be distinctions between names,
and the most important distinction is between something and nothing;
therefore there is disputation about 'the dimensionless'. Right is sometimes
wrong, the admissible sometimes inadmissible; this is named 'admissibility
of both alternatives'. There are differences even between the similar,
similarities even between the different; it is this which is called 'disputation
22
about the same and the different'."
2 0
Cf. G(2) 177-179.
2 1
Cf. G(2) 180.
2 2
For the emendations, cf. Hu Tao-ching (1934), 82.
L u Sheng's contemporary Ssu-ma Piao (died A.D. 306) hesitated

between the right answer, that chien-pai 'refers to disputation about a hard
stone or a white horse' ( f f i l l S &$!;£$§№), and a wrong answer, that it
23
was a dispute over the tempering of swords. The latter explanation, still
24
current during the fourth and fifth centuries, is interesting only for its
implication that chien-pai was already being mistaken for a particular
problem. This was also the assumption of the unknown author who, at
some time between 300 and 600, wrote the last three essays of the extant
Kung-sun Lung tzu. I have elsewhere examined his method of working up
misunderstood scraps of the Canons, in which he did not recognize the
25
head characters. Examples of such borrowings in the 'Essay on hard and
white' are the following, in which I underline characters corresponding to
head characters in the Mohist Explanations :
Kung-sun Lung tzu, Ch'en, 167 (cf. B 15) °
Ch'en, 177 (cf. B 4) J I J S ^ J l [ J3*J1] It °
Ch'en, 188 (cf. B 37) KE—1& » Efi—iMMEJfcS ° »
g°
From these we can see which Mohist passages caught the writer's eye
and how he misunderstood them:
B 15: "Without the hard one finds the white."
B 4 (a passage ending with the words 'hard and white'): "The seen and
the unseen are separate."
B 37: "As for the stone it is one; hard and white are two but in the stone.
Therefore it is admissible that there is something known in it and some
thing unknown."
It was evidently these misunderstood passages which suggested to him
that Kung-sun Lung's mysterious 'separation of hard and white' was the
separation of the white stone which one sees from the hard stone which
one touches. By writing an essay on this theme in the name of Kung-sun
Lung he deluded us all for 1,500 years (in my own case, for 10 more years
after recognizing his essay as a forgery) into looking at every pre-Han and
Han reference to the separation of hard and white through his own
astigmatic eyes.
2 3
Ssu-ma Piao, ap. Chuang-tzü shih-wen ch. 2 (Kuo 77 n. 14).
2 4
Cf. G(2) 178f.
2 5
Cf. G(2) 156-164.
1/4 177
1/4/4 C H I H Jh A N D H S I N G ft
Chih i t 'stay': intransitive, of names staying confined to objects (A 43, 50,

51, 78, 93. B 68, 82). Transitive, 'fix' (A 75, 96, 97. B 1, 2, 32)
Hsing ft 'proceed' (B 1, 72. N O 10, 11, 12)
Both chih and hsing are used occasionally in their primary senses, of
physical motion (chih 'stop' A 12(?), B 26, N O 16: hsing 'go' B 27, 63, 64,
N O 10). There is also a reference to the Five Elements ( £ f f , B 43).
In ethical contexts we find hsing 'conduct' (A 10, definition: 14, 18,
80, 89. B 6. E C 3, 7, 9), the behaviour by which a man is morally judged;
it includes actions which he does not personally perform (wet j*§) but has
commanded (A 18).
In logical contexts both chih and hsing are used technically. Intransitive
chih (defined in A 50) is used of a name staying in objects throughout their
duration and being confined to them, as in A 78 ^ ^ i b i h S ^ S K i f e "This
name stays confined to this object" (cf. B 68, 82). The verb can stand
alone without a directive: A 43 *ffclt "It stays fixed of all": A 93 Mlh-tfe
"The circle stays fixed": B 82 J / F * l t " T h i s ' does not stay fixed".
Transitive examples are A 96 Jh T *£A J "fix 'black man' " * i h f SLA J
"fix 'love of man'": A 97 ± 0 "fix the criterion": B 1 l h « "fix the kind":
A 75 Jh§?& "fix which you prefer".
Its opposite is hsing 'proceed', which both in A 97, B 1 and in N O 10
is used in conjunction with the words lei 'kind' and tao M. 'road' (the only
26
examples of tao in the dialectical chapters). By fixing the criterion for
judging something to be so of one object we establish the road or course
(tao) by which it 'proceeds' to the other objects of the same kind. In
discussions of the problems of transferring a name from one thing to
another we find both "Treating that as 'that' it stays confined in that"
(B 68 r $ J f f i i t S ^ ) and "What I call it does not 'proceed'" (B 72
In Names and objects, as in the Canons, to 'proceed' is to pass from

what is so of an instance to what is so of the kind, but for the first time this
is seen as the translation of one proposition into another, of "He loves
Jack" ( g K ) into "He loves man" ( 3 I A ) . "The proposition is something
which 'proceeds according to the k i n d ' " (NO 10 &Mft%&). To
test whether something is "so (in general) if the one instanced is this
(kind of thing)" (NO 13 7Wik[Tn$t), we explore the parallelism of apparently
similar propositions, "lay propositions side by side and they both 'proceed' "
26
Cf. § 1/5/14.
(NO 11 Ji^ffniicfT). If the parallelism is false, "as they proceed they

become different" (NO 12 ffifnil).
For examples in other texts, cf. Chuang-tzii ch. 18 (Kuo 622/1
"Names stayed in objects", Hsun-tzu ch. 12 (Liang 158/4) Wtftffc&ft
"Kinds cannot proceed of themselves".
Hsing seems once like chih to be used causatively: B 1 ft A "make
27
the man proceed".
Chih, the temporary validity of names fitting transitory objects,
belongs to the first Mohist discipline, the logic of consistent description,
while/)/ *j& 'necessity', the unending validity of the implications of defined
names, belongs to the fourth, disputation. The two words are contrasted
in A 50, 51, B 32.
1/4/5 C H I H Ifc
Chih H 'hold' (A 93, 94. E C 1, 2, 3, 8. N O 9, 16)

Except for one instance where the graph seems to represent chih S
'hibernate' (B 39) the word can always be taken as chih 'hold'. The main
difficulty is the frequency of unfamiliar nominal uses. Chih appears in two
kinds of context:
(1) Verb, 'uphold (one of two contrasted claims)': noun, 'the
position one upholds'. The verbal use is common throughout Mo-tzu:
"hold that spirits do not exist" (ch. 31, Sun 144/3), t ^ i f r "hold
that Destiny exists" (ch. 34, Sun 169/-3). Chih seems always to imply
taking one side in an issue, cf. Mo-tzu ch. 16 (Sun 76/2) #<Sfl . . .
'uphold separation . . . uphold universality', ch. 48 (Sun 385/2) fk^lMfifc
'maintain that there is/is not/the auspicious and the inauspicious'.
But in the dialectical chapters the word is used not only verbally (EC 1,
N O 9, 16) but nominally:
A94...ffJ**«C...№J*»So
(Canon) . . . then seek his reason. (Explanation) . . . then seek the
standard for the one he upholds. (Cf. also A 93, E C 1.)
(2) Verb, 'control (conditions)': noun, 'conditions'.
EC 8 * № ^ # A ± § f S t t b -
"What he chooses is under the control of others."
EC 2 nmmm®&&m °
"Cases when desire and dislike are born in the conditions one encounters
(lit. 'what one encounters being in control')."
2 7
Cf. p. 159 above.
1/4 179
EC 3 №mmm&m% °
"No external condition (what is under external control) can make me more
,,
beneficial.
1/4/6 CH'lNGift
CWing It 'the essentials of X , what is essential to being X ' (EC 2, B 81,

N O 9)
Ch'ing fi§ is in all cases written with Radical 149, a common usage
throughout Mo-tzu (cf. Sun 50/6). The word is used throughout pre-Han
literature of the facts of a situation, the genuine in contrast with the false
(wet <S) or with mere reputation (ming wen H , sheng S£). In pre-Han
28
philosophy, as I have argued in detail elsewhere, it has a precise meaning
with much in common with the Aristotelian essence. The ch'ing of X is
all that is conveyed in its definition, everything in it without which it would
not be a genuine X , conceived as something behind its hsing 'shape* and
mao t& 'looks' (the latter is discussed in § 1/4/20). Hsiin-tzu and other
Confucian ritualists use ch'ing especially of the passions which are essential
to being human yet must be refined and ordered by the mao, the outward
demeanour imposed by the rites; it is by descent from this usage that the
word later came to refer specifically to the passions.
Chuang-tzu ch. 13 (Kuo 479/1) ^ f f i B » T If MfalBfc^ J ° 0 »T4 1
» » J o
"Lao Tan said: 'May I ask what you mean by benevolence and righteous
ness?' Confucius answered: ' A loyal concern for the happiness of others,
universal love without selfishness, these are the essentials of benevolence
and righteousness'."
Ch. 11 (Kuo 390/5) ' ^MftW » $&#C(=0)i£ °
"Do not ask about their names, do not pry into what they essentially are;
it is inherent in things to be born of themselves."
Ch. 2 (Kuo 55/-1) Wa&in » f f i M f t ^ ° Wffiffni$J£ o
(Of the Way) "That it can be walked is true enough, but we do not see its
shape; it has identity (ch'ing, something it essentially is) but no shape."
Han Fei tzu ch. 20 (Ch'en 366/2) HMtM ' ^ $ P r ^ » » MM
"As for the essentials of the Way, it is indefinite and unshaped, it yields
with the times, it and the structures of things correspond to each other."
2 8
G(9) 259-265.
There is a very striking example of ch'ing at the end of the Te ch'ung fu

chapter of Chuang-tzu, where it is traditionally but surely mistakenly
taken to mean the passions. It should be read in the light of Chuang-rzu's
doctrine that man should spontaneously follow Heaven like other beings,
but in stead makes the mistake of distinguishing between right and wrong
29 30
alternatives (shih fei&$p) and preferring (hao 0 'liking') the former:
Chuang-tzu ch. 5 (Kuo 217-221) U S A S ^ X » XSSJBA ° WA£ffi5» 5EA
B » A#C3Ett№ ° K ^ B » ^ o » Affi^ffi * W E t S I ^ A ° J » B »
ttn^n»* s i # ^ i i ^ A ° K ^ B · K i i . ^ A » mn^vt o ffi^
B · A*5»ffflffi* ° ^ w a ^ f f l * » j t A ^ j a f f f f i f t f l S K * ·
"Having received your food from Heaven, what do you need from Man ?
Have the shape of a man, be without what is essentially man. Have the
shape of a man, and so flock with men; be without what is essentially man,
and so right and wrong will not be found in your person."
The sophist Hui Shih now enters and raises the obvious objection:
" 'May a man really be without what is essentially man?"
'Yes'.
'In that case, how can one call him a man?'
'The Way gives him the look, Heaven gives him the shape, how can one
not call him a man ?'
'Granted that we do call him a man, how can he be without what is
essentially man ?'
'Judging between right and wrong is what I mean by 'what is essentially
man'. What I mean by 'being without what essentially man' is that a man
does not inwardly wound his person by likes and dislikes, constantly follows
the spontaneous and does not add to what grows in him'."
Since Chuang-tzu disapproves of the passions it is easily assumed that
it is passion that he means by ch'ing. But there is more than one objection
to this interpretation:
(1) Nowhere else in Chuang-tzu is ch'ing used in this sense. It is
assumed throughout that the passions are undesirable disturbances, but
ch'ing (unless qualified as the ch'ing of something bad, as in this case that
incurable rationalist and moralist, Man) is self-evidently good, the state of
perfect genuineness which the sage recovers. Compare such phrases as
MXfiqfn "flee Heaven and turn one's back on what one essentially is":
29
Chuang-tzu ch. 2 (Kuo 55-79).
80
Ut sup. (Kuo 75/1).
1/4 181
Stjfcfe, UMn, t * # "flee what they have from Heaven, part

from their nature, extinguish what they essentially are, destroy their spirit":
31
IHftfn, "alter what they essentially are, change their nature".
(2) Distinguishing right from wrong is hardly passion; but it is, in
the Ch'i ucu ^'$4 chapter, what primarily differentiates man from things
which spontaneously follow Heaven.
(3) The man without ch'ing has only the 'shape' and 'look' of a man
(hsing, mao, terms commonly contrasted with Ming 'the genuine').
(4) It is no business of the sophist Hui Shih to express opinions such
as that a man without passions would not be a man, which any of us could
contribute to the discussion. His role is to point out the stunning self-
contradiction in Chuang-tzu's thought, and so involuntarily demonstrate
the futility of all logic; and the way in which he worries at the point, three
times pointing out the contradiction which always reappears in Chuang-
tzu's answers, shows that this is what he is doing.
From H u i Shih's questions we may deduce a definition of ch'ing. The
ch'ing of X is what X cannot lack if it genuinely is X . Mencius uses the
word in just this sense, although for him what is essential to M a n is
self-evidently good while for Chuang-tzu it is self-evidently bad. When
Mencius wishes to show that the incipient impulses towards the four
cardinal virtues belong to man's ch'ing he goes through them one by one
3 2
and says of each "all men have it" ( A t i r W ^ ) ; elsewhere after a similar
list he repeats it in the form "Whoever does not have X is not a man"
3 3
(MX*A&).
We find the same method of establishing what is ch'ing in the Ch'ing yii
ffi$C section of the Lii-shih ch'un-ch'iu. According to a doctrine preached
34
by Sung Hsing in the 4th century the ch'ing yii are few, meaning not
that 'the passions and desires' are few, but, as the Lii-shih ch'un-ch'iu
account makes plain, that if we sort out from our many desires those which
belong to our ch'ing, the essential desires without which we should not be
men, we find that they are few and easily satisfied. Ch'ing, in contexts
where it is translatable by 'what X genuinely is', approximates very closely
to the Aristotelian 'essence', as may be seen when it is used in definitions.
31
Chuang-tzu ch. 3, 25, 29 (Kuo 128/3, 899/1, 1005/5).
32
Mencius 6A/6.
33
Ut sup. 2A/6.
34
Chuang-tzu ch. 33 (Kuo 1082/4 I f (= flf) £ j ^ J § 3 £ "Provision for the
essential desires they took as their main point". 1084/1 iffSKBIEETYÎRJEÎ
"The essential desires are surely sufficiently provided for with rive sheng of rice".
F r
1084/4 £tfrc$C3S§i1i§i*I " <> private life their principle was 'the essential desires
are few and slight' Hsun-tzu ch. 18, 20 (Liang 251, 316/2).
(But there remains the difference that it is not connected with any copulative
verb like Indo-European 'to be\ The ch'ing of X is everything in it without
which it would not/** the name 'X'.)
Lii-shih ch'un-ch'iu ch. 2 (Hsu 2/8A/8-8B/2) ^ ^ A f f i ] f t * A W ^ ^ # t n ' 0
» 1 ·
"When Heaven generated man it caused him to have hankerings and
desires. Among desires there are the essential, for the essential there is
measure. The sage cultivates measure in order to check the desires, and
therefore acts only on the essential.
Now the ear's desire for the five sounds, the eye's desire for the five colours,
the mouth's desire for the five flavours, are essential. These three noble
and mean, wise and foolish, worthy and unworthy, are as one in desiring,
and even the sages Shen-nung and the Yellow Emperor are the same in
this as the tyrants Chieh and Chou."
1/4/7 CH'(il,S,I,ft
CA'wIEmoun, 'delimited area' (A 48, 63. B 22)

Ch'ii m (A 73), t*6 (B 11), m (B 12), m (B 63), B (NO 6): verb, 'demar
cate, divide off from others, group apart'
The verbal ch'ii is confined to logical contexts and is distinguished by
radicals which have been multiplied by later attempts at graphic standard
isation, iU Bfe, fil. Since in B 12 the second of these is corrupted as head
y
35
character to ft, the original graph was probably the 4B of B 63. The form
corrupted to £ in B 11 may be identified as $8, since these two graphs are
36
confused once elsewhere in Mo-tzii Names and objects writes the verbal
ch'ii with a different graph, B (NO 6). The interchange of chYw/*K'ltJG B
37
and ch'ul*K'l\J E is well attested; in other texts either graph is used in
38
the phrases ch'ii kai 31, ch'ii yii ? and chin % ch'ii.
Judging by the composition of the graph the basic meaning of ch'ii
39
is to store away a collection of articles in a receptacle. The Mohist uses
36
Cf. § 1/2/1/2/7.
36
Mo-tzu ch. 52 (Sun 324/3) ft j £ . "stick out spurs" (ch. 53, Sun 334/1;
ch. 56, Sun 337/3 tÜE(=}g)).
37
M . 33 def. 9.
3 8
Chu Ch'i-feng 1867, 1261. M . 167/120, 150.
3 9
Tuan Yü-ts'ai J&5&|$ on Shuo wen ch. 12B, 641A/9.
1/4 183
verbal ch'u of grouping certain things in contrast with others, putting them
in separate compartments, as in a well-known passage in the Analects:
Analects 19/12 ' °
"It may be compared to the herbs and trees, they are separately grouped
,,
and differentiated.
A 73 №9 9
1№°
" A l l oxen, and non-oxen as a separate group, are the two sides."
B 12 9
o
"Things grouped separately are one unit."

It may be noticed that ch'u, unlike lei 'kind', does not imply that
things are grouped on grounds of similarity, only that they are different
from something else. The eight types of sameness distinguished in N O 6
include both 'sameness in being of the same kind' ( f ^ S i l l R l ) and 'sameness
in being grouped apart* (JXIRI).
40
Since ch'iu or ch'il (Hi, US) as a word for a group of 16 'well* settle
ments is a specialisation of the same usage we may identify the Mohist
ch'ii with the ch'iu li J i l l mentioned by Chuang-tzu, in a section which
attacks disputation, discusses infinity, and uses the stock examples of horse
and dog and the technical termsyi M 'idea' and chia © 'loan-name* as they
are used by the Mohists :
9
Chuang-tzu ch. 25 (Kuo 909/1-3) T M f l £ S > 6 J . . . T
"What do you mean by 'ch'iu and li' sayings?" . . . "Ch'iu (a group of 16

'well' settlements) and li (a group of 25 families) put together 10 surnames
and 100 personal names and establish a convention. We put together things
that are different and deem them the same, disassociate things which are
the same and deem them different. If now you point out from each other
the hundred members of a horse's body and don't find the horse, yet the
horse is there tied up in front of you, it's that at the level where we put all
the members we call them a 'horse'." (For 'levels', cf. p. 363 below.)
Verbal ch'u is once used of demarcating space:
B63 &~^»mm
"That which being demarcated cannot be referred to without referring to
that from which it is demarcated is Space." (Cf. the Han and post-Han
41
cliche E'-? 'bounded space', applied to the Empire.)
4 0
M . 33 def. 9.
4 1
M . 2691/4, 5.
This connects with the nominal ch'ii (written without a radical), which
is found only in the geometrical and scientific sections. It means a delimited
area, as in Huai-nan-tzu ch. 6 (Liu 6, 7A/7) P^^BMkiM "does not
step outside the limits of 100 acres or an acre".
B 2 2 « H "Reduced area."
The phrase B 7 t (A 48, 63), 'delimited void', will be discussed in the
2
introduction to the geometrical Canons.*
1/4/8 CH'UAN«
Ch'iian m 'to weigh' (A 84. E C 7, 8) 'leverage' (B 25, 26)

Except in scientific contexts ch'iian is verbal, 'weigh against each other'
(cf. Mencius 1A/7 H , fl&f&ftUSS "Only after weighing does one know
light from heavy"). As a term used in ethics it is defined in E C 8 " T o weigh
light against heavy among the things treated as the 'units' is what is meant
by 'weighing' ( f e § f « i : * i l D « e f i > I H r « J ). Desire after weighing
benefit and harm (ch'iian) is contrasted with desire which is direct or
immediate (cheng IE A 84, E C 8). Mencius uses the term when he says that
"man and woman not touching when they give and receive is propriety,
using your hand to pull up your drowning sister-in-law is ch'iian" (judging
another consideration weightier, Mencius 4A/18).
The noun ch'iian 'leverage' will be discussed with the terms used in
43
mechanics.
1/4/9 F A N {&
Fan iR 'converse' (A 73, 74, 83. B 3, 30, 72)

The graph J5, otherwise unknown, survives uncorrupted only in the
summings-up of two Canons (B 30 t&ffiffiiftjK "Explained by taking the
converse of its price", B 72 ffcfiE'fK "Explained by the converse"). It may
be presumed to be a technical term derived from fan K 'turn round' and
distinguished by the 'man' radical so common in graphs invented by the
44
Mohists. If we refer to the Explanations we find that 'taking the converse
of its price' sums up an argument that if money is the price of the grain
the grain is also the price of the money; and that the 'converse' of B 72
is it being inadmissible that the name of X may be extended to Y unless
4 2
Cf. § 2/4/1/4.
4 3
Cf. § 2/4/2/3.
4 4
Cf. § 1/2/1/2/3-11.
1/4 185
we grant the converse, that the name of Y may likewise be extended

to X .
Since the graph was obsolete after the end of the Mohist school it
would be especially liable to corruption. A probable instance is the puzzling
example in B 3 of the graph f-R (the left hand of which is abbreviated in
grass script to
"If it is said of this, it's that inherently this is beautiful; if it is said of

another, it is not the case that this is beautiful; and if it is not said of
something the converse applies."
(Cf. Mo-tzu ch. 52 Sun 315/—1, ch. 71 Sun 381/-2, where one is tempted
to emend $8 to the/>an ItS of the immediate context, although Sun and
Ts'en 26, 142 read in the first case and accept the text in the second.)
On closer inspection we find that fan is actually one of the terms
defined in the Canons. It is surely the word in A 73, 74 corrupted to ft in
the Canons and to № in the Explanations, and given a definition which
agrees exactly with the use of fan in B 72, of a claim being admissible of X
only if the converse is admissible of Y :
A 73 (fe)MK » *«Jffi*«J-fe °
"Being the 'converse' of each other is if inadmissible then on both sides
inadmissible."
It is again corrupted in A 83:
"Those which are the converse of each other, apply on both sides, not on
one without the other."
In A 74 disputation is defined as 'contending over converse claims',
(4? ({&)*{&); the example given is contesting which are oxen and which
non-oxen of two classes into which everything is divided. 'They are oxen'
applying to one reverses to become 'They are non-oxen' applying to the
other. Similarly the Lu-shih ch'un-ch'iu uses fan (without the radical) of the
judgment that one disputant is right having as converse the judgment that
another is wrong:
Ch. 10/3 (Hsu 411/4) » RUffiS °
"Therefore they turn one way to judge each other wrong, turn the other
way to judge each other right."
Ch. 15/8 (Hsu 675/7) A J i l g ^ » R^ffl.iJI ° XTZm^B °
"Men because they judge themselves right turn the judgment the other
way round to criticise each other. Among the scholars of the world there
is too much disputation."
1/4/10 F A N G ft
Fang ft 'be relative to' (A 88)
The word fangl*J*[WANG 33 is exceptionally rich in divergent but

related meanings, syntactic mobility, and semantic and graphic cognates.
In the dialectical chapters we find it as an intransitive verb ('be square',
nominalised 'the square'), noun ('method', N O 12) and temporal particle
('at the time in question' A 33). A l l these usages are familiar; but there is
also a puzzling example of fang I*PIWANG $C (A 88), which compels us
to look more closely at a word-family of some importance in Chinese
philosophical vocabulary.
The basic sense of fang ~fi according to the Shuo wen definition is to
45
lash two boats side by side, as in the phrase 'lash boats'. Phonetically
it may be connected with such words as p'ang/*B'WANG 3? 'side',
/>//#/*B'IENG $fc 'side by side' and/>7*/*Pl3R I t 'put side by side, compare'
#
B ' I ^ R 'be side by side'. Let us try to map some of the branching meanings
relevant to philosophy.
(A) Intransitive verb, 'run side by side, run parallel in the same
direction':
(A 1) 'Be square' (run along parallel lines).
(A 2) Causatively, J3№ 'lash boats' (aligning them side by side), cf.
also pi fang itl^j 'compare' (as in the example from Hsun-tzu quoted
§ 1/4/37), in which pi may be taken as static (put things side by side) and
fang as dynamic (develop their characteristics side by side).
(A 3) Adverbially, 'proceeding side by side', as in Mo-tzil ch. 52
(Sun 309/-6) ^/SB 'rise up side by side'.
(B) Transitive verb, fang X 'run side by side with X , take its direction
from X ' :
(B 1) Fang № 'depend on', common in Mo-tzu in the phrase
46
'depend on'.
(B 2) Nominalised, 'guideline, method' (that from which one takes
one's direction). Cf. N O 12 " 3 " S a y i n g has many guidelines".
(B 3) Nominalised, 'direction'. Cf. B 33 ~№15 "The southward
direction".
(B 4) Adverbially, 'coinciding with/proceeding with an event/a time'.
Cf. Chuang-tzu ch. 2 (Kuo 104/1) ^ f t l ^ i f e "While he is
dreaming he does not know that he is dreaming". The temporal fang may
also be objectless: A 33 Jjf&tffB* "The just now so too is about to be".
45
Shuo wen ch. 8B (Tuan 408B/-1). M . 13620/229.
4 6
Sun 12/-5, 73/5, 75/-4.
1/4 187
Verbal fang ~f3,ft'take its direction from/depend on* (B 1 above) is

used in philosophical contexts of one of a pair being relative to the other:
Lu-shih ch'un-ch'iu ch. 10/3 (Hsu, 411/4) «CRl^ffl№ » R d f f i A ° K 8 f #
"Therefore they turn one way to judge each other wrong, the other way
to judge each other right. It's that what they judge wrong is relative to
what they judge right, what they judge right is relative to what they judge
wrong."
Cf. Chuang-tzu ch. 12 (Kuo 427/1) W A » g f f i f t » » °
"Take the case of a man who deals with the Way as with correlatives, who
treats the inadmissible as the admissible and the not so as so."
5
Canon A 88 IRIH ° : * « »<^ » *afc ' ' -ft^ > °
"Being the same or different. By interplay they become relative: 'having
or lacking', 'more or less', 'departing or approaching', 'hard or soft', 'dead
or alive', 'elder or younger'."
1/4/11 F u «
Fu « 'compound' (B 11. N O 9)
In both of its two occurrences in the dialectical chapters this word
4 7
appears to be fu \%. ( = ^ ) 'double', referring to the compounding of
names. The graph ffi (marking the basic meaning as 'doubled, lined
garment') is not found in Mo-tzii, but there are examples of fu tao %M.M
48 49
'double path' (a two-level path) written with the present graph for /w.
The two passages in the dialectical chapters are illuminated by comparison
with Kung-sun Lung's White Horse:
£ T I S Jliil" £ J j o
"Putting together 'horse' and 'white' one compounds the names as 'white
horse'."
B 11 » j £ « < sfc > i5 °
"Together or one: in the former case one compounds, in the other not."
N0 9 ^Bf»(=W)*iE°
50
"Names depending on the facts are compound or single."
Hsiin-tzu's Right use of names also distinguishes between single and
double names, but uses a different terminology (tan •¥> 'single', chien Ife
4 7
M . 10183 def. 24.
4 8
M . 34417/42.
49
Mo-tzu ch. 70, 71 (Sun 374/-3, 386/4).
5 0
For cheng, not otherwise attested of single words, cf. § 1/4/2.
'double'). His commentator Yang Liang SHsK (preface dated A.D. 818)
explained his term for double name by fu ming {Hsiin-tzu S P T K ch. 16,
6A/1 r * J » This use of fu has indeed never died out, surviving
in the still current fu-hsing ^ t t 'double surname*. The author of Yin Wen
tzii (c. A.D. 200), reflecting the revival of interest in the sophists at the end
of the Han, analyses two-word phrases which combine one's own attitude
with a reference to the object (UK 'keep close to the worthy', #F*4V^»/A
'like oxen/horses/men'), and several times uses fu:
Yin Wen tzu ( S P T K ) 3A/1 (cf. 6, 7) » °
"Self being in conjunction with other, one compounds with a further
name."
1/4/12 H A I ft
Hat iS 'interfere with, be an objection to' (A 20. B 5, 45, 51, 65, 73, 75)
In ethical contexts hai is 'harm', opposite of li M 'benefit' (defined

together in A 26, 27). But a more basic sense of hai is 'obstruct, interfere
with', as in B 27 s E i S i l "Something interferes with it" (similarly li is
'smooth, unimpeded', as in N O 9 'fluency of the tongue', or li 'sharp', of
what cuts through smoothly). In logical contexts hai is the ordinary word
for one claim being an obstacle to accepting another. Cf. Hsiin-tzu ch. 23
(Liang 335/4) S I ^ S S , №1$"ÎfiUSB "Even if one is unable to become
a Yu, that is in no way inconsistent with the possibility of becoming a Y i i " .
(In an argument distinguishing the ability from the possibility.)
a
1/4/13 H S I E N 5fe, H S I E N C H I H 5fe$a AND W E I K ' O C H I H * 7 f t l
Hsien 5te 'a priori' (EC 2, B 38). Hsien chih 3fcfc1 "know 'a priori' " (A 93,
B 57). Wei k'o chih *oJ$lJ "not knowable 'a priori' " (A 75,
B 58, 73)
The phrase hsien chih 'know beforehand' is common in pre-Han

literature, of knowing ahead of other people how a situation will develop:
Kuan-tzu ch. 51 (BSS 2/110/-4) gA3fc»lRi7& ° 4 E f « » i » 1 5 *
"The sage knows things in advance before they take shape. Since just now
I did not know it until after it took shape, it is not the case that I am a sage."
(Cf. also Lu-skih ch'un-ch'iu ch. 20/8 (Hsu 20, 29A/-4), as well as ch. 16/1
Hsien shih 3fcfll 'Being aware in advance'.)
But the Mohist summa uses this and similar phrases only in logical
contexts, of knowing something 'a priori' without having to observe it.
1/4 189
Indeed the word hsien is always translatable as 'a priori' except for the cases
contrasting with hou %k 'afterwards* in B 63, 64. That the opposite of hsien
chih is wei k'o chih "not knowable 'a priori* " (to be distinguished from
pu ^k'o chih 'unknowable', B 10) can be seen from the contrasting Canons
of B 57, 58. The 'a priori* is three times associated with the stock example
of a wall with unknown things on the other side (EC 2, A 75, 93), about
which we know certain things without having gone over the wall to observe
51
them. Another phrase found only in association with hsien is chih shih
fclA 'know what it is* (A 93, B 38).
The moral virtues are "desired 'a priori* '* (EC 2), the circle is "known
'a priori' ** (A 93). As we have shown in detail in § 1/1/2/5 and 1/1/2/6,
the Mohist justifies both claims by systems of definitions, deriving the
moral concepts from desire and the circle from likeness. It appears then
that the 'a priori* is conceived as what we know about a thing from the
definition of its name, even if it is hidden from sight on the other side of
a wall. We are also told that of the hardness and whiteness of a stone only
one is "referred to 'a priori' " (B 38), and that the idea or mental picture
is "known 'a priori' " in the case of a pillar but not of a hammer (B 57, 58).
Unfortunately the Mohist does not supply his definitions of concrete things,
presumably taking it for granted that they are common knowledge. But in
§ 1/5/7 and in the commentary on B 57,58 we shall examine Han definitions
of stone, pillar and hammer, which confirm that the hardness of the stone
belongs to its definition, and suggest that pillar and hammer would be
defined in terms of function, which imposes a visualisable shape in the case
of the former but not of the latter.
1/4/14 K u R3 AND K u Sic
Ku H 'inherently' (B 3, 38, 39). #t 'the thing as it inherently is' (EC 12,

A 88, B 7)
The adverbial ku is like pi № 'necessarily' in implying certainty, but
the basic distinction between them shows up very clearly in the dialectical
chapters. Pi (defined A 51) is the certainty of logical implication, ku of
being rooted in what is basic to the object or its situation; in terms of
Mohist disputation the former belongs to the realm of names (like yin H
'criterion'), the latter to the realm of objects (like ku SSt 'reason, cause').
Etymologically adverbial ku descends from the noun (with which graphically
52
it interchanges), in the sense 'the thing as it basically is/originally was'
5 1
Cf. § 1/5/8.
6 2
P'ei Hsiieh-hai 307-328.
used in connexion with hsing tt 'human nature' in Chuang-tzii and

hZ
Mencius. The essential identity can be seen in an example where ku is
adverbial but written with the graph of the ancestor noun:
Hsun-tzu ch. 23 (Liang 329/-2) / L H » * > S £ » 1 B A £ « » 2 № £ f e A £
" A l l propriety and righteousness is born from the sage's artifice, it is not
that basically they are born out of man's nature."
In the dialectical chapters adverbial ku is rare, and the only interesting
examples are in a section where the Mohist is explicitly distinguishing the
objective fact from the way we describe it:
B 3— » IBffiilSbUb °
"We dismiss one or other, but inherently the thing is what we call it."
This connects with an instance of the noun ku in the same sequence:
B 7 m^rnw^ o mm. °
"Dismissing one or other none is reduced. Explained by: the thing as it
inherently is."
The relation between this ku and the commoner ku 'reason, cause'
will be considered under the definition of the latter in A 1.
What X is in itself contrasts with what it is deemed to be (wet 1^):
cf. E C 12 "What Jack is in himself resides in Jack", A 83
tei^S, fitfe "What Jack is deemed to be is a matter of appropriateness".
1/4/15 K u o 31
Kuo m 'pass beyond' (A 5, 50, 98. B 10, 22, 24, 33, 40, 41, 58, 69)
Kuo is common in the difficult contexts which tempt editors to

arbitrary emendation, but a collation of all examples suggests that with the
possible exception of B 69 all uses derive immediately from its basic sense,
'pass beyond'. It is simply that the Mohist's special concern with the
problem of the transient conditions of knowledge causes him to use kuo
in unfamiliar connexions. In all usages kuo is the counterpart of tang H
'be plumb with', although there are no examples of the two contrasted in
a single sentence (as for example in Chuang-tzii ch. 6 (Kuo 226/2)
Sllfffl^fte, Hiffn^P§ flMfei " A man like that when he errs does not regret it,
when he hits the right course isn't pleased with himself").
Tang is the fitting of name and object, description and fact. Kuo is
'going beyond' in the sense either of 'missing, erring' as in the Chuang-tzii
53
Chuang-tzu ch. 19 (Kuo 658/1,2), Mencius 4B/26. The relation of ku and
hsing is discussed in G(9) 216, 251-254.
1/4 191
quotation (B 40, 41) or of 'exceeding' (A 98 « £ ( = • - ( £ ) , B 58 SIfr 'exceed

the matching'). In relation to time there is a similar contrast with tang.
Name and object fit only while the object exists (A 44, 50). If knowing
were the same thing as perceiving "knowing when prolonged would not fit
the fact" (B 46 ^ P A ^ ' l l f ) . Knowing is being able to give a description after
"having passed the thing" (A 5 ® %). Raising the question "Is it supposing
the already ended to be so?" ( J ^ B i ^ ^ i f e M ) is the class of doubt due to
kuo 'having passed' (B 10). While most names fit objects, the word huo l&
'one-or-other' transfers freely from object to object, is a 'name which passes
on' (B 33 » £ ) .
There are also a few cases of passing in space: A 50 "pass a
pillar", "pass a bridge" (get beyond the end of the bridge), B 22, 24
SISJE "pass the plane" (of a shadow receding into a mirror).
1/4/16 L i gf
Li M 'link' (A 79, B 3)
Both the examples of this word seem to refer to the connecting of
names in phrases or sentences, as in Hsun-tzu's Right use of names:
Hsiin-tzu ch. 22 (Liang 318/2) £ M r f n K « » £:£ffi-ib ° WMtiLX £ZM 9
"The object being conveyed when the name is heard is the use of names.
Making compositions by stringing them together is the linking of names.
When its use and links are both grasped we are said to know the name."
The forger of the Kung-sun Lung tzu who seems to have known the
t
uncorrupted text of B 3 (cf. B 3 n. 275), several times uses li H in the

same sense:
(Ch. 4) Ch'en 143/2 if ft . . . '^M ° 150/1 ° 157/1 °
"The blue and the white . . . are not linked." "The blue is linked with the
white." "The linked colours appear from it."
1/4/17 Li®
Li S 'pattern, arrangement' (A 75, B 78, N O 9, 10)

Li is the patterned arrangement of parts in a structured whole, of
things in an ordered cosmos, of thought in rational discourse, and in Names
and objects of words in a completed sentence. Its emergence in the Sung
dynasty (A.D. 960-1279) as one of the central concepts of Neo-Confucianism
54
was the culmination of a long development. In pre-Han philosophy it
5 4
Cf. T'ang Chun-yi Jtgjft (1955).
attracts attention especially in the Interpreting Lao-tzu of Han Fei tzu,

who uses it of the specific configuration of properties ("square or round,
55
long or short, coarse or fine, hard or soft") in each kind of thing. In
Names and objects it is the organisation of names in the sentence and the
interrelations of the objects to which it refers (NO 9 "scrutinise the li of
names and objects", N O 1 0 "The proposition . . . becomes full-grown
according to a /*"); a mutilated passage seems to identify it with the
structure common to parallel sentences and to compare it to the articulation
of members in the human body (NO 10). In the summa itself li is used only
twice, of the ordered relations of rational thought, in accordance with
which we accept or reject judgments (B 78) or practical alternatives (A 75).
In B 7 8 it is used with lun Ira 'sort out, grade*; our sorting of admissible
and inadmissible claims may or may not accord with their li, the objective
pattern of their logical relations.
Cf. Lii-shih ch'un-ch'iu ch. 18/4 (Hsu 8 3 6 / - 2 ) Wi^W%Wm » » i f
"Therefore what in the distinctions of disputation does not coincide with
li (the orderly relations of thought) is false, what in the clever man does
not coincide with li is deceitful. . . . Li is the ancestor (the ultimate source)
of judgments of right and wrong."
1/4/18 L I A N G M, P ' I E N M, C H O U Ml A N D Y I %
Liang M 'in both cases' (A 6 4 , 6 8 , 7 3 , 8 3 , 8 8 . B 2 5 , 2 7 , 3 9 , 4 8 , 54, 7 0 .

N O 18)
P'ien M 'in one case but not the other* (A 4 6 , 8 3 . B 3, 4 , 7, 6 3 , 6 6 .
N O 12, cf. A 4 9 )
Chou m,№ 'in all cases' (B 82. N O 13, 17, cf. B 5, 6 9 )
Yi % 'leave out of the argument* (A 75. B 3 9 , 7 4 . E C 11)
Liang as adjunct marks off a pair from other things (A 6 4 M^.^PHI

"between a couple of pieces of wood**, B 2 7 Mfra "a pair of wheels"). It
is not equivalent to the numeral erh ^ in the same position, which tends
to contrast two with one (A 8 6 "two names for one object"):
N O 18 ^ffiSffnEgJE-fe . . . o -Jgffn&e-fe °
"It is not that there are four legs to a pair of horses (taken together). . . . It
is that of two horses one is white.**
Apart from this use of Hang, the words Hang, p'ien and chou generally
function adverbially (B 4 8 "know it in both cases**, B 66 M^MM^
5 6
Cf. B 57 comment.
1/4 193
"in one case have and in the other not have any", N O 17 " i n every
case love men").
Of these terms liang is used especially of the converse claims of
disputation, "These are oxen" and "Those are non-oxen":
A 73 * iR * nJ-tfe o
"Being the converse of each other is if inadmissible then on both sides
inadmissible."
Chuang-tzu, as an enemy of disputation, advises us to "forget on both
sides" (MS), and to "proceed on both sides" (Mff). 66
In a few cases liang, p'ien and chou function verbally. Verbal Hang
equivalent to adverbial Hang before a null verb suppliable from the context:
Lu-shih ch'un-ch'iu ch. 15/2 (Hsu 628/-1) *PF«IW . . . > fliJA
"It is impossible to get one benefit without sacrificing the other. . . . If

you do not sacrifice the lesser benefit the greater will not be got." (Liang =
liang te 'get in both cases'.)
In the dialectical chapters the null verb after liang will be k 'o 'admissible'
or pu Wo 'inadmissible'; the use of Hang as the verb allows the Mohist to
cover both the positive and the negative possibilities:
A 83 » Wffn^fiS ·
"Converse claims, allow-or-disallow on both sides, not on one without
the other."
B 39 9 t « : J 5 » « M * ·
"What is left out: when sophistry is unable to allow-or-disallow on both
sides." (If instead of ox and non-ox we offer ox and horse as the two sides,
we are not bound to allow that if one is ox the other is horse, and vice versa:
all that is neither ox nor horse has been left out. Cf. B 35.)
"Explained by: not being applicable to everything at once."

Both verbal liang and 'leave out' appear in the account of the doctrine
of P'eng Meng 1&tsi and his school in the last chapter of Chuang-tzU
anticipating Chuang-tzu's own attack on disputation:
Chuang-tz* ch. 33 (Kuo 1086/1, 2, 3, 4) M ^ M ^ W . . . »H±
« « . . . · ±MM&2MXMM± · « r » . . . <
> mm
" . . . going whither things tend and not 'applying on both sides' . . . never
choosing between things but travelling with them. . . . The Great Way
64
Chuang-tzu ch. 2, 6, 26 (Kuo 70/6, 242/2, 930/-2).
can embrace them but not make the distinctions of disputation between
them. Know that for all the myriad things there are standpoints from which
they are allowable, standpoints from which they are not. . . . As for the
Way, there is nothing that it leaves out."
1/4/19 L U N Ira AND L U N L I E H fa3\
Lun m 'sort out* (A 6, 89. B 34, 78. N O 6, cf. A 9)

Lun lieh fan 'grading* (EC 9, 10)
Lun tm in pre-fian usage is thought or discourse which sorts things

and puts them in their proper positions and grades. In the dialectical
chapters it is used of sorting out the known and the unknown (B 34), the
admissible and the inadmissible (B 78), the long and the short (B 78).
This verb is closely related to the noun lun fa (written fra in A 9),
the general term for the differentiated modes of behaviour between persons
prescribed according to their relative social positions (ruler and minister,
father and son). In Expounding the canons it appears in the combination
lun lieh 'arranging by grade*.
Cf. Lu-shih ch'un-ch'iu ch. 18/7 (Hsu 864/8-865/1) T J ^ M f t f M ^ I U °
1 0 » r i S K J ° X 0 » r * » J M * J ° I B » T # S J . .. o0 » TM
" 'Which counts for more, Ho-nei or Liang?*

The king said, 'Liang counts for more.*
'Which counts for more, Liang or you yourself?*
The king said 'I myself* **
"Ho-nei is at the bottom of the three gradings, you yourself are at the top
of the three gradings. . . . " · '
1/4/20 M A O IS AND M u *
Mao IS: noun, 'looks, visible characteristics* (A 32, 47, 48, 71. B 22, 65,
N O 2, 7). Mao (A 5), U (A 95): verb, 'describe*
Mu ¥ : verb, 'describe* (NO 11)
In ordinary Chinese mao is a familiar word for a person*s features,
what he looks like (cf. Huai-nan-tzu ch. 17 (Liu 17/15B/7)
#£@ "What the spirits look like does not appear to the eye*'). It refers to
the surface characteristics while hsing TfJ 'form* refers to the three-dimen
sional shape. The combination hsing mao 'shape and features* recurs in
Names and objects, and we find mao t'ai til 'features and posture*
1/4 195
in B 22 (of a man reflected in a mirror). But the technical sense of the word
for the Mohist has a much wider range; it is his only term for the perceptible
qualities of a thing, corresponding to the chuang of Hsiin-tzu's Right
use of names. Things "named according to shape and features" include
mountains, houses, swords (NO 2, 7). The circularity of a rotating wheel
or the squareness of a block of wood are their mao (A 48, B 65). It seems
57
probable that the term is intended to pair with ch'ing 'essentials', although
we never actually find both in the same context; when we compare the mao
(the figure) of an object with the standard for a circle before calling it
circular (A 70, 71), we would be judging whether it has the ch'ing, that
which is essential to being circular. This is suggested by the dialogue of
Hui Shih and Chuang-tzu quoted in § 1/4/6, where Chuang-tzu maintains
the paradox that the sage has the mao and the shape of a man and is there
fore to be called a man, although he lacks man's ch'ing, the essentials of
humanity. The pairing of mao and ch'ing is common in Confucian literature,
but in a different context of ideas (which is perhaps the reason why
Hsun-tzu prefers chuang). It belongs to the rationale of the It Wt 'rites,
manners', a sphere in which there is a profound contrast between the mao,
the external demeanour, and the ch'ing, the real man, the passions without
which he would not be human, but which must be refined and disciplined
58
by the rites.
The graph 3ft has a surprising variety of forms. In li-shu it is written jft
59
or IB. The Wu manuscript of Mo-tzu regularly writes it without the
radical as Ji,, a form which is likely to be original. It seems that at one
stage in the transmission it was easily mistaken for US, an old graph for
60
min S 'people'; for in A 32, 47, 71 we find min where we should expect
mao:
All r»jft«»ts«a*o
"Being 'so' is the characteristics being like the standard."
(Cf. B 65 »<RW«[fn#l»ifofcRS » ( »
Ufa)» °)
"If the characteristics of the square are complete, that both things have the
standard but are different, one wood and the other stone, is not inconsistent
with their tallying in being square. Anything of which they complete the
characteristics (as in the case of being square) is so of both."
5 7
Cf. § 1/4/6.
6 8
Cf. G(9) 263-265.
6 9
Ku Ai-chi 4/58B.
6 0
M . 1760. Ting Tu T g g (A.D. 990-1053) reports it as an ancient graph for
min (Chiyiin WYWK 251/2).
A 4 7 (fit) ° *^Jt&-tfe o A 48 ( * M ) . . . tttt °

(Circling round) It is the figure of a curve.
(Rotation). The figure is constant.
A32ttrHJ*#»»P«B£ltitfi#fe« r t f i £ « »«HbJBte*
"Therefore 'saying* is an uttering of descriptions for which all speakers
share the ability. 'What has a figure like the drawing is a tiger* is saying.*'
(Cf. A 5 T » J ffe# » m S » S ^ r f o t B S S ± . " 'Knowing': having by means
of the consciousness experienced a thing one is able to desciibe it after
wards.")
Because this technical use of mao is unfamiliar, previous editors have
6 1
emended it away in A 48 and B 6 5 and proposed other emendations for
62
the corrupt S in A 3 2 and A 4 7 . But it can easily be appreciated that the
Mohist would feel the need for a word to cover all that is so (jan #&) of
a thing, and mao would be conveniently adapted for the purpose.
In Names and objects the verbal mao 'describe' is replaced by mu :
N O ii ip§n^±&»mimn±& °
"By describing summarise what is so of the myriad things, by assorting
explore which are comparable among the various sayings." (Cf. Chuang-tzU
ch. 2 3 (Kuo 810/5), a passage with close parallels to A 4, 5 quoted under
our translation of the latter: for the Mohist's mao it has mu 1^, taken as mu
63
* by Ch'ien M u . )
For the graph ft, cf. § 1/2/1/2/15.
1/4/21 M I N G £ A N D S H I H ft
Ming %\ 'name'
Shih If 'object'
Etymologicaily m w £ / * M l £ N G 'name' is related to falling-tone
#
wttW£/ MlANG # 'to ordain', which is to name either something to be
brought about (a sense distinguished by another cognate word, lingj
* L l £ N G ^ 'command') or an already existing thing. A n object is some
thing which is shih 'filled out, solid, real' (opposite of hsu Hfi 'empty,
tenuous, unreal'). Unlike wu % 'thing', which is commonly used generally,
64
of the thing which an object is, an ox or a horse, shih is used only of the
concrete and particular. This point is especially clear in A 78 and in Hsun-
tzu's Right use of names :
6 1
Sun 213/-3, 238/-4.
6 2
Sun 212/4,5, 213/-4.
6 3
Ch»ien Mu (1951) 193/7. Cf. p. 268 below.
«* Cf. § 1/4/33.
1/4 197
Hsiin-tzu ch. 22 (Liang 315/5-8) # W f i « * i f n » l № f » # » R i F n H 9 f # · "J
"There are things (wu) with the same characteristics but in different places
and things with different characteristics but in the same place; they are to
be distinguished. Things with the same characteristics but judged to be in
different places, although they may be put together, are said to be two
objects (shih); a thing which alters in its characteristics without objects
dividing and being judged different is said to be transformed (hua: cf.
A 45 T o r example, a frog becoming a quail'), and what is transformed
without division is said to be one object."
A name is used to refer to an object or to objects of one kind. Its
typical examples are therefore nouns, 'thing', 'horse', 'Jack' (A 78). But it
appears that 'white' and 'big' (NO 1), even chii ft 'all' and to & 'much'
(B 3), come under the heading of 'names', and in any case the Mohists
have no other term corresponding to our 'word'.
There is no presumption that each name refers to a separate object or
kind of object; on the contrary there is a technical term (ch'ung H ) for the
identity of objects referred to by more than one name (A 86). It is obvious
in any case that an object or objects indicated by 'white' or 'all' will also
be 'thing' and 'Jack' or 'horse'.
The basic sense of shih, 'full, solid, real' is nowhere exemplified in the
dialectical chapters. (We do find hsii 'empty', A 64, and ying S 'fill',
defined in A 65.) Except for one case of shih 'fruit' (NO 18) the word has
no other sense than 'object'. In A 11, at first sight an exception, the word
6 5
turns out to be a Sung substitute for the tabooed ch'eng M .
A practical concern with the relation between names and things (or, in
Taoism, a metaphysical concern with the limitations of naming) is common
to all the major pre-Han schools. It is against the background of this
general concern that the School of Names and the later Mohists ask their
deeper questions about the logic and semantics of naming. Admittedly
some inquirers have assumed the reverse, that the theoretical questions
preceded the practical, and have even refused to believe that the saying
about cheng ming JE£ 'right use of names' in the Analects of Confucius
87
can be earlier than the period of the sophists. However we might expect
65
Cf. § 1/2/1/4/4.
·· Analects 13/3.
67
Waley (1938) 172 n. 1. Waley finds the style of the passage too elaborate for
the earlier sayings of Confucius, and points out that Mencius never mentions the
right use of names. Creel ((I960) 321,322) adds that Hsiin-tzu in the Right use of
that one of the first consequences of the decay of Chou institutions would
be a conviction among conservatives that things are no longer being called
by their right names. Confucius laments that the ritual vessel now accepted
as a ku № is no longer a real ku, and calls for a return to the old order in
which a lord was a lord, a vassal was a vassal, a father was a father and a
68
son was a son. There is nothing fanciful about the statement in the Han
69
bibliography that the original source of the School of Names was the
office of rites; ritualists would be the first people to form the habit of
meticulously comparing the verbal prescription with the ceremony or
ritual object which can no longer be trusted to accord with it. But later
with the progressive breakdown of traditional standards a minister's
conduct comes to be compared, not with the ideal embodied in the name
of his office, merely with the wording of his ruler's decrees. In Legalist
thought the ruler 'names' (ming # 'ordains/names') a task, waits for the
performance to assume a definite shape (hsing by which it may be
judged, 'checks and matches' (ts'an wu i£ffi) it against the phrasing of the
70
decree, and rewards or punishes accordingly. This is the concept of
hsing ming ?1£/Jflj£ 'shape and name', translated by Creel as 'performance
and title', which by the 3rd and 2nd centuries had replaced cheng ming
'right use of names' except in the Confucian and Mohist schools (Creel's
suggestion that hsing has a special sense, 'accomplishment/performance',
seems unnecessary to his own argument; even in the Li-chi H I 5 examples
71
quoted by h i m hsing seems to have its ordinary verbal sense, 'assume a
fixed shape'). Creel argues persuasively that it originated with Shen
Pu-hai (died 337 B.C.) and that it underlies the techniques of
personnel control by which the Han solved the new problem of administrat
ing a reunited and bureaucratised Empire. We can see from the Han
72
accounts of the School of Names that the extinct tribe of logicians was
remembered simply as the lunatic fringe of people engaged in the serious
business of cheng ming and hsing ming.
names might be expected to quote the saying if he knew it, and that the reference to
punishments has a Legalist sound to it. Although these considerations have some
cumulative weight, I am more impressed by Fung Yu-lan's argument that the right
use of names is an integral part of early Confucian conservatism (Fung (1934)
84-89).
68
Analects 6/25, 12/11.
69
Han shu (ch. 30) 1737/2.
7 0
Cf. Han Fei tzu ch. 7, 8 (Ch'en 111/-2—112/2, 121/-2—122/2, 122/-4—1).
7 1
Creel (1970) 85. His examples of hsing interpreted as 'performance' are
criticised in Lau (1973), 122, 123.
72
Shih chi (ch. 130) 3291/-2, Han shu (ch. 30) 1737/2.
1/4 199
Another non-philosophical usage is the application of the terms ming

shih to a man's reputation and the achievements which substantiate it,
found in Mencius 12/6, Chuang-tzu ch. 4 (Kuo 139/2f).
1/4/22 MING#
Ming ffr 'to name' (A 20, 78, 79. N 0 1,2, 7)

The distinction between the noun ming %x 'name' and the verb ming
'to give a name to* is observed rigorously in the Mohist dialectical chapters,
as well as in the genuine essays of Kung-sun Lung and the Right use of
names of Hsiin-tzu. Ming 'command* is consistently avoided in favour of
ling ^ (A 18, 77, B 55) except in the sense of 'decree of Heaven, destiny*
(NO 16).
1/4/23 PEIS*
Pei № 'self-falsifying, illogical' (B 8, 34, 71, 73, 76, 77, 79)

The word pei 1$ is used of illogical behaviour or talk, which
discredits itself by its own inconsistency:
Lu-shih cWun-ch'iu ch. 18/4 (Hsu 839/-2) * » # « ± * - t b °
"The verbal formulation is the manifestation of an idea. It is illogical to

contemplate the manifestation but overlook the idea."
Huai-nan-tzu ch. 16 (Liu 16, 9B/2) » A « 1 ± ^ H & » » №£ °
(Of a man stealing a bell who covers his ears so that people will not hear
it ringing.) " T o dislike others hearing it is all very well, but to cover his
own ears is illogical."
Kung-sun Lung uses it in the White Horse: " T o judge that a yellow
horse is not a horse, yet in the case of a white horse judge that there is a
horse, this is flying things going underwater and inner and outer coffin
being in different places, it is the most illogical saying and garbled formula
tion (f#"8l3Lf$) in the world". In a surviving quotation from the Pie lu
78
№l$k of L i u Hsiang SfllRj, Kung-sun Lung in his turn has the word used
against him by Tsou Yen, who accuses sophists of "elaborating propositions
,
to prove each other illogical" (ff|if№Kffl(1$)* №. The corruption of the
phonetic may also be observed in B 34, 79).
According to B 71 "to be self-falsifying is to be inadmissible" (f£,
^ "Ufa), but this does not imply that pei is synonymous with 'inadmissible'.
7 3
Quoted p. 20 above.
The Mohist is criticising as self-contradictory Chuang-tzii's thesis that we

cannot allow a proposition from one point of view without disallowing it
from another: "to deem all saying self-falsifying is self-falsifying" ( K f
JiSlSf^, f$). He uses pet similarly of the two other positions which he
exposes as self-contradictory, teaching that learning is useless (B 77) and
denying the legitimacy of denial (B 79).
1/4/24 S H E N G 8&
Sheng B 'win, conquer' (A 74, 89. B 25, 35, 43, 70)
Sheng in logical contexts always refers to the victory of one party in

disputation (A 74, 89. B 35, 70). In scientific contexts it refers to one force
prevailing over another (B 25, 43).
1/4/25 SHENG S
Sheng % 'sound' (A 78, B 53, N O 9. Cf. A 11)

In other pre-Han texts sheng often has the sense of 'reputation', which
might seem to be the meaning in B 53. But in the dialectical chapters it is
used only of the vocal enunciation of names, except for a single reference
to the sound of metal in A 11. Cf. Lu-shih ch'un-ch'iu ch. 16/8 (Hsu 16,
28B/-1) Jiff] ( = 7 £ ) £^3feffoSW^!Bfk "This is the shape substantiating
something different from the name, and the sound referring to something
different from the object", as well as the programme of disputation quoted on
p. 20 above: 91 A S ^ ^ # R 3 t ; i ; "stretching what he literally says (the sound
in contrast with the meaning) so that he cannot get back to his own idea".
1/4/26 SHIHS
Shih g 'room, house' (A 86, 88. B 31, 42, 70. E C 6, 12)

In spite of their different initials there seems to be a phonetic as well
as semantic connexion between
(1) 5AiA/*Sl£T 'room', which like the English word, suggests a space
to be filled, and
(2) S / M ' A / * D ' I £ T H , 'full, solid' (opposite of hsu JSL 'empty, tenuous').
u 75
The Shuo wen defines No. 1 by No. 2. The Shih ming repeats and
amplifies the definition: T S J Jtfe. A ^ W S I * * ^ " ' R o o m ' is what is
filled. Because men and things fill up its interior".
74
Shuo wen ch. 7B (Tuan 341B/5).
75
Shih ming (ch. 17) SPTK 40A/7.
1/4 201
Both graphs have the 'roof radical, so that No. 2 is visualised primarily
as the filling of rooms or house. The Shuo wen indeed analyses the graph
as goods under a roof.
Since No. 1 is the ordinary Mohist word for a room or house, and
can always be understood in this sense without too much forcing, the
proposal of a technical usage is offered with some reserve. The Mohist
uses No. 2 only as a noun, once (NO 18) in the sense 'fruit* (which unlike
leaves and flowers fills out and becomes solid), otherwise of particular
objects (solids filling spaces). Now it is remarkable that he seldom uses
shih 'object' except in contrast with ming %x 'name'; in spite of his interest
in the parts of wholes and in hardness and whiteness he does not use such
expressions as ffi$f ' i t is in the object', 'inside the object'. But we do
find ffiil, 1*14* (B 70). The possibility arises that the natural way to speak
of an object's constituents may be to say, not that they are in the 'filling'
(No. 2), but that they are in the room which is filled (No. 1).
The Mohist, with his strong interest in geometry, analyses the object
in strictly spatial terms, using the words ch'u jS (verb 'occupy', noun
'position') and3>i»£ S (verb 'fill'). A n object "occupies a position in space"
(B 13 J S ? , cf. B 33). "Different positions do not fill each other. Not being
each other is excluding each other" (A 66 JSjfi^+lSL ffl2felrfl^"ife).
But such qualities as hardness and whiteness "fill each other" (B 15 t@S),
"do not exclude each other" (A 66 /FtB^r-). When ch'u 'occupy' is used
of two things it is implied that they share the same place (A 22 I" ^fe J *
TflJjSiSljS'tfe " 'Life' is the body sharing the same position with the
consciousness"). Ch'u is put in the same sentence with shih 'room' in one
of four pairs of definitions of sameness and difference:
A 86, 87 » . . . ° JÛST · °
"Both sharing positions in the room is the sameness of 'being together'. . . .

Not sharing the same place is 'not being together'."
The first sentence is a variation on the idiom ch'u shih 'live under one
roof, of which there is an actual example in A 88. We might take it as
referring to people in the same house. But in that case it merely illustrates
the concept of being together; we should rather expect another abstract
definition like the other seven, approximately equivalent to 'sharing the
same place'. This expectation will be satisfied if the Mohist has in mind
qualities which are together in that they share positions inside the room
filled by the object.
The same suspicion arises when we compare the related sections B 37,
42, 70. According to B 37 we can know that a stone is hard without
knowing whether it is white because hardness and whiteness are two and
"are in the stone" (ffiS). According to B 70 if told that a colour is like

that of something we know which is white, we will know that it is white.
Here there is no mention of a stone; the phrase is "the colour of that
which is in the room" (ffill#;£fe), contrasted with "that which is outside"
(Ê^#). Although it is possible that the Mohist is thinking of an unknown
object hidden inside a house, one again suspects that he is referring to the
room filled by the substance of the stone. The question of being in (tsaiÊ)
and belonging in (ts'un # ) the stone in B 37 leads on to a discussion of
tsai and ts'un in B 42. This offers as an analogy children being in and
belonging in different rooms in a house (shih Vang I ^ S 'rooms and hall').
There is no doubt here that the house is an actual one; but rooms Would
be the most obvious analogy for the rooms filled by different objects and
jointly occupied by hardness and whiteness.
If this analysis is correct there is a rather interesting difference between
the Mohist's conceptual apparatus and our own. The Western philosophical
tradition has conceived qualities as inhering in something underlying them
called the 'substance'. But the Mohist seems to conceive the object as the
composite 'filling' of a space, and to think of its mutually pervasive
components (shape and consciousness, hardness and whiteness) only in
relation to the space filled.
1/4/27 TAI?#
Tat 'require' (A 77, 78. B 49. N O 17)

Tai 'wait on, depend on' appears only twice (A 35, N O 10) outside
the constructions J&W 'necessarily requires' and X rfO^S Y 'requires X
J
and only then constitutes Y \ In these it implies having as a necessary

condition; Y has to 'wait' for the arrival of X before it can come about.
1/4/28 T A N G IS'
Tang m 'be plumb with' (A 14, 44, 50, 74. B 12, 35, 46, 71)
Tang is the fitting of name and object (cf. Kuan-tzu ch. 55, BSS 3/14/—1
^ W B I O ' / p , /fvHJMfiL " I f name and object fit there will be order, if not
there will be misrule", Lii-shih ch'un-ch'iu ch. 17/1 (Hsu 17, 3B/2)
^1SiK1ifff5V^^1Sftffi^ "There are many cases where a name does
not fit its object and a deed does not fit its function"). We may distinguish
between intransitive and transitive uses:
(1) Intransitive tang pronounces that a claim in disputation, or a
saying (A 14, B 71), or knowledge (B 46), fits the fact. According to A 74
1/4 203
"winning in disputation is fitting the fact" (M№?> Hi til); one party calls
the object 'ox', the other 'non-ox', and "they do not both fit" (^RHHi).
The tang of this usage is the main Mohist term for expressing factual truth.
(2) When the verb is transitive the relation is reversed and the
reference is to the object fitting a name (A 50 H» T J T J "fit 'ox'
and 'non-horse' ", A 74 % T ~X J "fit 'dog' ", B 12 gf T ^ J T M J "fit 'ox',
'horse' "). It may be noticed that since the language has no affirmative
copula corresponding to the negative copula fei W the Mohists say "fits
'ox' and 'non-horse' " where we should say "is an ox and not a horse".
Tang is once used concretely, of the shadow in the mirror being plumb
opposite (B 22). Otherwise the graph occurs only for the temporal particle
ch'ang # (B 38, 61).
1/4/29 T i S
#
Ti ^ 'complement' (A 22, 51, 85, 89. B 5)
A ti Wi is a party which confronts one as an equal. Its most familiar

use is of the enemy in war; but even in this usage wu ti means 'to be
76
matchless, have no equal', not 'to have no enemy'. Ti is also used of a
peer in social intercourse, as in Kuo yii (Chou B) S P T K 2, 12B/4 $StBIiO
"When a guest arrives from a state of the same rank", 2, 18B/4 %EWt y
iSfci&Hlfl "According to the rites one must defer three times to an equal",
(Gh'u B) 18, 9A/4 B J c S K t f i T F , I 0 # t £ "Moreover in the case of an equal
or inferior the vendetta is allowed". Ti in the sense of 'the enemy' is
common in the military chapters (Mo-tzii ch. 52-71), generally written
with the graph as far as ch. 62 but with the standard graph from ch. 63.
Originally it may have been written without any radical, as in the bronze
77
inscriptions.
In Taoist contexts the Way, which as the One transcends all dicho
tomies, is described as wu ti 'without complement' (with the implication
'matchless, unrivalled'). The graph is found with either radical:
Huai-nan-tzu ch. 11 (Liu 11, 8A/3) ; f c - ^ f M f t > &igj&5TF °
"The One is the utmost in value, it has no complement/rival in the world."
Ch. 14 (Liu 14, 9A/5) -~tfe^ » » mib±№& °
"The One is the root of the myriad things, is the Way without complement/
rival."
7 6
Cf. Mencius 1A/5.
77
Tuan Wei-yi 328.
In the dialectical chapters forms or corruptions of the graph may be

recognised in a number of puzzling graphs which appear in interrelated
contexts:
(1) * .
(2) (A 89), the form common in the military chapters.
(3) IS (A 22) identified as No. 1 by L i u Ch'ang (op. cit. 156). The
Tz'U hat f№#$ entry for No. 1 cites from the Ming dictionary Cheng tzu
t'ung J E ^ S a passage ending with the sentence " A t present shang is
vulgarly written for it, wrongly" (î&Mtfp T i8 J » ^ ) . The entry for No. 1
in the K'ang-hsi edition of the Cheng tzu t'ung contains the passage but
without this sentence, and I have failed to trace its source. The two graphs
are in any case often indistinguishable in bronze forms, and in li-shu the
78
graphs 21, S i are also written aS, SS.
(4) H (A 51, 85), which three times in one entry in the Fang yeti
79
appears to Chu Ch'i-feng to be a corruption of No. I . T o explain
the confusion Chu points out that the Shuo-wen graph for No. 4 is equiva
lent to S (abbreviated if& over M ) . This is in fact the graph used in the
Sung edition of the Fang yen reproduced in the Ssu-pu ts'ung-k'an, and also
(in the vulgarised form M . 30158) by the Mao edition of Mo-tzu in
A 51, 85. It is the common ancestor of the standard graph and of variant
forms M ( M . 5339), S ( M . 5533).
(5) M (B 5), for which the Sung dictionary Chi yun notes an ancient
graphic variant M ( M . 30164), known to have been often confused with
80
No. 4. P i Yuan, wrongly on the present hypothesis, emended No. 4 to
81
No. 5 in A 51.
A l l the passages are obscure except A 89, where Sun Yi-jang already
perceived that the word can only be ti in the sense of 'complement* (#!$&).
But when collated they turn out to be interconnected by references to elder
and younger brother, by the words pi 'necessary* and wet 'deem', and by
the correspondence of sheng 'engender' (in its raw state) (A 39) and shu
'ripened/cooked' (A 51). The latter remind one of certain images of organic
growth in Names and objects, which refers to 'the sameness of having the
same root' ( t ^ l S i t ^ ) and describes the proposition as "being engendered
by means of ku and growing up according to a pattern" (GUft^fe, KIMSk
N O 6,10). At first sight however ti seems to be translatable as 'complement'
only in A 89, B 5:
7 8
Tuan Wei-yi 32, 668. Ku Ai-chi 5/45A.
7 9
Chu Ch'i-feng 2642, on Fang yen SPTK 2/3B.
80
Chi yun (WYWK) 1362, M . 30161 def. 12. Chu Ch'i-feng 1389.
8 1
Pi YUan 10, 7A/8.
1/4 205
A 8 9 52J&* r « ^ J & °
"In the case of elder and younger brother, 'both being complements'."
A 51 (&) o n « « f i # - i f a o o
"('Necessary'). It is said of cases where the ti is ripened/cooked. (For
example, younger brother and elder brother.)"
A 39 (№) ° - A f f i i « f t f t « * ° < A 22 №±%M » * * T & * ° >
"('Agreeing'). They are two men but both see that this is a pillar. The
pillar's engendering of the ti in its raw state is not to be treated as necessary."
# #
A 8 5 ( * ) * + l * r # J t e < > S i » rtrjifeo
(Of senses of wei which do not imply making, such as 'deem', 'cure'.)
" O f hitting on a ti, 'leave as it is'. Of illness, 'get rid o f . "
B 5 » o
(On the 'seeing and appearing', 'one and two', 'length and breadth* of B 4.)
"Whether being deemed a complement is one-sided or double is not the
responsibility of the understanding."
Comparison of the five passages suggests several conclusions:
(1) Ti 'complement' does not require a two-way relationship (B 5),
which explains why we are told in A 89 of elder and younger brother that
'both are complements'. This agrees with common usage; granted that I am
my enemy's enemy, on the battlefield only one side is called the ti 'enemy'
of the other.
(2) The difficulty of thinking of ti as 'complement' throughout comes
simply from the fact that we are thinking in terms of propositions, the
Mohist in terms of names. In stead of saying that ' X is the elder' and ' Y is
the younger' entail each other, he says that elder brother and younger
brother are both complements; and in such a sentence as B 2 jifc&cl&£&$&
"If it is so of the instance here it is necessarily so of the thing it is judged
to be" he will be thinking of 'being so of the thing it is judged
to be' as the complement of llfcS^&ife 'being so of the instance here'
(phrases which we take from B 1). In B 3,4, it can be seen that he is thinking
of implication in terms of pairs of words one of which can or cannot be
dismissed without the other (p'ien ch'ti S i ) . All one-way or two-way
implications will be seen as one-way or two-way complements. It is there
fore natural that he should describe 'necessary' as said of complements
(A 51).
(3) The Mohist is a nominalist who holds that one names an object X
'pillar' and then "for 'like the object one necessarily uses this name" (A 78
r£*Jft#»&BU&£ifa). Y being deemed (wei ft) apillar is then necessary
as the complement of its similarity to X (A 85). There is however no neces-
sary relation between the deeming and the ku #C, Y as it is in itself, detached
from its similarity to X (for the contrast of ku and wet, cf. § 1/4/14). What
we say of Y is in the first place "engendered by means of Y as it is in itself "
(NO 10 EiSfc^fe), but "the pillar's engendering of the complement is not
to be treated as necessary" (A 39). It is when the complement is 'ripened'
by confirming its similarity to X that it becomes necessary (A 51).
1/4/30 TSAI ffi AND TS'UN #
Tsai ffi: with noun of place, 'be in' (EC 6, 10 bis, 12. A 37. B 14 bis, 20,
32, 33, 37, 70 ter). Before verbal phrase, 'depend on' (B 19 bis).
Intransitive, 'be present, exist' (EC 2. B 3, 42). Nominalised,
'presence, existence' (B 17, 41). Causative, 'locate' (B 16 ter).
Ts'un # 'belong in' (B 37, 42 nine). Intransitive, 'stay as it is, remain'
(be where it belongs A 46 bis, 85 bis, 88). Causative (keep where
it belongs), 'preserve' (EC 6 bis, 8)
The transitive verb tsai/*DZ'9G and the intransitive fa'w»/*DZ'W3N

are related semantically, phonologically and graphically and share the same
opposite wang Xl 'be absent, perish' (cf. A 85, 88, B 3). Their relationship
is important to the understanding of B 37, 42, which use both. Ts'un has
the narrower sense, implying not merely the presence of X but the involve
ment of its existence or character in its situation. In other texts we find
ts'un used of the members of the body, which are not merely in it, but are
constituent parts of it:
9
Mencius 4A/16 ff ¥ A # ^ A i f t M ^ °
" O f the constituents of a man none is more excellent than the pupil of
the eye."
Chuang-tzu ch. 2 (Kuo 55/-1) W f £ ^ 7 \ i < »Jgrfn#S§ °
"The hundred members, nine apertures and six organs are in him as his
constituents without anything missing."
In the dialectical chapters we do not find ts'un with the directive
preposition but do find so @f ts'un 'that in which it is a constituent' (B 42).
In B 37 tsai is used of the hard and the white 'being in the stone'
(ffiS); the summing-up of the Canon is 'Explained by: belonging in it'
(tftffiff). In B 43 it is shown that even if we start from a property as tsai
'present', as though it could exist by itself, we can always ask 'In what may
it be treated as belonging?' (SB "Iff). Throughout the corpus tsai without
object implies mere existence (EC 2, B 3, 17, 41, 43), unlocalised unless
an object is added.
1/4 207
1/4/31 Tz'rj fft
Tz'u 8? 'wording/verbal formulation* (EC 1). 'Sentence/proposition* (NO

10-12)
Outside the realm of philosophy tz'u is a general term for deliberately

composed utterances, phrased for ceremonial, aesthetic or persuasive effect.
In philosophical literature a tz'u is the verbal formulation of an idea, to be
inspected closely in case it is misleading:
Lu-shih ch'un-ch'iu ch. 18/4 (Hsu 839/9) ° HXUffBH*
"The wording is the externalisation of the idea. T o mirror the externalisa-

tion but discard the idea is nonsensical."
The Great appendix says of the Changes that "its meanings are
far-reaching, its wordings polished* * ( ^ e S l , № № X ) , but also that the
verbal formulations contain less than the symbols; "saying does not exhaust
ideas** ( I S ^ i f t K ) , the sage "established the symbols in order to exhaust
the ideas** and "attached wordings to them in order to exhaust what may
88
be said about them** ( A f t H f t j S . . . J M f i B B i f t K f i " ) .
Since the sophist*s art in putting words together can disguise the idea
he pretends to expose, references to tz'u in connexion with disputation are
nearly always pejorative. The programme of disputation quoted on p. 20
above ranks "correctness in wording** (f$lE) as the least important of the
disputant's tasks, and condemns "ornamentation of wording (!&§$) in order
to make nonsense of the other's case**. Kung-sun Lung*s 'White horse*
derides a "disordered formulation** (SLS?). Mencius claims to be able to
see through no less than four kinds of misleading tz'u, the 'one-sided'
(pi f$), the Vicious' (yin §=), the 'heterodox' (hsieh 3?) and the 'evasive'
B
(tun M).* The most important of these are yin tz'u, in which elegance and
subtlety of language exceed their due measure (yin, used primarily of
excessive rain and floodwater, also of sensual indulgence) in order to serve
some hidden end. The commentator Chao C h ' i fitifc (died A.D. 201)
describes them as "excessively beautified and untrustworthy wordings"
(Mencius S P T K 3, 25A/7 &fkfcS±tfc)> as in a royal concubine's out
wardly innocent proposal which concealed a malicious design. Kuan-tzu
refers to men who "devise vicious wordings and contrive vicious subtleties,
in order to flatter the ruler above and delude the people below" (Kuan-tzu
ch. io, 1/45/7 mnm» · fls«g*5 » a ± « a ± i f i . T « w B : ) .
82
Changes, Great appendix A 12, B 5.
83
Mencius 2A/2, cf. 3B/9.
The Lii-shih ch'un-ch'iu has a chapter entitled Yin tz'u which begins:
Ch. 18/5 (Hsu 841/2-4) IHMKfilffifti' * £ ( = « £ ) * « № ·
"Without wordings we have no means of sharing our intentions; but if we

indulge ourselves too much in wordings they become unruly. Inside an
unruly wording there is another wording, by which I mean what is in the
heart. If saying is not unfaithful to what is in the heart, we shall not go
far wrong."
It would appear from the examples given that "vicious wordings" are
characterised less by floridity of style (as Chao C h ' i evidently supposed)
than by verbal trickery and specious argument. The chapter includes
Kung-sun Lung's sophistry "Jack has three ears" ( W H S ) , and a series of
answers to a man who wants to know what time of day it is:
"Chuang Po, minister of Cheng, asked his father to see where the sun was.
'It's in the sky.' 'See how it looks.' 'It's exactly round.' 'See what time
it is.' 'It's now.'"
Expounding the canons starts with a section on ch'in-yin chih tz
( S S ^ f S ) . One is at first sight tempted to take this as merely a synonym
of yin tz'U, since the Shuo wen actually uses the phrase ch'in-yin ('soak into
and drench/permeate little by little', M . 17505/5) to define the word j t i i ;
Ch. 11A (Tuan 556A/-4) r S J · « g W » * . . . ° — H AMB p g j °
9
"Yin is permeating little by little following the lines of least resistance.

11
One authority says: Prolonged rain is called yin.
However there appears to be a contrast between the senses of yin
'go to excess' and ch'in 'encroach little by little' (written as 8 when it
refers to water soaking in and as ft of a state making inroads on another's
territory or a minister infringing on the powers of his ruler). What exceeds
its own limits (yin) at the same time encroaches on the limits of something
else (ch'in). The contrast can be observed in certain passages in Kuan-tzu :
Ch. 46, 67 (2/94/lf, 3/53/llf) A * K t e I ± J & m f e * * °
"Therefore the former kings in governing a state did not permit their ideas
excesses outside the laws."
9
Ch. 45 (2/90/9) %M T£#T&«&SL£-{fe °
"Selfishness is the reason why subjects encroach on the laws and are
seditious against their lords."
Both words are combined in reference to the grading or ranking
(lun Ira) of appointments according to merit:
1/4 209
ch.31 (2/32/-4-3) mmmzh^fc... ° » · ± m m

"If the rankings are encroached upon the meritorious will be endangered.
. . . The clear-sighted ruler is thorough in forbidding excess and encroach
ment. If above there are no rankings which exceed and encroach (one man
getting more than his due at the expense of another), then below there will
be no thoughts of seeking favour elsewhere.''
Kuan-tzU, it may be noticed, combines the words as yin ch'in 'go to
excess arid encroach', the Mohist as ch'in yin 'encroach and cause to go to
excess' (with the yin causative as in the first Kuan-tzu quotation, yin yi
'let the ideas go to excess'). Ch'in-yin chih tz'u should therefore be 'word
ings which encroach and vitiate', 'vicious wordings' which infect other
propositions and convert them into vicious wordings.
The theme of the first section of Expounding the canons is in fact the
claim that a man's nature may be criminal, which (since one's nature is
ordained by Heaven) radically affects the concept of the Will of Heaven.
Since the phrasing of formulations is the putting together of words,
the tz'u may be seen as contrasting, on the one hand with the idea it
verbalises, on the other with a word inside it:
Mencius 5A/4 tttt№S · · TOSf*fe °
"Therefore the interpreter of an Ode does not let a graph (that is, the word
it represents) interfere with the understanding of a phrase, does not let a
phrase interfere with the understanding of the intention."
It is from this usage that tz'u becomes the sentence in contrast with
the isolated name, in Names and objects and also in Hsun-tzii's Right use
of names.
1/4/32 W E I ft
Wei ft (level tone) 'is to be deemed as, counts as' (EC 1, 10. A 70, 83, 85,
88. B 2, 3, 5, 30, 37, 44, 60, 76. N O 5, 6, 17, 18)
(falling tone) '(be) for the sake o f (EC 1, 2, 5, 7, 9-12. A 7, 75)
The falling-tone wei is defined in A 75, and six senses of level-tone
wei are distinguished in A 85. The latter do not include the most common
sense of all, 'do, make'. Wei 'do' in fact requires no comment except that
in contrast with hsingft'conduct' it is ethically neutral (cf. A10,16-18, 80).
Level-tone wei is important in logical contexts; it affirms that a thing
satisfies the conditions for being deemed ' X ' or 'non-X' (B 2 ftlS "are to
be deemed milu deer", E C 10 ^cft# "is not yet to be deemed filial piety").
It is used with extraordinary syntactic mobility, without a complement and

84
even nominalised. Its relation to the transitive wet of the formula
' H X S Y ' ('deem X to be Y ' ) is that of intransitive to transitive uses of
verbs in general, and it is not obvious that either need be regarded as
primary. Yi wet 'deeming, supposing' may be mistaken, and is contrasted
with knowing in A 23, 24.
This sense of wei is distinguished in A 85 by the phrase which on
# , 1
grounds of systematic corruption we emend to r and understand as
"hitting on the complement" (the being called ' X ' which is the complement
of similarity to the object initially named ' X ' ) .
Falling-tone wei is the key word in the ethical reasoning in terms of
means and ends developed in Expounding the canons and in A 75. This wei
is primarily a transitive verb, 'do for the sake of. . .', which may be the
main verb of the sentence, as in Mencius 1A/7 I S ^ S S c l "How can Your
Majesty be doing so for the sake of this?", although more often wei and
its object are subordinate to the main verb. According to the Mohist
ethical system, one sacrifices a finger for the sake of an arm, an arm for
the sake of the man (whether wei chi 'for one's own sake' or weijen
%K 'for the sake of another'), the man for the sake of the world (wei
t'ien-hsia S X T ) .
1/4/33 W u ^
Wu & 'thing' (A 5, 6, 78. B 2, 9, 12, 65, 80. E C 2. N O 12, 13)

The later Mohists use wu when they wish to refer generally to the
thing meant by a name, not to a particular object which the name may
single out. However, the word has a particular reference at least once
(B 9); wu is not exclusively universal as shih 'object' is exclusively particular.
Wu may refer to such abstract entities as love and benefit (EC 2).
Similarly in older chapters of Mo-tzu we find objectors dismissing universal
love as an 'impracticable thing' (ch. 15 Sun 70/5 ^f~^ff^^) and recom
mending aggressive war as a 'profitable thing' (ch. 19, Sun 95/-1
1/4/34 WuE.ff
Wu 'to match point-by-point': £ (A 98): fF (B 58, 76)
A circle perfectly matches (wu E ) the standard for a circle (A 98),
which may be the idea or mental picture (yi M)> another circle, or the
compasses (A 70). When we try to imagine a hammer in use by imagining the
8 4
Cf. p. 118 above.
1/4 211
decorating of shoes, the content of each idea (yi) is 'in excess of the
matching* (kuo wu Mfp B 58). T o describe the two eyes either as inside or
as outside the face, and not exclude one eye from the face and include the
other, is 'matching with the face* (ftM B 76).
The wu E of A 98 is not the numeral '5', as Sun Yi-jang and more
recent editors have assumed, but the wu ffi defined in the Shuo wen as
9
'to check and match with each other* (ch. 8A, 377B f~ffi J ffi#ffiife).
The phrase ts'an wu 'check and match* is a Legalist term for comparing
commission and performance:
9
Han Fei tzu ch. 48 (Ch'en 1017/1) # f f i ± S [ ft^&M^ » g t f f i J ^ X * °
"By the principle of checking and matching one conducts checks in order
to plan for the best results, estimates whether the performance matches in
order to lay the responsibility for failure.'*
Wu 'match* is also found by itself:
Han shu ch. 21A (965/-1) A A f t f f i °
"The two series of eight match each other.*'
8 5
The wu of ts'an zvu is often written without the radical ( # E ) . W u
Yu-chiang has noticed, without seeing the relevance of the point to the
wu E of A 98, that everywhere else in Mo-tzu the radical has been supplied
by later graphic standardisation. The name W u Yuan ffi M (ch. 3, Sun 9/7)
is quoted with the radical missing in the T'ang manuscript of the Ch'un-shu
chih-yao In Mo-tzu ch. 70 the radical of wu ffi 'platoon' is
missing throughout in the W u manuscript and sometimes in the Mao and
86
other editions.
That the graph fp of B 58, 76 represents the same word is attested by
the account of the later Mohists in the last chapter of Chuang-tzu:
9 9
Chuang-tzu ch. 33 (Kuo 1079) tBiflBOS K K e i R № £ S W H t f
"They call each other heretical Mohists and upbraid each other in disputa
tion about hard and white and same and different, answer each other with
87
sentences at odds and evens which do not match."
It may be suspected that the more familiar wu f f 'oppose' descends
from wu 'match' in the same way as ti & 'enemy' from ti 'equally matched'.
For other examples of the interchange of graphs with the reading «?w/*NGO,
cf. M . 2703 def. 7 (3£, ^ ) , 409 def. 3 (ffi, ft), 435 def. 8 (ffi, £ ) , 2703
def. 2 (ft,**).
8 6
Chu Ch'i-feng 1799.
8 6
Wu Yu-chiang 15, 10A/8, Appendix 1, IB/4-6.
8 7
Cf. the comment on wu of Ma Hsu-lun JHi£{j% 21/3-5.
The proof that only a single word is involved is the phrase kuo wu
'exceed the matching', in which the wu is written with either graph (A 9 8
» £ , B 5 8 MM*).
1/4/35 WuYiftta
Wu yi 'lack what it is distinguished by' (A 7 3 , B 34)
Wu yi is certainly a technical term since it stands alone three times,

once as the whole summing-up of a Canon (B 34). The problem is to decide
what kind of verbal phrase is implicit after it. Probably it should be
connected with the formula yi X shuo Y 'argue Y on the grounds of X *
(B 1 , 6 6 ) :
B 66 K ^ W ® · m^m»»*z#m mî ·
"It is inadmissible to argue that oxen are not horses on the grounds that
oxen have incisors and horses have tails."
A 73 9
Ufa ° №&. # i b o
9 9
" A l l oxen, and non-oxen separately grouped, are the two sides. What lacks
that by which (one judges it to be an ox) is non-ox."
9
B 3 4 mZ. # » * I B H b o
"When we sort them out, the non-knowledge lacks what knowledge is
distinguished by."
1/4/36 X Y u m Y T ' U N G (SHUO) I R I ( £ )
X $i Y IRI (ift) 'Doing X is required by doing Y ' (B 58, 67, 68, 82)
The formula 'Xyii Y t'ung* with X and Y representing actions occurs
in three Canons (B 67, 68, 82) and one Explanation (B 58), 'Xyii Y t'ung
shuo' with X and Y as propositions in one Explanation (B 82). In all cases
the meaning is 'Doing X is required by doing Y ' , 'You cannot do Y without
doing X * .
B 6 7 ^*m±#*m^±m ·
" Y o u cannot allow that an ox and a horse are not an ox without also
disallowing it."
B82*r*ihJHr«^*itJHIft*«
"Therefore you cannot show that 'this* is not confined without showing
that it is."
In the second case the fuller phrasing of Y is explained by the fact
that it is always Y that is primary and X that is involved in it.
1/4 213
1/4/37 Y l M A N D H S I A N G ffi
Yi ft 'idea* (A 14, 70. B 38, 41, 57, 58. N O 3, 9, 10)

Hsiangffi'image' (B 38, 57)
The yi 'idea* is the thought which words convey from speaker to
hearer (A 14, B 38, 41, N O 3, 9, 10). The tendency is to think of it as an
image resembling the object and conveyed by the name (cf. A 70, B 57, 58,
N O 3). In Names and objects it is used verbally, 'to conceive/imagine/form
an idea o f (Jftjfi^SJ'^ffa " T o conceive a pillar is not to conceive wood").
It is used verbally with the same meaning in Hsiin-tzu's Right use of names:
Hsun-tzU ch. 22 (Liang 312/-4) J l ^ * f * l f f i # » ftXtiljfcft-feBI » & t b #
"With whatever is the same in kind or the same in ch'ing (what a thing is
in itself), the image of the thing which the senses present is the same
(more literally, 'the senses* image-ing of the thing is the same). Therefore
when put side by side and compared, being confusable and alike they
interchange. This is why they are given a conventional name in common
and so determine each other.**
8 8
There is also a more specific word for 'image*, hsiang ffi ( = S ) ,
also used both verbally (B 38, 'imagine*) and nominally (B 57 'image'). In
both cases hsiang is used with yi, in the latter in the combination yi hsiang
'idea or image*, which appears in Han Fei tzu used verbally:
Ch. 20 (Ch*en 368) A * J i £ * - t b » f f n l # ^ * ^ # ' & * B U « X £ m ° tt
"Men seldom see a live elephant, but when they find a dead elephant's
bones they resort to its picture to imagine it alive. Therefore everything
which men use to form an idea or image is called a hsiang (elephant/image)."
For the preoccupation of Mohists and others with the danger of ideas
being confused by their verbal formulations (tz'u 8$), cf. § 1/4/31. Accord
ing to N O 11 one "uses sentences to dredge ideas" (Ei%$¥fM). We find the
phrase shu yi 'dredge ideas' also in the programme of disputation quoted
on p. 20 as well as in Han shu ch. 36, 1932/1 -HfiSlfc "dredge out once
and for all my foolish idea". The choice of the rather rare word shu is
striking; although the phrase may later have become a cliche for expressing
one's ideas it must have begun as a living metaphor. The word shu f?, ff
'bale out/clear away' is used of cleaning a well or latrine: Kuan-tzu ch. 53,
3/9/2 (of spring cleaning) t f #l?tfk "clean out the wells and change the
water": Mo-tzu ch. 70 (Sun 379/10) ^ f f l l "order to clean out the
8 8
M . 23151 def. 18.
latrines". (Cf. also Huai-nan-tzii ch. 16 (Liu 16, 11B/-1) ifM "clean out
the latrines".) The only other example in the concordanced pre-Han texts
is Tso-chuan Wen 6/7 JS£i&ff^ "the difficulties will certainly be cleared
away". The point of "One uses sentences to dredge ideas" is presumably
that one should use them to detach ideas from all confusing accretions and
present them in their pure state. The phrase shu ch'ingffr/ifWi (M.11884/6,
14513/6), first attested in Chiu chang ft* N O 1 (CWu tz'u SPPY 4,
2A/2), may similarly be understood as 'dredge out one's true feelings'.
The use of yi for aims or intentions, common elsewhere in Mo-tzu,
is not found in the dialectical chapters.
1/4/38 Y i *
Yi % 'substitute, exchange' (A 45, 48, 85. B 23, 29, 30, 57)
The Mohists are rigorous in observing the distinctions between hua ft
9
'transform' ('change into , A 45, 85), pien £ 'alter' ('chang to\ B 7, 29, 30)
and yi 'substitute, exchange' ('change /or'). Cf. A 45 ["ft J * ffcSr-tfa
" 'Transformation' is the distinguishing characteristics being replaced",
B 29 %WM%k%3 "Without it having altered the name is replaced".
In two cases yi has the unusual sense of 'change round\ referring to
rotation (A 48) and to the inversion of the mirror image (B 23). The point
is presumably that the two sides exchange places.
1/4/39 Y I N H
Yin H : verb, 'depend on, have as grounds' (EC 1, N O 9). Adverbial, 'by
this criterion' (NO 2, 17). Nominal, 'criterion' (A 71, 97. B 3, 15)
Except for the single example in the earliest of the documents,
Expounding the canons (EC l) yin has a consistent technical use contrasting
y
with fa 'standard'. A man is judged to be black or to love mankind by

comparison with something taken as standard. But his agreement with the
standard may be only partial; we have to decide whether to take as criterion
the parts which are black or those which are not, the men he loves or the
men he does not love:
A 96 mm· IQKRS: . . . ° \>x\±^m^m^\t r MA J »m&M
ÂW^SÂ^jiirgAJ >SfAl>
(Canon) "Where the standard differs, observe which consideration is
appropriate. (Explanation) . . . Using what is black in a man or what is
not black to fix 'black man', using love of some men or lack of love for
other men to fix 'love of men', of these which is appropriate?"
1/4 215
A 97 i t H . . . ° & « & * J a J S j l t K & f I i > fl9»**fcfifnlKI£ °

(Canon) "Fix the criterion. . . . (Explanation) If the other man, referring
to a case where it is so, judges it to be so of the instance here, refer to a
case where it is not so and ask about it." (Cf. N O 17 ^JSS:, S J i ^ S A
"He does not love without exception, and by this criterion is judged not
to love men".)
B3 rejfir*n>j . . . HH*£»sTOiss-tk»tafta°
" 'Is white* and 'is blind* . . . we give up saying one or the other, but the
thing really is what we called it. Explained by: the criterion.**
(cf. No is ±m±&np» m*m±m*№... ° . mm±*m °
"If this horse's eyes are blind we say that this horse is blind. . . . If this
ox*s hairs are yellow we say that this ox is yellow.'*)
By collating these passages we can identify several examples of such
criteria:
"He does not love mankind" . . . one unloved man.
"This horse is blind'* . . . the eyes.
"This ox is yellow" . . . the hair.
Since it appears that the Mohists used yin in a highly technical sense
contrasted with fa 'standard*, and even considered the term explicit enough
to sum up the whole argument of an Explanation (B 3,15 IftffiH "Explained
by: the criterion*'), it may occur to us to wonder why they did not include
it with fa among their defined terms. Now it happens that the definition
of fa is in fact coupled with the definition of an obscure term never used
anywhere in the dialectical chapters (although another word, separately
defined, is written with the same graph № in A 15):
A 70, 71 & » rrSffn^* ° m » 0f&-tb o
"The 'standard* is that which if it resembles it is so. The 'erh' is where it
is so.**
There is a very strong presumption that erh is a corruption of a term
common in the dialectical chapters, since nearly all defined words belong
either to the ordinary terminology of early Chinese thought or are technical
terms actually used in the Canons and Explanations. The definition is
exactly that of yin; a thing is judged to be black by a fa ('that which if it
resembles it is so*, something black taken as a standard for comparison) and
by ayin ('where it is so*, the black part of it). But how could such dissimilar
graphs be confused? We have noticed the Mohist practice of marking
words adapted to a technical use by the 'man* radical, with the result that
later scribes often either dropped the radical or mistook the graph (cf.
§ 1/2/1/2/3). We may guess that the Mohists distinguished yin 'criterion* by
a graph ffl peculiar to themselves, which in one of itsti-skuforms would

89
be 1H and that the scribes responsible for the graphic standardisation of
Mo-tzu generally omitted the radical but in A 71 thought they recognised
the current graph \%.
1/4/40 Y u / W u S H U O W/fciift
Yujwu shuo W / &£t& 'have grounds to offer/have no grounds to offer'
(A 99, B 32)
Shuo 'explanation' is defined in A 72 as 'the means by which one makes
plain' (§fK№ffe). It may refer only to clarification of a statement, but is
often used of offering proof (NO 10 KtSftftSC 'by means of explanations
bring out reasons', cf. A 79, B 1, 2, 66, 82). This is the case in the formula
yu\wu shuo:
Chuang-tzu ch. 13 (Kuo 491/1) "I »ftEi&BQft°
" I f you can justify what you say, well and good; i f not, you shall die."
8 9
Ku Ai-chi 1/59B, 60A, 77B.
1/5
THE STOCK EXAMPLES
1/5/1 INTRODUCTION
T H E use of stock examples, many of them animal (such as the ox,

horse and dog) is very obvious in Mohist disputation. These illustrations
deserve a more systematic examination than they have yet received. The
dialectical chapters are full of unexplained references to milu deer, pillars,
stones, walls, circles, which editors have often tried to expunge by arbitrary
1
textual emendation. But closer examination suggests that each illustrates
a specific point which can be identified, with varying degrees of confidence,
after collating instances in this and other texts.
Since the Mohist definitions are all of abstract concepts, none of the
stock examples is defined except for the square and circle. The earliest
definitions we know are those of the Han dictionaries; it is possible that
the Shuo wen t & X (completed A.D. 100) and the Shih ming (c. A.D. 200)
may owe something directly or indirectly to the writings of the School of
Names (which were still extant), although I have noticed no unmistakable
borrowings from the Canons. The Shuo wen entries are too brief and loose
to be useful, but there are definitions in the Shih ming which are more
informative, in spite of the artificial etymologies on which they are based.
Unfortunately dictionary-makers do not bother to develop full definitions
of such familiar things as the ox and the horse.
1/5/2 Ox A N D HORSE
In discussions of whether or not a name fits an object the standard

examples, here as elsewhere in pre-Han disputation, are the ox and the
horse. The two frequently appear together (A 50, B 12, 35, 66, 67, N O 18).
When only one appears, 'ox' illustrates the principle that an object either
is X or is not (A 73, 74, cf. B 67), while 'horse' is the typical class name
(A 78), applying to different objects of the same kind which may be
1
Cf. Sun 212/6, 8, -1 : 220/-2 : 221/8, 9.
218 The Stock Examples
distinguished by such phrases as 'white horse', 'horse from Ch'in', 'two

horses' (B 3, E C 1, N O 2, 18. But cf. A 50 " 'horse' and 'non-horse' ").
Elsewhere the most familiar example of 'horse' as the typical class
name is Kung-sun Lung's sophism that " A white horse is not a horse".
In the Lii-shih ch'un-ch'iu the typical instance of misusing names is to call
2
an ox a horse and a horse an ox. We also find these and other animal
illustrations of the sophists in Chuang-tzu, where their significance may
sometimes be missed: " I f we use this (the Tao) for disputation, and
compare it with a dog or a horse, we shall fall a long way short of it",
" A t one time he deemed himself a horse, at another deemed himself an
3
ox". The latter sentence means " A t one time he deemed himself the
logicians' ' X ' , at another the logicians' ' Y ' ", as can be seen from another
passage which refers explicitly to names and objects:
Ch. 13 (Kuo 482/-1, 483/1) ^ > ^nfSMMbiffl T i i F J ^ J » * M M i f n
r m±m j o i s w j t i t · A S i * f f l i n i s » n ^ m °
"Formerly, when you declared me an ox I 'called it an ox', when you
declared me a horse I 'called it a horse'. If there is the object, to refuse the
name which someone gives to it is to make twice as much trouble for
yourself."
The phrase 'call it an ox' appears also in the Explanation of the
definition of 'disputation' (A 74); presumably it was an accepted formula,
and this is why Chuang-tzu switches so abruptly from the 1st to the
3rd person.
1/5/3 D O G
The dog is the stock example of an object with two names, ch'iian
X and kou $J (A 74, 79. B 8, 35, 39, 40, 54), and therefore of ch'ung B
'duplication/identity' (B 40, cf. A 86, 87).
The older name ch'iian had remained standard down to the 4th
century, but in Chuang-tzti (c. 300 B . C . and later) kou is already the com
moner word. We can observe this development inside the Mohist corpus
itself; a dog is nearly always called ch 'iian in the older parts of the book
(ch. 1-39) but kou in the later (ch. 40-71). But it appears that kou (originally
4
'puppy') was for some time narrower in scope than ch'iian. In its sporadic
appearances in 4th century texts (Tso-chuan, Mencius) kou refers to the
watch-dogs and eating-dogs of the people; the hunting hounds of the
2
Lu-shih ch'un-ch'iu ch. 17/1 (Hsü 17/3B/-1).
3
Chuang-tzü ch. 7, 25 (Kuo 287/-2, 913/-1).
4
M . 20345.
1/5 219
y
nobles are still called by the more general ch iian. In the dialectical chapters
two points deserve notice:
(1) When the two names are not being contrasted the name is kou
(A 79, B 8), and it is implied that not everyone knows the name ch'iian
(B 40, cf. 39).
4
(2) Nevertheless it is said that 'a kou is a ch'iian" (B 54), not the
reverse, and it is once implied that the synonymity does not apply to all
cases (B 40).
It would seem that for the later Mohists (who are not aristocrats who
hunt with hounds) a dog is primarily a household dog and is called kou.
5
(There is in fact only one reference in Mo-tzu to the hounds of the nobles.)
They recognise however that the name applies only to household dogs and
that the more general name (not necessarily known to the uneducated) is
ch'uan. In translation 'dog' is the only suitable equivalent of ch'iian, and
also of kou when it stands alone; when they are contrasted we shall use
'whelp' for kou.
The example of the dog was thus a vivid proof both that an object
has no fixed name and that names for a single object are not necessarily
synonymous. It was used by sophists outside the Mohist school, as may
be seen from two of the sophisms listed in the last chapter of Chuang-tzu:
" A whelp is not a dog" (#0?fcfc: rejected by the Mohists, B 54) and " A dog
might be judged to be a sheep" ( A " J J i l S ^ ) .
1/5/4 CRANE
The mysterious graph St (A 88, B 6, 8, 72, cf. B 53) was emended

by Sun Yi-jang, rather arbitrarily, to j £ hu 'tiger' (Sun 224/7). But he
admitted the possibility of an older suggestion that it should be f t hao
6 7
'crane'. Since L i u Shih-p'ei S0Ei№' noticed that in its single occurrence
8 9
elsewhere in Mo-tzu hao is written with the graphic variant i t nearly all
editors have preferred the latter suggestion (among others, T'an, Kao, W u ,
L i u Ts'un-yan IPPfffcl). With this as with so many other graphs the
dialectical chapters are peculiar only in that later scribes failed to supply
the radical. The clearest example is in B 6 (on the absurdity of comparing
things of different classes): "Which is higher, a deer or a crane?".
5
Mo-tzu ch. 20 (Sun 104/6).
6
Chang Hui-yen, B, 3A.
7
Liu Shih-p'ei B, 7B.
8
Mo-tzu ch. 19 (Sun 97/9).
9
M . 47470.
Borrowing the name 'crane' for a dog is the stock example of chia $i
'loan-naming* (B 8, cf. A 88, B 72). The words used for a dog's name are
shih R (B 8) and hsing #4 (A 88). One would expect a dog's name to be
simply its ming but the Mohists do not use ming for the personal name
(A 78 "For example, surname and style-name"), no doubt to
avoid confusion with the names of objects. The legendary Han L u , the
swiftest hound in ancient China, appears in various texts with his name in
such forms as ttft^Jt,
The dog in any case attracted attention, as we noticed in the previous
section, as an animal without a single fixed name; such a sophism as
" A dog might be judged to be a sheep" would win further support from
the fact that a dog may be called (and answer to) the name of another
animal. The Mohists go to some trouble to show the difference between
calling a dog 'Crane' and calling dogs cranes (B 72).
1/5/5 M I L U DEER
The milu or Pere David's deer (mi SI) is distinguished from other
deer by the Chinese as by modern zoologists, who exclude it from the genus
cervus as the elaphurus davidianus. It is remarkable that while other pre-Ha
texts refer to deer either as lu 8S or as lu mi 'cervi and elaphuri' the
dialectical chapters use only the word mi (B 2, 6, 11, 45). This suggests that
the animal is another member of the Mohist menagerie of logical illustra
tions. In B 2 (on the difficulties of extending from one object to others of
the same kind, t'ui lei tfeiS) the milu deer is introduced abruptly in a way
which defies explanation unless it had some significance which readers were
assumed to know and which is nowhere explicitly mentioned in the
dialectical chapters. It is fortunate for us that the Shuo shan chapter of
Huai-nan-tzii (which also explains the significance of the louse for dialec
ticians) has a section on t'ui lei which twice borrows from other parts of
11
the Mohist dialectical chapters and mentions the popular belief that the
cavities under the eyes of the milu deer contain additional eyes with which
12
it sees at night.
Huai-nan-tzii ch. 16 ( L i u 16, 19B/4) ^m^MM^m » MHffnÊ3a °
"If a pregnant woman sees a hare her child will be hare-lipped, if she sees
a milu deer the child will have four eyes."
10
Shuo yuan 11, 4B/7. Chan-kuo-ts'e ifcglSg BSS 1, 89/2: 1/96/3.
1 1
Huai-nan-tzii ch. 16 (Liu 16, 7A/7-9, 19B/5), from EC 8, NO 18.
1 2
M . 47625 def. 1. Schafer 253.
1/5 221
Here Huai-nan-tzu has passed from observations about argument from

analogy borrowed from the dialectical chapters to theories about the
interactions of things of the same kind which are very alien to the Mohists.
But we may infer that the peculiarity of the milu deer for dialecticians was
that it has four eyes. The Mohist reference (B 2) is in fact in a context
dealing with 'sets of four' (ssu 85) and animals having four legs. It becomes
intelligible on the assumption that the milu deer is the stock example of a
thing with a unique differentia; a thing with four legs may be any animal,
a thing with four eyes is necessarily a milu deer.
There is another mysterious reference in B 45, "Like the wounded
milu deer lacking a haunch", on the Canon "It is no objection that some
is lost". There the point may be that what something is judged to be is
unaffected by loss of a part; animals have four legs, but the milu deer which
has lost a leg is still an animal, since its four eyes distinguish it as a milu
deer and a milu deer is an animal.
1/5/6 LOUSE
While the dog is the stock example of an object with two names the
louse (shihI*SlzT S , IS) is the example of an object with the same name
as another object. This is stated in the Shuo lin chapter of Huai-nan-tzti,
which frequently echoes the Mohist dialectical chapters:
ch. 17(Liu 1 7 , 1 2 A / - 4 ) s t r 8 J r ^ t j
9
zmnmiko
"The *SIeT on the head (louse) and the *SIeT of hollow wood (zither)
are an example of the name being the same but the objects different."
This illustration would be especially apt because of the rarity of the
sound; Karlgren's Grammata serica records no other graph with this
reading. In the dialectical chapters it appears in B 6, in a series of illustra
tions of the absurdity of comparing things of different kinds:
. . . »)#33ft9 o
"Which is longer, a piece of wood or a night? . . . Which is more *SIsT,
a *SIeT (louse) or a *SIsT (zither)?"
The graph &], found here and in A 88, is otherwise unknown, but
can hardly be anything but a mistake for some form of the graph 2*. That
it had the same reading is confirmed by the mysterious phrase written
in A 49 and 3Si3J in A 88, in both contexts having something to do with
movement. This is one of the phrases for which nearly every editor offers
a new conjecture, and most do not even offer an explanation to cover both
contexts. But in spite of Sun Yi-jang's objection that the two graphs must
represent different words since they are distinguished in B 6, there is a
13
very strong presumption that if the graphs are not actually interchanged,
one must be a mistake for the other. The most promising course is to
consider whether the fact that it has the same name as a zither exhausts the
possibilities of the louse as a stock example.
The louse has a further significance, as the obvious example of a
creature which travels with its host:
9
Chuang-tzii ch. 24 (Kuo 863/4, 5) ^ f i £ f t <> . . . im®M jtt£ltt£ °
"The louse on a pig is an example of this. . . . Here it advances with its
surroundings, here retires with its surroundings."
In the puzzling phrase JfeCO the identification of the second graph as
shih 'louse* suggests as the next step the addition of a stroke to the first,
making a phrase t'u shih 'louse on a hare*, like the chu shih 'louse on a pig*
of Chuang-tzii :
A88(3a*3j)*3i»fliaH(=iBtt)»r*sfcjife°
(In a series of opposites either of which is true according to standpoint.)
" A louse on a hare moving this way and that is 'both leaving and
approaching*."
9 9
A 49 ffi«( = ffll£)*i(fc# F№ ( & S I ) * & * S i °
"Things which shift everywhere on the border: the hinge of a door, the
louse on a hare." (Two kinds of motion without change on either side of
a constant border.)
Since the Mohists use the example to illustrate kinds of motion, the
choice as host of the hare (for the Chinese, as for us, an especially swift
animal), seems inherently likely.
1/5/7 STONE
When different properties of a thing are considered the stock example

is a stone, which may be both hard and white (B 37), both white and big
(NO 1). According to B 38 one of the two properties hardness and whiteness
is "referred to 'a priori* ** (hsien chii 5fc$). This would be the hardness,
which is contained in the definition of the name, as we find it in a Han
dictionary:
Shih ming ( T S C C ) 12/3 OifllB T 5 J ° T E J tt(= ° SffJfrfe °
"The substance of a mountain is called 'stone*. 'Stone* (shih) is lo; it is
hard and kan-lo (resistant).** (The definition in the Shuo wen ch. 9B
(Tuan 453A/-3) is simply T S J lilSife " 'Stone* is the stone of mountains**.)
1 3
As in the proper name written in Chan-kuo ts'e (Han 2) BSS 3, 42/2
and in Shih-chi (ch. 45) 1873/1. The interchange is rejected in Sun 220/1.
1/5 223
14
The Mohist term closest to our 'properties' is tnao H i 'features';
and when it is explained how a name presents us with a new object by
conveying that its mao is like that of already known objects the illustration
is a stone (A 31, 32) and the quality is the colour, black or white (B 70).
A reason for the choice of stone would be that (unlike an ox or a horse,
for example) it has no fixed shape, and is therefore better conveyed by a
description of its qualities than by a picture. Stone does in fact appear with
wood as a material which can be given any shape ( ^ S 'square stone',
'square piece of wood', B 29, 65, N O 1).
1/5/8 WALL
The significance of a wall (whether of house, grounds or city) is that

we cannot know by direct observation what lies on the other side of it. We
find this image everywhere in pre-Han literature, with the observer
sometimes inside and sometimes outside the wall:
Analects 19/23. "Compare it to having a wall round you (^HSf). M y wall
is shoulder-high, you can peep over and see the prettiness of the house.
M y Master's wall is many yards high; if you do not find the way in through
the gate you will not see the beauty of the ancestral shrine and the richness
of the buildings."
Lao-tzu 47. "Without going out of the door he knows the world, without
peeping through the window he sees the Way of Heaven."
Hsun-tzu ch. 12 (Liang 171/8). "Things beyond the wall the eye
does not see, news outside the village the ear does not hear."
In the 1st century A.D. Wang Ch'ung (A.D. 27-C. 100) develops this
image when deriding claims that the sage has knowledge which is indepen
dent of experience:
Lun-heng ch. 78 (Huang 1071/6, 7) A&KflSSC » » ttlBAR
"Suppose a man stands east of a wall and we tell him to give a shout, and
suppose the sage listens to it from west of the wall; is he able to know
whether the man is black or white, short or tall, his district, village,
surname, style-name, and who his ancestors were?"
In the same context Wang Ch'ung frequently refers to the fore
knowledge which he denies as hsien chih 5fe^P 'knowing beforehand'. The
Mohists use the same phrase in a more sophisticated sense, 'know "a
1 5
priori" ', the opposite of wei k'o chih 'not knowable "a priori" ' :
1 4
Cf. § 1/4/20.
15
Cf. §1/4/13.
B 57 5fel?,*ffiife o B 58 ffi(= ·
(On the pillar) "We know the idea or image 'a priori'." (On the hammer)
"The idea of the hammer is not knowable 'a priori'."
Both terms are found together with references to walls; presumably
the 'a priori' is conceived as what one knows about an object imagined as
hidden behind a wall, simply by considering what is implied by its name:
A 75 m^zmm^KitiL ° m±w^n°...
"Whether there is benefit or harm beyond the wall is not knowable
'a priori'. If by hurrying for it you get the money. . . . "
A 93 tttt A i b l b · ffi^ffi* · "I ·
"When we go over the city-wall the circle 'stays' (is confined to certain
objects). By the things which follow from each other or exclude each other,
we know 'a priori' what it is." (Cf. also the fragmentary A 14 < S > Jftflhfe
" < ? go over ? > the city-wall and get the money".)
Similarly in E C 2, which contrasts the eternal moral principles
derived from what the sage "desires or dislikes 'a priori' for the sake of
men" (5feHA$^?§) with the concrete benefit of particular individuals
which varies from day to day, it is observed that "Yesterday's wall to the
intelligence (£0%) is not today's wall to the intelligence".
1/5/9 PILLAR
Two of the three passages about the pillar mention the 'idea' of it
(yi M), and imply that the idea is a mental picture (B 57, N O 3). The
third says that two men "both see (chien H ) that this is a pillar" (A 39),
presumably because they share the idea. According to B 57 when one sees
(chien) a round pillar the roundness is external to the idea but does not
alter it. The idea of a pillar is "known 'a priori' " but the idea of a hammer
is not (B 57, 58). It seems that the pillar is the stock example of something
of which the visualisable figure is known 'a priori' from the definition. In
commenting on B 57, 58 we shall offer evidence from Han dictionaries that
a pillar would be defined in terms of vertical weight-bearing without side
support (a concept developed in B 29); if so, the most efficient figure for
a pillar would be implicit in its definition in the same way that the figure
of a circle (also "known 'a priori' ", A 93) follows from its definition as
what "has the same lengths from a single centre" (A 58).
There is some reason to suspect that sophists may have used the pillar
to defend the claims of H u i Shih and Chuang-tzu that all distinctions are
illusory. B 58 shows that we have no 'a priori' idea of a hammer, because
we cannot visualise it in use without 'exceeding the matching' (kuo zvu
1/5 225
including the thing which it is hammering. It might well be argued

that even the idea of a pillar must include the ground it stands on and
perhaps the load above it. Chuang-tzii mentions among dissimilars made
one by the Tao "a stick and a pillar" (ch. 2 (Kuo 70/1) S U t t ) . "The
pillar has the ox" (SW^fc) appears in an obscure list of sophisms in
Hsiin-tzu's Right use of names (ch. 22 (Liang 316/6)). It may be worthwhile
to use our interpretations of stock examples to make a guess as to its
significance. Hui Shih's method according to the last chapter of Chuang-tzu
was to "tabulate the ideas of things" (M%£-M), such as the Greatest One
and the Smallest One, and prove that they lead to paradoxes. The stock
example of an 'a priori' idea is that of the pillar. Disputation classes things
as ox or non-ox, and a pillar is non-ox. "The pillar has the ox" would then
be equivalent to saying that every idea includes the idea of everything else,
that all distinctions are illusory.
The term yi 'idea' is examined in § 1/4/37.
1/5/10 T H E CIRCULAR A N D T H E SQUARE
Being circular (A 70, 93, 98. N O 8) and being square (A 80, B 65,
N O 1) are the representative examples of respects in which an object may
be judged to be 'so* (jan #&) by comparison with a standard (fa £fe). Both
are defined (A 58, 59), and in the case of the circle the Mohist takes the
same care as with ethical terms to define the terms used in the definition
(cf. § 1/1/2/6). The clarity and simplicity of geometrical standards makes
them especially suitable as stock examples. What it is to be circular may
be known 'a priori' (A 94); more than one standard is available, idea,
compasses, another circle (A 70); and agreement with standards is exact
(A 98). That a square thing will not rotate is the typical example of some
thing known by demonstration (shuo fS£, A 80).
These examples are already used in the older chapters of Mo-tzU:
Ch. 27 (Sun 133/2-4) ^ 5 f c t t A $ f t * S · mi*m&XT 2MS%fimtik
:
· T+
· *+5№»ai±?OT J o > tf^WfffcHb ° it
"Now the wheelwright takes up his compasses to estimate what in the world
is circular and what is not, and says 'What coincides with my compasses I
say is circular, what does not I say is not*. In this way both what is circular
and what is not he can get to know. Why is this ? Because the standard for
the circular is plain." (A parallel passage about the square follows.)
It may be noticed from this example that yuan H , I t and fang are
negated by the pre-verbal pu not by the pre-nominal fei Both are
primarily stative verbs, 'is circular', 'is square', and belong to the sphere
of what is so (jan ?&) of an object, not what it is (shih A).
YuanWl, H 'circular', which is two-dimensional, is to be distinguished
from the three-dimensional t'uan (itt, ffl), 'spherical, cylindrical' (B 24,
57, 62). On the other hand fang may be used of cubes as well as squares
(NO 1).
1/5/11 ILLNESS
Illness is the stock example of an event which may have more than
one possible cause. It appears explicitly in B 9, where the cause is a wound,
and implicitly in A 77 ( M , ife-ife "Dampness is a cause") and B 10 (drinking
to excess, the heat of the sun). References to dampness as a cause of illness
are common in pre-Han texts:
Chuang-tzu ch. 2 (Kuo 93/1) S l f f l B S S ^ °
"When people sleep in the damp their waists pain them and part of them
goes numb." (Cf. also Hsiin-tzu ch. 21, Liang 304/5: Han Fei tzii ch. 2,
Ch'en 37/2.)
The most striking development of this illustration is at the beginning
of the first of the three chapters on Universal love in Mo-tzu:
Ch. 14 (Sun 65/1—3) "The sage is the man who takes the government of
the Empire as his task. He can govern it only if he knows the sources of
disorder (fiL^lSf US), cannot if he is ignorant of the sources. He may be
compared to a physician curing a man's sickness, who can cure it only if
he knows the source of the sickness, cannot if he is ignorant of the source."
1/5/12 T S A N G 8$ A N D HUOffi?('JACK A N D JlLL')
Tsang is the typical proper name (A 78, cf. A 83, B 53, E C 4, 6, 7,13);
where two are required the second is Huo (EC 2, N O 14). Where a woman
is implied the name is Huo ( N O 15). For convenience we choose the
equivalents 'Jack' and 'Jill'.
1 6
Tsang and Huo were abusive names for bondsman and bondswoman ;
the basic meanings were presumably 'hideaway' and 'captive'. They are
used in other pre-Han texts for 'any Tom, Dick or Harry':
Han Fei tzu ch. 21 (Ch'en 407/5, cf. 407/4, 1085/4, 1093/3) E * — A # · ffJ
№M9№» °
"Using only one man's strength, even Hou Chi would lack; if they follow
the course of nature, even Jack and Jill have more than enough."
16
Fang yen SPTK 3/1B.
1/5 227
In the dialectical chapters one reference to Tsang implies that he is

in fact a runaway slave (B 53), and the references to loving Tsang and Huo
in discussions of universal love seem always to assume that they lack any
qualities deserving of love except their humanity.
'Tsang* as the typical proper name appears also in a sophism ascribed
17
to Kung-sun Lung: "Jack has three ears".
1/5/13 N O R T H A N D SOUTH
The characteristic direction of motion is southward, as for example in

Hsun-tzu ch. 22 (Liang 323/8f) №ZfiAW&l%№& "Suppose there
is a man who wants to go south no matter how far. . . ." The conventional
way of saying that space is infinite is "The south has no limit" (cf. B 73).
This point is important to the understanding of B 33, where the direction
of movement is not explicitly stated.
1/5/14 T H E ROAD
The word tao M 'way, road*, appears only in A 97, N O 11. It is a

stock example, or rather a metaphor, central to the art of description. By
applying a name to one object I am obliged to apply it to all similar objects;
I have established a road by which I proceed (hsing fi) to all objects of
the same kind (lei S§). The metaphor is fully developed in N O 11: "Now
a man cannot proceed without a road; even if he has strong thighs and
arms, if he is not clear about the road he will soon get into trouble. The
proposition is something which 'proceeds* according to the kind; i f in
setting up a proposition you are not clear about the kind you are certain
to get into trouble**. It is also implicit in A 97, B 1, which use all the words
'road*, 'man*, 'proceed*, 'kind*.
Richard Bodman has pointed out to me another word connected with
this metaphor. T o fail or refuse to describe an object as you have described
a similar object is to 'turn off course* (chuan ij$). Propositions "become
different as they proceed, become dangerous when they turn off course'*
(NO 12 ffffBJS, $|ffnJa). The Canons on the art of description similarly
warn us against the sophist whose "description takes a subtle turn** (A 94
tfrPifl, cf. A 95). In the original Teng Hsi tzu, a document of the School
of Names the preface of which survives in the extant work of that title, one
of the two chapters was called Chuan tz'u %$Wf 'Propositions which turn
off course*.
17
Lu-shih ch'un-ch'iu ch. 18/5 (Hsu 18,16B/5). (Cf. Hsu's note: the unintelligible
jj|(jEi5? is emended from the parallel in K'ung ts'ung-tzii 1, 76A/5-76B/4.)
Hsiin-tzu's summary of the art of description in the Right use of names

uses the same metaphor:
Hsiin-tzu ch. 22 (Liang 310/-4) tt3E#£ffl£ » £5gffnSJ§r · Mft\R\J&m ·
"Therefore when a true king institutes names, by the fixing of the names
objects are distinguished, by the roads being proceeded along the meanings
are universally intelligible.''
However Hsun-tzii, although he twice uses hsing in conjunction with
18
lei 'kind', thinks of the general applicability in a wider sense, embracing
the universal acceptance of the name:
Ch. 18 (Liang 250/9)g3E«fSr£ > ± * * B U t t B > WA&ITf ·
"The sage king establishes it as a standard, the high administrators as a

road to be followed, the lower officials as a practice to be observed, the
people as an accepted custom."
18
Hsün-tzü ch. 9, 12 (Liang 108/10, 158/4).
1/6
T H E ORGANISATION O F T H E 'CANONS'
A N D 'EXPLANATIONS'
1/6/1 T H E F I V E DIVISIONS OF T H E DEFINITIONS A N D T H E PROPOSITIONS
IT has always been apparent that the Canons tend to fall into sequences,
for example that A 1-75 consists of definitions and that B 17-29 deal with
optics and mechanics. Scholars since Sun Yi-jang have become more and
more aware of interrelations between adjacent sections. However, it is still
taken for granted that the order of the Canons is to a great extent random,
and that it is legitimate to guess at the theme of an obscure section without
considering its context. If this were so, our inquiries into the textual
history, grammar, technical terminology and stock illustrations would still
not altogether dispel the suspicion that the study of the Canons involves
too many imponderables for confident interpretation. In Chinese even
more than in other languages, a passage without context in a document
recognised to be corrupt gives too much scope for imaginative inter
pretation.
But in fact every Canon has a context which establishes its general
theme, except in the final series of problems in disputation (B 32-82),
many of which do exist in a contextual vacuum which allows us a dangerous
freedom to interpret as we please. Even in this series we can recognise
Canons which pair (B 44 and 45, 57 and 58, 63 and 64) or fall into sequences
(B 37-42, 73-75). The central dividing line is between the definitions
(A 1-87) and the propositions (A 88-B 82), the traditional division between
Parts A and B being, as we noticed in § 1/2/2/4 above, an arbitrary break
made probably at Stage 2. The two halves of the document are parallel in
their arrangement: we shall for the moment give only rough descriptions
of the content:
DEFINITIONS PROPOSITIONS
A 1-6 'Reason', 'unit', 'knowing' A 88-B 12 Procedures for consis
tent description
A 7-39 Conduct and government
230 The Organisation of the 'Canons' and 'Explanations'
A 40-51 Spatial and temporal con- B 13-16 Spatial and temporal con-
ditions of knowing ditions of knowing
A 52-69 Geometry B 17-31 Problems in optics, me
chanics and economics
A 70-75 Disputation B 32-82 Problems in disputation.
(A 76-87 Appendix: 12 ambiguous
words)
Between B 12 and 13 we should expect a series of ethical propositions
parallel with the definitions of A 7-39. Indeed one might have supposed,
since the interest of the Mohists is primarily in morals and government,
that this would be the most important sequence of all. But there is a simple
explanation of the omission. The Mohists have already dealt with the theme
in their first experiment in the canon/explanation form, Expounding the
canons. They do not go over the ground again, just as in A 1-75 they do not
repeat definitions from Expounding the canons or from the lost commentary
on the 10 theses of Mo-tzu (as we shall see in §1/6/2 below).
It has previously been assumed that we are not bound to take every
Canon in A 1-75 as a definition, and there did indeed appear to be one
obvious exception (A 66 MI&^fvffi^Mil, understood as "Hard and white
do not exclude each other", a proposition challenging Kung-sun Lung's
separation of hard and white). But now that this too has turned out to be
1
a definition, a geometrical definition like the rest of the series (A 52-69),
we need no longer hesitate to reject 'a priori* any interpretation of a Canon
in A 1-75 which treats it as a proposition. We can similarly lay it down as
a firm principle that all propositions in A 88-B 12 deal with logic, in a
discernible order which was partially obscured for earlier scholars by the
2
intrusion of a Names and objects fragment as A 89-92. From this we can
infer, for example, that B 5 ("One is unable, but it does no harm") refers
not to human limitations in general, as commonly supposed, but to the
"One cannot be dismissed without the other" of B 4. When T'an Chieh-fu
treats the obscure Canon B 11 as a proposition about mechanics we can
answer that the theme must be logic, unless the Canon has been transposed
from the sequence on mechanics (B 25-29).
The grouping of Canons is especially significant in the sequences to
which we give the heading 'Spatial and temporal conditions of knowing*
(A 40-51, B 13-16). Many of these have remained unintelligible because
they have been studied piecemeal, and make sense only when we appreciate
1
Cf. § 1/4/3.
2
Cf. p. 109 above.
1/6 231
that when the Mohist offers definitions of 'space* and 'duration', 'beginning*
and 'transformation*, declares that space moves and that the validity of
'Yao is good at ruling' depends on the time when it is said, defines
'necessary' as 'unending' and mysteriously introduces the logic of 'is an ox*
and 'is not a horse* in explaining a definition of 'staying', his basic concern
is always the same—how a proposition can have the validity for all times
and places claimed for the pronouncements of the ancient sages.
What are the principles of classification behind the five divisions?
Although we can hardly renounce the convenience of speaking of Mohist
logic, ethics, geometry, optics, these of course are Western categories
imposed by ourselves. It is especially important to know why the Mohist
has two categories which for us come broadly under the heading of logic;
he must be looking at logical problems from a different viewpoint which it
would help us to locate. That the logical theses of the first and last division
are different in nature is clear enough; A 88-B 12 is a close-knit sequence
laying down procedures for deciding what is so of objects; B 32-82 is a
series of miscellaneous propositions shown by logical analysis to be
admissible, self-refuting, consistent, unnecessary, inadmissible unless a
condition is fulfilled. But how does the Mohist see the distinction, which
should apply to the definitions of A 1-6 and A 70-75 as well ?
We may note in the first place that the third of the divisions, 'spatial
and temporal conditions of knowing', seems anomalous; it is the only one
which does not look as though it represents anything which could be classed
as a discipline or science. Now in B 10 four sources of doubt are classified,
3
which co-ordinate neatly with the first four divisions. The missing theme
is the fifth, disputation, precisely the one which does not admit of doubt.
The order follows that of the Canons, except that the anomalous third item
appears last, summed up as the question "Is it knowledge? Or is it suppos
ing the already ended to be so ?". This invites attention to A 80, which lists
three sources of knowledge (chih £P), report, explanation and experience,
and four objects of knowledge, names, objects, how to relate them, and
how to act (a pair of nouns, ming & and shih , and a pair of verbs, ho 'a
and wet Hi ; before the latter chih would be translatable as 'knowing how').
It seems inherently likely that the Canons would be organised according
to the four branches of knowledge; and of its three sources they are
concerned solely with explanation (shuo t £ ) . We may therefore try the
experiment of fitting the first and last pairs of divisions to the four
categories
3
Cf. the commentary on B 10 below.
Explaining names,
Explaining objects,
Explaining how to relate them,
Explaining how to act.
The five divisions now sort themselves out:
(1) The art of explaining how to relate names to objects: the pro
cedures for consistent description of changing phenomena laid down in
A 88-B 12.
(2) The art of explaining how to act: the procedures for considered
choice in changing situations already laid down in Expounding the canons.
(3) The problem of establishing unchanging principles behind the
changing. The definitions of this division conclude with the words chih i h
'staying* (A 50), the fitting of name to object for a limited duration, which
gives the judgments of the first two disciplines their temporary validity,
and pi 'necessity* (A 51), the unchanging validity of judgments in the
coming pair of disciplines.
(4) Examples of explaining objects, by establishing the causes of
physical phenomena.
(5) Examples of explaining names, by analysing the implications of
their definitions.
When we compare the order to that of A 80 we see that in both
arrangements the members fall into the same pairs, but in A 80 the order
is determined by convenience of listing (names, objects, how to relate them,
how to act), in the fivefold division by the project of advancing from the
temporarily valid to the eternally necessary.
There is however one place where the Canons do not quite fit this
classification. The difference between Divisions 1 and 5 is clear enough in
the propositions but remains nearly invisible in the definitions. But on
inspection it can be seen that the two disciplines that we would lump
together as 'logic* do share a common terminology, and the two words
which especially distinguish them, 'staying* and 'necessity*, are defined in
Division 3. What seems to have happened is that the Mohist took advantage
of Division 1 to dispose of terms fundamental to all four disciplines, such
as ku #C 'reason* and chih £fl 'know*.
The fourfold classification reappears in the account of the purpose of
disputation in N O 6, consisting of three rhymed sentences in each of which
the crucial phrase comes first:
("Explaining names**, the art of deciding whether X is or is not an
ox or a horse) §J3:)ik2fc£# "clarify the portions of 'is-this* and 'is-not* **.
1/6 233
("Explaining how to relate names to objects", the art of naming the

similar similarly and the different differently) K I R I J R ^ ] ^ "clarify points
of sameness and difference".
("Explaining how to act", the art of distinguishing the beneficial from
the harmful) AMIS "settle the beneficial and the harmful".
These follow the order of A 80 (names, objects, how to relate them,
how to act). "Explaining objects", which is outside the scope of disputation,
is relegated to a fourth sentence marked off by the conjunction yen M
'only then* and outside the rhyme scheme, beginning with the phrase:
"by describing summarise what is so of the myriad things".
It happens that we have a lucid although simplified digest of the four
disciplines adapted for Confucian purposes in Hsiin-tzu's Right use of
names. This is very probably based directly on the Mohist summa and on
Names and objects,* although the loss of nearly all the rest of the literature
of disputation makes it impossible to be certain. The Right use of names
begins like the Canons with a series of definitions, and proceeds with three
sections on:
(1) "The purpose of having names" ( §r SI'S* 3=1), which is to com
municate the similarities and differences between objects. " I f noble and
base are not clarified, if the similar and the different are not distinguished,
we are inevitably in trouble with the intention not being conveyed and
9 9
action being hampered and frustrated" ( J » » ^ ? 8 P I U ^ B f l tO&№&&
9
M9>&%№№±№).
(2) "What we depend on to recognise similarity and difference"
( @ f ^ £ i t ^ ^ ) , which is observation of objects by means of the five senses
in conjunction with the mind.
(3) "The pivotal requirements for instituting names" ( f t 0 & £ l t i l £ ) :
the compounding of names when a single one is insufficient, classification
of names at different levels of generality, agreed conventions (yiieh $J) for
the use of each name.
These correspond to three of the four Mohist objects of knowledge,
and follow the same order as the Canons:
Knowing how to relate names to objects,
Knowing objects,
Knowing names.
After general reflections on the uses and limitations of the study of
names and objects Hsiin-tzu concludes with a section on the weighing of
desires. At first sight this seems to have nothing to do with the rest of the
4
Cf. pp. 63, 64 above.
essay; it is only after referring to the Mohist scheme that we see that
Hsiin-tzu is winding up the essay with the one remaining discipline,
Knowing how to act.
The theme is in fact the grounds of considered choice, discussed in much
the same terms as in E C 7-9.
Hsun-tzu lists without explanation examples of faulty argument
belonging to his first three disciplines. (Fortunately we have his detailed
refutations of two of them at the end of his Cheng lun jEfra.)
9
Hsun-tzu ch. 22 (Liang 315/-4) T S f f t ^ J ["Â^gSJ »r&
" ' T o be insulted is not disgraceful', 'The sage does not love himself,
'Killing robbers is not killing people', these are cases of disordering names
by confusion in the use of names. If you test them by the purpose of having
names, and observe which alternative applies generally, you can forbid
them."
Why does Hsiin-tzu think that to refute these claims it is sufficient
to appeal to the purpose of having names ? Because the objection to them
is not that they are factually wrong but that they are confused as descriptions,
"disorder names by confusion in the use of names". The purpose of having
names is to convey similarities and differences, but " T o be insulted is not
disgraceful" obscures the distinction between 'moral disgrace' (114?) and
5
'social disgrace' ( I N ? ) , as we are told in the Cheng lun. The term hsing f f
6
'proceed' belongs to the Mohist art of description, although we here
translate it 'apply generally', since it is not clear that Hsiin-tzu uses it in
7
the strict Mohist sense. "Killing robbers is not killing people" is defended
by the Mohist as consistent description in N O 15.
(Liang 316/1-6) r i l r # ! № J r f f i $ C S J · r PMÎnM A J l ^ f J o ^ J
9
9 9
" 'Mountains are level with abysses', 'The essential desires are few',
'Fine dishes do not improve the taste, the great bell does not improve the
music', these are cases of disordering names by confusion in the use of
objects. If you test them by what one depends on to recognise similarity
and difference, and observe which alternative accords, you can forbid
them."
In these cases the description is wrong because the facts are wrong,
which is to "disorder names by confusion in the use of objects". We refute
5
Hsun-tzu ch. 18 (Liang 249/1—250/-4).
6
Cf.§l/4/4.
7
Cf. p. 228 above.
1/6 235
them by appealing to the evidence of observation by the five senses. The

Cheng lun criticises "The essential desires are few" on factual grounds, that
8
the desires essential to man can be seen to be many.
(Liang 316/6-317/7) r^rfnUJ · ["1&#^J » rS^JR-feJ »№MtkHi£
" 'You introduce yourself by what is not your name' (?), 'The pillar has the
ox* (?), ' A horse is not a horse', these are cases of disordering objects by
confusion in the use of names. If you test them by the convention for the
name, and use what one accepts to show that what one rejects is fallacious,
you can forbid them." (7V# 'refuse* is the opposite of shou 'accept*. Cf.
ch. 21 (Liang 297/1) " I f it thinks them right it accepts
them, if it thinks them wrong it refuses them**.)
This type of proposition, unlike the first, is mistaken in fact; unlike
the second it derives not from bad observation but from bad logic. By
analogy with the other two, we should expect this type to be tested by
"the pivotal requirements for instituting names**. But only one of these
requirements is relevant, the "convention for the name'*. You refute
Kung-sun Lung*s " A white horse is not a horse" (the only intelligible
example) by the fourth of the Mohist arts, disputation proper, which is
strict deduction starting from the definitions of names.
1/6/2 T H E L O S T DEFINITIONS OF WORDS I N T H E T E N THESES OF M O - T Z U
Among the 75 definitions which take up most of the first part of the
Canons there are certain omissions which deserve consideration. The most
surprising are the words chien ife 'collective* and at 5t 'love*, with which
Mo-tzu formulated his central doctrine of chien at 'universal love*. The
counterparts chien 'collective* and t'i I t 'individual* are common through
out the Canons and Explanations; almost at the beginning of the Canon
t'i is defined in terms of chien (A 2), although chien itself has not been
defined. Only a little later jen t 'benevolence* is defined as t'i at 'individual
love* (A 7), although both at and chien at remain undefined. The omission
becomes still more remarkable when it is noticed that the later Mohist
ethic is based on a compact system of definitions in which at is the only
missing member, a system so coherent that we can make a good guess as
9
to what the definition of at would be.
'Universal love* was one of the ten doctrines of Mo-tzii, formulated
in the two-word titles of the ten triads of chapters which are the core of
8
Hsiin-tzu ch. 18 (Liang 251/4-252/3).
• Cf. p. 48 above.
10
Mo-tzU. The sixteen words in the ten titles are never defined, although
some are of crucial importance in Chinese thought (t'ien % 'heaven',
wing fa 'destiny'), and there are others which are themselves used in
definitions (chih iS 'intent', A 20; ming ?fl 'be clear about/make clear',
A 6, 29, 30).
It seems safe to infer that the Canons presuppose a document which
defined the words used in formulating the early Mohist theses. This
document has disappeared, together no doubt with a great deal of Mohist
literature which we should have been grateful to possess. We may guess
that it would have done something to fill the gap between the relative
crudeness of early Mohist thinking and the logical sophistication exhibited
in the Canons. There is some evidence of this in the case of shang t'ung
'doing the same as those above', the second of the ten doctrines of
Mo-tzu. The section on ambiguities in the Canons includes an item
differentiating four senses of t'ung 'same' (A 86), which does not count as
a definition since the same section distinguishes seven meanings of chih %H
'know' (A 80), which has already been defined (A 5). There is however a
definition of a specially coined term t'ung № 'do the same' (A 39), which
seems to be precisely equivalent to the t'ung of shang t'ung. Presumably
the lost definition of t'ung 'same' was no longer adequate for the authors
of the Canons, who had reached the pitch of refinement of distinguishing
two words one of which had four senses. Later still Names and objects
increases the four senses of t'ung 'same' to eight ( N O 6).
We may notice similar gaps in a series of definitions in Names and
objects ( N O 5), but in this case fortunately the document that it presupposes
has not been lost. Huo 'some' is defined in terms of chin 8& 'all', hsiao §fc
'copy* in terms of fa ft 'standard', but the definitions of chin and fa which
seem necessary to complete the series are both in the Canons (A 43, 70).
No term defined in A 1-75 is anywhere re-defined in Names and objects.
1 0
Cf. pp. 3, 4 above.
Part II
T E X T AND TRANSLATION
2/1
EDITORIAL CONVENTIONS
(i) The following symbols are used:

*X X is an emended character. (Y)*X Read X for Y .
<X> Insert X . [X] Delete X .
tX X is a corrupt character. | X ] Y | Transpose X and Y .
(X) X is head character of
the Explanation.
(X) X is parenthetic and a
suspected gloss.
[ · • . A l ] Transpose to A 1. <A1 > Transposed from A 1 .
(2) The text is based on the Taoist Patrology edition, except that in
the one case of systematic corruption in the Taoist Patrology it consistently
follows the Mao edition in reading the phonetic Щ- for Щ- (p. 75 above). The
few other cases where a variant is preferred are indicated in the notes
(A53 n. 152, B56 n. 485, B73 n. 571, B76 n. 577).
(3) In the following cases of systematic corruption in all editions, the
corrupted character is omitted and the substituted character marked with
an asterisk, generally without further comment:
§ 1/2/1/2/13 X, £ § 1/2/1/3/4 К, Ж. fi.Ä
1/2/1/3/5 flijft 1/2/1/3/6 В, H
1/2/1/3/7 Я , in 1/2/1/3/8 А, А
1/2/1/3/9 fê.fé 1/2/1/3/10 ±, .Ù
1/2/1/3/11 в , m 1 /2/1/3/12?,^
1/2/1/3/131
Л 1/2/1/3/14 Ш, №
1/2/1/3/15 » , m 1/2/1/3/16 ft, Я
щ/щп га, m 1/2/1/3/18 Ш, Л
1/2/1/3/20 m, ш 1/2/1/3/21 Ä , Я
1/2/1/3/22 ·¥-, >f» 1/2/1/3/23 X , Tft
1/2/1/3/24 * M K 1/2/1/4/6 ff&.^l JßJ.^J
1/4/20 S, %
(4) In all other cases the traditional reading is given together with
the emendation, (Y)*X, except when the corruption is repeated in the same
section.
2/2
THE TEN THESES OF MO-TZU
AS C H A P T E R TITLES V A R I A N T S AS R E P O R T E D I N
MO-TZU CH. 49 (Sun 299/4-8)

Mo-tzu
ch. 8-10 (= - L ) K Elevating the worthy
n-13 nm Conforming to superiors
14-16 fkSt Universal love
17-19 Rejecting aggression
20-22 flfffi Economy in expenditure
23-25 if IF Economy in funerals
26-28 ^38 The will of Heaven 9-% Honouring Heaven
29-31 08«. Elucidating the spirits Serving the spirits
32-34 im Rejecting music
35-37 Rejecting destiny
The words in these theses were defined in a document now lost
(§ 1/6/2), and the definitions assumed throughout the surviving documents.
We list them for convenience of reference.
2/3
T H E FRAGMENTS OF
'EXPOUNDING T H E CANONS'
THERE are firm reasons for thinking that the mutilated remains of
Expounding the canons are the oldest component of the later Mohist summa
Its theme is ethics, the central preoccupation of this moralistic school, and
one would expect the compilers to dispose of ethics before proceeding to
such disciplines as logic and physics. This is confirmed by the absence of a
sequence on ethics among the propositions of the Canons (although among
the definitions it has the longest sequence of all), an omission which would
be inexplicable if the work had not already been done in the 13 propositions
1
of Expounding the canons. The Canons do not define two basic terms in
ethical reasoning, ch'iu # 'seek' and ch'iianfll'weigh/, yet use the former
to define lii M 'think* (A 4 ) and chih 7P 'order* (A 28) and the latter in
distinguishing the senses of yii 'desire* and wu SI 'dislike* (A 84); both
have been defined already in E C 7, 8. Above all, the consistent and
unmistakably deliberate grammatical restrictions observed throughout the
2.
rest of the corpus are not yet visible in Expounding the canons.
Like Names and objects and the chapters of the Analects and of part
of Chuang-tzii, the document takes its title from the opening words of the
text. In Chuang-tzu (although not in the Analects) the titles are self-sufficient
phrases, not necessarily read according to the syntax of the text; the chapter
called Ta sheng ft^k 'Fathoming life* begins with the words SS^kil
1re^f . . . "The man who fathoms what life truly is . . .". Expounding the
canons follows the same principle; its opening words are fn*
"Expounding as being a canon . . .**, and refer to one of the 10 theses of
Mo-tzu, the Will of Heaven. Evidently the 180 Canons of the later Mohist
corpus do not yet exist, and by ching 'canon* we are to understand authorita
tive statements of Mohist doctrine, like the ching 'classics* of Confucianism.
As H u Shih noticed, the T'ien-hsia chapter of Chuang-tzu must be referring
to debate over the fundamental doctrines of Mohism, not over the Canons
1
Cf. p. 230 above.
2
Cf. p. 113 above.
244 The Fragments of 'Expounding the Canons'
which formulate the techniques of disputation, when it says that the later
sectaries "all chanted the Mohist canons but diverged in opposite directions
3
and called each other heretical Mohists" ( ^ I I S « [ f n f e r l ^ ^ l , f f i f i S l J ^ ) .
In early Mohism "the Will of Heaven is the canon of righteousness"
(Mo-tzu ch. 28, Sun 141/5 %iM^±M^). Heaven, who rewards the
good and punishes the wicked, is the ultimate sanction of all morality. But
a fundamental difficulty has arisen, which is presented in the introductory
section. Individualists are now arguing that since man's nature (hsing tfe)
is ordained for him by Heaven, it is by following the dictates of his nature
4
that he obeys Heaven. T o 'expound as a canon' the Will of Heaven merely
encourages the selfish man to indulge his natural egoism. This is the first
and last reference to human nature (and to the Will of Heaven) in the entire
corpus. The metaphysical crisis of the 4th century, the opening of the
fissure between nature and morality, between the spontaneous order of
Heaven and the contrived order of human society, has forced the Mohists
to look for new foundations. From now on their 'canons' will be the
definitions and the demonstrable propositions of disputation.
3
Hu Shih (1919) 185.
4
Cf. §1/1/1/2.
2/3 245
EC i i
(mm ±&)
T C 2B/9-10 f £ g o R f t f e » # S »
2
3
» 4
£ H tt
5
» o 1B/9-2A/3 8 « A i f T ? c £ ' J fl№& »
fife o * A < £ 1 > 9

r «> J »a A # « ft&fe »iffift*
1
Ch'in-yin chih tz'u, 'proposition which encroaches on and vitiates',
converts another proposition into a yin tz'u 'vicious proposition'
(§ 1/4/31).
2
Yu chingyeh 'expound as being a canon*; for the syntax cf. pp. 155, 156.
3
"Reject the 'white horse* in it but uphold the 'colt* in it.** The reference
is to two sophisms, " A white horse is not a horse** (&M3¥M Kung-sun
Lung tzu ch. 2) and presumably a parallel " A colt is not a horse**
(unattested, but cf. <f&J^fc£ " A whelp is not a dog**, Chuang-tzu ch. 33,
Kuo 1106/3, 4).
4
Transpose on grounds of parallelism.
5
Sun emends wu 'dance* to wu № 'there is not* (Sun 255/8); but with the
reconstitution of the passage it can be seen to connect with the ko
'sing* of the next fragment.
6
* N G J O may be the *NGj.O 'expound* which is elsewhere in E C 1
written with its regular graph fp (but the interchange is unattested).
7
T'ien-chih 'Will of Heaven*, chih M (originally fi&) written without the
radical as as sometimes in the T'ien-chih chapters (Sun 123/2).
8
For nominalised chih 'the alternative upheld*, cf. § 1/4/5.
9
Restored on grounds of parallelism.
10
Wei wo 'egoism*, the teaching of the individualist Yang Chu S§/fc
(Mencius 7A/26).
11
Jen fei 'other men's disapproval* ( M . 344/410). Cf. B 3 'other
men's approval*.
("A proposition which encroaches on and vitiates others.")

"Expounding the canons. When expounding as being the canon, if I maintain
of the colt what I deny of the white horse, should I seek to prove it he will
dance to my song and prove the one which is wrong, should I by exposi
tion ?) magnify it he will dance to my song and magnify the one which is
wrong. If on behalf of a criminal I expound the ' W i l l of Heaven* as
being the right one of the two, and his 'nature' as being criminal, I sing of
the Will of Heaven as being the wrong one of the two. If among the
proposed alternatives there is already one that he is for, and I propose the
alternatives on behalf of him, the alternative that he is for will find a basis
in the one that I am for. If among the proposed alternatives there is not
yet one that he is for, but I propose the alternatives on behalf of him, the
proposed alternatives will find a basis in the one I am for. The criminal
will think that "every man for himself" is the Will of Heaven and that
the one which men condemn is the right one of the two, and his nature
will be incorrigible."
A tz'u (the word later adopted in Names and objects for the sentence) is
the verbal form we give to an idea; a yin tz'il, Vicious proposition', is one
which plays tricks with words; a ch'in-yin chih tz'u, a 'proposition which
encroaches and vitiates', should probably be understood as a proposition
which by its form of words infects others and turns them into vicious
propositions (§ 1/4/31). If so, the example is "It is one's nature to be a
criminal". T o admit it changes the significance of one of the ten canons of
Mo-tzu, the Will of Heaven; it implies that, since our nature is ordained
for us by Heaven, the criminal obeys Heaven by following his own nature.
The admission is like undermining one's defence of common sense against
Kung-sun Lung's " A white horse is not a horse" by conceding that a colt
is not a horse. This is the only reference in Mo-tzu to the problem of human
nature which so obsessed Confucians. But there is other evidence of
Mohist concern with the 'egoism' (wet wo) of Yang Chu, who taught that
one should "keep one's nature intact and protect one's genuineness"
( i f c t t ^ K Huai-nan-tzii ch. 13, L i u 13, 9B/10). The story of a debate
between Yang Chu and Mo-tzu's chief disciple Ch'in K u - l i
apparently from a Mohist source (G (4) 296 f.) survives in Lieh-tzu
(ch. 7, Yang 146/2-11), and Mo-tzu contains stories of the founder himself
refuting egoism in ch. 46 (Sun 272/5-14) and at the beginning of ch. 47
(Sun 275/1-5).
Yii 'expound' also appears in a five times repeated formula in a story
of Mo-tzu describing how he expounds his ten doctrines (Mo-tzu ch. 49,
Sun 299/4-7: one example is WMZMJKMfa ". . . then I expound to
them reverence for Heaven and service to the spirits").
E C 2 (ffiAifcJSXTte)
TC2B/4~5BAJB0^ » » » '^hZMfo » # 4 A / l - 2
mSMfc o tgffiA0f3fc«SA» 4B/5-5A/2 Sg# » A C * ) * # K£ 1 2
2/3 247
o E £ * n » * # 4 " B £ * ] * t e o 2B/10-3A/lHtt»&A ·
u
1
^ f * j e m ^ 3 A / 6 - 8 X T i : f i J W ° ffiAffaWofciBUI* 0±a-tfe · 71
1 2
Emended on grounds of parallelism.
13
Ch'ing (=Wf), what X genuinely is in itself, what is conveyed in its
definition (§ 1/4/6).
1 4
Nominalised chih, 'conditions' (§ 1/4/5).
15
Unknown graph on which Sun does not speculate. T'an suggests on
slender evidence that it is equivalent to yu W , making the phrase fu yii
( M . 11947/2), 'cherish and foster* (Tan (1958) 221).
i6, 17 Tsang, Huo: abusive names for bondsman and bondswoman
(§ 1/5/12).
18
Chih 'the intelligence' (defined A 3). The radical which distinguishes
chih I? 'know' (§ 1/2/1/2/2) is never dropped in the Ta-ch'u.
19
Chih ch'iang 'walls of the intelligence', cf. § 1/5/8.
20
'Three things': love, thought and benefit. For wu used of abstractions
such as love, cf. § 1/4/33.
21
Hsien in the sense of 'present, at hand' ( M . 34796 def. 7). Cf.
§1/2/1/2/11.
("The sage's concern being for the whole world.")

"That the sage dislikes critical illness but not dangers and difficulties, and
sits straight without moving, is because he desires the benefit of others, not
because he dislikes losing his own health.
In the case of everything that the sage desires or dislikes 'a priori' on behalf
of men, men necessarily learn from him by considering what the thing is in
itself; but in the case of desires and dislikes born from the conditions they
encounter, they do not necessarily learn from him by considering what the
thing is in itself. The sage's fostering care is benevolent but without the
loving which involves benefiting. The loving which involves benefiting is
born of thinking. Yesterday's thinking is not today's thinking, yesterday's
love of man is not today's love of man. The love of man involved in loving
Jill is born from thinking about Jill's benefit, not from thinking about
Jack's benefit; but the love of man involved in loving Jack is the love of
man involved in loving Jill. Even if to get rid of the love of them would
248 The Fragments of ' Expounding the Canons'
benefit the world one cannot get rid of it. Yesterday's 'walls of the intelli
gence' are not today's 'walls of the intelligence'. Only when the three things
['love', 'thought', 'benefit'] are present together are they sufficient to
generate the enjoyment of benefit in the world. The sage has love but
does not have pronouncements which benefit current situations, that is,
pronouncements about the transient. If there were no men at all in the
world, our master Mo-tzu's pronouncements would still stand."
In this fragment the Mohist can be seen to have already the conception
of a rationalised ethical system independent of appeals to the Will of Heaven
which is developed in the Canons. Logicians investigate what is knowable
'a priori' by imagining something hidden behind a wall and considering how
much we know about it merely by knowing its name, without going over the
wall to observe it. (For hsien 5fe 'a priori' and the image of the wall, cf.
§ 1/4/13, I/5IS). What do we "desire and dislike 'a priori' on behalf of a
man", without knowing anything about the man behind the wall except that
he is a man ? The moral principles laid down by Mo-tzii are the answers to
this question. They disclose themselves as necessary when we take moral
concepts and examine their ch 'ing, what they are in themselves. (Pi 'necessity'
is defined in A 51. For ch'ing, which approximates to the Aristotelian essence
conveyed by a definition, cf. the example from Chuang-tzU of "the ch'ing
of benevolence and righteousness" ( t H ^ W i ) quoted in § 1/4/6.) These
principles would be valid "even if there were no men in the world" to
apply them to. However, we cannot act on them without knowing what
actually benefits individuals in their concrete situations, which will differ
with different individuals and times. T o have examined the ch'ing of 'love
of man' does not enable me to benefit others in practice unless I pass over
the 'walls of the intelligence' to explore what in A 75 will be called the
"benefit and harm beyond the wall" (B$V£MJS). The ethical definitions
of the Canons will substantiate the claim that Mohism has established what
is "desired and disliked 'a priori' on behalf of men" by building a system
of interrelated concepts in which each is ultimately defined in terms of the
words 'desire' and 'dislike' themselves. We have analysed it in § 1/1/2/5.
The cornerstone of this 'a priori' system would have already been laid in
the crucial definition of 'love' in the lost document on the 10 theses of
Mo-tzu (§ 1/6/2). The author of Expounding the canons has seen the implica
tions at least of defining 'love' in terms of 'desire' and 'dislike', although
the possibilities of further definitions may not yet have been explored.
From this point he will be solving specific problems of Mohist ethics,
and also in E C 7, 8 laying down the principles of practical reasoning.
2/3 249
EC 3 » Jfc*IXT*H;B)
9
T C 3A/2-3 iS*»» ° 3B/4 ° 5A/2-5
The syntax of j^/r requires restoration of a verb at this point (cf. p. 154
above) which might however be It M 'benefit* (preferred T'an 222,
Chang Ch'i-huang 3A/3).
("The one who lives long and the one who dies young are equally beneficial
to the world.")
"In the matter of righteousness being beneficial and unrighteousness
harmful, one must distinguish between intent and achievement. Intent and
achievement cannot be assumed to coincide.
One exalted to the throne of the Empire is not more beneficial to man than
any ordinary fellow. Of two sons serving their parents, if one chances on
a good harvest and the other on a bad but they serve their parents equally,
it is not that the former's conduct is improved, not that anything is added
to it. No external condition can make me more beneficial."
E C 4(— 0ffnWiH^» № № )
TC 5A/5-6*»»*5EiffiXT*f 9
W^№&M№ » S*Miteffljn» °
The name Tsang (cf. § 1/5/12) is once disguised by the 'grass' radical,
twice written correctly.
("Even if a million were born in one day one would not be loving more.")
"Supposing that the whole world would be harmed if of all men Jack were
to die, I would make a point of caring for Jack 10,000 times more, but would
not love Jack more."
EC 5 (ftH1ft*J!t*ifn«ilitt*B*)
9
T C 3A/4-6 9t[m*]f№m9t&1&te2i ° g^ffilliig
« i f t - * ^ - B t o[ A ( t e ) £ * * # A t e 9
°
9 27
2A/9-2B/3 f§?TF№ffi U S <ZA> ife ° * X T J » * S · Till
2 4
Delete following Sun 255/-2.
25
Yu X 'also' is regularly written with this graph in the dialectical chapters.
2 6
This is a Names and objects fragment, accidentally written in here but
also preserved in its proper place and uncorrupted in the Hsiao-ch'ii
(NO 18).
27
Restored from what seems to be a parallel phrase in the next sentence.
2 8
Restored on grounds of parallelism (Sun 254/-2).
("In loving two generations we love some men more or less than others,
yet we love the two generations equally.")
"Love of many generations and love of few are equal, and equal also to the
love of every one of them. Love of past and future generations are alike
equal to love of the present generation.
To think Y i i important for the sake of the world is to be for the kind of
man Y i i was; loving Y i i more for the sake of the world is loving for the
sake of the kind of man Y i i was. We think it important that Y i i was added
to the world but do not extend the importance from Y i i to the world, just
as we dislike robbers being added to the world but do not extend the
dislike from robbers to the world."
Love of a man is "for the sake of the man he is" (EC 11 H^tAtfe);
we therefore love the sage Y i i more than other men, "for the sake of the
man Y i i was" ( J i B ^ A ) . But this does not imply that we love the men of
Yii's age more than we love contemporaries. Our love is not extended to
the other men of his time, any more than our dislike of a robber is extended
to all other men.
EC 6 · I B W - A )
T C 3B/8-4A/1 W f r l i l ' **S1ft ° »
»m... 3A/9-10 o x&*>&#№ite o Jims.

A » IB/4-7 3p№^T*B3g » I M o nmm-»°
31
T<-fe> °
29
Delete following Sun 257/1.
30
Chuan sha "kill on one's own authority", cf. Mencius 6B/7.
3 1
Restore on grounds of parallelism.
2/3 251
("Loving them equally, we choose one among them and kill'him.")
"We know that there are robbers in this generation, but love the whole of
this generation; we know there is a robber in this family, but he is not all
this family; we know that one of two men is a robber, but he is not both
the men; and even the one of them who is a robber, if we do not know
where he is . . .
. . . is not to kill Jack, to kill a robber on your own authority is not to kill
a robber. Speaking in general, when we learn to love men, if they are
equally beneficial to the world there is nothing to choose between them;
if a man's life or death are equally beneficial there is nothing to choose
between them. Killing one man to save the world is not killing one man to
benefit the world; giving one's own life to save the world is giving one's
life to benefit the world."
According to the Mohist doctrine love should vary with the worth of
its object (EC 5, 7), so that the man we kill although we love him as much
as others can hardly be the robber (as supposed in G(5) 29). The issue
appears to be the problem of killing innocent individuals, oneself or others,
for the sake of the rest (as in self-sacrifice or the execution of a robber's
whole family; that the Mohists approved the latter for serious crimes is
confirmed by the list of prohibitions and penalties in Mo-tzu ch. 70). But
the text is so mutilated that the sequence of the argument is lost. Possibly
the point was that it is beneficial to the world for the ruler to kill the
kinsman of a robber who has escaped but harmful for a private person to
defy authority by killing the robber, so that it may be better to kill the
innocent than the guilty; but this is no more than a guess.
Since self-love and love of others are granted quite equal status (EC 10)
it is at first sight hard to understand why "killing one man to save the world
is not killing one man to benefit the world" although "giving one's own life
to save the world is giving one's own life to benefit the world". But it
appears from the second of the three fragments that the theme at this point
is "killing a robber on your own authority". A private person may sacrifice
his own life to benefit the world; but if he kills another man to save the
world (instead of handing the criminal over to the proper authorities) he is
harming the world by disrupting public order.
EC 7 ( / h £ H * f c ? T № f B ; g )
TC IB/7-8nm&z^ffimnn&nr&i °M>#<nz>™&>
IB/8-9 o 1A/3-9 ^£gAte*îgA;£gA-iii»
33
Ä 5 W / h A * » 3 » / h A ± W ^ A f t · ülKÄÄ^-tbffiig±[*] gÄ
Restore on grounds of parallelism (Sun 254/5).

The parallelism of the last pair of arguments requires this deletion (Sun
253/-6).
("The conduct of the slightly benevolent and the highly benevolent are of
equal worth.")
"Weighing light and heavy among things to be done in practice is what is
meant by "seeking how to act". T o seek to do something is not to do it, to seek
to act righteously is not to act righteously. Heaven loves man less than the
sage does, but benefits him more than the sage does. The great man loves
the small man less than he is loved by the small man, but benefits the small
man more than he is benefited by the small man. Loving Jack on the
assumption that he is one's kin is loving one's kin, but benefiting Jack on
the same assumption is not benefiting one's kin. Desiring music for one's
son on the assumption that it benefits him is loving one's son, but seeking
music for him on the same assumption is not benefiting one's son."
Sun Yi-jang took the jen tl of the proposition as equivalent to jen A

'man'. But the reconstructed exposition does not seem to be defending
the claim that "The conduct of the small man and the great man are of
equal worth". It may be noticed that although jen 'benevolence' is a virtue
for Mohists as for Confucians, the single instance of the word elsewhere in
the Expounding the canons is in a reference to being "benevolent but without
the loving which involves benefiting" (EC 2). The Mohist judges a man,
not by his degree of personal benevolence, but by the extent to which his
benevolence is realized in effective action. He therefore points out that
there is no necessary proportion between loving and benefiting. Thus a
man is not really benefited by music (an art which Confucians valued but
Mohists condemned); to desire music for one's son is to be benevolent
without effectively benefiting.
EC 8
T C IA/9-IB/4 tkm»±^mmmn±m r* j ° m#*&fc»##i§
2/3 253
fti+Jfe/h [ i t ] » #Jfc«-fc » №RHb · № ? f J f c * A £ l l T » i b o S

3 4
«EAifii»f(»)*K»fiia# • · KaMSA»*-fe o K % H V f K »
IB/8 * £ 4 » # / j N < - f e > * ° 3 A / 8 - 9 * » E i f n ® C : £ »
± i f e ] 2A/3-6 ffniE38(^)*ifa o ffl2.*fr* » # * » E t t · S £
57
»:W#Effe ° B f * # M f t i « » * ? i J ± * ^ A i b o ^9rlEWifn
3 4
Unless this yeh is emended to che (Sun 253/-1) we must delete it, as
absent in parallel phrases. Yeh is not used after subjectless nominalised
clauses (p. 154 above).
3 5
The sequence of the argument suggests that the sacrificed limb should
be not the finger but the arm (wan, written in this section with both
the graphs IB6 and Bi = I ? ) .
3 6
The two scraps which we have fitted in at this point will completely fill
the gap if we make this and another minor emendation (n. 38 below).
But we can have no confidence that the text is not more seriously
mutilated.
3 7
Deleted following Sun 256/6.
3 8
Cf. n. 36 above. The same two characters are confused in E C 5, Appendix
5, N O 9 (cf. § 1/2/1/2/13).
("Promoting the beneficial is eliminating the harmful.")
"Weighing light and heavy among the things treated as the units is what
is meant by Veighing\ The wrong when weighed turning out to be the
right, and the wrong condemned as the wrong, are the judgment after
weighing and the judgment which is direct. Cutting off a finger to save an
arm is choosing the larger among benefits and the smaller among harms.
Choosing the smaller among harms is not choosing the harmful but choos
ing the beneficial; the choices open to you are under the control of others.
When you encounter a robber, to save your life at the cost of an arm is
beneficial, the encounter with the robber is harmful. Cutting off a finger
and cutting off an arm are choosing the smaller among harms; one desires
them because one has no alternative, it is not that one desires them directly.
Choosing the larger among benefits is not a matter of having no alternative,
choosing the smaller among harms is a matter of having no alternative.
Choosing between things you do not yet have is choosing the greater among
benefits; renouncing one or other of things you already have is choosing
the smaller among harms."
The weighing of relative benefits and harms was a common concern o

the Mohists and of individualists such as Yang Chu; its typical illustration
is generally the sacrifice of one member of the body to save another
(Explanation A 75, Mo-tzu ch. 47 (Sun 253/7-254/3), Lu-shih
ch'un-ch'iu ch. 21/4 (Hsu 1011/6-8, 1014/5-1015/2), Lieh-tzu ch. 7 (Yang
146/2-11)).
In E C 7, 8 the Mohist defines two terms which will remain important
in the vocabulary of the Canons, ch'iu 'seeking' (the counterpart of te #
'getting, succeeding', A 4, 28), which in Western terms is the choice of
means, and ch'iian 'weighing', the choice of ends. The distinction between
the primary desires which are direct or immediate (cheng) and the con
sidered preferences after weighing benefit and harm is reaffirmed in A 84.
The situation in which one prefers to lose a finger is reconsidered in A 75,
which introduces the question of the limitations of knowledge at the time
of choice.
The distinction between cheng and ch'iian may also be observed in a
discussion of the ethics of weighing in Huai-nan-tzii ch. 9:
(Liu 9, 32A/-2Q ftfcffft » ° £#ISlE • « t # S t t » &°
"Therefore the benevolent course and the wise course are incompatible,
but at times they tally. The choice when they tally is immediate, the choice
when they are incompatible is a matter of weighing; but in being the
righteous choice they are one."
It is used also, but with changed significance, in the art of war:
Ssu-ma fa ^ I S S c (SPPY) A , 1A/3, 4 >J^fcfl* > > ±M
"In ancient times, to treat benevolence as the basis and manage affairs
according to righteousness was what was called the 'direct' course. If the
direct course did not succeed they 'weighed'. Weighing (deciding which is
stronger) derives from war." (Mohist influence on the terminology is
suggested by an example of chien at 'loving everyone' shortly afterwards,
ut sup. IB/1.)
A point deserving of attention is the description of weighing as between
"things treated as t'i" (Sffi). A t'i is primarily a member of the body, and
connects with the example of sacrificing a finger to save an arm. But in
Mohist terminology it is the general term for anything countable as one,
whether as part of a whole or as member of a class, t'i 'unit' as the counter
part of chien ^ 'total'. (Cf. A 2. For the causative use of technical nouns
after so Sf, cf. § 1/3/13.) The use of t'i in the definition is important as
the one indication in the surviving fragments of the principle behind the
2/3 255
weighing of alternatives; one loves the totality of men (chien at 'love of

everyone') and in each choice prefers the t'i which contributes most to the
chien, preferring the arm to the finger for the sake of the whole body and
the beneficial man to the harmful for the sake of the world. We can find
the same subordination of t'i to chien in Mencius:
Mencius 11/14 A ± 3 f c H & » £ 0 f * ° * § r S » №Jfc0T*-fe ° . . . »
"Man's attitude to his body is that he loves every one of its members; and
if he loves every one, he tends every one. . . . Among the members there
are noble and mean, lesser and greater: do not harm the greater for the
sake of the lesser, the noble for the sake of the mean. . . . A man would
be a lunatic if he knew no better than to tend one of his fingers at the cost
of his shoulders and back."
EC 9 (mm^wa\ftmft)
T C 2A/6-9 m^mmz » m^mm± > m r mi J ° mnm±^&m
&» ik%®x№& o H a n i > ^tam o mmm»mmrn > mmn^
mm o rmwtmnwn j 4B/3-4 B » * » K » · m%m» s
* g » o 2B/7-8 mA±m >nt *»m» °
-te» J ^ £ ± I I i ® °
4 1 4 2
3 9
Cft'ioi^ 'encourage' ( M . 9815 def. 2/4).
40
Wang (=26) 'forget' (Sun 255/4). Cf. Chuang-tzu ch. 14, 29 (Kuo
498/-1, 990/2) S £ 'forget parents'.
41 Li ( = | l ) , 'ceremony' ( M . 45291 def. 23); we find the reverse substitu
tion of graphs in Explanations A 45.
42
Chieh ( = » ) 'exhaust' ( M . 17788 def. 2/2).
("Do not justify doing more for parents than for others by their conduct,
but do pay attention to their conduct.")
"Doing more for those for whom duty requires more, less for those for
whom duty requires less, is what is meant by "arranging according to
grade". Men whose acts deserve gratitude, rulers, superiors, the aged, one's
elders, are all persons for whom one does more. Doing more for a man
because he is elder, one does not do less for a man because he is younger.
One does more or less according to degrees of kinship, as far as the remotest
degree which does not impose a duty. " D o not justify doing more for
parents by their conduct, but do pay attention to their conduct" encourages
sons to recognize what is deep or shallow in their parents, to improve them

where they can be improved, honour them where they deserve honour.
By the standard laid down by the sage you forget your parents when they
die, for the sake of the world. Doing more for parents than for others is
your portion, and finishes with the rites of death and farewell."
EC 10 (^n&±^m-&)
T C 2B/8-9 nmtimm&mmizmm · s a * 2B/3^ s t w t a ·
affljr*£4* ° a ^ g f g » g t o j ^ a » j t & № g a g A - & ° 3A/1-2
w±ga*ssaÂ-tfe °* ^ a » » s a ^ n - t f e 4 3
o
3B/7-8 w ^ - w * * * * ° # * £ f c » * « a £ « ^ * ·
4 3
Fei hsien yeh "is not for the sake of one's worth". Cf. Shih-tzu SPPY,
A 10A/8f < CI > S № i , # J R B t e "Parents' care for children
is not for the sake of their worth or strength".
("That one may learn not to be concerned for oneself alone.")

" T o do more for some than for others in promoting the beneficial, but
without the promotion of benefit which arranges according to grade, is
being concerned for oneself alone.
Love of men does not exclude myself, I am myself among those I love.
Being myself among those I love, my love being extended to myself, is the
love of self and others which arranges according to grade . . . Jack's love of
himself is not for the sake of the others who love him. Those for whom
one does more do not exclude oneself, but the love is neither more nor less;
giving special attention to oneself is not on account of one's worth. . . .
Knowing one thing of benefit to one's parents is not sufficient to be judged
filial; it falls short of knowing the benefiting of parents which is not for
the sake of oneself."
From E C 10 onwards the expositions are more and more seriously

mutilated. The introductory passage can be fixed, because it follows on to
E C 9 on one fragment. Three more fragments appear to belong with it;
the last has a parallel in E C 11 with which it was juxtaposed by the compiler
of the Ta-ch'ii. It may be guessed that they were parallel conclusions of the
two adjacent expositions.
Apparently the Mohist is arguing that one can learn from self-love
how to love others; it is a pity that textual mutilation allows us only a
glimpse of this important argument. It seems that self-love serves as a
model for love of others in the following respects:
2/3 257
(1) Self-love is not for the sake of others who love one (and similarly
love of others should not be for the sake of oneself).
(2) The special care owed to one's person does not imply greater love
(any more than the greater care owed to ruler or parents, E C 9-12).
(3) Special care of oneself is not on account of one's excellence (just
as the care of parents is independent of their moral conduct, E C 9).
E C 11 ( * A # f t » i b )
45 9
T C 3B/4-7 ^ ^ A f b S ^ A * © A ^ ^ | * f X |
0
W*fcH«A
[ » A ] ife o f£A1%n&n o S « [ * * J - A ^ f t K * * J A 4 * ^
4 6
4 4
The first words of the phrase are lost (since a nominalised clause with
yeh requires a subject with chih cf. p. 154 above).
4 5
Transpose on grounds of parallelism (Sun 256/-2).
4 6
Delete the two repeated graphs.
4 7
as in Canon B 39. Cf. also A 75 "overlook none
of the harm in it".
("Love of men is not for the sake of praise.")

" . . . benefiting a man is for the sake of the man he is. Enriching a man is
not for the sake of the man he is, one enriches him for some special purpose;
and governing men is partly for the sake of the spirits. Benefiting one man
for the sake of reward or praise is not benefiting men for the sake of reward
or praise; it falls short of applying to men without exception."
E C 12 ( * A £ « £ * & « )
T C 2 B / 5 - 7 . . . UA£ftft ° S A ^ S f t S ° «£*to@Ei&l* ° I S A ^ I f
("Love is as much for others' parents as for one's own.")

". . . is because he dislikes harm to other men. The sage's concern is not
for his own family, Jack as he is in himself resides in Jack. The sage does
not get to perform the services of a son. . . . "
E C 13 ( * $ f f i 3 g ° - g f f i ; g )
("Love of everyone and love of an individual are equal.")
The worst damage was evidently to the last strips of the document.
There is only one fragment which is plausibly placed among the last two
expositions. Even this is so disjointed that one is tempted to divide it

further and distribute the bits among earlier sections.
E C (Appendix)
T C 5B/6-6A/7
1 mm±m· smw&tm °
2 igA(te)*^ ISXTte »
4 8
o
3 aaiis*» » ? i j ? c № ( f £ ) * f f i 3 g » » * f f i ( * ) * # 6 o ^ o
4 9
7 /ht*At >nmtm>M:mfc%K<>
8 H?]l»S-ttl » J « £ M M 5 1
°
9 ff^fflff rfn(«)*H tr ' J W 8 £ r X ± #
52 0
11 S A ^ g f f t ^ l l E i i o
12 * A ± a « * x a > jygffiWfs °
4 8
Emended (following Chang Ch'i-huang, 7B/1) on the evidence of the
syntax of yeh, not found after a noun subject (p. 154 above).
4 9
Emended following Sun 259/-3.
6 0
Emended following Sun 259/-3.
5 1
= g ( S u n 260/4).
5 2
Emended from the parallel in E C 9.
5 3
For the phrase lieh t'u 'hunter* in a Mohist fragment, cf. Sun 491/2.
5 4
Repetition deleted with Sun 260/8.
5 5
= attested here as a variant in the T'ang-ts'e-hsien edition (Wu
11,10B/-1), which however is a poor authority (§ 1/2/1/1). Cf. Canon
B 39 'the spring snakes', where the second graph is convincingly
identified as she 'snake' by Kao Heng (Kao 158).
1. " A proposition which encroaches on and vitiates others: analogy,'...' "

2. "The sage's concern being for the whole world: analogy, 'going after
someone who has lost his way'."
3. "The one who lives long and the one who dies young are equally
beneficial to the world: analogy, 'the poisonous stone'."
4. "Even if a million were born in one day there would not be more love:
analogy, 'disliking harm'."
2/3 259
5. "In loving two generations we love some men more or less than others,
yet we love the two generations equally: analogy, 'the markings on a snake'."
6. "Loving them equally, we choose one among them and kill him:
analogy, 'the rat down a hole'."
7. "The conduct of the slightly benevolent and of the highly benevolent
are of equal worth: analogy, ' . . . ' "
8. "Promoting the beneficial is eliminating the harmful: analogy, 'the
leaking pot'."
9. "Do not justify doing more for parents than for others by their conduct,
but do pay attention to their conduct: analogy, 'the well by the river'."
10. "That one may learn not to be concerned for oneself alone: analogy,
'the hunter'."
11. "Love of men is not for the sake of praise: analogy, 'the inn'."
12. "Love is as much for others' parents as for one's own: analogy, ' . . . " '
13. "Love of everyone and love of an individual are equal: analogy, 'the
dead snake'."
The list of analogies was put by the compiler of the Ta-ch'u at the end
of the chapter. This suggests that it was either the conclusion of Expounding
the canons or an independent document; the compiler would recognize the
final strip, if intact, by the fact that it was not filled, and perhaps by some
scribal note. It is not likely that this is the list of the 13 propositions which
we would expect at the head of Expounding the canons, since the expositions
have nothing to do with the analogies. Presumably it was a memorandum
of analogies useful for defending the propositions in debate. The parenthetic
analogies introduced by jo 3=r 'Like . . .' throughout the Explanations
might also come from such a memorandum, since they seem to be glosses
(§ 1/2/2/6).
It would seem that Expounding the canons did not contain the 13
propositions themselves (just as the Explanations of the canons in Mo-tzu
ch. 42, 43 are separated from the Canons). But very fortunately for us the
propositions had to be quoted (possibly in abridged form) in the appended
list of analogies. These analogies remain mysterious; in the absence of
further information there is little point in discussing them.
2/4
T H E 'CANONS' A N D 'EXPLANATIONS'
2/4/1 T H E DEFINITIONS (A 1-87)
T H E art of definition continued to develop and refine itself throughout

the pre-Han period, from its primitive beginnings in the Analects of
Confucius to its culmination in Hsiln-tzu and the Interpreting Lao-tzu
chapter of Han Fei tzu. (An odd consequence of this technical advance is
that the Legalist Han Fei tzu, an enemy of Confucian moralism, produced
at the very end of the period the best definitions of the Confucian virtues.)
The present series of 87 definitions is by far the longest in pre-Han philoso
phical literature.
We may lay down certain principles for the interpretation of the
definitions:
(1) A l l Canons in A 1-75 really are intended to be definitions
(cf. p. 230 above); any interpretation of a Canon which overlooks this basic
fact is unworthy of consideration.
(2) The grouping of definitions deserves close attention, not only the
five major groups distinguished in § 1/6/1 but the sub-grouping and pairing
of Canons inside them. One is at first sight surprised that thinkers capable
of such neat geometrical definitions as "The 'starting-point' is the unit
without dimension which precedes all others" (A 6 1 ) and " 'Next' is
without interval but not coinciding" (A 69) should, although their primary
concern is with moral questions, be satisfied to define 'righteousness* and
'conduct' by the single words 'benefiting' and 'doing' (A 8, 10). But on
closer inspection one sees that the definitions which seem vaguest generally
belong to sequences or pairs of contrasted Canons (A 3 - 6 , 7 - 1 0 , 11 and 12,
4 7 and 48, 6 2 and 63, 7 0 and 71). In Aristotelian terms we might say that
the group establishes the genus and the single Canon the differentia; a more
Chinese way of putting it would be that words grouped together are
assumed to be alike except in the respect specified in the definition. (Cf. the
story in Shuo-yiian ch. 11 of the sophist H u i Shih defining a tan 33 by
analogue and differentia: " A tan in form is like a bow but with the string
made of bamboo", quoted under B 70.)
262 The 'Canons' and 'Explanations'
(3) In general a defined term will be either a current word in pre-Han

Chinese or a technical term put to use elsewhere in the dialectical chapters.
If we are right in our identifications of terms in A 15, 39, 47, 71, 73 the
only exception is the otherwise unknown geometrical term in A 64.
(4) The definitions sometimes interrelate to form systems, which
throw light on what the Mohist understands by hsien 5fe 'a priori'. We
appreciate why the ethical principles of Mo-tzu lay down "what the sage
desires and dislikes 'a priori' on behalf of men" (EC 2) only when we
observe that all ethical terms are ultimately defined in relation to 'desire'
and 'dislike' (§ 1/1/2/5). The system of definitions culminating in the circle
(A 58) substantiates the claim in A 93 that the circle is "known 'a priori' "
(§ 1/1/2/6). Some apparently vague definitions turn out to be precise when
we notice words in them which have themselves been defined, for example
the ch'iu ^ 'seek' of A 4, 28. The missing definitions of words in the 10
theses of Mo-tzu (which may not be the only losses) would have improved
our understanding, for example, of the recurring mingty]'clear, clarify'
(A 6, 29, 30, 72). We shall therefore note under each Canon in A 1-87 the
words which either have their own definitions or are analysed as ambiguous
in A 76-87.
Why are the ethical definitions, the most important of all, so much
shorter and less explicit than those of Hsiin-tzu and Han Fei tzu ? (Cf. the
examples quoted under A 3-6, 7-9, 10.) Probably because of the Mohist's
need to define or be able to define every word in them. This is suggested
by comparison with two other series of interrelated definitions, in Chuang-
tzu ch. 23 (quoted § 1/1/2/8) and at the beginning of Hsiin-tzu's Right use
of names. Chuang-tzu's series has verbal parallels with A 4, 5 (quoted
under A 3-6); the definitions are shorter than Hsiin-tzu's, have fewer
undefined words, and more nearly approach the ideal of a self-contained
system. The conclusion directly alludes to the interconnections: "The
names are opposed but the objects take their courses from each other"
(£ffiRffi1№JB-t&). On the other hand Han Fei tzu, whose definitions in
Interpreting Lao-tzii are even longer than Hsiin-tzu's, makes no attempt
to systematise them. Evidently the Mohist's system-building imposes
conflicting demands, for definitions long enough to be adequate yet short
enough to remain within the scope of preceding definitions.
2/4/1/1 Description (Al-6)

The first of the later Mohist disciplines is the art of consistent
description, the procedures of which are laid down in the first series of
propositions (A 88-B 12). But it shared most of its terminology with the
2/4 263
fourth discipline, disputation proper, the definitions of which are in

A 70-75. As we noticed in establishing the divisions of the Canons (§ 1/6/1),
the difference emerges distinctly only in the two sequences of propositions,
and here the Mohist seems especially concerned with the fundamental
terms which the first discipline shared with the rest, first of all the word
ku ASt 'reason'.
A i WL'rmWik™^*
55b
(*) ° /h*»»te.z.'j&fit& ° < > «•& ° (
5 5 a
ErA-Aon 'only after that' (§ 1/3/11).
5 5 b
The Explanation is badly mutilated, no doubt because the beginning of
a document is especially liable to damage. The mysteriously brief
phrase t'iyeh "It is a unit" is followed by the illustration "Like having
a starting-point" (yu tuan, cf. A 68, B 19). I conjecturally restore three
words from the definition of 'starting-point'; A 61 ffi, *№
ffnSHll^'tb "The starting-point is the dimensionless unit which
precedes all others".
56
The missing graphs were restored by Sun Yi-jang (Sun 209/3). The
Explanation by itself would allow us to take ta ku 'major reason' as a
sufficient but not necessary condition, and restore pu pi pu 'not
necessarily not' in stead of pi pu. But the Canon confirms that a ku
'reason' is always a necessary condition.
57
Cf. B 4 M M Sift "Seeing and appearing are apart", where the corruption
of the Canon shows that hsien 'appear' was originally distinguished by
the 'man' radical (fH).
C. "The ku (reason/cause) of something is what it must get before it will

come about.
E. 'Minor reason': having this, it will not necessarily be so: lacking this,
necessarily it will not be so. It is the unit <which precedes all others (?) >
(Like having a starting-point.)
'Major reason': having this, it will necessarily <be s o > : lacking <this,
necessarily it will n o t > be so. (Like the appearing bringing about the
seeing.)"
Ku is phonetically and graphically descended from ku "tS* 'ancient

times, the past'; it is that which lies behind something, in its past or at
its basis, that from which it originates. It may refer to
(A) The object as it is in itself, the facts as they inherently are, as

the basis of what one says (A 88, B 7 Ku 'the thing as it is in itself, E C 12
'Jack as he is in himself). The adverbial equivalent is ku @
'inherently, basically' (cf. § 1/4/14).
(B) The further facts behind the facts, as reasons which justify what
one says (A 94 ;fctt#C "seek his reasons", A 96 TStfc "ask about the
reasons", N O 10 JJHStBfe "by means of explanations bring out the
reasons").
(C) The facts behind the facts, as causes of the latter coming about.
Cf. A 77 "Dampness is a ku (of illness). It is necessarily required that what
it does will come about".
A minor ku is a necessary condition, a major ku SL necessary and
sufficient condition, for the occurrence or the affirmation of something.
Even usage A (in which ku is not translatable as 'reason' or 'cause') satisfies
the definition, since the facts as they inherently are serve as the necessary
and sufficient condition for the affirmation. Cf. B 3 ^ ' ; £ J № i W £ f B ; £
"Only when there is this object is something said of it", which connects
with the adverbial ku of its Canon (gRifSHSfb " I n itself it is what it is
said to be") and the nominal ku 'the thing as it is in itself of the related
Canon B 7. There is no reason to suppose that usages A and B are dis
tinguished in N O 10 ^S№J^#C^fe . . . "The proposition is something
which is engendered according to facts/reasons".
Ch'eng 'come about' is the word used for a thing or event completing
the process by which it becomes what it is called. In A 77 it is used of the
coming about of the effect. But it is unlikely that the Mohist is thinking
only of the realisation of events, which would reduce the Canon to a
definition of ku in the sense of 'cause'. Ch'eng appears to be used of confirm
ing a statement in A 88 "proved or not proved", A 94 f&m "a proved
statement". If so, the definition will apply equally to the facts and reasons
behind a proposition and the causes of an event.
The terms 'major' and 'minor' reason never reappear in the dialectical
chapters. The minor reason is a necessary condition allowing more than
one possible consequence, like a starting-point from which more than one
measurement can be made (cf. A 60,67,68). The major reason is a necessary
and sufficient condition ensuring the occurrence of the event or validity
of a claim, as the appearing of something ensures that it is seen (cf. B 4).
It may be noticed that in the ordinary formula of implication, 'p tse
№\ q' ('If p, then q')> the premise p would not count as a ku, since it is a
sufficient but not a necessary condition. Implication is discussed in quite a
different terminology (§ 1/1/2/4/1). A premise is no more a ku than it is a
2/4 265
'reason* in English. One may of course in the case of a sufficient condition

cite the premise as one's reason for asserting the conclusion (or in the case
of a necessary condition the conclusion for asserting the premise), but then
it is a necessary condition of the asserting.
A 2 ft > ^ t ô ^ f f e «
9
(fft) ° K£ttHb) °
(Defined: chien, lost (§ 2/2) )
y
C. " A t i (unit/individual/part) is a portion in a chien (total/collection/
whole).
E. (For example, one of two, or the starting-point of a measured length.) "
Of the very important pair t'i and chien only the former is defined in
the Canons, the latter being among the terms in the lost definitions of the
10 theses (§ 2/2). In the older Mohist documents chien is a verb 'put together'
(Mo-tzû ch. 47, Sun 277/-2 fê«f№*t "Present a white thing and
a black thing together and ask a blind man to pick them out from each
other"), or 'do to every one of them' (ch. 4, Sun 13/5 J^X^ÏÏfl3t£,
^Ïï5$ljiltii "Because (Heaven) loves every one of them, benefits every one
of them"), often used adverbially as in chien ai 'love every one'. T'i on the
other hand is a noun, used only of the members of the body (ch. 6, Sun 21/5
i S f ê a S M "Strengthen their limbs and suit their bellies"). The Canons,
adopting the two as counterparts, on occasion use chien nominally (as in
this definition) and t'i adverbially (A 7).
T'i and chien are primarily numerical, 'unit' and 'total' (cf. B 12, 67),
as can be seen from the first example in the Illustration, 'one' and 'two'
(two being throughout the dialectical chapters the typical number which is
more than one). They are used of all countable units, whether part and
whole, individual and class, sub-class and class (for none of which are
there special terms), as well as things only accidentally together (B 37).
Any chien may in turn be treated as a t'i and counted as one in a new total
(B 12, 59):
T'I CHIEN
A 1, 2, 61 Starting-point Measured length
B 12 Finger The fingers
B 12 The fingers (The man)
N0 7 Head, fingers The man
A 7, B 73 A man A l l men
B 67 Oxen, horses Oxen and horses
B 12 Oxen and horses (Animals)
The Mohists noticed that we cannot say of two things 'They are both
two' (ft—), but that if something is so of both of them we may say that in
this respect 'They are both one' (ft—*) and count them as one in a new
total (B 3, 12). That things are the same in 'kind' (lei SI) is therefore a
reason to treat them as a chien on one level and a t'i on the next (in our own
terminology, treat them as a 'class').
The Mohist geometry recognises no points except the tuan 'starting-
points' (or end-points, meeting-points) of measurements, which are
countable, for example, when one speaks of the two ends of a stick.
(Cf. p. 58 above.) Starting-points are therefore the only units which are not
further divisible, the only ones which cannot be treated also as totals. This
explains why, as the most apposite example after "one of two", the Mohist
chooses "the starting-point of a measured length". At the other end of the
scale the infinite is the totality of things (B 73) but does not satisfy the
present definition of a unit. Presumably the Mohist would reject as
illegitimate H u i Shih's "Heaven and earth are one unit/count as one"
(Chuang-tzu ch. 33, Kuo 1102/-1 №&).
9
A 3 %\ °
A 4 1« » * t e °
(J*) ° r « J : &Kfc№*-BJ » mfc>№±Z ° (5gBS) °
60 9 8 1
A 5 ftl S * °
(2y) ° rftiJ -Hi*: &M№™^mm™z ° ( & M ) ° :
6 4
A 6 S ' °
(Defined: E C 7: ^ lost (cf. § 2/2))
58
Ts'ai(literally 'timber'), natural resources, for example the 'five resources'
of the crafts, metal, wood, leather, jade and clay ( M . 257/40, 7/1).
A man's ts'ai (=^*) is his talent, his personal capacities; but these too
may be enumerated: cf. Liu t'ao S P T K 3/16A/10 BfdHEW^f,
My t, fit, "What one calls the five resources (of a military
commander) are courage, wisdom, kindness, trustworthiness, loyalty".
59
Yeh che after a single word marks it as a quotation from the Canon,
exposed at the head of the sentence (pp. 116, 140f above).
6 0
This word was distinguished from the chihftlof A 3 by the graph H?,
eliminated as far as B 9 of the Explanations by later editing (§ 1/2/1/2/2).
2/4 267
Chieh 'make contact, come into touch' cf. Huai-nan-tzü ch. 1, (Liu
1, 6B/7) ^ l l Ä ^ S c l f i J & t t ^ i ^ "When consciousness comes into touch
with things, likes and dislikes are born from it".
Kuo 'pass': after passing the thing and no longer having it in front of
one's eyes one still knows it. Comparison with other examples of kuo
shows that Sun's emendation to yü S 'encounter' only obscures the
point (§ 1/4/15).
The mao of a thing are all its describable characteristics; the transitive
use of the word ('describe') is found only here and probably in A 95,
where it is distinguished by the 'speech' radical (§ 1/4/20).
This peculiarly Mohist graph, several times corrupted to & , represents
the chili 'wisdom' commonly written W (§ 1/2/1/2/2).
The use of ming in no less than three of its senses ('eyesight', 'clear
headed', 'clear sight') is curious; possibly as with chih its senses were
originally differentiated graphically.
C. "The chih (intelligence/consciousness) is the capability.

E. The 'intelligence': it being the means by which one knows, one
necessarily does know. (Like the eyesight.)
C. Lii (thinking/forethought) is the seeking.
E. 'Thinking': by means of one's intelligence one seeks something, but
does not necessarily find it. (Like peering.)
C. Chih (knowing) is the connecting.
E. 'Knowing': by means of one's intelligence, having passed the thing
one is able to describe it. (Like seeing.)
C. Chih (understanding/wisdom) is the illumination.
E. 'Understanding': by means of one's intelligence, in discourse about
the thing one's knowledge of it is apparent. (Like clearness of sight.)"
The four definitions are contrastive, and the four illustrations point
up the contrasts by presenting parallels between knowing and seeing. They
make it plain that the Mohist distinguishes knowing from perception but
also conceives them as analogous. One of the theses, B 46, says explicitly
that we do not know by means of the five senses, and that if we did knowing
would end with the disappearance of the object from view (Sfcfi'M
"Knowing when prolonged would not fit the fact"). Here in A 5 the test
of knowing a thing is that after experiencing and leaving it behind (kuo,
'passing') one is still able to describe it. (For the contrast of tang 'fit' and
kuo 'pass' cf. § 1/4/15.) The definition of knowing as chieh 'connecting,
being in touch', which was current outside the Mohist school, also served
to distinguish knowing from perceiving. The Lu-shih ch'un-ch'iu has a
chapter entitled Chih chieh #Pfic "In touch by knowing", about people who
are acquainted with the facts but do not chieh 'connect, catch on'.
According to Hsiin-tzu's Right use of names the five senses perceive a
thing, but we know the thing only when the mind (hsin >it) recognises what
the thing is (cheng Wi 'recognise by the distinguishing marks' cf. A 45).
The Canons and Explanations avoid the word hsin (so common in Menciu
Chuang-tzii and Hsun-tzu), although the older chapters of Mo-tzu use it
and Names and objects revives it; possibly their absolutely extroverted
moralism made them suspicious of a word which encourages an unhealthy
interest in the inner life. But since he conceives knowing as analogous to
perceiving, the Mohist does find it necessary to postulate a chih 'intelligence'
by which we know, as we see by the eye and hear by the ear. He distinguished
the word from chih 'know' by adding a radical to the graph of the latter,
eliminated by an editor throughout the first three dialectical chapters
(cf. n. 60 above).
A consequence of postulating a special faculty of knowing (in Western
as in Mohist thought) is that since by definition its operation is knowing
it cannot be in error. The Mohist does commit himself to this conclusion
in A 3, but in A 4 explains that it is thinking by means of the intelligence
which is liable to error; the tool is perfect but its employment is not. His
word for 'thinking' is lii (cf. also E C 2, B 50), which generally implies
forethought, worrying about practical problems. It is not clear whether it
is intended to cover the pure exercise of the intelligence in demonstration,
but presumably it must include false reasoning of any kind. For the
'supposing' which may be mistaken he elsewhere usesjy/ wei (cf. E C 7,
A 24, B 10, 71, 77).
We use the intelligence, not only to know things in the sense of being
able to describe them, but also to sort them out, grade them, arrange them
in coherent discourse (lun). There was already a current derivative of chih
describing the man who can do this, the falling-tone chih W 'wisdom',
which the Mohist writes with the 'heart' radical.
In two series of definitions in other texts there are items comparable
with the present sequence:
Chuang-tzii ch. 23 (Kuo 810/5) ° ft]:£gK (=M) * ° ftl^£0f^
"Knowing is the connecting, knowledge is the representation, what know

ledge does not know is as though peered at."
2/4 269
Hsun-tzu ch. 22 (Liang 310/5-7) § r R * n £ # A # » §B£»1 ° ftlWW^ » I I
"The means of knowing which is within man is called the 'intelligence';

his intelligence tallying with something is called 'knowing'. The means
of being able which is within man is called 'ability'; his ability tallying with
something is called 'being able."
In this passage (although not elsewhere) Hsiin-tzu distinguishes the
graphs for chih as the Mohist did. The parallel distinction between ability
and being able is also assumed in A 8, 13.
2/4/1/2 Action (A7-39)

The second discipline is the art of judging between practical and
ethical alternatives in changing situations, already introduced in Expound-
ing the canons. The definitions begin with the traditional moral terms
(A 7-14) and end with the government of the state (A 34-39); but the
consistency with which the themes of conduct and administration are
sustained throughout only emerges when we sub-divide:
Sequence 1. The virtues (A 7-14).
2. Leader and follower (A 15-21).
3. 'Nurture of life' concepts (A 22-28).
4. Saying (A 29-33).
5. Ruler and subject (A 34-39).
The difference between Nos. 2 and 5 is that the former concerns only
low-level organisation, presumably within the Mohist community (with its
teachers, craftsmen and soldiers). A l l the terms defined except chiian
'scrupulous' (A 16) are common in the military chapters, including the
ssu P\ 'be in charge' which we conjecturally restore in A 15. Although the
Explanation of li 'strength' (A 21) connects with the Mohist interest in
mechanics, it owes its place in the series to being one of the qualities of a
good soldier.
Sequence 3 examines terms common to the Mohists and their polar
opposites, the individualists who preached the 'nurture of life' (yang sheng
at^fe). Mohists and individualists agreed that both self and society are 'put
in order' (chih ?o) by achieving benefit and avoiding harm, but it was
important to the Mohists to detach these concepts from the individualist
principle of caring primarily for one's own bodily needs. The apparent
digressions (A 23 Sleep, A 24 Dreaming) have the purpose of clarifying the
definition of 'life' in A 21.
Sequence 4 contains definitions very important in Mohist logic. But
they are placed here because for Mohists as for others it is by his words
(yen If) and his deeds (hsing ff)> and the agreement of his deeds with his
words, that a man is judged.
As we have noticed in § 1/6/1 above, the whole series is related to
Expounding the canons in the same way that the other four divisions of the
definitions are counterparts of the four divisions of the theses.
Almost all the terms defined in A 7-39 are quite familiar. In the sphere
of moral behaviour the Mohist is concerned only with defining words in
ordinary use; it is only in logic and the sciences that he needs new technical
terms. Two of the terms are corrupt (A 11, 15), but the first at least turns
out when corrected to belong to the ordinary vocabulary of moral discourse.
Only once, in A 16, do we find a word which is at all rare.
The Explanations are the worst mutilated part of the Cations and
Explanations (cf. § 1/2/2/7/3). There are two long lacunae (A 22b-25a,
28&-3la) and several misplaced fragments (A 12, 14, 31, 39).
2/4/1/2/1. The virtues (A 7-14).

86
A 7t » Si îfe o
(t) o « a # » * « f f l B i b o pRg*H) o psgiJijtt o
(it) o ^ X T S ^ 6 8
mmm^mz ° °
A 9 I f ' SH& °
7
(it) o »ia#«»M«â«3K»^jr^ °& o
(Defined: t'i, A 2: lost (§ 2/2): li, A 26)
66
T'i 'individually', adverbial in contrast with chien 'collectively': (p. 265
above).
6 7
These are a misplaced repetition of the last three characters of A 6.
68
Fen 'portion, lot' (cf. E C 9) distinguished by the 'grass' radical, a usage
unattested except here and in A 13 (T'an, Kao, Wu, Liu).
6 9
Here and in the parallel in A 13 the first neng is nominal, the 'ability'
which Hsun-tzu distinguished from being able (p. 269 above).
70
Lun ( M . 35658, def. 12 = $a), modes of behaviour proper to different
family or social relationships. A man's fen is the allotted field within
which he has duties, for example to his own parents (cf. E C 9).
C. " T o bejen (benevolent/humane/kind) is to love individually.

E . Love of oneself is not for the sake of making oneself useful. (Not like
loving a horse.)
2/4 271
C. To be yi (righteous/dutiful/moral) is to benefit.
E. In intent, he takes the whole world as his field; in ability, he is able
to benefit it. He is not necessarily employed.
C. Li (manners/courtesy) is respect.
E. The noble are addressed as 'Sir', the base by their given names, but
in both cases one may be either respectful or rude, because modes of
behaviour are different for different ranks."
The definitions of the virtues are contrastive, selecting the attitude to

other men which distinguishes one from another. A 7 and 8 dispose of jen
and yi with a brevity which would hardly be intelligible unless the fuller
discussions of Expounding the canons were already available. But the two
Canons provide the crucial components in the remarkable system of inter
related definitions on which the later Mohist ethic is built (discussed in
§ 1/1/2/5 above). A l l the words in both definitions are themselves defined
terms, t'i 'individual' in A 2, ai 'love' in the lost document on the 10 theses
(§ 2/2), li 'benefit' in A 26. But in A 9 It 'manners' is left outside the system,
being defined by a word which is itself undefined. Indeed if it were not for
this Canon we should hardly guess that the Mohists counted li as a virtue
at all. Although all parts of Mo-tzu take for granted the existence of rites
and customary rules called li, the chapter Against the Confucians objects
to their over-elaboration by Confucians, and a bare reference to 'lack of
manners' ($Rit, ch. 49, Sun 299/7) is the nearest to a moral use of the word
to be found anywhere else in the book.
Adherence to the principle of chien ai 'loving everyone' assumes moral
value only as the love of individuals (EC 2), which accounts for the defini
tion oijen as 'loving individuals'. In the Explanation one may suspect the
loss of some words linking love of others with self-love, but the point is
clear enough in the light of E C 10. Self-love is one example of loving a man,
and shows that love is for the sake of the man himself, not for his usefulness
to some further end.
The definition of yi in A 8 does not tell us whether the righteous man
is the man who actually benefits others (which would depend partly on
accidents of opportunity), or the kind of man who is beneficial to others.
The Explanation clears up this ambiguity. A man is righteous if he has the
aims and ability to benefit the world, even if he does not achieve high office.
The exactly parallel Explanation of A 13 (Filial piety) makes the same point,
and similarly in A 14 the test of trustworthiness is the words agreeing with
the thought, not with the fact.
The Mohist conceptions of jen and li are not significantly different

from those of other schools. Mencius for example says "The benevolent
love men without exception" (7A/46 fc^ft^pk-fe) and that "the heart
respecting and revering is manners" (6A/6 I5§5c;£'frif fil). But in ordinary
usage conduct was righteous (yi H ) when it was appropriate (yi Si) to the
social relationships of the persons involved. The one-word normative
definition " T o be righteous is to benefit", exposing all ideas of the moral
to a single utilitarian test, must have sounded to other schools more like a
subversive slogan than a definition. It is interesting to compare the
definitions ofy^w mdyi, the former similar but the latter totally different,
in Han Fei tzvi's Interpreting Lao-tzu:
"Jen refers to a man's whole-hearted and joyful love of other men, his
delight in their good fortune and dislike of their misfortune. It is something
which the living heart cannot check, it is not because he expects reward
for it."
" Yi belongs to the service of minister to lord and inferior to superior,
the degrees of father and son or noble and base, the dealings of friends and
acquaintances, the portions of kin and stranger as insider and outsider.
When minister serves lord and inferior cherishes superior, son serves father
and base respects noble, friends and acquaintances help each other, kin and
stranger are treated as insider and outsider, each in the manner appropriate
(yi *&) to him, yi is said of the appropriateness." (Han Fei tzu ch. 20,
Ch'en 329, 330.)
A 10 ft » °
(ft) o > ffik » · *5ib ° (5g«fiE) °

(Analysed: wet, 84)
7 1
The Canon suggests that wei should be level-tone (so wei 'what one does'
cf. A 77) rather than falling-tone (so wei 'one's ends' cf. A 75).
7 2
For shan ming cf. the story in Han Fei tzH cf. 30 (Ch'en 546) of the king
accused of being "too benevolent, too merciful" who answers in
amazement "Aren't they good names?" ( ^ H £ ? P ) .
C. "Hsing (conduct) is doing.

E. What one does, neutrally named, is one's conduct. Giving a good name
to what one does is trickery. (For example, to committing robbery.)"
Hsing is conduct by which one is morally judged, while wei in the

sense of 'do, perform' is morally neutral. The point of the brief definition
2/4 273
is presumably that one's conduct is what one actually does, pace Confucians
who pride themselves on their right conduct without actually doing any
thing useful.
The shan ming ("good name") of the Explanation is intelligible in the
light of A 29-31, which distinguish praising, blaming and referring, and
A 79, where the different kinds of 'calling' (wetffl)include chu IS 'referring'
and chia fifl 'applying'. Chia is used elsewhere in Mo-tzii in combination
with 'good name':
Ch. 28 (Sun 136/-2) * / L f i W j t f c f » ffi*H& » fcffrfe · J&K-ib » · S
"Therefore all who conduct affairs in this way are 'sages' and 'wise men',
are 'benevolent' and 'righteous', are 'loyal' and 'gracious', are 'com
passionate' and 'filial'. This is why one collects the best names in the world
to apply to them."
To judge a man's conduct we need to know what his actions actually
are, referred to by their own names; if he uses moral terms to describe
them, prior to our moral judgment, he is throwing dust in our eyes.
For a fuller definition of hsing, cf. Hsiin-tzu ch. 22 (Liang 310/3f):
Uifljrfnjfl, M£M. jEitrffiSi, t l i l f f "Acting on considerations of benefit
is called 'business', acting on considerations of righteousness is called
'conduct' " .
The illustration, "For example, to committing robbery", may be
related to the argument of the chapters 'Against aggression' (Mo-tzu
ch. 17-19), that military aggression is glorified and called righteous, yet is
no different in kind from the robberies of private individuals which are
universally condemned.
74
A ii (») *m™ > * -tb °
76
A 12 & » Kf§$№3S(ffi)* ffi te *
(Defined: li, A 26: jen, A 19)
7 3
This is plainly one of the many places in Mo-tzti. where shin replaced
ch'eng to avoid a Sung taboo (§ 1/2/1/4), since (1) The definitions in
A 7-14 are of standard moral terms, to which shih does not belong.
Shih in the dialectical chapters is an object to which we give a name.
(2) Ch'eng 'sincerity' pairs very neatly with the chung 'loyalty' of the
1
next Canon. In two cases of a phrase very common in Mo-tzil 4 Jf,
Ji&W 'loyally and sincerely', the original reading ch'eng survives in the
Y i i manuscript (Luan (1957) 155). (3) Shih 'fruit' and jung 'flower' are
common as opposites (cf. Chuang-tzu ch. 14, Kuo 507/2f) and it is
startling to find one defined by the other. (4) The Explanation is
clearly a description of ch'eng, cf. Lii-shih ch'un-ch'iu ch. 26/1 (Hsu
26,2B/5) nm±Mt W » ¥ A ' k # , Mifa "The spirit of a
thoroughbred horse, the high aims of swan and falcon, showing in a
man's heart, are sincerity".
74
Jung 'to flower' refers to what is within a man spontaneously showing
itself outwardly. L i u Shih-p'ei (op. cit. 36) has collected several good
examples: Kuan-tzu ch. 49 (BSS 2/102/1 f) "The
essence being preserved grows of itself; outwardly it then flowers".
Hsiin-tzu ch. 25 (Liang 346/-1) , B J 1 » , "Think, and it will
be refined to the essence, the flowering of intent". Ch. 26 (Liang
357/-5) Iftlîi^-tfa, S S ^ S H f a "It is the essence of the energies of
the blood, the flowering of intent and idea".
75
Chin sheng 'ring of metal' is attested as the ring of bells ( M . 40152/650,
651), although one wonders whether for Mohists, based in the trades
and crafts, it might suggest the ring of money. Yii fu is presumably
fu yii, the jade pendants on a cap ( M . 14345/32 cf. fu def. 7). Its
significance is shown by Shih chi (ch. 56) 2054/2 ¥ £ # . ^ 3 ^ , #pg
;
3 £ 5 , K4*^4& fif'tfe "Although P'ing is a handsome fellow he's only
like the jades on a cap; you can't be sure about what's inside".
7 6
Sun reads ® 'lord'. I prefer to take it as the jen 6: of A 19, since JS is
written as HI on Han inscriptions ( K u Ai-chi 1/49B). L i u Shih-p'ei
(op. cit. 36) avoids an emendation by taking ti as ti ffi; this is attractive
but makes the Canon an observation about loyalty, not even a partial
definition ("Loyalty is being firm in resisting when one judges some
thing to be beneficial").
7 7
The whole Explanation looks like a misplaced fragment. The use of
chiang 'about to' in stead of ch'ieh JL is contrary to the grammar of the
Canons and Explanations (§ 1/3/1). Possibly it is a scrap from Expound
ing the canons, but if so I have failed to place it.
C. " T o be ch 'eng (sincere/whole-hearted) is to reveal oneself spontaneously.

E . The manifestation of his intent and his zeal enables others to know him.
(Not like the tinkle of metal or jade pendants.)
C. T o be chung (to serve loyally/do one's utmost) is to be energetic in
sustaining the responsibility when one deems something beneficial."
E. . . .
2/4 275
The definitions of the two kinds of whole-heartedness are contrastive.

Sincerity shows itself spontaneously, loyal service in persistent effort.
There seems to be a direct allusion to the definition of loyalty in a
Lii-shih ch'un-ch'iu story illustrating how an ambiguous form of words, in
y
which the verbal formulation (tz u SI) diverges from the idea (yi M), can
make an absurd argument seem irrefutable. A man refuses to die for his
lord and makes a sophistical defence (unusual among Chinese arguments
in verging on the syllogistic form):
Lii-shih ch'un-ch'iu ch. 18/4 (Hsu 839/6) J\MK ' T J ^ S W J ife ° °
" A l l service of another is because 'one deems something beneficiaF. T o

die is not beneficial. Therefore I do not die."
The context in the chapter Li wei ('Diverging from what is
referred to') shows that the speaker is not starting from the cynical premise
that one serves others merely to benefit oneself, but shifting the reference
of 'beneficial' from benefit to another over to benefit to oneself.
The dialogue continues:
" 'Can you still see people?' (Can you still look them in the face?)
'Do you suppose I could see them any better if I were dead?'
Right answers took turns from one moment to the next ( S ^ f M ' f f ( = ^ )
cf. A 95 n. 256)."
The narrator comments: "Not dying for one's lord or leader is sup
remely undutiful, yet on the verbal level he was irrefutable ( f t f f f M ^ "JBK)'\
A 13 # ' WBfo °
(Defined: It, A 26)
78
Chih restored from A 8, which is perfectly parallel.
C. " T o be hsiao (filial) is to benefit one's parents.

E. In intent, he takes his parents as his field; in ability, he is able to
benefit them. He does not necessarily succeed."
The virtue which when extended to the whole world isyi 'righteousness*
(A 8) is filial piety when applied to one's parents.
A 14 fit » a ^ S i f e o
7 9
(in) o A » < A29 ± » ^ f s & · ft a £ « » & A
8 si
* z ° > ° (... °)
(Defined: yen, A 32. Analysed: ho, A 83)
79
Hsin 'open', entered in the Shuo wen as a different word from hsin Jft
'cheerful* (Tuan 415B, 507B). There is no example in any concor-
danced pre-Han text, but the Shuo wen provides one from SsU-ma fa
îMlfk (an otherwise unknown passage): Rr3K;£JI
"The good open up what is good in the people, shut up what is bad
in the people". Although one does not expect rare words in the
dialectical chapters, this one is confirmed by the parallelism with A l l
^ S ^ ^ J a L i f e " H i s aim and his zeal being manifested" (in words and
conduct).
8 0
This fragment, unintelligible as the Explanation of A 29, fits very neatly
here. In place of the strange "send a man to look at a city wall and get
gold" it gives us the four-word phrase ftAft^l "enables others to
look at them" (look at his words, cf. Chuang-tzu ch. 18, Kuo 618/If
№№§T"6 "When I look at what you say"), contrasting with A 11
$ ! A *£0B "enable others to know himself", and parallel with ft A
HF;2l "enable others to verify them" (Tu 'scrutinise, verify* is used in
Legalist contexts of inspecting the performance to see that it accords
with the words, cf. Han Fei tzu ch. 48, Ch*en 1029/3 f f f f t f f l
"listen to the words and verify how they are put into practice**). For
the construction pu yi . . . yeh, cf. A 20; long instrumental clauses
placed before the verb are characteristic of the late Mohist grammar
(§1/3/12/4/6).
8 1
The last three words look like the remains of an illustration parallel with
the one at the end of A 11 (which also has the word chin 'metal*, gold').
Cf. A 93 "jump the wall", and the example of going over a wall
(JIT) to "get a coin" (f#*73) in A 75. A possible reconstruction would
be < ^ 3 = f ® >Mt4^ "Not like finding the money when you go over
the wall" (The trustworthiness of the speaker is not judged by whether
his suggestion does turn out to fit the fact).
C. "Hsin (good faith/trustworthiness) is the words agreeing with the

thought.
E . It is not by his words fitting the fact that he makes others pay attention
to them; it is because they are sure of his conduct. The frankness of his
words enables others to verify them. ( . . . the city wall and get the
money.)"
2/4 277
2/4/1/2/2. Leader and follower (A 15-21).

2
A 15 (\%) *m* > SfNfe*
84
A 16 I I > fMffe °
8
A17Bi»f£t#-& «°
87
(«) o atfM8±.»»ft*|M|W| °
A18^-^S.9f№ibo
(^) ° ##$?r ° 88
Erh 'to second, assist' does not fit the definition, and another word
written with the same graph is defined in A 71. Probably a scribe has
wrongly identified the unknown graph SB of A 17, on the assumption
that it has the Shuo wen radical ssu On the principle that any term
defined in the series on behaviour will be a word in ordinary use,
I propose to take it as ssu ^9 'be in charge', written with the 'ear'
radical. (The interchange of ssu/*SloG M , is attested, cf. Chu
Ch'i-feng 0541, M . 10462 def. 4.)
Since all four definitions are in terms of tso 'initiate' W u Yii-chiang is
probably right in taking this unknown graph to be hsiin tfi 'follow'.
Cf. Mo-tzu ch. 39 (Sun 186/11) » r f i P № "Follows but does not
initiate". In support of the present punctuation, cf. Mo-tzu ch. 62
(Sun 344/1) / C + l i l i ( = i ) A S "Meet up with the enemy inside the
tunnel". Jen chung 'the mass of men' is also a common phrase in
Mo-tzu (Sun 55/9,10, 311/11).
One of the few points of general agreement in this obscure sequence is
that Sun is right in taking this unknown graph as chiian 'scrupulous,
squeamish'. Cf. Analects 13/21 M^ffi^^fe "The scrupulous will
not do certain things" (quoted Mencius 7B/37), Kuo-yu (Chin yu 2)
S P T K 8,3B/2 /h'Offl^Ptfcfrtb "He is careful and squeamish and
does not dare to go" (cf. also 8,13A/6, 18,12A/-3).
In the vocabulary of the dialectical chapters the most likely word for the
one corrupted would seem to be hat 'interfere with' (Kao).
'To initiate the wrong' is hardly an acceptable definition of lien 'honest'.
The curious negative definition of 'commanding' in A 18 suggests that
it is designed to contrast with this one, which may have been something
like "(To be honest) is to perform what one has not
initiated" (lien being a virtue of subordinates).
8 7
In the dialectical chapters sui Stt 'although* is generally written tffi, "ft,
and t'o ffiJ 'other* always without the radical (§ 1/2/1/2/12). Chih ch'i
t'o 'know the rest of it* is a common phrase, cf. Chuang-tzu ch. 28 (Kuo
985/-1) SFfcUMfe-fe "I don't know anything else about him". The
one emendation which is certainly necessary is the transposition of so.
8 8
"It is not that he is not himself the agent" (an action one commands is
part of the conduct by which one is judged). For the construction,
cf. Mo-tzu ch. 48 (Sun 291/2) P g ^ , £r>j&ff& "What the mouth
says the person must act on".
C. " T o ssu( ?) (be in authority( ?)) is to act on one's own initiative.

E. When he confronts another, the multitude follows his lead.
C. T o be chiian (scrupulous/squeamish) is to be deficient in initiative.
E. He does the right things; but when doing the right thing interferes
with some other thing, he will not do it.
C. T o be lien (honest/conscientious) is to ( . . . ?) initiates.
E. Although he performs it himself, it is under the authority of someon
who knows the rest about it( ?).
C. T o ling (command) is not to perform what one initiates.
E. It is not that he is not himself the agent."
This is a curiously isolated sequence, and nothing in the rest of the

dialectical chapters throws much light on it. A l l editors heavily emend the
text, and some follow an entirely different track, taking the tso as tso
'ashamed' (Liang, L i u Shih-p'ei, Kao). The present interpretation is
tentative, but undertaken in full confidence that all four words are defined
in terms of tso 'originate, initiate'. The tzil tso 'initiate oneself of the first
definition is too familiar a phrase to admit of doubt (cf. M . 30095/226-230),
and occurs elsewhere in Mo-tzu:
Ch. 7 (Sun 24/2) B 3 t e i ± ^ » XUfrm °
"On the basis of the music of the former kings, they also composed music
themselves.'*
The Mohists strongly contested the Confucian claim that "the
gentleman follows and does not originate" (fJ^tftrfiJ^fF ch. 39, 46, Sun
186/11-187/6, 271/13-272/4). The same claim, in a Taoist rather than
Confucian terminology, is criticised in B 69.
A 19 ' ± » a i f f i £ 0 f JS-fe °
(Defined: sun, A 46: yi, lost? (A 47 n. 131): wet A 75)

2/4 279
C. "Jen (bearing the weight of a responsibility) is an officer working to his

own loss but to the advantage of those on whose behalf he acts.
E. He performs what in his own interests he dislikes in order to bring
about what others need."
Jen is to carry the burden of an office. Cf. Mo-tzu ch. 9 (Sun 33/9)
"If a man of worth found an enlightened lord to serve, he devoted the
strength of his four limbs to bearing the burden of service to his lord
( J ^ B r S ' i l ^ ) , and never tired all his life". The basic reference is to the
sustaining of a load; in Mo-tzu ch. 49 for example it is used of a five-inch
pin "sustaining a load of 50 shih" (Sun 303/1 flEH+S^*).
A 20 m » &£0rKtft-fe °
A 21 ~JJ » 7PJ £0f fiM °

91
(^j) o n±m<-& > 92

°T* 9 3 93a
* ' (*)** 'ife o
(Denned: chih, lost (§ 2/2))
89
Ming 'to name'; only ling ^ is used for 'command* (§ 1/4/22).
90
Yi X hat Y 'let X interfere with Y ' cf. Hsun-tzu ch. 21 (Liang 294/10)
^ G t B r E l $ v § 0 ? $ r S "Not letting what is already stored interfere
with what one is about to receive".
91
Using M . 1886 def. 13), 'shape, body', as in A 22.
9 2
A final yeh should probably be supplied, after the analogy of N O 8
(§ 1/3/12/1/8).
9 3
The radical is generally omitted in chii * 'lift* (§ 1/2/1/2/14).
93a Corrected from the Canon.
C. "Yung (courage) is that by which intent is daring.

E. One names a man 'brave* because of what he does dare; one does not
raise as an objection something else that he does not dare.
C. Li (strength) is that by which the body exerts itself.
E. It is of weight that it is said. Lifting a weight from below is exertion.'*
Courage and strength are defined as counterparts, the two qualities of

the soldier. But the Explanation of the latter concerns rather the concept
of strength as it is used in the sequence on mechanics (B 26), where the
vertical descent of weight serves as the main explanatory principle
(B 27).
In Mo-tzu as elsewhere pu kan 'not dare* regularly implies that one

does not venture, would not be so bold, as to do something open to criticism.
A man's courage is not in question when he does not dare to act against
moral principle.
2/4/1 /2/3. 'Nurture of life' concepts (A 22-28).

A 22 £ »JflJMftlStii o
94
( £ ) ° [ f f i ± £ i S ; ^ "J#vfe ° A 39]
A 23 $k > ftftifcHb *
(jft) · . . .
A 24 ^ » EîfnJ^S^-tfe °
(f)°...
A 25 *P » ^ O t t ^ - f e o
( ¥ ) ° . . . f&f* °
(Defined: chih, A 3: c/wA A 5: wo, A 23: Analysed: yii A 84: aw A 84)
9 4
The Explanation is another fragment unintelligible in the light of the
Canon. It was probably placed here because fortuitously it contains the
word sheng. The area of fragmentation extends from here to A 39,
where there is a suitable place for it.
C. "Sheng (life) is the body being located with the intelligence.

E. . . .
C. Wo (sleep) is the intelligence not knowing of anything.
E. . . .
C. Meng (dreaming) is supposing to be so while asleep.
E. . . .
C. P'ing (calm) is the intelligence neither desiring nor disliking anything.
E. . . . tranquil."
Sheng ^ (be born/live/engender) and its phonetic and graphic

derivative hsing (the nature of a thing) were the key words in the
profoundest of all the philosophical issues of ancient China. It was common
ground that the birth and growth of the living body follows a principle
which is its nature, with which we must accord if we are to live out the
full term of our lives in good health. Man's nature is from Heaven, by
definition, since the test of whether anything is from Heaven or from man
is whether or not it is within our voluntary control. Yang Chu and the
2/4 281
individualists drew the conclusion about 350 B.C. that the way to accord
with Heaven is to nourish our vital needs even in defiance of social
obligations. We have already noticed the metaphysical crisis which this
doctrine precipitated and the significance of the point that Expounding the
canons begins with the admission that appeal to the will of Heaven can
be distorted into an argument for egoism (§ 1/1/1/2, pp. 243-246).
Hsiin-tzu begins the sequence of definitions in his Right use of names
with a definition of nature in terms of life. A similar sequence in Chuang-tzu
ch. 23 goes a step further back by defining life in terms of te 3s8, which is
the Way as it is embodied in each thing animate or inanimate: № l ^ / p f # B
~£ML$& "It is the motions which are without choice that are called te",
&%iWi£.yt-& "Life is the lighting up (the emergence into consciousness?)
of the te" (cf. § 1/1/2/8). The Canons on the other hand refuse to admit
nature into the philosophical vocabulary and define life as "the body located
,,
with the intelligence (the consciousness which will survive it, the ex
istence of spirits being a basic Mohist doctrine expounded in the 'Clarifying
the spirits* chapters: Mo-tzu ch. 31 is quite specific that "the dead have
consciousness" ÊAWftl, Sun 147/7 cf. 145/2). The formula excludes from
consideration all vital tendencies which can be nourished or thwarted, and
makes nonsense of such slogans as yang sheng ii^fe 'tending life* and
ch'iian sheng 'keeping life intact*. The importance of the formula to
the Mohist can be seen in the care with which he adds finishing touches,
in definitions of sleep, dreaming and calm which momentarily distract him
from the ethical considerations paramount in A 7-39. In order to avoid the
objection that by his definition to sleep is to be dead he shows that in sleep
the intelligence is still present although quiescent; although knowing has
lapsed, we still "suppose things to be so** in dreams.
The issue was especially sensitive since there was much in common
between the Mohists and the individualists. Both schools subject traditional
morality to a utilitarian test; for both the setting in order (chih 7P) of
personal and social life is a matter of weighing benefit and harm, sacrificing
a finger to save the arm and the arm to save one's life. Mencius* horror of
the word li ffl 'benefit* reflects his revulsion against all who judge traditional
values by their practical consequences, whether Mohist or individualist.
In the remaining Canons in the sequence the Mohist continues to deal with
concepts shared with the individualists (A 26-28).
A 22-39 is the one great area of fragmentation in the Explanations
(§1/2/2/7/3), and we have nothing for A 22-25 except two words at the end.
The Canons however are fully intelligible, and closely related to each other
and to others elsewhere. In Western terms no doubt the natural form of the
first definition would be "Life is the intelligence being located in the body".
But in Chinese terms a man 'has shape* as he has colour or has
consciousness, and hsing may be verbal, 'assume shape/be shaped*. A man
is alive as long as his shape is located with, shares the same space as, his
intelligence or consciousness (the chih £0 of A 3 which has the place of
mind in the system). The choice of the word ch'u 'dwell/be located* con
nects with A 86, where "all being located in the room" ( ^ i f i i ' i i i ) is the
description of one of the four kinds of sameness, "sameness in being
/
together" (ho n). Shape and consciousness then are the same man in that
they are together as occupants of the same space (not identical, not parts
of one whole, and not of a kind). Both presumably are in the man, as
hardness and whiteness are in the stone (B 37).
A surprising weakness of the definition is that it does not cover
vegetable life. According to Hsiin-tzu ch. 9 (Liang 109/-4) "herbs and trees
have life but not consciousness" ( ^ ^ W ÊlTuMftl), and the Mohist himself
in B 2 treats the living as a wider category than the animal.
The issue, fundamental to the Mohist, of whether an ethical system
can be based on the increase or reduction of life, and be independent of
desire and dislike, is considered in B 44, 45.
A 26 M » 0fftfffiSiii °
(M) ° nmm > m-M№ ° jMriwmtfe °
A 27»»mmm-fc °
(Analysed: wu, A 84)
C. "Li (benefit) is what one is pleased to get.
E. If you are pleased to get this one, this is the beneficial one, and the
harmful one is not this one.
C. Hat (harm) is what one dislikes getting.
E. If you dislike getting this one, this is the harmful one, and the one
which is beneficial is not this one."
These are the foundation definitions in the ethical system, and

righteousness (A 8), loyalty (A 12), filial piety (A 13) and achievement
(A 35) are all defined in terms of them.
It was important to the Mohists to provide definitions which detached
benefit and harm from the concepts of life and nature. In the individualist
chapters of the Lu-shih ch'un-ch'iu benefit and harm are as prominent as in
2/4 283
Mohist thought, but they are the nourishing or thwarting of the vital
impulses:
9
Ch. 1/2 (Hsu 1, 8A/3) » Sf&ttM
"Therefore the sage's attitude to music, beauty and flavours is that if

beneficial to his nature he chooses them, if harmful to his nature he discards
them. This is the way to keep one's nature intact."
Li 'benefit' and hat 'harm' are concepts which easily connect with the
natural tendencies of life since they suggest unimpeded or thwarted
activity; a knife is It 'sharp' if substance offers no resistance, speech is li
'fluent' (cf. N O 9) when it runs freely, and the Mohist himself constantly
uses hat of one assertion interfering with, being inconsistent with another
(cf. § 1/4/12). He detaches them from the concept of life by defining them
in terms of yii ®C 'desire' and wu SB 'dislike', the senses of which he analyses
in A 84 (although, since desire precedes achievement, he has to replace
yii by hsi 'be pleased' in the definition of 'benefit').
A 28 fp » °
9 9 5
([£) ° A# ^Jt*>
(Defined: CAVK, E C 7)
95
Yu(=%) 'again' (Sun), always written with this graph in the dialectical
chapters.
9 6
The syntax of the second clause is obscure, and there is no reason to
assume that the sentence is complete; the lacuna which reaches to the
middle of A 32 may begin at this point.
C. " T o chih (deal with successfully/put in order) is to achieve what is

sought.
E. One's own affairs having been put in order, others also putting in
order North and South. . . . "
The individualist chapters of the Lii-shih ch'un-ch'iu present the

nurture of life as the way not only to 'put oneself in order' (?n#) but to
'put the state in order' (7nIH) and 'put the world in order' (7n?cT). Cf.
ch. 1/3, 2/3 (Hsu l,12B/-4, 2,10B/1). Chih is of course the ordinary word
for right government throughout Classical Chinese, but it is to be noticed
that the Mohist is defining it in relation to the principle of the weighing of
benefit and harm which he shares with the individualists. ChHu 'seeking'
has been defined in E C 7 as "Weighing light and heavy in the sphere of

practical action". To seek is to ponder the means to benefit ourselves and
the world; to succeed (te 'get', contrasted withch'iu as in A 4) is to have set
in proper order both our personal and family affairs and the government
of the state.
Not much can be learned from what is left of the mutilated Explanation,
except that our first duty is to get our own affairs in order. In any case that
would be all that a Mohist of the Warring Kingdoms period would wish
to be judged by. The righteous man has the intent and the ability to benefit,
but "is not necessarily employed", "does not necessarily succeed" (A 8,13).
2/4/1/2/4. Saying (A 29-33).

9 7
A 29 * » M J I 4 o
( f ) o [A 14 Z<J№ffm£*fttA*£] ·
9 9 8
A 30 » BJK * o
(Defined: ming, lost (§ 2/2))
97
Met, and
9 8
O ('beautiful* and 'ugly*), of fair or ugly repute.
9 9
These eight characters are repeated from the preceding fragment, itself
misplaced*
C. " T o yii (praise) is to make plain what is honourable.

E. ...
C. T o fei (blame) is to make plain what is dishonourable.
E. . . .**
These are the first of five Canons connected with saying. In A 79

kinds of 'calling* (wet ffl) include chii 13 'referring* and chia #fl 'applying*.
Chii is the subject of the next Canon (A 31), while praising and blaming
are kinds of chia:
9
Mo-tzu ch. 26 (Sun 126/-3,127/1) & * ^ T S t £ f t l £ ! B £ I B £ <>. . . jft$
9
? T F f f l £ J n ± . o
"Therefore they picked out the most honourable name in the world to
apply to them, and called them 'sage kings*. . . . Therefore they picked
out the most dishonourable name in the world to apply to them, and called
them 'tyrants*.**
2/4 285
A 31 * » « « » 1 № o . . .
1 0 1
<A34 r « 5 J | * f E | < ^ > K ^ * # i b o >
A 32 a » tti*ib<>
1 0 2 103 1 0 4
((»)·* ° *S) 'SWItA » * » f № o ft r g J - t i l * ·
100
Ni, for which there is no neat English equivalent, is to show what X is
like by offering something for comparison. Cf. Shuo yuan, S P T K
12,3B/7 « f f i £ # 0 > r K W I J . . . 0» r « A ^ * t e » i t
^Sffi»felSf S i ^ J "Looking round he pointed at his courtiers and said
'With which of these is your lord equal in height ?\ . . . 'When you
show what a man is like by offering someone for comparison it must
be someone of the same rank. A prince has no equals; there is no one
I can offer for comparison to show what he is like* ".
1 0 1
The Explanation was lost with the rest of A 28b-32a, but this misplaced
fragment is surely a relic of it. It can be corrected from a very close
parallel in A 78 T £ W J » # B U i £ - t b "For 'like the object* one
necessarily uses this name".
1 0 2
Sun emended to chii alone, and since Liang discovered the head char
acters it has been taken for granted that it is the heading of A 31. But
all that follows is certainly from A 32; in the second sentence yeh che
marks yen 'saying* as a quotation from the Canon (n. 104), and ku
'therefore' follows on to the first sentence. The scribe who copied the
badly fragmented A 22-38 restored a perfect chain of head characters
even when the Explanations were lost and he had to write them in
immediate succession in A 23-25 (cf. § 1/2/2/7/3). Evidently he did the
same here, and they were later squeezed together and mistaken for
one graph.
103
Kaoyichih ming 'inform about this name', not 'inform by means of this
name'; the Mohist avoided this ambiguity by always putting instru
mental phrases before the verb (§ 1/3/12/4/7).
104
Yeh che after a single word marks it as a quotation from the Canon
(§ 1/3/10).
105 gu n emended this graph (and also ~S) to £ ; for the grounds for
proposing a systematic corruption of mao, 'characteristics', 'describe',
cf. § 1/4/20. Cf. also § 1/2/1/2/17.
1 0 6
Editors since Pi Yuan have generally taken this as hu 'tiger' expanded
with a radical. This is plausible, although the 'man* radical so favoured
by the Mohists is in this case surprising; possibly the Mohist is deli

berately choosing a rare word for his example of defining a word
(as may be suspected in B 41). But a tiger would be a good example of
a thing most conveniently shown by drawing a picture (cf. Han Fei
tzu's example of the elephant, quoted in § 1/4/37).
1 0 7
For the parenthetic placing of yu shih 'as with the stone', cf. § 1/2/2/6/4,
and p. 138 above.
108
Chih Sfc ('cause to arrive') is used of the speaker successfully conveying
the object to the hearer, and its counterpart chih 35 ('arrive') of the
hearer apprehending the object. Cf. Hsiin-tzu ch. 5 (Liang 58/7) Xfffl
Wifn«iH, -Mi±-MTZm%-& "Elegantly expressed but convey
ing the object, wide in learning but committed to the orthodox, such
is the disputation of the gentleman". Chuang-tzu ch. 17 (Kuo 572/-2)
^Mmm%, VaZfflZ. ^\mm$ y "Those which can be
sorted in words are the coarser things, those that can be conveyed as
ideas are the finer things"; ch. 22 (Kuo 731/5) a P F " J R ,
"The Way cannot be conveyed, the Virtue cannot be arrived at";
ch. 33 (Kuo 1106/2) "Pointing out does not arrive at" (a
paradox of the sophists, discussed in § 2/4/2/5).
C. "Chii (to refer to/pick out by name from others) is to present the
analogue for the object.
E. [Example, 'stone' explained by pointing out a stone?]. . . . For 'like
the stone', one necessarily uses what is like the name.
C. Yen (to say/speak of/words) is to emit references.
E. [Example, 'tiger' explained by a picture?]. . . . To inform about this
name is to refer to the other object. Therefore 'saying' is an emitting of
something's characteristics of which any speaker is capable. 'If the charac
teristics are like the picture, it is a tiger' is saying. T o say that which it is
called (as in the case of 'stone'), is to communicate it."
The long lacuna in the Explanations, which does not end until well
into A 32, leaves us with no more than a glimpse of a basic item in the
Mohist logic, the account of referring and saying. We do however have the
intact definitions of the Canons, which are fully intelligible in the light of
the nominalist account of naming in A 78.
Throughout pre-Han philosophy chii 'to refer' is to pick out an object
from others by name (cf. N O 10 Jil&IPBf "by means of names refer to
objects"). In B 38, 53 referring by name is contrasted with directly pointing
out the object; in A 79 it is classified as one kind of 'calling' (wet »1), and
2/4 287
distinguished from chia }]\\ 'applying', which includes the praising and
blaming of A 29, 30. According to A 78 we name an object (that is, a
particular), and then use the name as the equivalent of 'like the object',
extending it to all which are of the same kind. In B 70 gaining information
about an unknown object by hearing its name is compared to learning what
colour it is by being shown something else of the same colour, or using
a foot-rule to measure an unknown length. Here, in A 31, referring is
defined as presenting the norm or analogue for an object, in implicit
contrast with directly pointing it out. If someone refers to X as a stone,
what he conveys to me is that it resembles the particulars which I know
as stones.
A consistent nominalism has to extend its principle to the particular
utterances of the name itself; I pronounce the sound 'stone' over X and
afterwards convey that Y is like X by pronouncing a similar sound. The
surviving fragment of the A 31 Explanation seems to be making exactly
this point (cf. A 78 n.).
Referring is merely picking out an object by name, saying is making
a statement about it. Although the Mohists did not reach the stage of
distinguishing the sentence from a string of names until Names and objects,
they were already quite clear about the difference between referring and
saying. Trustworthiness is saying being in agreement with the thought
(A 14), judging all saying to be fallacious is fallacious (B 71). The definition
of saying as 'emitting references' suggests the emission of a stream of
references, as implied in the common phrase ch'u k'ou fli P 'being emitted
by the mouth' (A 78 cf. the examples in M . 1811/116), and in any case the
emission of one would be equivalent to referring. The last sentence appears
to treat the emission of one as the marginal case; 'What looks like the
picture is a tiger' is saying, while 'stone' belongs to a category within saying,
the 'saying which calls by name' (wei yen), which is the conveying of an
object (chih %X cf. n. 108).
The proposition about the tiger is too specific to be a casual example
of saying, and must have come in before the end of the lacuna as an
example contrasting with the stone. I will make a guess about what the
Mohist was saying in the missing part. Referring without 'saying' confines
us to the names of objects within our experience; it is 'saying' which makes
available to all the names of things outside shared experience, by explaining
by means of other names and comparable objects which we do know. A
man may not have seen a tiger, but after being shown a picture and told
"What looks like the picture is a tiger", he will receive from the name the
description of the thing. This guess would account for the first sentence
of the fragment, "Informing of this name ('tiger') is referring to the other

object (the picture)", and explain why in the next sentence the understand
ing of descriptions by all speakers is presented as a consequence of what
precedes ("Therefore . . .").
A 33 a » ( & ) * u < M > - * ? M i 1 0 9 °

(JL)° @jTO0&> i « 0 B « J » a » <A 37<S>%»tt> 1 1 0
°
(Defined: A 32)
109 ]y[ t editors have been content to delete one ch'ieh (Pi, Sun, Liang,
os
Yang, Kao, L u , Liu); some indeed find the emendation too obvious
to be worth mentioning. Others prefer to transpose the second ch'ieh
with yen: "With 'to be about to' one says it is about to be so" (Hu
Shih, Chang Ch'i-huang, T'an). But neither solution makes the Canon
2L definition. I prefer to restore tzu ch'ien from the Explanation.
1 1 0
Judging by the strict parallelism of A 35 and A 37 these two words
should be a two-word illustration with the introductory jo missing.
But they are unintelligible in their context and appropriate here.
C. "Ch'ieh ("will/about to") is declaring something to be so before the

event (?).
E . Before the event one says 'about to', after the event one says 'already'.
The just now so too is about to be. (Like 'I'm afraid it will. . .'and 'Let's
for the moment. . .' (1).)"
The emendation of the Canon and the placing of the illustration are
conjectural; interpretation must start from the Explanation and from the
placing of the Canon in the sequence. Why should being about to happen
stand among topics all of which have a direct or indirect ethical significance,
and why in the sub-division concerned with saying? Ch'ieh is the temporal
particle used in discussing a matter of the greatest ethical importance,
fatalism (B 51, N O 16). By defining it as a word put in front ofjan when
speaking before the event (turning 'It is so' into 'It will be so'), the Mohist
removes it from the realm of objects to the realm of names. It is not that
an event is inherently about to happen, and no endeavour can stop it;
ch'ieh implies only that the saying precedes the event. This interpretation
fits B 51, where one is at first sight surprised to find a criticism of fatalism
beginning with the words J L < $ t > i f t $ t , I L E i & B " I f it will be so it is
necessary that it be so, if it will come to an end it is necessary that it come
to an end". But the point, which becomes clearer as the argument proceeds,
2/4 289
is not that inherently (ku H ) the event is bound to happen, but that it is
logically necessary (pi) that if 'It will be so' is true then 'It is so* will
be true.
The statement that 'the just now so too is about to be* may be con
nected with the Mohist conception of the commencement as durationless,
so that there is both a momentary and a continuing time of movement
(A 44). At the moment of commencement the movement is both 'just now
so* and 'about to be\ This was the basis of H u i Shih's paradox that the
sun at noon is also declining and a thing is both alive and dead, accepted
by the Mohist in A 50 (a horse at death is both horse and non-horse,
cf. also A 88).
The illustration, if we are right in placing it here, puts ch'ieh firmly
in the realm of names by very aptly comparing it with two words which
point forward to the future but cannot be supposed to say anything about
the event, tat which expresses a fear and ku which proposes an action. Cf.
Mencius 2A/2 "Let's be done with this", 7B/23 "I suppose
it can't be done again?". Tax seems actually to be used to mark an opinion
as on the plane of names in B 53.
2/4/1/2/5. Ruler and subject (A 34-39).

A 34 31» EWiiîfe o
[A 31 5 B E # 4 ] ( I ) 1 1 1
o
1 1 1
The scribe who restored the head characters after the fragmentation of
this part of the Explanations has planted this one right in the middle
of a misplaced scrap.
C. "The chiin (ruler) is the common knot tying ministers and people.
E. ..."
The sequence on ruler and subject is unexpectedly meagre; only three

out of six have Explanations, perhaps because the rest were lost, but it is
not obvious that they ever existed. There remains enough to show that the
Mohist conception of government is still the fully centralised authoritarian
ism of the Shang t'ung chapters. Shang t'ung $1 ( = J l ) fÎ 'conforming to
superiors' is the subject of the concluding item (A 39) and is assumed here.
According to Mo-tzu ch. 11-13 its purpose is to avoid the anarchy which
results when villages, districts or states recognise duties only to their own
members. A n administrator at any level should 'harmonise the moral
duties' of everyone below him (—'PJ . . . ZML) by imposing conformity.
This conformity, however, is not to himself but to the authority at the level
above him (in the case of the Emperor himself, Heaven), so that he strives
to harmonise his own subjects with those of others in a wider social unity.
From the point of view of the people, conformity is owed to authority at
every level, except when it conflicts with still higher authority (in the case
of the Emperor, with the authority of Heaven). The ruler of a state is
therefore the common bond uniting all his subjects, whether ministers or
people, in the sense that he is the single authority for the morality which
unites them and obliges them to obey a power higher than himself.
A 35 $3 » «
A 36 Jt » _h¥BT£SJte o
(*)·
A 37 H » 3G«tii o
(#1) « » f i M P o [ < ^ > ^ f e i i 4 A 33] [ ± ^ T £ « H f e ] 1 1 5
*
A 38 ft » ± # T ± i № o
(ft) ° [ ± ^ T £ i № ] 1 1 6
«
(Defined: li A 26)
t
1 1 2
Restored by comparison with the strictly parallel A 37.
1 1 3
The Explanation, with four characters already missing and the illustra
tion already incorporated, has accidentally been written in twice.
1 1 4
The jo is restored from the strictly parallel A 35, but this illustration
seems to belong to A 33.
us, ne The Canons of A 36, 38 have been accidentally repeated in the
Explanations, no doubt by the scribe who was restoring the sequence
of the head characters in the dislocated Explanations by comparing
them with the Canons.
C. "Kung (achievement) is benefiting the people.

E. If not at the due time, even if it is beneficial there is no achievement.
(Like summer and winter clothes.)
C. Shang (reward) is requital from above for achievement below.
E. . . .
C. Tsui (crime) is violation of a prohibition.
E. If not covered by a prohibition, even if it is harmful there is no crime.
C. Fa (punishment) is requital from above for crime below.
E. . . . "
2/4 291
The illustration of doing things at the right time is the very common
one of wearing furs in winter and bean-cloth in summer. The evidence for
taking yi as a light coat in contrast with ch 'iu 'furs' is a Mohist slogan noticed
by T'an: "In summer a bean-cloth coat, in winter a deerskin" (JIB J t S ,
Shih-chi ch. 130, 3290/-2).
A 39 IRI » Hfffi« <PI > 1 1 7

ft£—* °
(M) o - A f f i f t J i f i t t U l o o < A 22 ftZtm™ »
-tfe >120 o
(Analysed: yi, A 87: t'ung, A 86)
1 1 7
Although chii is used as a full verb ('associate with') in some texts, in
the dialectical chapters it is invariably adverbial, 'both, all'. I therefore
restore the verb t'ung, on the analogy of a recurring formula in
Mo-tzu ch. 11-13: J8} ( / ± ) I ^ ^ ( / ? ) X / X - ? / a S "Conforming to
Heaven/the Emperor/the ruler of the state who is above". The t'ung
which is defined is marked as a different word by the radical, eliminated
in the Canon but preserved in the head character.
1 1 8
The pillar is the stock example of a thing recognised by comparison
1
with a mental picture (yi M). Cf. § 1/5/9. For the embedded ' Y yeh
phrase after Men 'see' ("See that this is a pillar") cf. p. 155 above. In
general pre-Han usage one would expect this only with the subject
rendered possessive, cf. Mencius 6A/8 AHLftlfelR'tfa "Other men see
that he is a beast".
11» Vulgar graph for til& 'complement' (p. 204 above). For sheng 'engender'
in logical contexts, cf. E C 2, B 53, N O 10.
1 2 0
I transpose this sentence from the other end of the area of fragmentation
(A 22-39); without it, the Explanation would have little point.
C. "T'ung (agreeing/conforming) is being different but both the same in

relation to this one thing.
E. They are two men but both see that this is a pillar. (Like serving a
ruler.) The pillar's engendering of the complement is not to be treate
necessary. (?)"
This t'ung is distinguished from the ordinary t'ung 'same' by the 'man'
radical, eliminated in the Canon by the scribe who standardised the graphs,
but surviving in the head character. The radical distinguished the word as
used in the phrase shang t'ung 'conforming to those above'; this is clear
from the illustration ("For example, serving the ruler"), from the placing
of the Canon at the end of the sequence on ruler and subject, and from the
absence of this important usage from the four senses of the ordinary t'ung
distinguished in A 86.
The aptness of the example of a pillar is only appreciated when it is
recognised that the pillar is the stock example of something known by the
'idea* (yi M) or mental picture that we have of it (§ 1/5/9). The two men
are different, but show that they have the same idea of a pillar by both
recognising pillars when they see them. The Shang t'ung chapters refer not
to conformity in ideas but to conformity in morality (yi H ) . But when the
same doctrine is approached from another direction in the Will of Heaven
chapters we find repeated references to 'ideas' (which in this context are
aims rather than mental pictures):
9
Mo-tzii ch. 27 (Sun 133/-4) W&M °
" T o accord with Heaven's idea is the standard of morality."
Ch. 28 (Sun 136/1, of the sages: 136/5, of tyrants) °
"They shifted their people's ideas over to their own."
Outside Mo-tzii we do find shang t'ung described directly as conformity
in ideas:
sun-tzii (ch. i )
9
1/3B/6 m% &B>HIM& °
SPTK
"The Way causes the people to conform to the ideas of those above them."
The fragment which we have transferred from A 22 is very obscure,
but we may propose a tentative interpretation, taking into account that:
(1) Being an X follows necessarily (pi) from similarity to the object
initially named ' X ' (A 78).
(2) Any necessary relation is between terms of which one is the
'complement' (ti) of the other (A 51).
Consequently agreement implies more than that both men see the
pillar; being a pillar is necessary only as the complement of a similarity,
and the two men might have different names for objects it resembles.
Approaching from another angle, we identify the 'root' (ken Wt) of
N O 6 as the uneliminable core of the sentence (in this case Ying yeh '(It) is
a pillar'). This is engendered by observing the ku iifc, the object as it is in
itself ( N O 10 ^cl№£l#C^fe . . . "The proposition is something which is
engendered by means of a ku . . ."). But the 'engendering' by the sight of
the pillar is unnecessary; there is necessity only when "the complement is
ripened" (*|£5?ft., A 51) by checking its similarity to the standard for a
pillar.
2/4 293
2/4/1/3 Knowledge and change (A 40-51)

A 40-51 define terms for space, time and change, among which one
is at first sight surprised to find the logical terms 'all' (A 43) and 'necessary*
(A 51, defined as 'unending'). When the definitions are co-ordinated and
compared with the parallel series of theses (B 13-16) it can be seen that the
basic theme is the question of what kind of principle is valid for all places
and times, raised by the Mohist rejection of traditional authority (B 16).
The whole series culminates in the definitions of chih i t 'stay' (A 50), the
temporary fitting of a name and a transitory object, and pi*j&'necessary'
(A 51). The first two disciplines, the arts of description and of choosing
how to act, allow only the temporary validity of chih: the two which are
still to come, physics and disputation, establish causal and logical relations
which are pi.
A 40 !K » m m № & o
^ (A) * ( ^ ) * £ 1 2 1
*JL H 1 2 2
o
1 2 3
A 41 ( ^ ) * ¥ » WJI§HJ°
(floats (W.)*m mto iu
(Analysed: yi, A 87)

1 2 1
Most previous editors have emended not the second but the first chin,
either reading ho (Hu Shih, Liang, T'an, Yang, Chang Ch'i-huang) or
simply deleting it (Wang Yin-chih, Sun, Kao, Liu). But the parallel
ism with A 41 throws doubt rather on the second chin, as L i u Ch'ang
MM (176-178) has appreciated. The first chin precedes the head
character, but there is nothing suspicious about this, since the head
character often entered the text after in stead of before the first
character of the Explanation (§ 1/2/2/5/2).
1 2 2
Tan mu ( = ¥ , M . 30178 def. 1) 'morning and evening', as in B 14.
1 2 3
Corrected from the head character.
124
Meng (=H!, M . 1584) 'cover over'. Except for L i u Ch'ang, previous
editors who have made this correction (Hu Shih, Liang, T'an, Yang,
Chang Ch'i-huang) have thought it necessary to transpose the meng
to the head of the Explanation.
C. "Chiu ('duration') is pervasion of different times.

E. 'Present' and 'past' combine mornings and evenings.
C. Yu ('space/extension') is pervasion of different places.
,,
E. 'East and West' covers North and South.
The later Mohists are unique in using for space and time the terms yii
and chiuj*Y>\X}G !K, not yii and chouj*T>'\C>G E£J ; we shall consider th
point in detail in the introduction to B 13-16. Their objection to the latter
pair is apparently that it tends to suggest the 'cosmos as it extends* and the
'cosmos as it endures' rather than the abstract concepts of extension and
duration. The present definitions make it quite clear that the Mohist terms
refer to duration and extension, to 'pervasion' of times and places, not to
'what pervades'. Examination of the other definitions of A 1-75 confirms
that the latter interpretation would require the particle che as in A 61.
(Cf. A 63 n. 167.) The Explanations repeat this crucial point. Duration and
space are pervasions of times and places in the same way that the past or
the present pervades mornings and evenings, that 'East and West' (as a
description of the whole of space divided horizontally) pervades North
and South; they are not entities of another order.
A 42 m J £ 9
125
# 126
M 1 2 7
^§Rft °
125
Huo, 'somewhere' (p. 129 above).
126
Yu ( = X ) , 'again'.
127
Ch'ien when verbal is 'move to a position ahead'. Cf. B 60 and Mo-tzu
ch. 62 (Sun 342/2), of the props set up in digging a tunnel: $i7t{RM
"They go forward stage by stage with the tunnel".
128
Mo, 'nowhere' (p. 129 above).
C. "The ch'iung (limit) is wherever at the next advance there is no room

for a measured length.
E . If somewhere there is no room for a measured length, it is limited; if
everywhere there is room for a measured length, it is limitless."
A 43 ft > m^^M o
(|§)° *№±o
C. "Chin (exhausting/applying to all/all) is none not being so.
E . Something is fixed of all of them."
Of the quantifiers, mo 'none' is taken as primary and is undefined.

Chin 'all' is defined in terms of mo; and later, in Names and objects, huo
'some' is defined in terms of chin (NO 5 T s£ J -fe^PPiHk " 'Some' is not
all").
2/4 295
The dialectical chapters use three words for 'all'; adverbial chieh
in ' X is Y ' sentences, adverbial chiiff:('together, ah") referring back to the
subject in verbal sentences, and chin both as verb ('exhaust, apply to all')
and adverbially referring forward to the object in verbal sentences (§ 1/3/6).
Since chin is syntactically the most mobile it is natural that it should be the
one chosen for definition.
Chih i t 'stay fixed', defined in the same series (A 50), is used of name
and object remaining together throughout the existence of the object. We
may determine an object as being 'circular' or 'square' (cf. A 93 j i l h
"The circle is fixed"), but if we use the word 'all' it is also determined of
every similar object (cf. B 65 ilift, Strife, "Anything of which
they exhaust the characteristics (as in being square) is so of all of the
things").
A 4 4 *& » flfPSte °
C. " T o shih (commence) is to be plumb with the time.

E. Of times of a movement, there is one which has duration and one
without duration. The commencement is plumb with the one without
duration."
This section is to be read in conjunction with A 50. We may speak of

something in advance, using ch'ieh IL 'about to' (A 33). At a certain moment
a thing is plumb with its time and commences to be X and non-Y (A 50).
The Canon does not make it plain that it refers to the moment of becoming
plumb, not to the succeeding period; the Explanation clears up this
ambiguity.
A 45 ft » i ^ f f e o
(ft)°(fSgS)
C. "Hua (transformation) is the distinguishing marks of one thing
changing to the distinguishing marks of another.
E. (For example, a frog becoming a quail.)"
The term cheng Hfc (the tests of what a thing is), especially common in
Hsiin-tzu, does not occur elsewhere in the dialectical chapters. Cf. Lii-shih
ch'un-ch'iu ch. 20/8 (Hsu 20, 29B/-3) *8§«£, B A № t P * T K « &
"Although the distinguishing marks change, although the signs are
difficult, one who is a sage will never waver". Huai-nan-tzu ch. 13 (Liu 13,
13A/7) SfclBAJMfcElKR* "Therefore the sage sees the transformations

and so watches for their distinguishing marks". Cf. also § 1/4/38.
A 46 ft · № H b *
129
(1) · r m J : m±(m) *m * ° X I I < * > " ° # · m%
1 2 9
These graphs are also confused in E C 12.
130 p o r occasional dropping of the second huo of a pair, cf. p. 129 above.
C. "Sun (reduction/loss) is the removal of some without the rest.

E . 'Some but not the rest* : it is a unit in a total. Of its units, if one is
removed and the other remains, we say that the one which remains is
reduced."
Tï and chien (cf. A 2) are countables which may be either member

and class or part and whole.
A 47 [ A â ] *I » Ï J M 6 < - t f a >
1 3 1 132
188 184
«
3
(11) o * ^ * & i H f e o
1 3 6 137
A 48 * « » ^ -tfe °
( * § ) o mn^mm™ » sam <>
At Stage 2 this fragment may either have preceded A 47 or followed
A 94. It is probably, as most editors suspect, from a lost definition of
yi S 'increase' following the definition of sun S 'reduce' (A 46), but
there is no means of reconstructing the missing Canon, which must
have lacked an Explanation.
Hsiian (=38, №) 'circle round'. Although it has the 'man' radical of
characteristic Mohist graphs it is attested elsewhere as a graph for
this word ( M . 1188 def. 5).
Morohashi records only the second of the two characters ( M . 25000
cMA/*TfeR, vulgar form of M . 24959 8£ 'rice newly ripened', which
however is read c/wA/*Tl£G). Sun (200/8) emended both Canon and
Explanation to 4R$£ " A l l are the base" (any point on a circle may be
taken as starting-point). But familiar graphs are unlikely to be
corrupted into rare or unknown ones; a pair of strange graphs sharing
the same radical are much more likely to represent a binome. It may
be noticed that this is the radical which has replaced the obsolete Shuo
2/4 297
wen radical No. 2 2 0 ( ^ , seal form rft, defined as "a tree's bending top
stopped and unable to rise"), and that three of the five graphs under
No. 2 2 0 are scarcely known except in binomes so written only in the
Shuo-wen definitions themselves (^l$Dt, WffiR). The former binome is
elsewhere written with a variety of graphs read chih-chuj*T\1iLG-Y^\\],
* K l £ G - K I U (Chu Ch'i-feng 1282 fltfO, as the
name of a tree with curling branches or as a verb 'to curve round', cf.
Huai-nan-tzu ch. 1 9 (Liu 19, 2 0 B / - 1 ) tt^JS, №%i№ "(The dancer)
soars like a dragon, swerves like a swallow". We may guess that the
two characters in A 4 7 represent a similar binome, perhaps even chih-
chii with the syllables reversed.
1 3 4
This is the only Canon in A 1-75 with the final yeh missing.
1 3 5
(0RJS) When the two systematically corrupted characters are replaced
(§ 1/2/1/4/6, 1/4/20) the Explanations of A 4 7 , 4 8 turn out to pair like
the Canons, both commenting on the tnao 'figure'.
136 This emendation of W u Yu-chiang is confirmed by evidence of
systematic corruption (§ 1/2/1/3/20).
137
Yi 'change round' seems to refer to the two sides exchanging places
( § 1/4/38).
138
Ch'ii hsiieh ('bounded hollow' ?) occurs also in A 63. Comparison between
the two contexts shows that it is a geometrical term, apparently for the
circumference.
139
SsU 'cut'. Ssu 'this' does not belong to the pronoun system (p. 1 1 2
above); indeed it is not found in Mo-tzu at all.
C. " T o hsiian (circle round) is to swerve (?).

E. It is the figure of a curve.
C. T o yiin (rotate) is to change round.
E. The circumference is like a cut, the figure is constant."
The Canons make a pair with contrasted definitions. There are close
#/
connexions with the sections on optics, where we find (!&) ^J 'curve' and
(№)*S 'rotate' in B 19 and yi J?r in the unusual sense of 'turn round' used
of the inversion of the shadow in B 2 3 .
Phonologically J W / I / * G I W 3 N 'rotate' is related to yuanl*GXWAN
M , H 'circle'. In Mo-tzii we find both words used of a potter's wheel
(ch. 35, 3 7 , Sun 1 6 9 / - 2 "a rotating potter's wheel", ch. 3 6 , Sun
1 7 4 / - 2 Atk). Here and in A 8 0 (although not in B 19) it is assumed that
the rotating object is itself circular.
1 4 0
A49№'l£ *ffi-til°
141 142
( » ) ° {B£* *fifcf » »(?62g)*a*S o
140
Huo 'somewhere' (p. 129 above). For huo hsi 'shift in some direction'
cf. B 13.
141 = a j R pien chi 'all over the border' (proposed Sun 213/-2). The first
interchange is well established ( M . 848 def. 17), but the dropping of
the radical in the second graph is unattested. Chi is not found else
where in the dialectical chapters; it normally refers to a border
between, not a border around.
1 4 2
For this correction see § 1/5/6.
C. "Tung (to stir/move) is to shift somewhere.

E. Things which shift all over the border: the hinge of a door, the louse
on a hare (?)."
The Explanation as it stands does not make a sentence and may be

further corrupt. Some editors (Liang, T'an, Wu) make the emendation
(îf)*;§ "Shifting everywhere on the border is like . . .". Its point seems
to be that a thing may be moving although there is no visible change on
either side of its borders, as with the 'rotation' defined in the preceding
Canon (A 48), or the motion of a stationary louse on a running hare.
1 4 8
A 50 i t > E t ^ *o
(jh)oî44^fCih»*r^J r^SJ 1 4 5
o(S*^ 1 4 6
5i®) 1 4 7
°#^
± ^ i h « « r S J T^SJ °(Â5S*)o
(Defined: chiu, A 40)

143
Yi chiu "endure as it was", cf. § 1/3/12/4/5(6).
144
Wu chiu 'durationless': the moment, referring to the commencement
(cf. A 44).
145
Tang niu fei ma "fit 'ox' and 'non-horse' ", cf. § 1/3/5/8.
1 4 6
( * ) cf. §1/2/1/3/13.
1 4 7
Cf. Yi-li MB (ch. 5) SPPY 10, 7A/8 (on archery contests) S i g f f l H
"The shooting is from between the pillars".
C. "Chih (to stay/remain fixed) is to endure as it was.

E. When the one without duration does not stay, the thing fits 'ox' and
'non-horse' (Like the arrow passing the starting-post). When the one
which has duration does not stay, the thing fits 'horse' and 'non-horse'
(Like a man having passed over a bridge)."
2/4 299
This section is to be read in conjunction with A 44, which shows that

'the one without duration* and 'the one which has duration* are times, the
moment when name fits object and the period during which it continues
to fit it.
Although the Canon presents a definition of the ordinary word chih
'stay, stop*, the Explanation shows clearly that the Mohist is thinking
primarily of a name 'staying* in certain objects and not in others, a technical
usage found in A 43, 75, 78, 93, 96, 97, B 1, 2, 32, 68, 82. There is in fact
no example of a non-technical use of chih 'stop* in the Canons and
Explanations except in one of the mechanics sections (B 26. Cf. also the
fragment in A 12 which seems to be displaced from another source). The
point which concerns the Mohist here is that name stays with object only
for a limited duration. A n object may be transformed (hua ft), assume the
distinguishing marks of something else (A 45), and its birth and death are
themselves transformations (cf. Chuang-tzu ch. 22 (Kuo 746/2) BftffnÊ,
XftffflÊ "Having been transformed we are born, being transformed again
we die").
The commencement, the fitting of name to object, is a moment which
does not stay, but the object continues to be X and non-Y until the moment
when it passes from being X to being non-X. It is assumed that at this
moment it is both X and non-X (cf. the comment on A 88). The stock
examples 'ox* and 'horse* are pure logical counters (cf. § 1/5/2); indeed if
we took them literally we should have to ask why what began as an ox
seems to have ended up as a horse.
A 51 & > ^ B i f e o
& "0-ft , S##vfe o

(Analysed: yi, A 76)
1 4 8
( S ) . Ti 'complement* (§ 1/4/29).
149
Shu 'ripe, cooked* ( = ^ ) . P i Yuan*s emendation to chih %K 'hold* is
accepted so universally that most editors forget to mention that it is
an emendation (Liang, T*an, Kao, Liu).
150
Pi pu pi 'necessary or unnecessary* is surely unintelligible in this
context; I therefore propose to emend the first pi to chih i h (which is
systematically corrupted to 'k\ § 1/2/1/3/10).
C. "Pi (the necessary/the certain) is the unending.

E . It is said of cases where complements are 'matured/ (For example,

younger-brother and elder-brother.) Of a thing so in one case, a thing not
so in one case, that it stays as such (?) is unnecessary, that it is this or is not
is necessary."
This is the last of the series on the spatial and temporal conditions of
knowing, and its culmination. According to the preceding section (A 50)
the name 'ox* stays (chih) in the object only for the limited duration of its
existence; but, as we learn here, that while it exists the object is or is not an
ox is 'necessary*. This would be because, if it is like the object initially
named 'ox*, it is necessarily called 'ox* (A 78). The necessary is 'unending',
unaffected by all the varieties of change considered in this series of
Canons; it frees us from the doubt that we are "supposing the already
ended to be so" (B 10, 33 filBfttt).
In ordinary Chinese usage adverbial pi 'for sure, certainly* has no
suggestion of strict necessity. The Mohist himself more than once uses
causative pi of being sure of contingent facts (A 14 'J&Uff "be sure of
his conduct**, B 32 "not sure whether he is alive or dead**), and
both causative and adverbial pi of making sure that one does something
(B 38, 41). He conceives all knowledge as pi 'certain* (A 3), presumably
because if a judgment is wrong one is not knowing but 'supposing* (yi wei
&s%, cf. A 24). This would apply to knowledge by observation or report
as well as by explanation (cf. A 80). But the summa is concerned solely with
knowledge by explanation, and its great innovation is that within this sphere
it recognises no certainty except the necessity of relations between two
terms. Relations (ho &) are classed as exact, appropriate or necessary;
" i f without something else something is necessarily absent, the relation
is 'necessary*.. . . The 'necessary*, accept and do not doubt** (A 83). In the
present Explanation a term implied by another is called its ti 'complement*
(cf. § 1/4/29). The example given is of two-way complements, elder and
younger brother* (mentioned also in A 88), the existence of either of which
necessarily implies the existence of the other. In the sentence "It is said
of complements which are matured**, we may connect shu 'ripe, mature*
with the sheng 'engender* of the obscure Explanation of A 22, "The pillar*s
engendering of the complement is not to be treated as necessary'* (Wii^M
>fCWf^Jl&). We suggested under A 22 that the Mohist thinks of the judg
ment 'It is a pillar* as generated by the sight of the object and matured by
comparison with a standard; the relation between pillar and naming is
unnecessary (in Western terminology, 'contingent*), between naming and
standard necessary.
2/4 301
Pi is generally adverbial, joining two clauses in the formula ' X pi Y ' ,

(If X then necessarily Y). Since the Mohist has not yet arrived at the concept
of the proposition developed in Names and objects, we may presume that
he thinks of all implication as a relation between two names of which at least
one is the complement of the other, like elder and younger brother; "It is
said of cases where the complements are mature" would then amount to
"It is said of cases where implication is strict". With the exceptions just
mentioned (A 14, B 32, 38, 41) all the 62 instances oipi in the corpus refer
either to logical or to causal relations. This is so even of the statement in
A 3 that one "necessarily knows/knows with certainty" (*&£P), in which
the translator has an open choice between the words 'necessary' and
'certain'. Knowing, defined as chieh 'being in touch' (A 5), is a relation
between knower and thing, and Hsun-tzii in his Right use of names uses the
word ho 'relation' of A 83 in defining it (Hsiin-tzu ch. 22 (Liang 310/5)
"The intelligence relating/tallying with something is
called 'knowing' "). Here the thing would be a one-way complement; if it
does not actually exist, necessarily there is no knowing, only supposing.
Where does the Mohist draw the line between the necessary and the
unnecessary? Like Western philosophers before Hume, he takes it for
granted that causal relations are necessary, as can be seen from the many
examples of pi in the scientific sections (pp. 54, 55 above). It seems pro
bable that even there he thinks of pi as belonging to the realm of names;
on the plane of objects if X is the cause (ku #t, A 1, 77) it is inherent in it
(ku H, § 1/4/14) that Y arises from it, and one can infer Y as necessarily
following from X . Apart from causal relations it is surprisingly difficult to
find in over 50 instances implications which a modern reader would refuse
to accept as necessary. " I f when your knowledge is less you refuse to learn,
it will necessarily be little" ( B 69) is presumably to be classed as causal
necessity (cf. also E C 2, "The three things must necessarily be present
together before they are sufficient to generate the enjoyment of benefit
in the world"). In B 48 K±B!li&B " I f you ask about it he will necessarily
say . . . " the question is conceived as allowing only one possible answer;
one may prefer to translate "he is obliged to say", but a mere "he is sure
to say" would be inadequate.
2/4/1/4 The sciences: (Geometry) (A 52-69)

The third discipline is the explanation of objects, by which we
establish the causes (ku #t) of what is so of them. The corresponding series
of propositions ( B 17-31) concern optics, mechanics and economics, but
the definitions are of terms in geometry.
In reading the geometrical Canons it is important to think in terms of

a craftsman measuring pieces of wood (cf. A 64) with a foot-rule, or, in
the case of the definitions of 'straight* (A 57 'aligned'), 'circular' (A 58
'having the same lengths from one centre') and 'square' (A 59 'circuiting
in four from a right angle(?)'), aligning two posts with the sunrise or
operating with compasses or L-square. (Cf. also the definition of a 'limit'
in A 42: 'wherever, at the next advance, there is no room for a measured
length'.) Only if we see the concrete situation can we estimate how far the
Mohist has got in conceiving abstract lines, points and space. In the case
of these definitions it seems clear that he is defining in terms of ideal and
not of empirical operations. Indeed, if we are right in our interpretation
of the Explanations of A 58, 59, he says as much.
The understanding of the geometrical sections has been hampered by
too facile an identification of ch'ih R (the measure approximating to an
English foot or the foot-rule itself, cf. B 27, 70) and tuan S (the end of the
thing from which you start to measure) with the line and the point. Cer
tainly it is immaterial to the Mohist whether or not we are measuring in
feet (in A 42 the limit does not allow room for any further measurement by
the foot-rule, however small), but in A 60, 67 it is relevant that 'a ch'ih and
a ch'ih' are the same in length. We therefore translate by 'measured length',
reverting to 'measured foot' where equality is implied. 'Line' as English
equivalent is especially misleading since it allows us to think of other than
straight ch'ih, for which there is no evidence. A tuan is conceived as
dimensionless and therefore a point; but it is the starting-point (or the
end-point) of a length, and is defined as such in A 61.
It is not however that the Mohist has failed to advance from the concept
of a starting-point to the concept of a point in general. H u i Shih had
already presented the paradox that the 'dimensionless' (wu hou &SJ¥) does
not accumulate, yet a length is the sum of the units into which it is ultim
ately divisible, which are dimensionless (cf. A 55 n.). This is certainly to
conceive the dimensionless as a point which is not a starting-point. The
Mohist on the other hand recognises no dimensionless unit except the
starting-point of a length (A 61) and also no durationless time except the
commencement of a period (A 44). This is not a regress from H u i Shih's
position but an advance from it, which does not allow his paradox to arise.
It suggests Plato's substitution of arche grammes, ('beginning of line') for
stigma ('point'), which he rejected as a geometrical fiction, (as reported
by Aristotle in Metaphysics 1,9/25). B 60 confirms that the Mohist thinks
of lengths as infinitely divisible. At each stage then in the dividing there
will be lengths with dimensionless tuan at either end, and it is self-contradic-
2/4 303
tory to suppose that division will ultimately arrive at a tuan, since it would
have nothing of which to be the starting-point.
In A 63 the tuan is paired with something else which seems to lack
at least one dimension, the mysterious ch'ti hsiieh EH/C, which appears also
in A 48 ('demarcated hollow'). Ch'ti ('the demarcated') is used alone of a
demarcated area in B 22. Hsiieh ('cave, tunnel') is used some 80 times in
the military chapters of tunnels dug in attacking a city, in contexts which
suggest that a tunnel would be an example near at hand of the measuring of
empty space (Mo-tzu ch. 62, Sun 343/3 7tift+R "The tunnel is ten feet
high", 345/-2 J^7tJWTffiKSK "the measurements to be according to
the height and width of the tunnel"). A 48 says of the rotation of something,
presumably circular, "the 'demarcated hollow' is like a cut, the figure is
constant". We may conclude that the ch'ti hsiieh is the outline of a body in
space, with the body abstracted; we cannot be sure that it would be clearly
distinguished from the empty space inside it. It is in fact a cave or tunnel,
with the ch'ii 'demarcated' showing that the enclosing earth or stone is to
be left out of account. This supposition is supported a little later when,
without further reference to empty space, we are told in A 65: (on 'filling')
"Along the measured length wherever you go you find the two". This
surely means that the foot-rule measures both the body and the space it
fills. There is no reason in any case to doubt that the 'space' (yii ?),
defined in A 41 as 'pervasion of different places', is conceived by the Mohist
in abstraction from the bodies which occupy it, although it is hard to
determine how far this is true of pre-Han thinkers less interested in geome
try (cf. pp. 365, 366 below).
Let us return to the concrete situation of the carpenter measuring
pieces of wood with a foot-rule, or the height and depth of a tunnel. He
measures in either case from an edge or from a starting-point chosen by
himself. He is aware that neither the edge nor a starting-point on the surface
adds to his measurement. His first abstractions then are of the outline of a
body (ch'ii hsiieh) and the starting-point (tuan) of a measurement. From
there it is easier for him to proceed to the concept of a point, which has no
dimensions at all, than to the line, which has one dimension but lacks the
other two. But in his concepts of the outline as well as the point, sophists
discern paradoxes. The outline of a mountain can be conceived as the
boundary of the enclosing sky, so that the mountain may be said to descend
out of a hole in the sky in stead of rising out of the earth (cf. A 62-64 n.).
The Mohist, who is both craftsman and thinker, sees that he can avoid the
paradoxes of the point by denying its existence except as the starting-point
of a measurement; he also finds reasons for denying that a body is between
the two ends of a measurement in the same sense that a void is between two
flanking bodies, and declares that "a measured length before it reaches the
outline and after it leaves the starting-point is not flanked by starting-point
and outline" (A 63). But his decision that there are no points but starting-
points will only confirm his presumption that there is nothing one-dimen
sional except the outline. (Lines would in any case raise the paradox of
points in a more complicated form.) Even the straight line drawn by the
foot-rule will be seen as the coinciding edges of the body and the ruler.
When we examine the optical Canons we shall find similarly that he is
thinking of the borders of light and shadow, not of lines or rays (B 17-24).
The remarkable manner in which the Mohist builds up the definition
of the circle from the undefined term 'like', as an exhibition of 'a priori*
knowledge, is examined on pp. 57, 58 above.
A 52 ¥ · NiS-tfe o
1 5 1
A 53 mtk * & * *B«r&°
(P) o «iM||ffi"»^:I^R-lb(^)* JEIM O
(Analysed: t'ung, A 86. Defined: chin, A 43)
151
Cheng (—IE), the first of the 18 examples in Mo-tzu of this graphic
variant. In scientific contexts, when used of edges or surfaces, it means
'even, straight* (§ 1/4/2).
1 5 2
A l l early editions except the Ming Taoist Patrology (which has the
idiosyncratic reading St, presumably a post-Sung error) have this
graph (Lu, Mao, Mien-miao-ko), or gt (Wu M S , Horyaku), the latter
attested as a graphic variant ( M . 15136 def. 4/2).
1 5 3
K'uang S 'square container*, with an unattested or corrupt radical
(cf. A 59). Cf. the description of a machine for shooting cross-bows
simultaneously in Mo-tzu ch. 53 (Ts*en 40): Sun 334/-4 ^ f f i W t t H
"The cross-bow handles are at both ends level with the frame**,
335/2,3 (punctuating with Ts*en 40/5) WSS, W/vd", KBrf, ftftlffi
"There are spurs, in breadth six inches by three, in length equal to
the frame*'. In the present context k'uang would be a door-frame
(T*an, Kao, Liu), a sense later distinguished by the 'tree* radical
(fg M . 14760 def. 2).
1 5 4
Corrected from the Canon. In most other instances of «fr in the Canons
and Explanations it is a corruption of i h (§ 1/2/1/3/10).
2/4 305
C. "P'ing (level/flat) is of the same height.

E. ...
C. T'ung ch'ang (of the same length) is each when laid straight exhausting
the other.
E. The same lengths of door-bar and door-frame are straight/*
The very first Canon is the only one which introduces the third
dimension (kao hsia 'high and low*, B 19, 81), if we except the
definition of the position of the sun at noon in A 56. We should probably
think of the Mohist as a man struggling with the three dimensions of
concrete experience and focusing his attention on two. Thus the geometrical
Canons consider the yuan M, M 'circle* (A 58), not the three-dimensional
t'uan IB, ^ 'sphere, cylinder* of B 24, 57, 62. The Canons as far as A 57
lay the groundwork for coming definitions. The first definition may be
taken as also a declaration that from now on we can forget the third dimen
sion, all measurements will be on a flat surface, in terms of kuang hsiu HtWr
'length and breadth* (B 4).
c 156 0
A 54 + · < @ f a > H l f t - t b
(*)o g f t f t » * B ; g t e o
1 5 6
A 55 № · W W * * o
(»)o'№<«> 1 5 7
№A°
(Defined: t'ung ch'ang, A 53)
1 5 5
Comparison with the definition of a circle (A 58) suggests the loss of
two characters here. The phrase so tzti, 'that from which* is attested
elsewhere in Mo-tzii (cf. ch. 14, Sun 65/1-4).
156 p o r y tax of so ta 'that than which it is bigger*, cf. § 1/3/12/3.
t n e S n
1 5 7
We propose this restoration on the grounds that only the tuan 'starting-
point* is recognised by the Mohist as wu hou 'dimensionless'
(A 61).
C. "The chung (centre) is < the place from which (?) > they are the sam
in length.
E . Distances outward from this are alike.
C. Hou (having bulk/thickness/dimension) is having something than which
it is bigger.
E. Only < the starting-point (?) > has nothing than which it is
bigger."
A 54 is a crucial member of the series leading from the undefined jo

'like* to the definition of a circle in A 58 (p. 57 above). We therefore
translate hsiang jo literally as 'alike', although it is the ordinary way of
saying 'equal' (cf. p. 138 above).
Hou 'bulky' (often used metaphorically of lavish treatment of some
body) is often translatable as 'thick' in other texts, but does not exclude
length and breadth as specifically as the English word. In the dialectical
chapters it is never confined to a single dimension. Wu hou (the dimension-
less, the point) was a central concept of the sophists, as in H u i Shih's
paradox:
Chuang-tzu ch. 33 (Kuo 1102/2) «iJ|CF«J8ttb » JWc=FS °
"The dimensionless cannot be accumulated, yet its size is a thousand
miles."
This has often been taken to refer to a plane of a thousand miles which
is without thickness (cf. Fung (1952) 1/197,198). But the present Canon
shows conclusively that hou J¥ and ta A were not contrasted by the
Mohist. Moreover ta is size as measured by circumference (cf. B 21 n. 341).
In any case the examples of wu hou (A 61, 65, 69) show that it is the point,
just as the wu chiu ('the durationless', A 44, 50, B 15) is the moment. H u i
Shih's paradox is that points cannot be accumulated yet a length is the
sum of the units into which it is ultimately divisible, which are points. For
evidence that in early geometrical texts chi 81 refers to 'accumulation' of
units on the same plane, cf. Solomon 12 n. 10.
A 56 0 * » fiSHfe °
A 57 I S * #-tfe 158
°
158
Chih 'straight' (never 'upright' in the dialectical chapters) is used of a
rope, the path followed by light, the direction of a pointing finger,
while cheng IE refers to the even or straight edge of a body (§ 1/4/2).
Ts'an 'align' (also B 38), a word used of aligning posts with the sun
in a procedure for determining the East-West axis, cf. Huai-nan-tzu
ch. 3 ( L i u 3,32B/-1) ^9 > » »
J ^ ^ M H f e t B J b K " T o determine sunrise and sunset, first plant one
post on the East side, retire ten paces behind the post carrying another
post, and align it with the North corner( ?) of the sun as you observe
it as it first rises".
2/4 307
C. "Jih chung (the sun at the centre/noon) is the sun being due South.
E. . . .
C. Chih (straight/on a straight course) is in alignment.
E. . .
'Straight* is defined as 'aligned* (ts'an), the standard term in Chinese

astronomy for the aligning of two gnomons with the observed heavenly
body. This definition is similar to Plato's. Cf. Heath 165,166 "The only
definition of a straight line authenticated as pre-Euclidean is that of Plato,
who defined it as 'that of which the middle covers the ends' (relatively, t
is, to an eye placed at either end and looking along the straight line). It
appears in the Parmenides 137E: 'straight is whatever has its middle in
front of (i.e. so placed as to obstruct the view of) both its ends'. This
definition is ingenious but implicitly appeals to the sense of sight and
involves the postulate that the line of sight is straight". Euclid's obscure
definition " A straight line is a line which lies evenly with the points on
itself" is apparently an attempt to free Plato's from its implicit appeal
to vision (Heath 168).
The first Chinese dictionary definition is: Ifi, jEJJUil " 'Straight' is
seen evenly" (Shuo wen ch. 12B (Tuan 640A/9)).
A 58 HI »— + l * l £ l k °
A 59 ft ' ( t t ) * K l t t 0
PSEg(li) «
# l t i i
'tb o
(ft) ° № H £ l f l 2
te °
(Defined: chung, A 54: t'ung ch'ang, A 53)
159, 162 P ' / * P ' U K is unattested except as a radical, itself composed of a

W
phonetic />«/*PUK h with the 'right hand' radical underneath, and

identified in the Shuo-wen (ch. 3B, Tuan 123B) as the />'w/*P'UK
'beat lightly' for which the standard graphs are St, (the latter with
the same phonetic and the 'hand' radical). The unknown graphs in
Mo-tzu are the ones which deserve the most confidence; there is a
very strpng presumption that this one is not corrupt, but either repre
sents an unknown word or is an unattested interchange with a known
graph, presumably with the phoneticpu V . The only suggestion known
to me is that of L i u Ch'ang (op. cit. 3,6A), that it is />'w/*P'tJK Wk, #
(the latter graph interchanges with Jh, M . 14428 def. 1/7). Although
this word when distinguished by the 'tree* radical refers specifically to
an uncut block of wood, it has the general meaning of a thing in its
rough state, uncut jade ( H , £r), a clod of earth (SI, *r), or simple,
uncultivated behaviour (cf. Mo-tzu ch. 36, 37, Sun 176/-2, 179/5
^ i l i m "the multitude of the foolish and simple"). The last usage is
, ,
not necessarily a metaphor from wood or jade; the word /> w/*B UK
St 'servant', which goes back to the Book of Songs, presumably derives
from it. Whether L i u Ch'ang's proposal is found acceptable or not,
Sun's emendation to 3£ (followed by most editors) goes against the
principle of preferring the lectio difficilior; how could a scribe twice
mistake a familiar graph for one he had never seen except as a
radical ?
1 6 0 r #
^ W w § / K ' I W A N G 'square container' or 'to square up, square off',
corrupted to in A 53. K'uang was part of the name of a Sung
Emperor; the Ming Taoist Patrology still taboos it, and in graphs where
it is phonetic drops the bottom stroke (§ 1/2/1/4/2), so that it would
be especially liable to corruption. L i u Ch'ang (op. cit. 3,6A), who is
responsible for this proposal also, notes that in the li~shu script
zvangl*'XWANG ft 'bent' has a form written with the vertical stroke
rising above the top horizontal stroke ( t £ K u Ai-chi 3/57B, 58A), so
that it would be easily mistaken for chu ft (the graphs are in fact
confused in Mo-tzu ch. 71, Sun 387/5). There are instances elsewhere
both of the graph E for wang 'bent' and of the graph ft for k'uang/
* G T W A N G ffi 'mad' ( M . 2606 def. 1/7, 14530 def. 2/2). K'uang yii
would presumably be 'a squared corner' (a right angle), although I know
no supporting examples. ¥or yii 'corner', cf. Mo-tzu ch. 52 (Sun 313/4)
"The city wall has four sides and four corners" ch. 71
(Sun 381/3) ¥ H H "The lookout posts are three-cornered".
1 6 1
Tsa 'make a full circuit' ( M . 42122 def. 7 = 7fT), as in Mo-tzu ch. 70, 71
(Sun 374/-4 H · ft "three times circuit", 381/13 fflt "twice circuit").
This emendation of Sun has firm support: Shuoyiian S P T K 19, IB/6
%mA±mm% *n№£(H)» \m, fi±s», WIOR**
"Therefore the sage in his sagehood is like the carpenter's square
circuiting in four, like the compasses circuiting in three, when the
round is completed they begin again, when they reach the limit they
return to base". This assumes a straight line taken as one side when
using a carpenter's square and as diameter when using the compasses.
Cf. Chao Ch'un-ch'ing £§f!#P (c. A.D. 200), commenting on Chou
peisuan ching JS#?JHS ( S P T K ) A , 2A/4 I B « - r F n J S H , rfnTTO
2/4 309
"The circle has a circumference three times the diameter, the square
has a circuit four times the diameter".
C. "Yuan (circular) is having the same lengths from one centre.

E. The compasses draw it in the rough (?).
C. Fang (square) is circuiting in four from a right angle (?).
E. The carpenter's square shows it in the rough (?)."
The second definition is corrupt. The present interpretation is the

first which combines Sun Yi-jang's emendation of the fifth character with
L i u Ch'ang's of the second (Liu Ch'ang himself took the fifth character as
kuan H 'observe'; " A door-frame and its corners observed four ways").
When combined the emendations can be seen to support each other,
making the Canons exactly parallel like the Explanations.
The circle is the stock example of perfect agreement with a standard
(§ 1/5/10), and like the ethical terms is established as 'a priori' by an
elaborate system of preliminary definitions (p. 57 above).
Both Explanations hinge on the unknown word/>'w which no one
has attempted to identify except L i u Ch'ang. He takes it as p'u tH, 'wood
(or jade, or earth, or a man) in its crude state', and interprets A 58 as "The
compasses draw on rough wood" (which is later carved into a round vessel).
But if he is right in identifying the word, the meaning is surely that com
passes and carpenter's square draw only rough approximations to the true
circle and square. This is not too Greek an idea for the Mohist, for he
certainly does have some of the Greek's fascination with the intellectual
precision of the square and the circle (cf. A 80, 93, 98). Moreover there is
wide agreement that this is the point of one of the paradoxes of the sophists:
Chuang-tzu ch. 33 (Kuo 1106/2) > S R ^ K H B °
"The carpenter's square is not square, the compasses cannot make a circle."
(Cf. Fung (1952) 1/218.)
Although fang is used characteristically of the true square it can be
applied to all rectangles, and the definition as we understand it does not
specify that the sides are equal in length.
A 60 ft » * o
( f t ) o -163 , ^ j j j £ ^ Jg , 164 -ftfctpub o
1 6 3
Cf. B 74 —, " I f they are two, we know their number".
1 6 4
Editors ever since P i Yuan have taken tuan as the heading of A 61. But
this assumption, which breaks an intelligible sentence into two
unintellible ones, is quite unnecessary (cf. § 1/2/2/5/6). Ch'u 'depart

from' as in Mo-tzu ch. 56 (Sun 338/1) "10 feet from the
wall".
C. "Pei (doubling) is making two of them.

E. When they are 'two', measured foot and measured foot both depart
from one starting-point, which is being nowhere the same."
'Straight* was defined in A 57 as 'aligned*. Wherever the Explanation

introduces two measured lengths their alignment is assumed (cf. A 67, 68):
A B
c —5
This figure is in the background of all the remaining Explanations, with
A B and C D sometimes touching and sometimes overlapping or separated
by an interval.
A 6i m»mz№*m™mmm%& °
(Defined: t'i, A 2: hou, A 55)
1 6 5
(Jf). Since in most editions this important correction is presented as
though on the same level as their plausible or implausible conjectural
emendations, it is worth repeating that this is an example of systematic
corruption throughout Mo-tzii (§ 1/2/1/3/14).
C. "The tuan (starting-point) is the unit without dimension which

precedes all others.
E. . . . "
This definition makes it clear that tuan is the word, not for the point
in general, but for the starting-point. It can also refer to the end-point;
whereas in English a stick has two ends, in Chinese it has two beginnings
cf. Mo-tzu ch. 63, Sun 349/8,12). The 'starting-point of a measured
length* appears as one example of a t'i 'unit' in A 2.
A 62 WfS » < ^ R > 1 6 6

*-feo
A 63 H 1 6 7 9
^PR^-tfe 0
° m <m > 168

o R w ^ E T t ^ MmWi» î&mmm
2/4 311
1 7 0
A 64 I I ' Rflffiib ° 171
( « ) o r a J ft* : m*±rs>awyii*#fe °
(Defined: chung, A 54)
1 6 6
Most editors try to force a meaning out of the text as it stands, which
is surely nonsensical ("Things with an interval between them are in
the middle"). W u Yu-chiang plausibly takes chung in the sense of
'empty* (Si), but A 64 shows that the interval is not necessarily empty.
In any case the Canons of A 62, 63 are parallel, with chung 'centre*
corresponding to p'ang 'sides* (cf. A 88 4*:$:, ^ 'centre and sides*).
The parallelism suggests the restoration of pu chi. The chi chi of A 63
will then be a reference to both Canons.
1 6 7
Falling-tone chien 'intervene, be between* (cf. Mencius 2/13 RUS^/iS
"It is between Ch*i and Ch'u"), as is shown by the absence of a
nominalising che # in the definition (cf. A 41, comment) and by the
parallelism with yu chien 'have an interval*.
1 6 8
The contrast with the parallel sentence in A 62 seems to require this
restoration.
169 p g
o r i f taking ch'u hsiieh (also in A 48) as 'outline*, cf.
t n e r o u n ( s or
p. 303 above.
170
Lu ( M . 28054), primarily 'thread', has no known sense which fits the
definition. The phonetic / M / * L O tends to be associated with some kind
of hollow container (J§ 'food vessel', J& 'house*, '№ 'stove*, ® 'skull').
1 7 1
The word marked off as a quotation from the Canon by yeh che is very
loosely exposed, as in A 3-6. Cf. p. 116 above.
C. "Yu chien (having an interval/discontinuous) is <not extending t o >

the centre.
E. It refers to the flanking ones.
C. Chien (intervening/in between) is not extending to the sides.
E. It refers to what is flanked. Lengths measured from starting-point to
circumference are not flanked by starting-point and circumference. The
two extensions are extensions of which it is not the case that they come out
level with each other.
C. Lu ( . . . ? ) is the interval being empty.
E. 'Empty': of the interval between the two pieces of wood, it refers to
where there is no wood."
The model according to A 64 is two blocks of wood with an interval

between them, which may or may not be filled by a third block:
The essential point for the Mohist is that an interval must have bodies
'flanking* it. There is no interval if the two blocks 'extend to the centre* :
The space is not an interval if it 'extends to the sides*, without bodies

flanking it:
In other words, the two extensions must not 'come out level with each
other* (ch'i jSf. Cf. the ch'i in Mo-tzu ch. 53 quoted in A 53 n. 153 above).
In the last case we are measuring from the centre to the circum
ference of a circumscribed space. We are not measuring an interval, because
the two measured lengths are not 'flanked* by their starting-point and the
circumference, they could only be flanked by bodies.
It would seem that there must have been sophists who built paradoxes
on the assumption that a body can be regarded as filling intervals between
points on its surface. One of the sophisms of the Second List in Chuang-tzu
ch. 33 is a probable example, "Mountains come out of holes** (UjfcB •).
One could think of the surface of the mountain as a hole in the sky, so that
it descends out of the hole in stead of rising out of the earth.
2/4 313
A 65 £ » J t P W i i i o
A 66 » WB*Mb o
(S) o o*B*iwd6ffi^* °
1 7 2
7e accidentally repeated.
1 7 3
The syntactic mobility of fei 'is not* in the grammar allow hsiang fei
'are not each other* (pp. 117, 118 above).
C. " T o ying (fill) is to be nowhere absent.

E. What does not fill anything is dimensionless. Along a measured length
wherever you go you find the two.
C. Chien pai (as hard to white/mutually pervasive) is not excluding each
other.
E. Different positions do not fill each other. Not being each other is
excluding each other."
Faith in the authenticity of the essay on Hard and white in Kung-sun

Lung tzu (one of the three essays forged between A.D. 300 and 600) has
made it impossible for previous editors to deal with these two sections
without arbitrary emendations and forced interpretations. But chien pai
is a technical term (mutually pervasive like the hardness and whiteness of
a stone), and was understood as such at least until the Han dynasty
(§ 1/4/3). If the Canon of A 66 were a proposition repudiating Kung-sun
Lung's separation of hard and white ("Hard and white do not exclude each
,,
other ) it would not be among the geometrical definitions, it would not be
among the definitions at all.
Different positions do not fill each other, by the mere fact of not being
each other. But "where the hard occupies the white they necessarily fill
each other" (B 15). Different qualities of the same object are however not
the only examples of mutual filling, and in A 65 they are not yet differenti
ated from the filling of a demarcated space by a body ("What does not fill
anything is dimensionless"). Along the edge of the foot-rule every position
is both a position on the object and a position in space (and is also, on a
stone, both hard and white). For the filling of space, cf. B 73 ( S M I ?
"fill the limitless").
174
A 67 81» *B# -ifa o
( « ) o RHR&ftm » «(^)*JS 1 7 6
i f f i i * « * » RS3<S >™l№&fi
174
Te 'get' in geometrical contexts is to occupy the same spatial position
as something else, cf. A 65, B 15.
1 7 5
Emended on grounds of parallelism (Sun). There is the same corruption
of conjunction yii (written ?) to zvu (written 3t ?) in ch. 48 (Sun
288/12), cf. § 1/2/1/2/14(3).
1 7 6
Restored on grounds of parallelism (Sun).
C. "Ying (touching/coinciding) is occupying each other.

E. Of measured feet, neither is wholly covered by the other. Of starting-
points, each is wholly covered by the other. Of the measured foot and the
starting-point, one is wholly covered and the other is not. The hard and
the white in coinciding wholly cover each other. Countable units in coincid
ing do not wholly cover each other."
Ying is ordinarily used of one body entering the space occupied by the
other, cf. Mencius 7B/23 (of a tiger) l££fiScS5 " N o one dared get in its way".
The Mohist uses it only of the coinciding of measured lengths (A 67-69).
Of two measured feet neither is wholly covered by the other, although
coincidence may be nearly complete:
A B
C D
Here it is especially important to remember that ch'ih are not lines but feet,
and therefore equal in length (AB and CD). If the equality were irrelevant
we could object that A D does contain the whole of B D . It is also important
that for the Mohist there are two ch'ih only if there are two measurements.
If we move back C D until it coincides perfectly with A B we have only one
ch'ih, so that it is still true that "of measured feet, neither is wholly covered
by the other".
The end-point of A B and starting-point of B D fully coincide. A is
contained in A B but A B is not contained in A . What is hard and what is
white in the body coincide all the way from A to D . Finally, the units (t'i)
which we count, parts of particular objects, the objects, their classes, will
never coincide except at touching borders. In terms of our diagram, the
units A C , C B and B D cover each other only at B and C. They are of course
fully covered by A D , but that is the chien i s 'collection, total', although it
may itself be treated as a unit at another level of counting (cf. A 2 n.). We
might object that according to A 2,61 starting-points are themselves t'i
2/4 315
(and indeed the only t'i which are not also chien, since they are not further
divisible). But coinciding starting-points count as one unit (cf. A 60 "both
depart from one starting-point"); it is the opposite ends of a measurement
which count as two (cf. A 61 п.).
177
A 68 ( Ю ) * № > #ШП« ' #*«1й№ о
1 7 8 179
[й] ( № ) о МЯ№ Ш)5~*1 о
1 М
А 69 # » Ш Ш * ( » ) * № Й Н Ь о
ш # 1 8 2 б
(*)° < й ! > ш ^ | Ш ; и " 1 о
(Defined: jyi/ig', А 67: chien, А 63)
1 7 7
Emended from the head character. Ssu 'resemble' is not used in the
dialectical chapters, which employ jo 'like' even in the negative, pujo
^РЗёг 'not like' (§ 1/3/8). Pi #L 'lay side by side, compare' is used both
of measurement (A 88 ttlffi, ^ ^ i f e " O f putting side by side and
measuring, 'more and less' " cf. Chuang-tzu ch. 2, Kuo 49/-1, iilYS
'the tubes side by side', the pan-pipes which are parallel but unequal
in length) and of comparing statements. The 'man' radical which
distinguishes the former here, and the 'mouth' radical for the latter
in В 6, may originally have been used consistently (§ 1/2/1/2/5).
178
Tuan appears to be missing both in A 67 and in A 69, and this misplaced
character may be either of them.
179
Yu tuan 'have a starting point'. Cf. A 1, В 19. Variants, p. 75 above.
1 8 0
Corrected from the Canon of A 68.
1 8 1
Parallelism with the Explanation of A 68 suggests a missing character.
We supply tuan (possibly the misplaced tuan in front of A 68) because
nothing else in the system of geometrical concepts is described а в ш
hou 'dimensionless' (A 61).
1 8 2
(I?-). A systematic corruption which is any case confirmed by the
parallelism with A 68 (§ 1/2/1/3/15).
C. "Pi (side by side/measured against each other/commensurate) is over

part of the length coinciding and over part of it not.
E. It is possible only if both proceed from a starting-point.
C. Tz'u (the next/adjacent) is without interval but not coinciding.
E. It is possible only because the starting-point is dimensionless/'
The model is again, as in A 60, two aligned measurements with a

meeting-point:
A B
C D
A B and A D are commensurate, because A B and A C coincide. But are not
A B and A C also commensurate, which conflicts with the requirement that
commensurate measurements only partly coincide? No, because there is
only a single measurement, not comparative measurements (cf. A 67 n.).
C D is next to A B , they are adjacent. They do not coincide except at
B/C, the common starting-point which is dimensionless.
2/4/1/5 Disputation (A 70-75)

The fourth discipline is the explanation of names; by the elucidation
of a name (typically, the name 'ox'), we establish that something is this
thing or is not (shih fei -jk^). This is the only one of the four disciplines
of which we can say with any confidence that we know what the Mohist
called it; it is the pien 'disputation' defined in A 74.
A 70 & ' W^ffi^-tfe o
A 71 ( ^ ) * H 1 8 4
» ©f^ ffe o
185
183
Chu 'all', not 'together' (p. 129 above).
1 8 4
For the grounds for this emendation, see § 1/4/39. The graphic confu
sion suggests that yin was written with the 'man' radical characteristic
of Mohist technical coinages.
185
Sojan 'that wherein it is so' (cf. § 1/3/12/3).
1 8 6
( S ) cf. § 1/4/20.
C. "The fa (standard) is that in being like which something is so.

E. The idea, the compasses, a circle, all three may serve as standard.
C. The yin (criterion) is that wherein it is so.
E . Being 'so' is the characteristics being like the standard."
T o judge whether something is circular (A 70), is square (B 65), is a

black man, is a lover of mankind (A 96) we compare it with a standard.
Different standards may serve to judge one thing, which may be recognised
as circular by comparison with the idea of a circle, with the compasses or
with an actual circle. Similarly the same standard may apply to different
things, for example square pieces of wood and stone (B 65).
2/4 317
In the case of a circle agreement with the standard is exact (A 98).

But there are also cases where agreement is only partial, and we have to
decide whether the criterion (yin) forjudging whether or not someone is
a black man or a lover of mankind is to be the part of him which is black
or the part which is not, the people he loves or the people he does not
(A 96, 97).
The idea (§ 1/4/37) is the thought which words convey from speaker
to hearer (A 14, B 38, 41, N O 9, 10), apparently conceived as a mental
picture (B 57, 58, N O 3). The idea of a circle would be that embodied in
the words of its definition, "having the same lengths from one centre"
(A 58). Presumably all the 75 definitions lay down standards for
names.
It may be noticed that in the Explanation of A 71 the combination
yeh che marks jan 'so' as a quotation from the Canon (cf. § 1/3/10). The
sentence is not a true definition (its relation to the definition of the standard
in A 70 is in any case circular) but an observation about the meaning of
jan as used in the definition of the criterion.
Expounding the canons once refers to "the standard laid down by the
sage" (EC 9), but the Canons, Explanations and Names and objects use
exclusively as a logical term. For the model set by the sage they use yi
m = m (B 53).
A 72 S i i f r & M l o
9
(Defined: ming, lost (§ 2/2))

C. "Shuo (explaining/demonstrating) is the means by which one makes
plain.
E. ..."
To 'explain* (shuo) in the dialectical chapters (EC 1, A 80, 98. B 1, 2,

32, 44, 66, 70, 82. N O 10) is nearly always to offer proofs; indeed there is
no other word for demonstration in the vocabulary. Cf. N O 10 J^lfttBSfc
"by means of explanations bring out reasons", and the phrase yujzvu shuo
'ff/^StSi 'have proof to offer/have no proof to offer' discussed in § 1/4/40.
The word does not distinguish between clarifying the meaning of a thesis
and making it plain that it is true, and the 'Explanations' of the Canons
themselves serve both purposes.
A 73 *i& 187
^ "J^ ^ oRfe o
9 1 8 8
(*f£) o № ffi 2№ Mife o №&L

9 189 90 9 1919
2№ o
1 9 2
A 74 M » ° MB ' m * °
o( ^ 1 9 3
t T A J ) o-
1 8 7
Corrupted to ifo in both Canons and to $E in both Explanations. The
grounds for identifying as the technical term /aw which survives in
B 30, 72 are explained in § 1/4/9. Previous editors have generally
followed Sun in accepting the pi 'that' of the Explanation, at the cost
of denying that the Canon presents a definition.
188
Liang 'on both sides' cf. § 1/4/18.
189
Ch'ii 'group separately' disguised by a mistaken radical (§ 1/4/7).
190
Fei niu 'non-ox' (§ 1/3/5/8).
191 Wu yi 'lack what distinguishes X ' (§ 1/4/35).
1 9 2
There is no other Canon in A 1-75 with more than one definition.
Possibly there were originally two Canons run together at Stage 2; but
the Explanation (written at Stage 1) seems to cover both definitions
and to be indivisible.
193 p j < t like'; the ordinary pre-Han sense 'not as good as' (preferred
u 0 no
by Leslie, op. cit. 12) is foreign to the language of the Canons (pp. 138,
162, 163 above).
C. "Fan (being the converse of each other) is if inadmissible then on both

sides inadmissible.
E. A l l oxen, and non-oxen marked off as a group, are the two sides. To
lack what distinguishes an ox is to be a non-ox.
C. Pien (disputation) is contending over claims which are the converse of
each other. Winning in disputation is fitting the fact.
E. One calling it 'ox' and the other 'non-ox' is 'contending over claims
which are the converse of each other'. Such being the case they do not both
fit the fact; and if they do not both fit, necessarily one of them does not fit.
(Not like fitting 'dog'.)"
The converse of applying the name 'ox' to one kind of thing is that
all other things must be called 'non-oxen'. What makes them the converse
of each other is that neither 'This lot are oxen' nor 'The other lot are
non-oxen' can be denied without both being denied. As the Explanation
of A 73 notes, it is essential that the combination of ' X ' and 'non-X' is
all-inclusive, so that everything which lacks the characteristics of X is
classed as non-X.
2/4 319
Contending over whether something is X or non-X is pien M dis

putation', a word intimately related to pien 'distinguish between,
discriminate' (cf. Mo-tzu ch. 11, Sun 47/-3 ~MLW$W*ZM "distinguishing
between right and wrong, benefit and harm", a phrase which recurs in
ch. 35, Sun 170/3 with the other pien). For the Mohists, as for the sophists
(pien che 'disputants') and for the anti-rationalist Chuang-tzu, pien
is the regular term for rational discourse, for disputation with the purpose
of distinguishing between right and wrong alternatives. The essence of pien
for Mohists is that the alternatives should exclude any third possibility, so
that one necessarily 'fits' (tang '©') the object and the other necessarily does
not. (Pi *j& 'necessarily' is itself defined in A 51.) As we learn from B 35,
it is not pien to argue whether something is an ox or a horse (it may be
neither), a whelp or a dog (it may be both). The illustration to A 74
("Not like fitting 'dog' ") refers to the latter example of imperfect
disputation, the dog being the stock instance of objects having two names
(§1/5/3).
It is to be noticed that although disputation might be described as
argument over contradictories ('It is an ox', 'It is not an ox'), this is not
the angle from which the Mohist approaches it. He sees 'ox' as a name
which divides all objects into two classes, and 'They are oxen' said of one
as the converse of 'They are not oxen' said of the other. The issue is not
whether or not X belongs to the former class, but which class it belongs to.
It would be decided (like the issue of whether someone is a black man or
loves men in A 96), by comparing the object with a standard (fa 7&).
Judging by the alternative standards for a circle mentioned in A 70,
comparison could be either with an actual ox or with the idea of an ox,
presumably as embodied in a definition.
In B 66 we are told that the ox having and the horse lacking horns is a
respect in which "the kinds are dissimilar" (iHÎÔ) but is not "the
dissimilarity between the kinds" (SHÎÎRI). Evidently the characteristics
which distinguish the ox were clearly formulated, but they are nowhere
mentioned in the corpus. (But cf. p. 438 below.)
The defense of disputation in B 35 confirms that the pien of the
Canons is something narrower than argumentation in general, its usual
meaning in pre-Han literature. It Would seem that after conceiving the
idea of logical necessity and distinguishing the four branches of knowledge
(§ 1/1/2/2) the Mohists came to regard only the fourth of their disciplines
as disputation proper. But it is notable that in Names and objects (NO 6)
discrimination between X and non-X is only one, the first, of three sorts of
disputation. Although there are no examples of the wider usage in the
Canons, very probably it never became obsolete; certainly there is no

evidence of a word replacing it.
A 75 &*to№M™Vk№&<>
K r ft»*»iifn«»« J * mm&#&&»KflTOft* · am
l2ilw| [ f t ] ™ * * * * · # K * o
(Defined: ch'iung, A 42: cAi*, A 5. Analysed: jw, A 84)
1 9 4
A special Mohist graph distinguishing one sense of hsiian i& 'hang*
(§ 1/2/1/2/10). Hsiian is frequently used in Hsiin-tzU of scales settling,
cf. ch. 22 (Liang 324/8) flWRftWJrfoAClJ**, ^Pfe&ffi
A ^ H H " I f the steelyard is wrongly adjusted, the heavier comes to
rest on the arm which has risen, and men think it the lighter; the
lighter comes to rest on the arm which has fallen, and men think it the
heavier". This usage belongs to the terminology of the 'weighing*
(ch'iian W) of benefit against harm, which is the theme of this Canon.
1 9 5
Unknown graph, variously identified (Sun W 'chop*, T*an H 'nourish',
W u H 'burn'). The context leaves little doubt that it refers to some
way of injuring a finger to avoid worse harm, cf. E C 8 B$ffiiJSt#Bl
"cut off a finger to save the arm".
196 This is probably the chih 35 'wisdom, understanding* of A 6 (cf. A 88
Jl: *3K 'sound understanding*) rather than the chih $Q 'consciousness*
of A 3; the original graph appears once later in the Explanation.
i»7 Li ( = IS), 'chance on (a mishap). Cf. the 'happening on robbers*
( S & A ) of E C 8, and Mo-tzu ch. 5 (Sun 18/2) B 8 t ; £ № "When a
state happens on bandits or enemies. . . **.
198
Sao 'stinking, putrid* = M ( M . 44935 def. 5).
199 This i g sentence is obviously mutilated; my restorations are guesses
o n
based on parallelism with the next sentence. For the pattern yi X

chih Y , 'on the grounds of X determine what Y is*, cf. A 95.
2 0 0
Ch'iang 'wall* = » ( M . 9509 def. 2).
2 0 1
Tao 'coin*, as in B 30 (where in four of five examples it is corrupted to #
in the Horyaku edition).
2/4 321
Following the emendation first proposed by Chang Hui-yen.

Note on the syntax. The syntactic difficulty of three of the long sentences
centres in each case on the problem of interrelating ' A tse JM B ' and
' A shih A B y eh "ife' constructions (cf. also the long opening sentence of
B 38). The dialectical chapters use shih very freely between verbal
units, and in all these sentences its presence does not affect the fact
that both units are inside the apodosis after tse. For the use of yeh
after the first unit, which further obscures the organisation of the
sentence, cf. p. 157 above.
C. " T o wet (to be doing something for the sake of . . ./to have as end) is
to give the most weight in relation to the desires, having taken account of
all that one knows.
E. If you prefer to cut of£ your finger, and the understanding does not
recognise the harm in it, this is the understanding being at fault. If the
consideration paid to it by the understanding overlooks none of the harm
in it, but you still prefer to cut it off, then that things have turned out
unhappily is as with eating dried meat. Whether putrid meat will benefit or
harm is unknowable in advance; if you prefer to eat the meat, and it is
putrid, then eating it is refusing to take the doubt as grounds for fixing
which you prefer. Whether there was benefit or harm 'beyond the wall*
was not knowable in advance; if by heading for it you could get money, then
refusing to head for it would be taking the doubt as grounds for fixing
which you prefer.
In the light of the principle that " T o be 'for* is to give the most weight
in relation to the desires, having taken account of all that one knows",
when you cut up dried meat it is not wisdom, when you cut off a finger
it is not foolishness. When what you are for and what you are against put
each other in doubt, you are failing to plan things out."
In disputation over benefit and harm the ultimate grounds for choice
are one's ends, what one is 'for* (falling-tone wet), a word which recurs
throughout Expounding the canons. The topic of the present Canon i
commonly taken to be 'action* (level-tone wei), although some scholars
(Kao, Liu) identify it as o M 'false', a word not used in the dialectical
chapters at all. But for anyone who appreciates the organisation of the
Canons as a whole, Chang Ch'i-huang (op. cit. 1/9A) is plainly right in
reading the wei on the falling-tone. If we understand it as 'action', the
Canon cannot be read as a definition, and ought to be in the series on
behaviour (A 7-39), not the one on disputation. Moreover when we look
at the structure of the system of ethical definitions (pp. 47-49 above) it

is evident that wet 'to be for* had to be defined in order to complete the
account of at $t 'love'. We may compare the Shen wet # ^ 'Be aware of
what you are for') chapter of the Lii-shih ch'un-ch'iu, which discusses the
weighing of benefit and harm from the point of view of individualism:
Ch. 21/4 (Hsu 21, 8A) » WSffe ° » 0 r K * - f l i ° #№[K]fSrfn@
"One's person is what one is 'for', the Empire is a means to what one is
for. Be aware of what you are for, and the light and the heavy will be
recognised. Suppose there were a man who cut off his head in exchange
for a hat, let his person be killed in exchange for a coat; the world would
surely think him deluded. Why ? Cap and coat are the means to adorn head
and person. If for the sake of the means of adornment you kill what they
adorn, you do not know what you are 'for'."
In ethical contexts we never find the word ku Afc 'reason'; ethical
thinking is in terms of ends, of what one is 'for'. Since wet is the basic
logical term in ethics it is defined with the same meticulousness as the
ethical terms themselves. Except for particles, the only undefined and
unanalysed word in the Canon is hsiian 'let the scales settle', and even this
may be regarded as covered by the definition of ch'iianfll'weigh' in E C 8.
Since the Mohist never repeats himself (unlike the authors of earlier
chapters in Mo-tzu), the Explanation merely takes up the situation des
cribed in E C 8 (escaping from robbers at the cost of a finger or an arm),
and develops a new point implicit in the definition, that a choice is to be
judged in terms of the knowledge available at the time. The man who loses
a finger did not make a mistake when he exposed himself to this risk; he
was pursuing the reasonable end of getting money (the Mohists, if we are
right in thinking of them as craftsmen, would be most likely to run into
robbers while travelling on business), "whether there was benefit or harm
beyond the wall was not yet knowable", and he "suffered it as a mishap"
($!;£,). The 'wall' is not to be taken literally; it is a recurrent metaphor for
the limits of factual knowledge (§ 1/5/8).
Hsiin-tzu's digest of the four disciplines in Right use of names concludes
with a section on considered choice in which the metaphor of weighing is
fully explicit. We can see from Hsiin-tzu's account how the hsiian (letting
the scales come to rest/giving the most weight) of this Canon is related to
the ch'iian (weighing) of E C 8 (cf. n. 194 above):
"When a man chooses, what he desires never comes to him unmixed; when
he rejects, what he dislikes never departs from him unmixed. Therefore
2/4 323
every move a man makes is to be accompanied by weighing (ch'uan). If

the steelyard is wrongly adjusted, the heavier comes to rest (hsiian fi) on
the arm which has risen and he thinks it the lighter, the lighter comes to
rest on the arm which has fallen and he thinks it the heavier. This is the
reason why men are deluded about the light and the heavy."
2/4/1/6 Appendix: words with more than one usage (A 76-87)

This series of Canons is not in sentence form. There is no final yeh
as in the definitions of A 1-75; different usages are simply listed after the
word itself. The distinctions analysed vary widely in nature. Thus A 78
distinguishes three kinds of name, A 80 three sources of knowledge and
four objects of knowing, A 85 six subsidiary meanings of the highly
ambiguous word wet % 'do, make*. But all the terms examined are
important in the Mohist logic, and the classifications add greatly to our
understanding of it.
Although the series is most conveniently treated as an appendix to the
definitions of A 1-75 it also serves as a bridge to the theses which begin
at A 88. The long Canon on sameness and difference which begins the next
series follows straight on to the analyses of sameness and difference in
A 86, 87.
The eleven sections which have Explanations may be observed to fall
into two main types. Some merely call attention to the ambiguity of a word
by giving a series of examples, unfortunately very elliptical; we have to
work out for ourselves what kind of sentences the Mohist has in mind
(A 76, 79, 83, 85). Others carefully distinguish and define the various kinds
of knowing, hearing and seeing (A 80-82) and of sameness and difference
(A 86, 87). In the cases of causation (A 77) and names (A 78) the Mohist
both gives examples and analyses the distinctions which they illustrate.
A 76 B ° JS » t °
C. "Yi (end/finish). T o bring about, to get rid of.

E. Of making a coat,: 'to bring about*.
Of curing an illness: 'to get rid o f . "
For yi used of curing an illness, cf. Lu-shih ch'un-ch'iu ch. 11/2 (Hsu
11, 6B/3) ££gfcfc*IE-tb "His Majesty's illness is certainly curable" (The
same story has three more examples.)
Yi is the Mohist's only word for ending (chung is never found). It is
important in theorising about time, as a verb generally implying cessation
of existence (A 51, B 10, 33, 51), but also used adverbially ('already', A 33,
B 61). The sense of completion would be relevant mainly to the latter,
cf. B 61 B#& 'already fully provided'. The crucial definition of pi i&
'necessary' is in terms of yi (A 51).
A 77 $ o f | , Jft o
(Analysed: A 79. Defined: k , A l )
2 0 3
Delete (Kao, L i u , Liang).
C. "Shih (to commission/to cause). T o tell. The cause.

E . T o give orders is 'to tell'. The thing does not necessarily come about.
Dampness is a cause. It is necessarily required that what it does comes about."
The distinction between shih 'to commission' and shih 'to cause' is
used to solve a problem in B 55. Examples of the two usages would be
(1) Mo-tzu ch. 47 (Sun 277/-2) ttflfflicg "make a blind man choose
between them". This is 'telling', and the act is not necessarily realised
(cf. Mencius IB/5 h%m&W№& "Everyone tells me to pull down the
Hall of Light").
(2) Lu-shih ch'un-ch'iu ch. 14/4 (Hsu 14, 17B/5-7) # * S f i 9 * * * ,
#^3E, "When the breath of spring arrives the herbs and trees
grow, when the breath of autumn arrives the herbs and trees shed their
leaves. Growing and shedding are caused by something, it is not that they
are so of themselves. Therefore if what causes it arrives, no thing is not
brought about; if what causes it does not arrive, no thing can be brought
about". (The use of wet 'do, bring about' may be compared with that in
the Explanation.)
The stock example of an event with several possible causes is illness,
caused for example by a wound or by dampness (§ 1/5/11). If dampness
does cause an illness the illness 'necessarily' does come about. Here it is to
be noticed that the Mohist uses the word pi *5k (necessary on the plane of
reason), not ku H (inherent in the situation). Cf. § 1/4/14. His point is that
unless the man exposed to the damp does fall ill the dampness has not
caused anything; he is not saying that dampness is certain to cause illness.
We may notice also that shih used in this sense implies that the agent is a
'major' and not a 'minor' cause, as defined in A 1.
2/4 325
A 78 £ o o
(45) ° T ^ J » « * o # f f . ^ • ^ ( ^ ) « ^ » * i b o ^ r J » J · *
ffft ° ' o (^JS(^)*^) 2 0 6

·
2 0 4
Corrected from the next two sentences.
2 0 5
Yeh che is used consistently as a quotation device (§ 1/3/10). One is
tempted to delete yeh, making a simpler sentence: "For what is like
the object we necessarily use this name". But the corrupt parallel in
A 31 has che yeh in the same position, with the particles mistakenly
transposed; we can hardly delete yeh in both parallels.
2 0 6
There is confusion of the same graphs in B 63. Cf. Mencius 7B/36 #4#T
PIHfe, 4a§r$i"iil " A surname one shares, a given name one has to
oneself".
C. "Ming (name). Unrestricted: classifying: private.

E. 'Thing' is 'unrestricted'; any object necessarily requires this name.
Naming something 'horse' is 'classifying'; for 'like the object' we necessarily
use this name. Naming someone 'Jack' is 'private'; this name stays confined
in this object. The sounds which issue from the mouth all have the name.
(For example, surname and style-name.)"
This section shows clearly that for the later Mohists (as for Hsun-tzu
in his Right use of names chapter) an object (shih Sf) is a particular, and the
function of common names is to be explained on purely nominalist princi
ples. Indeed, if we are right in giving the particle combination yeh che its
full weight, a common name is treated as an abbreviation of 'something
which is like the object', the object being the particular for which the name
is ordained (mingfacf. § 1/4/22). In every sentence except the first and last
the word shih 'object' either appears or is resumed by the pronoun chih HL.
A point of some interest is that common names are distinguished not only
from proper names but from the 'unrestricted' name 'thing'. 'Thing' and
'Jack' are both unlike 'horse' in that similarity is irrelevant to their use.
The passage may be compared with B 70, which seems to be the
answer to an objection that on this analysis a common name can only tell
us what an object is like, not what it is. We might also object that "Whenever
the sound issues from the mouth the name is present" does not explain how
different utterances can share the same name. But it seems from the corrupt
fragment which we locate in A 31 that the Mohists had a nominalist
solution to this problem too: "For 'like the stone' we necessarily use what
is like the name". We name the stone by pronouncing the sound 'stone*
over a particular, and afterwards use a sound like the sound for an object
like the object.
A 79 m ° &» m»ja o
(Defined: chii, A 31)
2 0 7
The technical term li 'connect (name with name in a phrase)', here
disguised by an added 'water' radical (§ 1/4/16). L i u Shih-p'ei, who
proposed this (op. cit. B, 6B), attached the li to A 78 (3gf& *M * ·
"Like the surname linked with (the given name)"). Most editors agree
in placing it in A 78, because li precedes the head character; but we
have seen that the intrusion of the head character on either side of
the first character is to be accepted as normal (§ 1/2/2/5/2).
208
Yi restored from the Canon, with L i u Shih-p'ei and L i u Ts'un-yan.
Most prefer to emend ming to yi (Sun) or yi to ming (Liang), but both
graphs and readings are too dissimilar for confusion to be plausible.
C. "Wei (call). Transfer, refer, apply.

E. Of naming by linking 'whelp' and 'dog': 'transfer'.
Of 'whelp', of 'dog': 'refer'.
Of hooting at a dog: 'apply'."
Wei is calling objects by names (cf. A 80), and may refer to a relation
between name and name, name and object, or name and speaker.
(1) Yi 'transferring' is not otherwise attested in the dialectical
chapters, but may describe the wet of the formula ' X chih wet Y ' ('It is X
that is called ' Y ' '), used in E C 7-10 in providing definitions. If so, it is
the Mohist term for definition; considering the enormous importance of
definition in their logic we can be sure that they had one, and the classifica
tion of varieties of 'calling' would be the obvious place to mention it.
Although ' X chih wet Y ' is the formula of definition in Expounding the
canons, the Canons abandon it for ' X Y yeti; "naming by linking 'whelp'
and 'dog' " would be forming the sentence ^ ^ i f e . " A whelp is a dog",
which we actually find in B 54. That the Mohist distinguishes this from
chii 'referring' can be seen from A 32, where saying "What looks like the
picture is a tiger" is distinguished from merely conveying an object by
saying 'stone'.
2/4 327
(2) Chii 'referring' (picking out an object by means of its name) is

defined in A 31. It is the wet of the formula jttillBife "It is this that it
refers to".
(3) Chia 'applying' is wet as used for example in the formula f i X 0
'addressing X say' (say to X ) ; hooting at a dog is an apt example:
Shih-tzu FT ap. Tai-p'ing yu-lan JS&WL S P T K 905/1A/-2, Yi wen
lei chii mxmm ch. 38 (Taipei 1960, 1043/-6) <frft^J& [" * J ° . . . $1A
"He named his dog F u ('Rich'). . . . The dog entered the room and Kuo
hooted at it, 'Get out, F u ! ' " (CAVA 'hoot at' is misprinted in the recons
tructed Shih-tzu, SPPY, B 10A/5).
But it appears that chia includes every kind of calling in which the
relation is of speaker to name, not of name to object, not only addressing
but praising, blaming, commanding. We have noted in commenting on
A 29, 30 that chia is used of praising and blaming several times in Mo-tzu
ch. 26-28. The definitions of praising and blaming immediately precede
the definition of referring in A 31, and are evidently to be distinguished
from it as chia. For wei used of commanding, cf. A 77.
A 80 » o * H £,fto
(») °« S £ » o > tSffe o #f£-£f 210 > o
0rHH » * f t o 9fSB ' * t e o . o^fj , «4 o
(Defined: chih, A 5: shuo, A 72. Analysed: A 81: ming, A 78: ho,
A 83: w«, A 85)
2 0 9
( * ) . Cf. § 1/2/1/3/20.
2 1 0
The final yen shows that this is not the transitive kuan 'observe' but
kuan yii № X ' (travel around observing X ) . Cf. also A 81.
C. "Chih (know). By hearsay, by explanation, by personal experience. The

name, the object, how to relate, how to act.
E. Having received it at second hand is knowing by 'hearsay'. Knowing
that something square will not rotate is by 'explanation'. Having been a
witness oneself is knowing 'by personal experience'.
What something is called by is its 'name'. What is so called is the 'object'. The
mating of name and object is 'relating'. T o intend and to perform are to 'act'."
Chih 'knowing' is the only word analysed in A 76-87 which has already
been defined. The section falls into two parts, how one knows and what
one knows:
(1) The sources of knowledge are report or tradition, explanation,

and personal observation. Examples of ch'in chih 'knowing by experience*
and shuo chih 'knowing by explanation* appear at the end of B 70. The
dialectical chapters are concerned solely with shuo, explanation or demons
tration, and it is only in A 80-82 that we are told how we know the facts
from which explanation starts. It is interesting that although there is no
evidence that the Mohists, or any other Chinese thinkers, developed
geometrical proofs like Euclid's, the example of knowing by explanation
is a geometrical one, that "something square will not rotate". (Similarly
in A 98 the illustration of WIS "explaining something" is "For example,
a circle is nowhere straight". Cf. also B 62). The most obvious example
would be a square wheel (cf. A 48, where the rotating figure is assumed
to be circular):
Mo-tzu ch. 49 (Sun 300/-5) JRJBffllB^ltt °
" A jaded horse and four-cornered wheels."
But in any case circle and square were stock examples of free and
hampered movement:
Sun-tzu (ch. 5) S P T K 5, 23A/-3 * P ^ * S # № l i t » IBflSff °
"Like a revolving thing of wood or stone. . . . If it is square it stops, if
it is circular it goes."
Kuan-tzu (ch. 31) BSS 2/35/2f 1S%$k °
"What is circular rotates. . . . What is square sticks fast."
(2) The fourfold classification of the objects of knowledge in terms
of name and object is of fundamental importance if, as we believe, it is
the clue to the organisation of the Canons (§ 1/6/1). It is to be noticed that
the first pair of words are nouns, ming and shih, and the second verbs,
ho and weu Chih 'know' before a verb is 'know how', so that chih ho would
be, not 'know about relations' but 'know how to relate'. We find the same
two pairs in the organisation of the Canons, with the disciplines that we
interpret as 'explaining how to relate (names and objects)' and 'explaining
how to act' preceding the bridging sequence on change and necessity, and
the disciplines 'explaining objects' and 'explaining names' coming after
wards.
The distinction between knowing names and knowing objects is
applied to problems of disputation in B 39-41, 48.
2 1 1
A 81 N ° ( « ) * # » Mo
(W) o > m& o #fui > m& o
2/4 329
Corrected from the Explanation.
C. "Wen (hear). At second hand, in person.

E. Being told by someone is hearing 'at second hand*.
Being a witness oneself is hearing 'in person'/'
In Mo-tzü ch. 16 (Sun 78/9) Mo-tzu claims to know that the ancient
sages practised universal love, not because "I heard their voices in person"
but through the writings which they "transmitted to their
descendants in later generations" ( ß J ä t ^ t ö : - ? ^ ) .
A 82 1 ° i » t °
2 1 2
(*) o (i$)# ^ » M o i » i 4 o
(Defined: t'i, A 2: chin, A 43)
2 1 2
A conjectural emendation of Sun Yi-jang (Sun 218/-2). There is no
other example of t'e in the dialectical chapters.
C. "Chien (see). Individuals, all.

E. One member of a pair( ?) is an 'individual*.
The two of them are 'all'."
What we see, unlike what we hear (A 81), is always known "by personal
experience" (A 80). The limitation of seeing as a source of knowledge is
that we may have seen not all but only some.
T'i (unit, individual, part) has been defined in A 2, chin 'all* in A 43.
The words are contrasted in the phrases "love individually" (A 7 $!§£)
and "love all" (B 74 ftX).
The dialectical chapters tend to use two as the typical number above
one, as in the definition of t'i (A 2 cf. A 60, B 75, N O 18). There is some
evidence that t'e ('a single one', primarily a single sacrificial victim) was also
conceived as one of a pair with its mate missing. Cf. Fang yen S P T K
6, 4A/2 %№№B№ " A thing without its mate is called t'e", Odes 45/2
J l Jtfflft "It is he who is my mate", 88/3 ^ M f f ^ P " Y o u seek your new
mate". (The Mao % commentary glosses t'e as p'i "E 'mate' in both.)
A 83 £ ° HT » S » & °
2 2 1 7
2
((-&)*£) 13 o , ^TlUifr » /feX ™ · j E 4 o « ± l § · fiffe o
218 2 1 9 2 2 0 221
m . Œrfno r&J » *J <ffi > ^ / M ° *iR # »
(Defined: A 51)
2 1 3
Corrected from the Canon.
2 1 4
For the emendation, cf. § 1/2/1/3/21. Yii li 'standing with' would be
being together (ho) on the same level (wet iiL) for purposes of counting,
for example the five fingers which are one when the parts of the body
are counted (cf. B 12, n.).
2 1 5
The fan of the last sentence with the radical dropped. Fan chung 'coincid
ing as the converse*. Cf. A 85 *1S 'coinciding with a complement',
N O 5 ^Sfc 'coincide with an example*.
216
Kung 'achievement* ( M . 8714 def. 13 = # 0 as in B 51. Cf. Mo-tzu
ch. 49 (Sun 297/6) 5 H £ f £ £ £ K S # I f f l K ® "I recommend your
lordship to make observations of them, fitting their achievements to
their aims**.
2 1 7
For wet 'constitute, be deemed* without a complement, cf. p. 118 above.
2 1 8
Most editors have followed Sun in emending to IE or !S, for want of a
better suggestion. The assumption is presumably that the Mohist
cannot possibly be saying that the words of the sages are not to be
treated as certain. But once it is recognised that pi refers to logical
necessity, and that for the Mohist the necessary is independent of
temporal changes while the example of a sage is not (p. 33 above),
there is no difficulty in accepting the text as it stands. Sheng che is an
unusual phrase, but since sheng is a stative verb (the sage is the sheng
jen A 'sagely man*) it is acceptable ('sagely things').
2 1 9
The quotation device yeh che is used only after pi because this is the
only word quoted from the Canon (§ 1/3/10).
220 This restoration is required because of the parallelism of the last three
sentences and because the strict use of fu and wu before object
less transitive verbs (§ 1/3/5/1, 2) makes k'o wu an impossible com
bination in the syntax of the dialectical chapters.
221 c f ^ 1/4/9. Previous editors have followed Sun in emending to W
(on the grounds that the graphs are confusable in grass script).
C. " i / o (relation/tallying/being together). Exact, to the one which is

appropriate, necessary.
E. The relations of sharing a level with ( ?), coinciding as the converse ( ?)>
aim and achievement, are 'exact'.
2/4 331
The relation to what Jack is deemed to be is 'to the one which is

appropriate'.
If without something else something is necessarily absent, the relation is
'necessary'. The judgments of sages, employ but do not treat as necessary.
The 'necessary', accept and do not doubt. The ones which are the converse
of each other, apply on both sides, not on one without the other."
Ho is the term in the Mohist's vocabulary closest to our own 'relation'.

It may refer to a point-by-point correspondence of X and Y (as in the tally
ing of different squares, B 65) or to X and Y being spatially or otherwise
together (for example as occupants of a room, A 86). The terms cheng
'exact' and yi 'appropriate' are used elsewhere of the tallying of an object
with its standard (fa Sc). A circle tallies exactly (A 98), but whether a man
who is partially black and loves some men fits the standard for 'black man'
and 'love of men' is a matter of appropriateness (A 96). The difference is
similar to that between desiring immediately (cheng IE A 84, E C 8) and
desiring after weighing benefit and harm (A 84, E C 8). Here the examples
of ho which is exact are, on the proposed interpretations:
(1) Putting things together to count them on one level or on another
(five fingers or the fingers as one part of the body), without any overlap of
levels (cf. B 11, 12, 59, 65).
(2) The converse of 'They are oxen' coinciding exactly with the
total of non-oxen (A 73).
(3) A man's achievement matched directly with his aim.
There is only one example of appropriateness, Jack being deemed to
be, for example, a black man although not all of him is black. The phrase
WiZ% "What Jack is deemed to be" may be taken as contrasting with the
« £ # C "Jack as he is in himself" of E C 12.
Ho is necessary (pi) if there cannot be X without Y . If there also
cannot be Y without X they are the converse (fan) of each other, for
example 'elder brother and younger brother' (cf. A 51, 88); the Mohist
warns us to draw the conclusion both ways, not on one side but not the
other.
The classification of varieties of ho has a very important place in the
Mohist's scheme. The word yi 'appropriate' appears no less than three
times as the summing-up of a Canon (B 43, 44, 51), in each marking a
relation as not exact and not necessary. According to Names and objects
"There is no necessary relation between name and object" (NO 1
/pift * 'o); as we have just noticed, exact and appropriate relations between
name and object are distinguished in A 96, 98.
A 84 ш о f& > mm»a - as о Ф » * * о

(Defined: сД'шт, E C 8 : A 2 6 : <Ж%А, A 3 3 : hai, A 2 7 )
C. "Yii (desire/be about to). Directly, weighing the benefit: be about to.
Wu (dislike). Directly, weighing the harm.
E. . . ."
This is the pair of undefined terms at the base of the whole ethical
system (p. 4 7 above). It is perhaps surprising that there should be no
Explanation; possibly one has been lost. However, the Mohist does not
repeat himself, and he has already defined every word in the Canon except
cheng, and in E C 8 distinguished the cheng (the directly, immediately
desired) from the desired after weighing. The distinction between the
direct and the weighed in ethics corresponds to that between the exact
(cheng) and the appropriate in logic, drawn in the preceding Canon
(A 83).
Sun Yi-jang deleted the ch'ieh 'about to\ But it would be worth the
Mohist's while to distinguish yii 'about to* from the yii 'desire* which,
with wu 'dislike', is the basis of the ethical definitions. It is true that this
usage is attested rather late, but the Canons are only a century or two
earlier than the Shih-chi, in which it is frequent (P'ei Hsueh-hai 63).
223
(S) о *ф*]й222 , о Щ » tl-fb ° MSI» S ib ° · ШЬ °
225 2 2 e
Щ Й » ° « t M » ft* о
(Defined: chih, A 2 8 : hua, A 4 4 )
222 (^-Ш), systematic corruption of chung ti 'coincide with a complement'

( § 1 / 2 / 1 / 3 / 2 2 , 2 3 ) . Cf. A 8 3 "coincide as the converse", N O 5
"coincide with an example". It may seem easier to accept the
t'ai 'terrace' (cf. Mencius 1 A / 2 %Ш "made a terrace"), and emend the
tsao to Щ (Sun), or (T'an), or W (Kao); but this item illustrates
the first of the pair ts'un wang 'leave as it is or abolish', not ch'engwang
'bring to completion or abolish' as in A 7 6 .
223 Y - < ip andjy/ 'exchange' reappear in В 3 0 , 3 1 .
u se
2 2 4 %
Sun and his successors take the first word as hsiao ffi 'dissolve', in which
case there is no concrete example; it seems better to take the second
as chin 'ashes' ( M . 2 3 0 2 9 def. 9 =*S).
225
Hsun 'instruct' ( M . 4 3 3 4 9 def. 9 =iU), as once elsewhere in Mo-tzu
(Sun 176/1).
2/4 333
Judging by A 45 the second word should be ch 'un W 'quail*; the taboo

ing of the graph in the Sung exemplar (§ 1 /2/1 /4/5) may have something
to do with its replacement here by a different graph. A suggestion
mentioned by Sun that ch'un was written with the graph M ( M . 28535)
and corrupted has been adopted by Kao and L i u .
C. "Wei (constitute/become/deem/make/cure/govern). Keep as it is, get

rid of, exchange, dissolve, govern, transform.
E. Of coinciding with a complement^): 'keep as it is*.
Of illness: 'get rid o f .
Of buying and selling: 'exchange*.
Of sleet or ashes: 'dissolve'.
Of instructing and leading: 'govern'.
Of frog and quail: 'transform'."
This section does not consider the falling-tone wet 'be for' defined in
A 75, nor the most obvious sense of the level-tone wei, 'do, make'. (However
we understand the corrupt first item, it is classed as ts'un 'keeping as it is',
while the wei yi 'making a coat' of A 76 was classed as ch'eng 'bringing
about'.) Its purpose seems to be to list six senses of this highly ambiguous
word in which level-tone wei does not imply that an object is made, brought
into existence:
(1) Wei 'deem/be deemed', its most important sense in disputation.
(Cf. § 1/4/32.) The item as we emend it is only intelligible in the light of
the observation we made in explaining A 39, that being X is conceived as
the 'complement' (ti), the logically necessary accompaniment, of being
similar to the object initially named ' X ' . Deeming something to be X is
therefore coinciding with this complement. Unlike wei 'making', it leaves
the object as it is (ts'un).
(2) Wei 'cure', as in Tso-chuan £ W Ch'eng 10/5 mb®&MM±..
. . . B , ^ ^ " T ^ f f a "The Earl of Ch'in sent the doctor Huan to cure
him. . . . He said 'The illness is incurable' ". The illness is not brought
into existence, but removed from existence.
(3) Wei used of trading, implying not that the goods are made but
that they are exchanged. In this case it is less easy to see what kind of
phrase the Mohist has in mind. Cf. Mencius 2B/10 " S ^ T l T - t k , KigkBrft
"In ancient times when they ran a market, they exchanged
what they had for what they did not have".
(4) Wei 'become', of snow becoming sleet or wood becoming ashes,
implies dissolution.
(5) Wei 'become', of a frog becoming a quail (A 45), implies

transformation.
A86PI-o»»«"&»*o
A 87 £ o - , < ^ >228f| , ^ , o
(II) o Z l ^ 2 2 9 . Z H f a o ^ ; j g J l , ^ | H b o ^ g | f 0^^) 230 ,
(Defined: *'ww£, lost (§ 2/2): A 2. Analysed: Zw, A 83)
227
T'i 'unit' and cAtt* 'total', cf. A 2.
2 2 8
Restored from the Explanation.
2 2 9
" I f the (names) are two the (objects) are necessarily different", the
implicit subjects being carried over from the parallel item in A 86.
230 yy t'ung ('nowhere the same', A 60) and yu pu t'ung ('somewhere not
u
the same') would both be false descriptions of pu lei 'not of a kind*;

the meaning of the words in this unexpected order is presumably 'not
the same in a certain respect', referring back to the yi t'ung 'having
a respect in which they are the same' of A 86. But the other examples
of pu yu (A 65, 83, B 44) do not throw light on this point.
C. "T'ung (same). Identical, as units, as together, of a kind.

E . There being two names but one object is the sameness of 'identity'.
Not being outside the total is sameness 'as units'.
Both occupying the room is the sameness of being 'together'.
Being the same in some respect is sameness in being 'of a kind'.
C. Yi (different). Two, not units, not together, not of a kind.
E . The objects if the names are two necessarily being different is being
'two'.
Not connected or attached is 'not units'.
Not in the same place is 'not together'.
Not the same in a certain respect is 'not of a kind.' "
T'ung yi ('sameness and difference') was one of the standard themes

of the sophists, together with wu hou M№ ('the dimensionless') and chien
pai^fe ('the mutually pervasive'): cf. § 1/4/3. It led to extravagances very
distasteful to the cool-headed Mohists, such as the claims of Hui Shih and
Chuang-tzii that from one point of view all things are the same and from
2/4 335
another all are different. The later Mohists continued to refine distinctions
between kinds of sameness throughout their history. The lost definitions
of the ten theses of Mo-tzu must have included a definition of the t'ung
of shang t'ung InJIÎ 'thinking the same as those above* (§ 2/2). In the Canons
this is turned into a separate word by the addition of the 'man* radical and
defined in A 39, and four senses of the basic word are distinguished in the
present Canon. Later the author of Names and objects discovered the
importance of the sentence and was no longer satisfied with the four cate
gories of the Canons. He listed them in N O 6 as types of 'sameness with
the same name* ( N ^ ^ N ) , and supplemented them with four types of
'sameness with the same root* (IRI#I;£[R1), the 'root* of the sentence being
apparently what we call the predicate.
The descriptions of the four varieties themselves use the words t'ung
and yi, presumably depending on the lost definition of t'ung.
Two of the four terms for types of sameness, ch'ung 'identical* and lei
'kind', have a regular place in the Mohist vocabulary.
(1) Ch'ung is used in its literal sense of 'doubling' in the optical
sequence (of one shadow reinforcing another, B 18), and of identity in
B 38, 40, 53, N O 6. The typical example is the identity of a whelp with a
dog, when the two names refer to one object (cf. § 1/5/3).
(2) The sameness of being units is renamed lien 38 t'ung 'sameness
in being connected' in N O 6. Neither term is put to actual use, but t'i
'unit' and chien 'total' apply to the counting both of parts in a whole and
individuals in a class. A n obvious example would be the head and finger
being the same man (NO 7), but there is no reason to suppose that it refers
exclusively to part and whole. Although oxen as a species are the same in
kind (Type 4), as members of the same herd they would no doubt be the
same as t'i.
(3) Sameness in being together is the obscurest of the four types.
The description "all occupying the room/house" seems on the face of it
to refer to a family (cf. A 88 JÎsirf "a child, a child and the mother
living under one roof"), in which case it is unexpectedly concrete, the only
one of the eight in A 86, 87 which merely illustrates. The opposite in A 87
is "not in the same place"; but being the same family would be a matter
of kinship rather than place. But there is some reason to suspect that shih
'room' was a technical term for the space occupied by a shih *M 'object'
(more literally, 'filling'); the sameness of being together would then be
that of shape and consciousness constituting the same man and the hard
and the white composing the same stone (§ 1/4/26). In N O 6 it is renamed
chii ^ t'ung, which may be taken as "sameness as components".
(4) Lei 'kind* is very common (EC Appendix. A 78. B 1, 2, 6, 66.

N O 6, 10, 12, 15, 16). Kinds differ in generality, 'living thing* covering
'animal' and 'bird', and 'animal' covering 'milu deer' (B 2). 'Horse' is a
name for a kind of thing (A 78); it differs from the ox in lacking horns,
but this is not what differentiates the two kinds (B 66). Lei differs from the
English word 'class' in that X and Y are inside a class but not inside a lei;
they are 'the same in kind' ( ! S P ) or 'not of a kind' ( ^ S ) . A class is one
sort of chien, 'total'.
The Mohist nowhere discusses the relation between t'ung and jo 35
'like'. But it can be seen that his first three senses of t'ung have nothing to
do with similarity, and even being 'of a kind' implies not merely that
X and Y are alike, but that the likeness is the grounds for classifying them
together, calling them by the same name, treating them as the same thing.
2/4/2 T H E PROPOSITIONS (A 88-B 82)
The propositions follow the same fivefold division as the definitions,

except for the omission of a sequence on ethics, already covered by
Expounding the canons. The division of the Canons into Parts A and B
which probably dates from Stage 2, was made in the middle of the first
sequence. The propositions differ in form from the definitions in conclud
ing with a summing-up formula, preserved only from B 1 onwards
(§ 1/2/2/4).
2/4/2/1 The Procedures of Description (A 88-B 12)

The first of the four disciplines, however we choose to name it, is not
to be confused with the last, disputation. Unlike disputation it explains,
not the implications of names, but the procedures for applying names to
objects; it establishes, not the unending relations called 'necessary' in A 51,
but the temporary fitting of names to transitory objects called chih i h
'staying fixed' in A 50. Its concern is not with whether something is or is
not (shihfei $kf£) but with what is so of it (jan #&), reflecting the distinction
in Chinese syntax between the ' X Y yeK and verbal types of sentence.
A point which emerges clearly from this close-knit sequence is that the
first of the Mohist arts has nothing to do with inference from known to
unknown facts. It assumes as known the 'objects' (B 3) as they 'inherently'
are (B 3 ku H , B 7 ku SSC); to judge by the classification of sources of
knowledge in A 80 they are known either by observation or by report. The
problem is how to describe them coherently, not only in particular but in
general statements (cf. N O 9 " B y describing we summarise
2/4 337
what is so of the myriad things"). If this Mohist art is to be called logic,

it is the logic of consistent description. Some descriptions, for example
'circular', fit objects as unequivocally as the name 'ox' which was the
original example of 'staying' in A 50, but in other cases we need rules for
fixing a name on an object. Throughout the sequence the intransitive chih
'stay fixed' is used only of the circle (A 93), and all other instances of the
verb are causative, 'fix' (A 96, 97. B 1, 2).
The Mohist begins in A 88 by distinguishing two types of pairs; in
the first, either word may be applied to the same object or objects, according
to one's choice of objects for comparison ('many/few', 'elder/younger'); in
the other, only one of the pair fits ('black/white', 'present/absent').
Difficulty arises not only with the first type but with some of the second,
since an object may have both black and white parts. The Mohist lays down
a procedure in four stages:
(1) The commitment (chih WK, A 93, 94). Having committed yourself
to X being the larger or heavier, you cite the commitment as your authority
for pronouncing Y the shorter or lighter. T o affirm the one is to deny the
other; "assent and denial are one in the benefit and the use" (A 93).
(2) The standard (fa A 94-96). In the case of colour, for example,
the commitment may be challenged; you then appeal to a standard. (Cf.
B 70, where a white thing is offered for comparison.)
(3) The criterion (yin H , A 97, B 3). If a man's blackness is only
partial the standard fits him only partially, and to describe him as a black
man is still challengeable. The question is whether the black parts or the
white are the appropriate ones (yi S , A 96), which are to be taken as
criterion. In the case of the circle on the other hand the matching is exact
(cheng IE, A 98).
(4) The kind (lei B 1, 2, 6). What is said of particular objects
must also be said of their kinds, provided that we fix the proper level of
classification. Thus it is 'animals', not 'living things' (which include birds)
which are to be described as four-legged.
The Mohist proceeds to complex descriptions, which he conceives as
linkages (li JSf) of names (B 3-7). Sometimes one member is implied by
another, cannot be dismissed without it (p'ien ch'ii Sife), such as length
and breadth. O n the other hand if we try to combine 'both' with 'They
fight' and 'They are two', it turns out that we have to dismiss 'They are
both two'. But the relations between the names which we apply to the
objects in no way affect the objects themselves, which continue to be two.
This example, the first in B 3, leads to the differentiation of describing
collectively and describing distributively in B 11, 12. Although we cannot
say 'They are both two', it is legitimate to say 'They are both one' (B 7,12),
in the respects which make them objects of the same kind, for example ox
and horse having four legs. At each level of classification we can count the
units and take the total as one at the next level, "like counting the fingers:
the fingers are five and the five are one" (B 12).
In punctuating the Canons of this sequence, it is to be noticed that many
begin with a two-word phrase, sometimes detached from the main clause.
With one exception (A 98) this pattern is sustained regularly as far as B 3.
2 3 1
A 88 fialPI ° S » № £ 4 > ' * 8 > E « » tttk > *'> ° :
2 3 2
' J £ * ' « S » # t : » fflft > Mm o >
2 3 3 2 3 4
(fjlf§) o ^ # ( ^ ) * № :ii * » ' o
(^)*m s #
M^ 2 3
» * M o *№fflffi3
6 2 3 7
>sg?-tb ° $](*)*# 2 3 8
M » 2 4 1
: SJitt »+ o £ , ft , [ftft] 242 ^ , n , o
2 4 3 244
J i »IS » J f c № o » ft® * o ^ i f c g f t * o f g , g
2 3 1
Fang 'depend on, be relative to', cf. § 1/4/10.
232 Previous editors have not appreciated the necessity of restoring these
25 characters from the Explanation (p. 93 above).
2 3 3
Emended from the Canon. A n important consequence of reconstructing
the Canon is that chiao te fang 'In interplay become relative* is revealed
as a heading contrasted with Hang chiieh sheng 'on both sides prevail
absolutely*. This saves us from the necessity of having to postulate four
head characters (with T*an, Kao, Liu); elsewhere there are never more
than two (§ 1/2/2/5/5).
2 3 4
Fu 'rich* = m, as in Sun 268/8, 289/9. The phrase fu chia %% 'rich
family* occurs in Sun 113/5, 279/6.
2 3 5
The Hang chih #P 'native understanding' of Mencius 7A/15.
2 3 6
For the evidence for taking this as t'u shih 'louse on a hare*, cf. § 1/5/6.
237
Huan-yiian/*G'WAN-GJWAN is surely the binome written with the
'carriage* radical ('uneven, irregular*, of a path, cf. M . 38551/1) in
Kuan-tzu (ch. 27) BSS 2/20/-4 J U ^ t f , jftft¥fti&H8tt;£lfe
"Leaders of troops must first make themselves aware of difficulties in
the irregularities of terrain".
2 3 8
Niao che 'swerve of a bird' is a military term for a tactical retreat when
attack meets resistance. Cf. Yen t'ieh lun ch. 38 (Wang 264/6) MSA
J^SlH, JRJMIIJfr (On nomad marauders) " I f they meet no resistance
they drag off their prey like tigers, when in trouble they swerve away
2/4 339
like birds". Yung-t'ungl*Dl\JNG-D'UNG looks like a binome parallel

with the huan-yilan of the previous sentence, but I have failed to
identify it.
239 f § i/2/1 /3/22. If chung 'hit' is accepted on grounds of systematic
c
corruption the second graph (of which this is the unique example in
Mo-tzu) should probably be fang: "Just as the sword strikes" (for
fang cf. A 33).
240
Ch'u shih 'living under one roof", cf. Mencius 4B/33 —*§S—'îffijSlI
"Live under one roof with a wife and with a concubine". Tzu mu
( M . 6930/759) "children and mother". It is unnecessary to delete one
tzil (with Kao, Liu), since the relativity of ages implies at least two
children. Cf. the phrase it^T 'daughters' ( M . 6036/131), in Li chi
1,49/2, 78/2: 30,37/2.
2 4 1
Sheng 'win in disputation' (§ 1/4/24). Cf. B 70 M S ^ J S l k , « "It is
as with 'white or black', with which does one win?"
2 4 2
Delete following Sun.
2 4 3
In the next section (A 93) we get nuo W 'assent' and in the next after
that both nan 'raise objections' and ch'eng 'established' (of a thesis).
Here we may give the same meanings to nan and ch'eng, and under
stand su 'stay overnight' in the light of Analects 12/12 -Ê&ftfSfg
"Tzu-lu never put off to tomorrow what he agreed to do". Mo-tzu
ch. 48 (Sun 284/1) " T o put off a good deed till tomorrow
is unlucky".
2 4 4
Sun already recognised that the graph is to be read as ti iSc 'mate'; it is
the one example of this technical term (§ 1/4/29) which has always
been recognised.
245 fc < original thing', as in B 7. Cf. the note on A 1. Wei 'constitute'
u tne
without a complement, cf. p. 118 above.
C. "Sameness and difference. In interplay the following become relative: .

having and lacking, more and less, departing and approaching, hard and
soft, dead and alive, elder and younger.
On both sides the following prevail decisively: being this or not being it
(/right or wrong), proved or not yet proved, both being complements,
present or absent, surname or the thing itself, dear or cheap.
E. In interplay the following become relative:
In the case of a rich family, of native intelligence, 'having and lacking'.
In the case of putting side by side and measuring, 'more and less'.
In the case of a louse on a hare moving this way and that (?), 'departing a
approaching'.
340 The * Canons* and 'Explanations'
In the case of retreating the better to attack (?), 'hard and soft'.
In the case of a sword just striking (?), 'dead and alive\
In the case of son, son and mother within a family, 'elder and younger*.
On both sides the following prevail decisively: white or black, centre or side.
In the case of discourse, conduct, learning, an object, 'being this or not
being it (/right or wrong)'.
In the case of raising objections to a proposal or putting off acting on it (?),
'proved or not yet proved'.
In the case of elder-brother and younger-brother, 'both being comple
ments'.
In the case of the body being here and the thoughts on something elsewhere,
'present or absent'.
In the case of what 'Crane' constitutes, 'the surname or the thing as it is
in itself.
In the case of a price being right, 'dear or cheap'."
The argument that all things are the same from one point of view and
different from another is familiar from Chuang-tzu, where it is used to show
the pointlessness of all disputation. Things differ in being big or small in
relation to each other, but all are the same in being big compared with the
point and small compared with infinity; and for Chuang-tzu this relativity
is characteristic of everything which can be said about a thing, including
shihjfei 'is X/is not X ' (cf. the quotations in § 1/3/12/3, § 1/4/10). For the
Mohist on the other hand there is a fundamental difference between long
and short, which are relative, and being this or not this, which are not
(B 78, 80). His two sets of examples have three-word phrases introducing
them as relative or absolute, in which the essential contrast is between
chiao 'intercourse, interplay' and chiieh 'absolutely' (more literally 'breaking
off from the other'), as in the common phrase chiieh chiao 'break off relations'
( M . 27407/22). The phrasing (although not the thought) may be compared
with the following passage, which also has Hang 'on both sides':
Kuan-tzu (ch. 11) BSS l/49/7f 5 f c * * £ f f i » A £ ° £ t t « r f n * i £ °
"The reputation and the substance have been enemies for a long time, and
therefore have broken off from each other and have no intercourse. The
wise man knows that he cannot hold on to both, and in stead chooses one
of them."
The first series illustrates the point that if there are more than two
choices, if we are comparing X to both Y and Z , then whether X has a rich
2/4 341
family and a clever mind, whether a louse moving on a running hare is

leaving or approaching, becomes relative both to Y and to Z . This is true
even of such a pair as 'alive or dead*. The example of this pair is unfort
unately corrupt, but if we have emended it correctly it refers to the moment
when life ends and death begins (cf. A50), and should be read in the light
of H u i Shih's sophism "The sun is simultaneously at noon and declining,
a thing is simultaneously alive and dead" (Chuang-tzu ch. 33 (Kuo 1102/3)
B ^ t f ^ S S , # J # £ # ? E ) " . We are consequently at liberty to say, either
that X and Y are the same but different from Z , or that Y and Z are the
same but different from X .
If however we have only two alternatives, then one either is X or is not
X . A n action is either right or wrong, an object is either ox or non-ox,
and whether or not a case is proved is independent of whether people raise
objections to it or put off acting on it. "Son, son and mother as members
of the household" allows Y to be older than X and younger than Z , but if
we take the two sons as a pair of complements then one either is or is not
the elder brother (cf. A 51). A dog may be named 'Crane' (§ 1/5/4), but the
thing in question is either a dog or a crane. Commodities, since there is
an exact and proper price fixed by supply and demand (B 30, 31), are
either dear or cheap.
The classification of yujwu 'there is/there is not' as relative surprises
at first sight. But although close to the Western concept of existence yu
does not coincide with it. It is translatable according to context as 'there is
or as 'has', without implying any conceptual difference: ; ^ T ; £ § T W "what
the world has/what there is in the world", S W f e "Horses have colour/in
horses there is colour". T o conceive yu as relative is not mere Taoist play
fulness ; one might say that the more properties or the more of a property
that a thing has the more of it the world has. The Taoist idea of yu and wu
as mutually dependent is criticised in B 49. But in the example here, having
(there being) a rich family or an excellent native intelligence is no less
relative than being rich or being excellent. The distinction he does see as
absolute is between ts'un and zvang, presence and absence, which have the
localising implications of 'exist'. The body either is or is not occupying a
certain place (ch'u), his chih (what his thoughts are centred on) either is
or is not in the same place; there is no question of the man being here and
elsewhere at the same time. (Cf. Chuang-tzu ch. 28 K u o 979/1
$ f c ± ± , i f r g ? S W H ± T " M y person/body is here by the river and the sea,
my heart dwells in the palace of Wei".)
A 89-92 [ H S ± № * o mtimffi$№&' °B'P£*№ ° tUfSI
This misplaced sequence of four sentences which entered the Canons at

Stage 1 we locate in N O 9 (p. 109 above).
A 93 R * ™ —*Jffl247 o
w j o i , ; ^ ® ' HiJfîm »
249
(Br) o jSttfiih-fe o 2 4 8
ffi-i ' jfcftlM
^ 250|^251o
246
Fou = •& 'No', the radical dropped as in B 52 and in Hsun-tzu ch. 31
t
(Liang 402/4) $viSlXli "assent or deny, choose or reject".
247
Li yung 'benefit and use' is already a common combination in pre-Han
literature ( M . 1932/185, 187). The Mohist, who does not use two
words where one will do, probably intends by It the benefit of what we
learn from an assent, and by yung the use to which we put a denial.
Cf. Lao-tzu 11 (on hub of wheel, vessel, door)
Sffl "Therefore we judge it beneficial that they exist, but find them
useful where they do not exist".
248
Hsiang ts'ung 'be implied by each other', as in E C 3.
249
Hsien chih 'know beforehand', know 'a priori' without having to 'jump
the wall' and make empirical observations (§ 1/4/13, 1/5/8). Chih shih
'know what it is', as in B 38, where it is again in the same sentence
as hsien 'a priori'.
250
Yuan 'adduce' (as evidence), defined in N O 11.
251
Chih 'uphold (one of alternatives)', used nominally ('the side to which
one is committed') here as in A 94 and in E C 1 (§ 1/4/5).
C. " A n assent and a denial are one in the benefit and the use.
E. When we 'jump the wall', the circular stays fixed. By the things which
follow from each other or exclude each other, we may know 'a priori' what
it is. For the five colours, long and short, before and after, light and heavy,
adduce the one to which you are committed."
Proceeding from the 12 pairs of opposites in A 88, the Mohist points

out that a claim that something has, is more, is this, is present, and a denial
that it lacks, is less, is not, is absent, give exactly the same information.
His word for 'assent', nuo, reappears in A 96; since it is phonetically and
semantically related to jo 3? 'like' and jan 'so', it may be taken as assent
to something being so, in which case it belongs to the terminology of
description rather than disputation.
2/4 343
The Explanation begins by disposing with 'circular', the kind of

description which raises no problem. The circle 'stays', confined to the
objects which exactly fit the standard for a circle, as unequivocally as the
name 'ox', which illustrated the original Explanation on staying (A 50).
We know 'a priori' what a circle is because it has a definition supported by
preliminary definitions (A 58), and when we go over the 'walls of the
intelligence' ( E C 2), beyond which nothing is knowable 'a priori', we can
recognise actual circles simply by whether they agree or disagree with the
standard. (For the circle as stock example, cf. § 1/5/10.)
Long and short on the other hand are relative; we commit ourselves
to which are long and "sort out the shorter in relation to the longer"
(B 78 J^Sfrafei). It is at first sight surprising to be told that the same
principle applies to the five colours, since in A 88 black and white were the
first example of descriptions which are not relative. But we learn in A 96
that a black man is not unequivocally black, since not all parts of him are
black. To describe things of uncertain colour I must commit myself to the
colour of some, and then describe others consistently with my commitment.
A 94 2 5 2
' SSWW" 8
» fl']**ftSt o
( K ) o mats ' 254
> mmxzf& °
Both Canon and Explanation are hard to punctuate, but except for one
graph there is no reason to assume corruption. Since most neighbour
ing Canons begin with a two-word phrase (A 88, 93, 95-97, B 1-3), and
at the end of A 93 chih seems to be nominal, 'the claim one upholds',
I take fu chih as 'devote oneself to the claim one upholds', cf. Chuang-
tzu ch. 27 (Kuo 953/1) IttrfiBKftl "zealous in his purpose and devoted
to knowledge", ch. 6 (Kuo 279/1) I ^ B R t f t "personally devote
oneself to benevolence and duty".
The unknown graph §1 is presumably mao 'describe' distinguished by
the 'word' radical (§ 1/4/20). Preserved in the Taoist Patrology text,
it is corrupted in the Mao and later editions to na 153, which modern
editors reproduce without even recording the original reading. For
3
ch iao chuan 'subtly turning', cf. A 95 and Kuan-tzu (ch. 20) BSS
J
1/110/-4 » A 5 W M ^ ( = » ) * J "His character has subtle twists and
is pointed and sharp" (cf. M . 1356/45).
Again punctuation is difficult. Both nan and ch'eng appeared in one item
in A 88, presumably in the same senses. Taking nan as 'object to'
(for an example which is also in the same context as wu Hr, cf. Lu-shih
ch'un-ch'iu ch. 18/4 (Hsu 18, 14B/5) § P # r S 5 « £ K ^ S ^ , №
%3£ "Teng Hsi made it his business to raise objections to him. . . .

He turned the wrong into the right and the right into the wrong"),
we can hardly understand ch'eng yen except as a phrase of the same
type as 'fully developed form', $M 'established state*, Jfofr
'fully developed heart', 'established custom' ( M . 11544/79, 107,
168, 207). I propose to take ch'eng as 'established, proved', as an
explanation which accounts for its use in A 88, 94 and is implicit in
the definition of ku 'reason' in A 1. Cf. Hsiin-tzu ch. 6 (Liang 59/-4)
S^f^WSSC, Hm£.f&$& "He has reasons for maintaining it, and
develops an organised case for saying it".
2 5 5
The corrupt graph is surely chih with most of its strokes obliterated.
C. "When someone devotes himself to a commitment, if the description

takes a subtle turn, seek his reasons.
E. If he raises objections to an established statement, and makes it his
business to establish this commitment, seek the standard for the commit
ment."
In A 93 the first issue requiring commitment was colour, and the next
concrete example will be the partially black man (A 96). The next reference
to colour after that is in B 3, "Most of a white horse is white, most of a
blind horse is not blind", which connects with an example in N O 18:
"If this horse's eyes are blind we say that this horse is blind, though this
horse's eyes are big we do not say that this horse is big". The kind of
commitment which is to be challenged by demand for the standard would
be something like the Second List sophism " A white dog is black" ( £ ^ M ) ,
the argument for which was reconstructed very plausibly by Ssu-ma Piao
(died A.D. 306) on the model of N O 18:
Chuang-tzu ch. 33 (Kuo 1111 n. 21) «j£BBJ> ' ° Sg£B;fc » *
"If a dog's eyes are blind we call it a blind dog; if the dog's eyes are big,
we do not say it is a big dog. Here in one case it is and in the other it is not.
If so, a white dog with black eyes can also be deemed a black dog."
Here "description takes a subtle turn", diverges from the course
(tao M A 97) we must follow if, for example, we are to call yellow-haired
oxen yellow (cf. N O 18). For the connexion between chuan 'turn' and the
metaphor of a road, cf. § 1/5/14.
In the Taoist Patrology and all other early editions the unknown graph
H in the Canon is followed by two characters written small, If M, the only
note printed as a gloss in the dialectical chapters. It may be a scrap from
2/4 345
the same source as A 89-91, mistaken for a phonetic gloss ("pronounced

U") and therefore written in after the next obscure graph. We locate it
with A 89-91 in N O 9 (cf. p. 110 above).
A 9s mm* mm&m °
A 96 » W№%&™ °
(S) ofcjitjMfe> №&m: ° \>ih±^m%^wM\t r MA j » *
256
Ch'iao chuan 'subtle turn', as in A 94. W interchanged also in Sun
109/10, where the note adduces examples from other texts. (Cf. the
example in the Lii-shih ch'un-ch'iu story quoted p. 275 above.)
257
Yi 'the one which is appropriate', cf. A 83.
2 5 8
In this long sentence, we may note the conjunction yii 'and* linking long
nominalised clauses as in B 9, the formula 'yu . . . yii . . .' for quant
ifying the object (pp. 131,132 above), and the causative use of chih 'fix,
confine within its limits' (§ 1/4/4).
C. " I f the standard is the same, examine what is the same in it.
E. Choose what is the same, and examine the subtle turn.
C. If the standard differs, examine what is appropriate to it.
E. Choose this and pick out that, ask about reasons and examine appro
priateness. Using what is black and what is not black in a man to fix 'black
man', and using love of some men and failure to love other men to fix
'love of man'—of these which is appropriate? "
A circular or square piece of wood or stone exactly fits its standard

(A 98). One has only to examine "what is the same in it" (its circularity or
squareness, not the wood or stone, B 65), and look out for sophistical twists
and turns. A white horse perfectly fits the standard for a horse, but that
did not stop Kung-sun Lung from arguing that it is not a horse.
If the standard fits imperfectly, one has to decide what it is 'appropriate'
(yi) to judge by, the parts which are black or the others, the persons he
does love or the others. Yi 'appropriate' and the cheng IE 'exact' of A 98
are distinguished in A 83 as two kinds of ho 'tallying', the word used in
B 65 of the tallying of objects which fit the standard for squareness. The
point which is appropriate is the^w H 'criterion' of the next Canon (A 97).
We use it to fix 'black man' and 'love of man', cause them to chih 'stay';
for unlike 'circle' in A 93 they do not stay unequivocally in the objects to

which we confine them.
The two concrete examples point forward to the fixing of a description
on an individual in A 97 and on a kind of thing in B 1:
(1) In the case of the black man the criterion is no doubt his skin
rather than his eyes; the eyes are adequate as criterion in the case of
blindness but not of blackness (cf. our note on A 94).
(2) The criterion for 'love of man' is laid down in N O 17, with yin
'criterion' used adverbially. It is required that he loves all men: "if he does
not love without exception, by this criterion he is deemed not to love man"
( E U S ^ S I A ) . In the case of riding horses on the other hand it is enough to
have ridden some horse: " i f he has ridden some horse, by this criterion he
is deemed to ride horses" ( 0 J S S J S & ) .
A 97 j h H a S O J i o
C. " F i x the criterion in order to 'separate the roads'.

E. If the other man, referring to a respect in which it is so, deems it so
in the instance here, refer to respects in which it is not so and inquire
about them. (For example, the sage has respects in which he is not, yet
he is.)"
This Canon makes a pair with the Canon of B 1, on 'fixing the kind'
(ihSH). Both assume a metaphor made explicit in N O 10, of a man at a
crossroads (cf. § 1/5/14). The same key words recur in N O 10, the man
(jen), the roads (tao), proceeding along them (hsing), and kind (lei).
Tz'ii 'the one here' refers to the concrete instance, contrasted in B 1, 2
and also in B 33 with shih & 'what it is'. (For the grammar of these
pronouns and of tz'ii ch'i, cf. pp. 120, 122f above.)
The Explanation continues with the first example in A 96. The other
man adduces parts of the man which are black to prove that he is black;
I answer by appealing to the parts which are not black.
As for the illustration, it was a commonplace by the 3rd century B.C.
that the sages did things which Confucian morality condemned; Yao
executed his own son, Shun married without his father's permission, T'ang
and King W u banished or executed their lords. The Lii-shih ch'un-ch'iu
quotes these stories with the comment "Judging by these examples, how
can anything be kept perfect?" (ch. 19/8 (Hsu 19, 26B/7) , 4^"J
^sfc), but also derides a man who used them as reasons for refusing to
2/4 347
honour the sages (ch. 11/4 (Hsu 11, 11B/2-12A/2)). But the Mohist's point
is probably a little different; not everything that the sage says has the
unending validity of the logically necessary, some of it becomes obsolete
(cf. A 83, B 16, 53), but he deserves honour none the less.
We have identified the yin 'criterion' as the term defined in A 71 as
"where it is so" (!???<&), immediately after the definition oi fa 'standard'
(§ 1/4/39). The criterion is the part or instance by which we decide that
an object either does or does not fit the standard. That you have ridden
one horse is the criterion for affirming "You ride horses", that you do not
love one particular man is the criterion for denying "You love men", as
we noticed under the last Canon.
A 98 i & » °
(IE) ©iLjfg-Hf * 2 5 9
o îmum ·{%m®U)™ » m%&№&
260
262
r£'£@f££ °
Wu (—(ft) 'match', the radical of which is always missing in some

editions where the graph occurs in Mo-tzu (§ 1/4/34). Chieh (=©)
'together' (chieh 'all' is not used pre-verbally, p. 127f above): cf. Kuo-yii
(Chin 2) S P T K 8, 5B/3 fgltif&A- "go out together and go in together".
Chih 'consciousness', the word which in the terminology of the Canons
replaces hsin 'mind' (cf. A 3). Cf. Chuang-tzu ch. 21 (Kuo 719/2)
fBft&Âfe'k "Dignities and salary do not enter his mind".
Yu shuo 'have proof to offer' (§ 1/4/40). Kuo wu 'exceed the matching',
as in B 58, where it is written S f f (for kuo, cf § 1/4/15).
The illustrations with jo 'like' which we bracket were probably marginal
glosses (§ 1/2/2/6/1), and this one is clearly parenthetic.
In spite of the jo the last phrase is not another illustration. The perfec
tive particle yi, used rarely and strictly (§ 1/3/12/7) and never found
in illustrations, shows that it is the main clause of the last sentence.
Cf. Kuan-tzu (ch. 2) BSS 1/6/-5 J M ^ S f * " I f you grasp
the Way of Heaven, affairs will be as though running themselves".
For jo before a verbal expression, 'as though', cf. B 22, 38. Tzu jan
'so of itself is used nowhere else; Mohists would be more likely to
use it of self-evidence than of the Taoist ideal of spontaneity.
C. "The exact nowhere is not.

E. The matching and the assent enter the consciousness together. When
something is explained, and you assent to more than that they match (for
example, to a circle being nowhere straight), or nothing is explained and
you assent on the basis of the matching, it is as though it were so of

itself."
For the other kind of tallying (ho &), which is not 'as appropriate*
(yi *M) but exact (A 83), the Mohist returns to the example of a circle in
A 93. It is immediately evident that the circle 'matches' (wu, § 1/4/34) its
standard point by point, whether the standard is the idea of the circle, the
compasses or another circle (A 70).
In A 80 the example of something known by explanation (shuo) was
"something square will not rotate"; here, when "something is explained"
(yu shuo), one knows that "a circle is nowhere straight", information
additional to the knowledge that the figure matches the standard for the
circle. We have elsewhere quoted examples from other sources of IE )3
'exactly square' (p. 170 above) and I E H 'exactly circular' (p. 208 above).
B i ±$mn A 2 6 3
o m>-&m™ ·
(it) o m&,it^M&^M o imitM:^M&^^ o
Sun emended jen to chih But comparison of the parallel Canons A 97

and B 1 with N O 10, where the metaphor is fully developed, show that
they refer to a man at a crossroads. Using 'proceed' is a technical term
contrasting with chih i h 'stay', proceeding from what is so of some to
what is so of all (§ 1/4/4). Since chih is frequently causative, we can
treat hsingjen in the same way, 'make the man proceed' (for causative
uses of technical terms, cf. § 1/3/13).
The first of the summings-up of Part B, probably eliminated from A
88-98 at Stage 2 (§ 1/2/2/4).
C. " F i x the kind, in order to 'make the man proceed'. Explained by: the
sameness.
E. The other, on the grounds that it is so of the instance here, argues that
it is so of the thing it is; I, on the grounds that it is not so of the instance
here, doubt that it is so of the thing it is."
In A 97, when considering how to describe one instance, we were like

a man standing at a crossroads (the metaphor developed in N O 10);
whichever criterion (yin) we chose, we were committed to describing in
the same way all objects of the same kind. By fixing what kind of thing the
object is, we 'make the man proceed' along his chosen road.
2/4 349
Traditional Western logic conceives the sentence as composed of

subject and predicate, and quantifies only the subject. But in the Chinese
sentence subject and object have equal status. For the Mohist, there is no
difference between proceeding from ' X has four legs' to 'Animals have
four legs' (B 2) and from yu aiyiijen 'loving some men' to aijen 'loving
men' (A 96).
B 2an 2 6 5
» m m m ± m ° mfe{2L)*±™ r J ™°
2 6 9
! i ( K ) < 0 > 268 £ , r « J H » T ^ J T * J K » r ^ a u m *
/h J ife * » 2
o7 0
2 6 5
Locating the divisions is a controversial matter both at the start and at
the end of this Canon. Sun located the start at this point (taking ssU
as the remains of a mutilated four-word phrase, 4^11 E9S), but most
prefer to take ssu yi shuo as the end of B 42, which immediately pre
cedes in the Stage 3 text ("The four are different. Explained by . . .").
But although ssu 'team of four horses' (also attested of sets of four
dragons or four men, M . 44683) may be written without the radical
( M . 4682 def. 5), no editor known to me has offered evidence of the
reverse, the numeral written with the 'horse' radical. I follow Sun's
division on the grounds that (1) an Explanation always has a head char
acter either at first or at second place, although, if the next character
happens to be the same, one occasionally drops out (§ 1/2/2/5/2, 4);
(2) the neighbouring Canons begin with two-word phrases.
266 Previous editors supply a lost character before variously identified
as £ (Sun, Chang Ch'un-yi), % (Chang Ch'i-huang), 1^ (Lu), «
(Kao, Wu, Liu). I prefer to emend ^ to i t , on the grounds that in
bronze script the graphs are often indistinguishable (Tuan Wei-yi,
14-16, 387), their confusion is very common (Chu Ch'i-feng 0124,
Chu Ch'ien-chih 84, 85), and the same corruption must be
assumed in B 82.
2 6 7
Since Stage 2 wu chin has stood at the head of B 3, and at Stage 3 it was
further separated from B 2 by the intervening Canon from the bottom
row. But comparison with the Explanation shows that the scribe at
Stage 2 made the division two places too early (§ 1/2/2/2/3).
2 8 8
There is no difficulty about taking ssU as the head character at second
place, since the dropping of one character when the next is the same
is attested in B 64, 69 (§ 1/2/2/5/4), and head characters which differ
in the radical are common in Part B (although not in Part A , where
A 39 is the only example).
269 p o r h double question with yii (interrogative form of yeh) in front of

t e
' Y yeh\ cf. the last sentence of B 10. The yii is not to be confused with
the conjunction yii 'and' of the concluding phrase, which may even have
been graphically distinguished (§ 1/2/1/2/14 (3)). Most editors follow
Sun in emending to ^rM 'oxen and horses'. But it is to be noticed
that when Hsiin-tzu discusses common names in his Right use of names
chapter his examples of the order below wu % 'thing' are 'animal'
(shou) and 'bird' (niao), and that sheng 'life' appears in another
context where he distinguishes between minerals ('water and fire'),
vegetables ('herbs and trees'), animals ('beasts and birds') and man:
ch. 9 (Liang 109/-4) "Herbs and trees have life but
not consciousness". For nominalised sheng 'living thing', cf. Analects
10/12 15'SHy^fe, " I f his lord gave him a living thing, he was sure
to rear it". Like the Mohist, Hsiin-tzu uses t'ui 'push, extend' of
advancing up or down the scale: tfkffti^^ "push forward and general
ise them", tf£ff5£lJ2l "push forward and subdivide them".
2 7 0
The last sentence may be taken as a contrafactual supposition, since
there appears to be no grammatical indicator of the contrafactual
(§ 1/3/H/6).
C. "When fours are different, explain the difficulty of extending from kind
to kind. Explained by: fixing 'the wider and the narrower', 'all of the thing'.
E. If we say it has four feet, is it an animal ? Or a living thing, or a bird ?—
'all of the thing' and 'the wider and the narrower'. If what is so of the
instance here were necessarily so of a thing that it is, all would be milu
deer."
This is an especially difficult section. We have offered solutions of the

textual problems on purely textual grounds, leaving the interpretation of
the argument out of account. There remains the problem of interpretation,
which centres on the ssu 'set of four horses' (or dragons, or men, cf. n. 265)
of the Canon, and the abrupt introduction of the milu deer in the Explana
tion. The milu deer is one of the more mysterious of the stock examples,
so many of which are animal (ox, horse, dog, crane), each with a constant
significance in disputation which the Mohist does not need to explain.
"Oxen and horses have four feet" (B 12), and so do all animals, but
not all living things. In Aristotelian terms, having four feet is the differentia
of animals. (Cf. Erh ya S № (ch. 17) SPPY 10/18A/3 r^SM^mZ l~$rJ '
0 £ f f № i ! 2 i T K J "The two-footed and winged are called 'birds', the
four-footed and hairy are called 'animals' ".) Presumably the significance
2/4 351
of the milu deer was also in its differentia. In § 1/5/5 we pointed out that
the milu deer has cavities under its eyes which were believed to contain the
eyes with which it sees at night, and that according to Huai-nan-tzu ch. 16
if a pregnant woman sees a milu deer her child will have four eyes (E9 @).
We may take it then that the 'fours' of the Canon are the four legs of animals
and the four eyes of the milu deer.
If I describe one milu deer as four-legged and four-eyed, I commit
myself to describing all animals as four-legged and all milu deer as four-
eyed. It is important to sort out the different levels of classification and to
establish the kind of thing with which one or other property is shared. If
"what is so of the instance here were necessarily so of a thing that it is",
then all animals, all living things, and therefore all birds (which are another
kind of living thing) would have the four eyes which are the differentia of
the milu deer.
Until recently I wavered between two false alternatives, that the t'ui
lei 'extending of kinds' of this Canon is either induction or inference by
analogy, which is its ordinary meaning elsewhere (Hsiin-tzu ch. 13, 22
(Liang 175/-2, 318/-2)). Lu-shih ch'un-ch'iu ch. 25/2 (Hsu 25, 3B/-2).
Huai-nan-tzu ch. 16, 17 (Liu 16, 19A/8f, 19B/7:17, 11A/2). The issue
centred on the contrast between the near demonstratives tz'ii and shih in
B 1, 2 and also B 33. Tz'u certainly refers to a particular, 'the instance here'
(cf. also A 97). But does one 'proceed'
(1) from the particular X to the particular Y {shih 'the one in question')
which is of a kind (lei) with it, or
(2) from the particular X to 'what it is' (shih), a kind of thing such
as milu deer or animal or living thing?
This is a point on which I have contradicted myself in print over the
past two decades (G(l) 288 n. 24, G(5) 19f, G ( l l ) 72). It is only since the
study of the pronoun system presented in § 1/3/4 that I have definitely
reached the conclusion that the first alternative is grammatically impossible.
If two particulars were being contrasted the pronouns would have to be
the near and far demonstratives tz'u and/)/ 'the one here' and 'the other
one' (cf. A 96, B 68). But t'ui lei does not seem to be induction either. The V
Mohist is concerned with consistent description, not with inferring from
the known to the unknown. Whether the cavities in the creature's head
actually contain eyes may not have interested him at all; the point is that
if you say of one milu deer that it is four-eyed you must say the same of all.
In B 76 he observes that, whether eyes are conceived as inside or outside
the head, it is k'uang chu 5f $ 'referring arbitrarily' to say that one is in the
head and the other not.
B3 2 7 1
[ « B 2 ]m £ » r — » B J » r ^ j » r * » J B J »reis(8a)
(§*)o«i(.^«-»r(^)*^ 5sHj-tbo-a Bfsii»·

m
: raj
ife:S^> rASii(te)*SJte 2 7 4
ofiJSB^fi >
SIR » *^ ( K ) 8 7 6
° « T * J >\>1
T A * J o ^ f t f * J · <J^> T * J ' ^ * 2 7 7
°f§
r*J »J^TM*J » S I
2 W
» r^HSMJ-tfeo
M i l * o 280 o lij&iijsHîfe · iSib»JA*n 281

»mm\
2 7 1
B 3 is variously treated by different editors, who take it as one, two or
even three Canons. But the problem was definitively solved by Luan
T'iao-fu, who demonstrated on textual grounds that the Canon is a
unity (§ 1/2/2/3). The phrase which has stood at the head of the
Canon since Stage 2 is the conclusion of B 2 (§ 1/2/2/2/3).
2 7 2
The graph B has variant forms M ( M . 23200, cf. 23199) and U
( M . 23213), and Kao Heng is surely right in emending to RJ> tniao
'short-sighted' in both Canon and Explanation (Kao 109). The latter
graph was apparently written unrecognisably at some stage in the
transmission of Mo-tzii, since the two examples in N O 18 (the only
other ones in the book) are corrupted to i&. The context requires a
stative verb parallel with pat 'is white', and the Explanation contrasts
'white horse' and 'blind horse' as 'big horse' and 'blind horse' are
contrasted in N O 18.
2 7 3
2 7 4
We correct one graph (*®) from the Canon. Pao 'wrap' must be the
name we are invited to combine with the words of both items, for it
cannot be a coincidence that in the Tribute of Yii (the patron sage of
the Mohists) in the Book of Documents (Karlgren p. 13, § 11,13) it is
combined with both words of the second item: J S M U S t t "their
wrapped oranges and pumeloes", ^ "the three-ridged mao-
grass that is wrapped and presented in bowls" (Karlgren). Combined
with the words of the first item however pao would be pao jfe 'hug',
cf. the proverb in Shih chi (ch. 55) 2045/-1 E M , * # l № i S "I have
heard that the son the mother loves gets hugged". This passage may
be corrupt, or merely elliptical. We have preferred not to risk a
translation.
2/4 353
Previous editors have taken the last four words liyiipao yeh as parallel
with the conclusions of the other items, and have therefore restored pao
at the same position in the Canon. But the whole context diverges from
the pattern of the first four items; the explanation of the next pair
('husband' and 'shoes') is introduced by jo 'for example', and follows
a closely parallel sentence to which there is no reference in the Canon.
Fortunately a key to the whole corrupted passage survives in the forged
T'ung pien M S I chapter of the Kung-sun Lung tzu, much of which is
patched together from misunderstood phrases from the Explanations
(p. 176 above). There we find It (written 8!) used three times of the
linking of names in phrases, a usage well attested in disputation
(§ 1/4/16); but the word paired with pao is not It but lei. When the
phrases pillaged from B 3 are identified it can be seen that a copyist
of Mo-tzii has twice mistaken ch V "K (a graph known to have been used
at one stage for ch V since some 80 times in Mo-tzii it is corrupted
to yi iff, Sun 194/2) for a partially obliterated It, written without its
radical as M ( M . 104), orfiB( M . 43): Kung-sun Lung tzu ch. 4 (Ch'en
154/1) S^ffe, g|§*¥. &HB¥ " I f you treat the
horse as yellow, is it of a kind with it ? If you treat the cock as blue,
does it conflict with it ?", B 3 HBj, Sfrfc * t f < | § § | > , *ftmB&
"When 'deemings' are linked, one cannot take it as necessary either
that they are of a kind with each other or that they conflict with each
other". (For the structure cf. A 29, 30 #Xfrfe, B 32 ^>&%M$L.)
Since wet is a technical term (constitute X or non-X) and is used
freely without a complement (p. 118 above), we need not hesitate to
treat it as nominalised.
It would appear from comparison with the Canon that linking of the
names X and Y is linking 'by means of (yi) X ' . The linking makes
the phrase jen shih fei 'others' approval or disapproval'. Cf. jen fei
'others' disapproval in E C 1.
The phrase is yung fu 'brave fellow' ( M . 2360/73); yi is restored on
grounds of parallelism.
The phrase is mat yi chii 'buy coat and shoes'. Cf. Kuan-tzU (ch. 80)
BSS 3/98/5 'sell their coats and shoes'. The Canon has a
different word for 'shoes' (lii).
For this use of wet after the verb (as in A 88 J&^c 'proved or not yet
proved', cf. § 1/3/5/4.
Fu < = Sc) 'flower, blossom' ( M . 32316) (T'an, Kao, Liu).
Wei shih 'refer to this (kind of) thing' (contrast wet chih 'refer to it',
used of a particular object in the previous sentence. Wei shih also B 72).
354 9
The 'Canons and 'Explanations'
Wei t'o (=flb, always written without the radical), 'refer to another*
(cf. § 1/3/4/8).
2 8 2
For this emendation, cf. § 1/4/9.
C. "With the same name, of 'are two' and 'fight', 'loves', 'eats' and 'sum
mons', 'is white' and 'is blind', or with others as linked to them, of 'husband'
and 'shoes', we discard one of the pair, yet inherently the thing is what we
called it. Explained by: the criterion.
E. 'They both fight', 'they are not both two' ('Are two' and 'fight').
. . . ('Loves').
. . . ('Eats' and 'summons').
'Most of a white horse is white', 'Most of a blind horse is not blind'
('is white' and 'is blind').
When deemings are linked, we cannot treat it as necessary either that they
are of a kind or that they conflict. What is deemed fei (not-this), if you link
jen shih with it (jen shih fei 'others' approval or blame) is not being deemed
not-this. For example, someone deemed a fu (husband), when you link
yung with it (yung fu 'brave man') is not being deemed a husband; but
something deemed chii (shoes), when you link mat yi with it (mat yi chii
'buy coat and shoes') is being deemed a pair of shoes ('Husband' and
'shoes').
Being two is lost with the one that is lost, does not remain with the one
that remains. Whether or not we are dropping 'They are two', only when
there are these objects is it said of them, without these objects it is not said.
It is not like 'flower' and 'beautiful'. If 'beautiful' is said of this, then
inherently it is this that is beautiful; if it is said of another, it is not the case
that this is beautiful; and if it is not said of it the converse applies."
This Canon anticipates the exploration of parallel sentences in N O

13-18, but the discovery of the sentence in Names and objects is still to
come. Here the Mohist is aware only of names (which for him include all
words) and their 'linkings' (li). If you add the same name to two descrip
tions of an object or objects, it may turn out that you have to drop one or
other. For example if you add the name chii 'both' to the names tou 'fight'
and erh 'two', you find that it is legitimate to say chii tou 'they both fight'
but not chii erh 'they are both two'. This sort of thing happens also
whenever a name is 'with others as linked to them' (BS£), a phrase which
is presumably intended to exclude words which are merely 'with' each
other in simple co-ordination, such as niu matfrWi'oxen and horses'. The
name fu 'husband', when combined With, yung to make yung fu 'brave man',
2/4 355
no longer calls the man a husband. This awareness that words change their
meaning in context is reflected also in the analyses of ambiguous words in
A 76-87, as well as in Hsiin-tzu's observation that "when its use and links
are both grasped we are said to know a name" (ch. 22 (Liang 318/2) fflM
But that a word is dropped or changes its meaning in combination

with others has nothing to do with whether it correctly describes the object.
What matters is still the yin 'criterion' of A 97, the part or instance which
entitles us to affirm or deny that the object fits the standard. If we remove
one of two objects, the one left does not entitle us to affirm 'They are two',
the one lost is the criterion for denying it. But as long as they are two it is
inherent in the objects that they are two, whether or not my choice of
words allows me to include 'two* in a combination. 'Two* is unlike
'beautiful', a relative word of the type discussed in A 88; a thing is inherently
beautiful only in relation to another thing which is ugly.
B 4 *njfli*if5— o ia^Bie(tf:)*fa »—n—»mm*m °

283
H. ( 5 ) 284
SSI» —r^yefi » ] R * f f t S £ o
2 8 5
283 This g p b survives only in E C 2, where it appears to be hsien M ( = 9 i )

ra
distinguished from chien HL by the 'man' radical common in Mohist

coinages. The correction is confirmed by the Explanation, where the
graph survives with the radical dropped as in A 1 M ^ I J ^ M "appearing
bringing about seeing".
284
The head character, as so often has entered one place too late. When
the Kung-sun Lung tzu forger borrowed this passage he failed to
recognise the head character (ch. 5, Ch'en 177 k>WF>M* [ H ^ S ] Ml,
——-vpfBS "Seeing and not seeing are separate, the one and the one
do not fill each other"). Cf. also Ch'en 181, which shows that (#)*H£
ksiu 'length* was not yet corrupted: MUt^viL, — S H , ^jlRiSrfiHS
Sife "Seeing and not seeing, the two and the three, fill each other like
length and breadth". Previous editors, assuming the authenticity of
Kung-sun Lung tzu, have supposed that the parallels confirm that the
Explanation has no head character and show that its topic is Kung-sun
Lung's supposed sophism of 'hard and white'.
2 8 5
For the syntax of chien pat (separate but mutually pervasive), cf.
§ 1/4/3.
C. "Even when one cannot be dropped without the other they are two.
Explained by: seeing and appearing, one and two, length and breadth.
E. Seeing and appearing are apart, one and two do not fill each other,
length and breadth are 'as-hard-to-white'."
2 8 6 28
B 5^tgM^* o ffrfiE*^ ? o
2 8 6
P u hat 'It is no objection', cf. § 1/4/12.
2 8 7
(*). Cf. B 82 Kffi^3Hi(=fll) "Explained by: not being applicable to
everything at once", B 69 Mffiffl "nowhere have a full range of
competence". Here the reference would be to the full range of human
faculties, cf. Huai-nan-tzu ch. 9 (Liu 9, 12A/7) ^ T S P W I , fBffiWW
/ f ^ i f e (of the deaf and dumb, who can do some jobs but not others)
"Because their bodies have limitations and their abilities have things
beyond their scope".
288
CM 'lift* is written without the 'hand' radical in the second instance;
the Horyaku edition omits it in the first instance also (JSJIL "lift a
weight", as in A 21). For the omission of the radical in chii 'lift', cf.
§ 1/2/1/2/4(1).
2 8 9
For this corruption, cf. § 1/4/29.
290 The Mohist's graph is unknown in any relevant sense ( M . 43524).
I follow Sun's emendation to ch'i as in ch'i ou ISfPi 'odds and evens'
( M . 35105/2). He takes the sentence to refer to a guessing game
("whether what is in a closed hand is odd or even"), wrongly on the
present interpretation.
C. "It is no objection that one is unable. Explained by: the full range of
our capacities.
E. T o lift a heavy thing but fail to lift a needle is not something for which
one's strength is held responsible. Whether being deemed a complement
is one-sided or double is not something for which the understanding is held
responsible. (Like the ears and eyes.)"
This section follows directly on to the last. It is objectively inadmissible

(pu k'o) to dismiss one without the other of ' X sees Y ' and ' Y appears toX',
' X is one and Y is one' and ' X and Y are two', ' X has length' and ' X has
breadth'; it is no criticism of one's intelligence that one is unable (pu neng).
A man has his full range of capacities (chou) if intelligence and strength,
ears and eyes, all perform their separate functions, but within the limits
set by objective conditions. (Cf. Mo-tzu ch. 6 (Sun 21/—5) B ^ t & f f i i i i ,
2/4 357
^^§fe№$k, U ^ f £ S № "The eyes are unable to look everywhere, the

hand is unable to grasp everything, the mouth is unable to taste every thing".)
The strongest man cannot lift a needle too fine for his fingers to grip, the
cleverest man cannot turn every one-way complement ( ' X implies Y') into
two-way complements ( ' X implies Y and Y implies X ' ) .
(*) ° » »mmnnmm^n * mmwm»*

jesses · (ra)*»aiiw№B«» o
This is the only instance in the text as we have it of pi 'compare'

distinguished by the 'mouth' radical, just as pi 'measure side by side'
is only once distinguished by the 'man' radical (A 68). The original
text probably distinguished them systematically.
Shih 'louse', stock example of a name applied to two objects (as identical
in sound with se 'zither'). Cf. § 1/5/6.
Although the last two examples may be radically corrupt, as commonly
supposed, it is remarkable that the key words in both have very rare
readings, and that the five words with these readings recorded in
Karlgren's Gratnmata sericafitvery neatly into the passage: * G ' A K
(1) 1117b, noun 'crane', (2) 1117b, stative verb 'sleek'. *SIsT (1) 506a,
noun 'louse', (2) 411a, noun 'zither', (3) 411a, stative verb 'glistening'.
C. "Different kinds are not comparable. Explained by: measuring.

E. Which is longer, a piece of wood or a night ? Which do you have more
of, knowledge or grain ? Which is the most valuable, aristocratic rank, one's
own parents, right conduct, a price ? Which is higher, a deer or a crane ?
# #
Which is more G ' A K (sleek), a deer or a G ' A K (crane) ? Which is more
*SIsT (glossy), a *SIsT (louse) or a *SIsT (zither)?"
The first example is directed against sophistries of the type of " A

tortoise is longer (/lives longer) than a snake" (Chuang-tzu ch. 33 (Kuo
1106/2) H S & 4 6 ) .
B 7 M&IOa'Jr * ift##t 294

o
294
Ku 'the thing as it inherently is', the noun corresponding to the adver
bial ku 'inherently' of B 3 (§ 1/4/14).
C. "When one or other is dropped there is no reduction of the things.

Explained by: the things as they inherently are.
E. In both being as one nothing has altered/'
This Canon is the bridge between the pu chii erh 'They are not both
two* of B 3 and the chiiyi 'Both being as one" of B 12. When we add chii
'both* to tou 'They fight* and erh 'They are two* we have to drop the erh,
cannot say chii erh 'They are both two'; nevertheless the objects remain
'inherently* (ku H ) two (B 3). We can however say chii yi 'They are both
as one* (in fighting), just as ox and horse are both as one in having four legs
(B 12). The verbal change still does not alter the fact that the objects as
they inherently are (ku #C) remain two.
B 8 ® ^ 2 9 5
' t e ? E ^ o
295
Pet 'self-refuting, self-contradictory* (§ 1/4/23).
2 9 6
Embedded 'Yyen' clause: 'loan-named as being a crane* (p. 155 above).
2 9 7
For shih 'clan-name* after a dog*s name, cf. p. 220 above.
C. " T o loan-name is necessarily self-contradictory. Explained by: not

being so of it.
E. What it is loan-named it necessarily is not, otherwise it would not be a
loan-name. When a dog is loan-named as being a crane, it is as when one
gives it the clan-name 'Crane*.**
Since the dog is the stock example of a thing with two names (§ 1/5/3),
and a dog named 'Crane* the stock example of borrowing the name of one
thing for another (§ 1/5/4), there can be little doubt that T*an Chieh-fu is
right in taking chia (literally 'borrowing*) as metaphorical naming. The
Tso-chuan (Duke Huan 6/5) distinguishes five kinds of personal names,
among which "those taken from other things constitute the 'loan-names* '*
(ffrlfe^M r J ); an example supplied by the commentator T u Y i itt5Bis
the name of a son of Confucius, Li fil ('Carp*). Cf. also Chuang-tzu ch. 25
(Kuo 917/4) J ffi№Mff " 'Way* as a name is something we
loan in order to walk it**.
The difference between naming a dog Crane and saying that it is a
crane is discussed in B 72. The loan-naming to which the Mohist objects
is the latter, a philosophical device of relativists of which a striking example
is the sophism " A dog may be deemed a sheep'* (Chuang-tzu ch. 33
2/4 359
(Kuo 1106/1) What is wrong with it is that a thing is judged

to be a dog, sheep, or crane by a standard, which is "that in being like
which it is so" (A 70 #r3=fffn #&-&), and whatever may be the characteristics
of the thing from which the dog is loan-named they "are not so" ( ^ $ t )
of the dog.
In Names and objects the chia 'loan-named* appears again (NO 5). If
our rearrangement of the fragments is correct at this point, it refers to the
borrowing of a place name for a thing no longer in the place, for example a
"Ch'in horse": "The loan-named is not so now" ^ v f ^ f f e ) . This
kind of loan-naming is tolerated as a 'term of convenience' (NO 2).
B 9 HxZSx» »iswmftA*n±»o °
(»)o*«^»r«fJ-lb«a±' J298Ô(^)*^299^, f J
-fee
2 9 8
The radical which distinguished chih %H 'know* from chih ftl 'the
consciousness', systematically erased in the Canons and up to this
point in the Explanations, is more and more common in the Explana
tions from B 9 onward.
2 9 9
Emended following Sun.
C. "Why a thing is so, and how I know it, and how I make others know
it, are not necessarily the same. Explained by: sickness.
E. That someone wounded him is why it is so. That I saw it is how I
know. That I tell them is how I make others know."
A 80 distinguishes three sources and four objects of knowledge. This

and the next Canon employ this classification, B 9 considering the sources
of knowledge and B 10 its objects. The sources are report, explanation and
observation. I may know that something is so by observation, and enable
others to know by report, while the cause of it being so is known by
explanation.
Why does the Mohist think it useful to tell us that why a thing is so,
and how I know it, and how I make others know it, "are not necessarily
the same?". Because the Mohist summa is concerned only with knowledge
by explanation, in which they necessarily are the same. We know by
explanation that "a circle is nowhere straight" (A 98). That a circle by
definition has the same lengths from one centre (A 58) is the reason why
this must be so, and also the principle by which I know it and by which
I make others know it.
B 10M°tfc£IÊ>«>® >$É3ooo
(K) - «f§&»j < a > ± ^ 3 0 1
» S#A* ° , (£)·»& o
3 2
^gij
3 0 0
The last four words all appear in corrupted forms in the Explanation,
where they are corrected from the Canon.
3 0 1
I propose to take the p'eng as p'eng-p'eng ( M . 31720/92) 'dense', which
appears with the radical dropped in Mo-tzti ch. 46 (Sun 266/-3 %k%k
"Densely the white clouds gather"), and the zoua&wuffc 'mist'
with the radical dropped. With the restoration of a yi the sentence
would then present a situation like the one in Hsiin-tzu ch. 21 where
a man in the evening takes his shadow for a ghost. (Liang
303/11-14).
3 0 2
Cf. the Shuo wen definition of lu 'shed, booth' (Tuan 447B/-8):
$Cîfe, # J C B " A lu is a lodging; it is abandoned in autumn and
winter, lived in during spring and summer".
3 0 3
Fei, 'stand something on the ground' ( M . 8492 def. 7), its only meaning
in the dialectical chapters (also B 27, 29).
3 0 4
Sun identified this graph as ffi ( M . 14496), 'shave wood'. Cf. the Shuo
wen definition: (Tuan 270B/8) t№trffe "to scrape off rough wood".
305 The illustrative gloss has entered the text six characters too late.
306
Pi ( M . 9644 def. 9), 'collapse from exhaustion' also written f£, 9k. The
latter characters are variants for it in Sun 304/13, 305/2.
* » > « o r ' (§1/3/11/8).
308 Yiyi zveijan "suppose the already ended to be so"; the phrase reappears
in B 33, also in connexion with kuo 'having passed'.
C. "Doubt. Explained by accidental, easy, coinciding or transient circum

stances.
E. In a dense mist mistaking a man for an ox (?), the dweller in a boo
being cold in summer: 'accidental circumstances'.
T o lift things when they are light and put them down when they are heavy
is not a test of strength (for example, stone and feathers), to shave wood
along the grain is not a test of skill: 'easy circumstances'.
Whether the fighter's breakdown is due to drinking wine or to the midday
sun cannot be known: 'coinciding circumstances'.
Is it knowing ? Or is it supposing the already ended to be so ?: 'transient
circumstances'."
2/4 361
At first sight the classification of doubt seems quite unsystematic, but

on closer inspection one sees that it co-ordinates with the four objects of
knowledge in A 80 and the fivefold division of the Canons. In the later
Mohist scheme there are four disciplines, of which the first two allow the
temporary validity of chih i l l 'staying* and the last two the unending validity
of pi 'necessity'. This section identifies the kind of doubt proper to each
discipline, in the order of the fivefold division of the Canons:
(1) Explaining how to relate names to objects. The procedures of
consistent description establish which objects are to be called 'ox' (for the
ox as the stock example, cf. § 1/5/2) and oblige us to extend the description
of any object to all objects of the same kind, to say that not only this summer
is hot but summers in general are hot. Doubt arises from accidental
circumstances, leading us to mistake a man for an ox or to an unexpectedly
cool summer.
(2) Explaining how to act. By the procedures of practical and ethical
judgment we decide whether a man is righteous, whether he has the intent
and the ability to benefit others. Doubt arises from easy circumstances
which do not put a man to the test, as when we remain doubtful of a man's
strength and skill if we have never seen him lift any but light weights or
plane except along the grain.
(3) Explaining objects. The causal relations between objects are
necessary; but doubt arises whenever possible causes coincide, as in the
case of illness (the stock example of an event with more than one possible
cause (§ 1/5/11)).
(4) Explaining names. For this discipline there is no corresponding
category of doubt, since the necessity of the logical implications of
definitions is transparent. "The 'necessary', admit and do not doubt"
(A 83).
Between the two pairs of disciplines the Canons have a further division
on the problem of the validity of knowledge in changing conditions, serving
as the bridge between the temporary validity of the first pair and the logical
necessity of the second. Here the doubt occasioned by transient circum
stances is mentioned last: "Is it knowing? Or is it supposing the already
ended to be so?".
B 11 £J&£— · ^ 3
<38> S o t f t f t ( f i ) « » o
0 9 3 1 0
# u
8 1 2
 <B43£±
814
· r j m r * J · ( 5 g § § « j s & ± « o · mm · >
315
B 12 Bfc fe—«·& ° tfrffft— 316
»t6S 3 1 7
°
r№AJ
318
( ( · ) · « ) ° T « - J ' « « 0 >1t r^J f"JHJ o
» fio^js- o * № B S » o (^JK» . f g i f f i s - o)
3 0 9
'add name to name', as in Kung-sun Lung's n T M J IS f S J *
/
T J "Putting together 'horse* and 'white' one compounds

the names as 'white horse' " (cf. § 1/4/11).
3 1 0
For other cases of the dropping of the second huo, cf. p. 129 above.
3 1 1
For the corruptions of ch'ii E 'separate off as a group' in B 11, 12,
cf. § 1/4/7.
3 1 2
Emended on grounds of the parallelism with the next pair of clauses.
Jan has a variant graph H ( M . 19623); possibly the scribe as his eye
returned to the page saw the preceding li as the top of it.
3 1 3
The 'six storehouses' (f\J($ M . 1453/518, 520), water, fire, metal, wood,
earth, grain. The phrase appears in Mo-tzu ch. 25 (Sun 112/4).
3 1 4
B 11 is missing at its proper place in the Explanations, certainly because
of textual mutilation; although many of the definitions of A 1-75 lack
Explanations, the rest of the theses of A 88-B 82 have them. A section
on the five elements (B 43) includes two fragments, unintelligible in
their context, which look as though they have been packed in at the
beginning and end because they refer to water, earth, fire and wood.
But the word fu 'storehouse' suggests that these are conceived as items
in the 'six storehouses' (the six kinds of natural resources at man's
disposal, which include all the elements), not the five elements them
selves. The fragments may be identified as belonging to the missing B
11 by the word ho 'put together', by the examples of 'adding name to
name', and above all by the illustrative gloss (incorporated in the text
three places too early), which pairs neatly with that of B 12.
3 1 5
C f . n . 311 above.
316
Chii yi 'both are as one', as in B 7, following on to the pu chii erh 'they
are not both two' of B 3.
317
Wei shih 'be specifically this thing' (p. 119 above).
3 1 8
Since tang is the fitting of name to particular object (cf. § 1 /4/28) tang
niu ma will be "fit either 'ox' or 'horse' ".
C. "Together or one thing: in the former case one adds name to name,
in the latter not. Explained by: marking off as a group.
E . You put together 'water and earth', and from 'a fire and a fire' separate
'a fire'; you put together these fu (water, earth, fire), and from 'the wood
and the wood' separate 'the wood'; it depends only on convenience.
2/4 363
(For example, noting the numbers of the deer and the fish.)
C. Things marked off as a group are one unit. Explained by: both being
as one, being this thing specifically.
E. 'Both being as one': for example, 'oxen and horses have four feet\
'Being this thing specifically*: they fit 'ox* or 'horse*.
If you count oxen as an item and horses as an item, oxen and horses are
two. If you count oxen and horses as an item, oxen and horses are one.
(For example, counting the fingers. The fingers are five but the five are
one.)"
In the Mohist terminology, counting is of units (t'i S ) in collections

(Men H&) which may themselves be treated as countable units (cf. A 2 n.).
Inside a particular object (shih Iff) we may divide and subdivide parts
(such as the five fingers which may be taken together as one part of the
body, which in N O 7 is contrasted with the head); we may also count the
mutually pervasive hardness and whiteness of a stone as two but inside the
one stone (B 37). On a higher level there are classes and sub-classes of
things (wu which are 'what (the object) is* (shihM), such as ox or horse,
animal, living thing (B 2); and on each level things may be counted as
together (ho £ ) or separated (It 81) as single things. If we can depend on
a couple of emendations, the Mohist explicitly describes his principle as
'establishing the level/the station* (B 59 S *ti£) within which things 'stand
with* each other (A 83, on ho 'together*: or are 'with each other*
(B 65 #=U§; both Canon and Explanation also have ffi * £ 'together with
each other*). This principle excludes sophisms based on counting together
things on different levels (B 59, note). The placing of whiteness on a lower
level than the object also disposes of an assumption behind Kung-sun
Lung*s ' A white horse is not a horse*, that 'white* and 'horse* are put
together (ho) as 'white horse* (cf. B 37 n.). Cf. also p. 183 above.
Why does the Mohist insist that to be one thing is to have one name,
although willing to recognise that "things marked off as a group are one
unit**? The crucial example of two names implying two things is niu ma
'oxen and horses*, which reappears in B 67. In Classical Chinese the line
between regularly co-ordinated words and a single compound word is very
fluid (the former having in fact developed into the bisyllabic words of
modern Chinese), and it would seem quite reasonable to take niu ma as
one kind of thing ('livestock*) at the level between oxen and animals. But a
particular object is an ox, is an animal, but is not an ox-or-horse. The whole
of B 12 is devoted to clarifying this point; oxen and horses "are both as
one**, for example in being four-footed, but a particular object "is specifically
this", fits either the name 'ox' or the name 'horse'. That this analysis was
relevant to some current issue in disputation is shown by B 67, which
refutes with remarkable subtlety a thesis which in our ignorance of its
context seems quite pointless, that "Oxen and horses are not horses". Very
probably sophists were assuming that niu ma is one kind of thing which
has four feet (cf. B 12 "Oxen and horses have four feet") and is obviously
neither a niu or a ma. The Second List of sophisms at the end of Chuang-tzu
includes " A yellow horse and a black ox are three" ( S ^ S S ^ H : yellow
horse, black ox and four-footed ox-horse?) on which Ssu-ma Piao com
ments: " 'Ox and horse' treats two things as three, 'oxen', 'horses' and
'oxen-and-horses'" (Kuo 1110 n. 20 ^ J U G d f i H » B [" ^ J ' B J »
H i " * * J).
In Western terminology we may say that the Mohist's problem is the
relation between the collective and distributive use of common names. It is
considered again at the end of Names and objects: 'Some horses are white"
(JisSSS) implies more than one horse, "Horses have four feet" ( H E S S )
implies only one. In B 3 tou "They fight" refers distributively, erh "They
are two" refers collectively, so that the distributive chii 'both' can be
introduced only before the former. "Oxen and horses have four feet"
refers distributively; it says nothing about the number of oxen and horses,
but pronounces them all as one in having four feet. Consequently any
number of particular objects which are as one in some respect can count
as one group or sub-group, 'oxen and horses', 'oxen', 'horses'.
2/4/2/2 Knowledge and change (B 13-16)

The transposition of the Canons of B 14b-21 and 22-24a at Stage 1
(cf. § 1/2/2/1/2) has dislocated this sequence in the standard text of Mo-tzu.
Scholars at first supposed that whole Canons had been transposed; Luan
T'iao-fu (Luan (1957) 14) was the first to locate the break after the second
character of B 14 and show that when the two fragments are joined B 14
turns out to be parallel and contrasted with B 15. Even when this step is
taken the whole sequence remains unintelligible on the traditional assump
tion that the chien pai 8 £ £ of B 14, 15 refers to the 'Hard and white'
sophism attributed to Kung-sun Lung. This misconception has forced
editors to ignore Luan T'iao-fu's conclusions and propose various other
ways of dividing both Canons and Explanations, backed by arbitrary
deletions, emendations and transpositions. But once it is recognised that
chien pai is a technical term for 'mutually pervasive' defined in A 66
(cf. § 1 /4/3) there is no longer any need to question the soundness of the
text, apart from a couple of cases of systematic corruption already corrected
2/4 365
by Sun Yi-jang (Sun 203/5, 226/4) and a widely accepted transposition of

two characters in B 16 (Liang, Kao, W u , Liu). If the text is accepted the
ordinary rules for locating head characters (§ 1/2/2/5) allow only one way
of dividing the four sections. One more preliminary step is required before
we can interpret the sequence with confidence, a recognition of the
organisation of the document as a whole (§ 1/6/1), within which this group
of theses and its counterpart among the definitions (A 40-51) share a single
theme, the spatial and temporal conditions of knowing.
The argument of the sequence is a criticism of the assumption that
space and time are mutually pervasive like the hardness and whiteness of
a stone. Its objection is that it is the moment which is mutually pervasive
with space, since at any moment the whole of space is present and all other
times are absent. Space itself moves, and its movement is duration. From
this seemingly very abstruse proposal an unexpectedly practical conclusion
is drawn in the last section (B 16); a method of government suitable at one
time may no longer be appropriate when space has moved on to another
time.
This argument was evidently important for the later Mohists, since
it is reflected in their consistent use of yii ? 'space* and chiu iK 'duration*
where other pre-Han sources use yii and chou Hf. It may be noticed that in
the other sources yii and chou do not seem to be clearly conceived as space
and time abstracted from the physical cosmos; they seem rather to be the
'cosmos as it extends* and the 'cosmos as it endures*. If so, they are
ultimately identical, like the stone as hard thing and the stone as white
thing. Both nouns are metaphorical derivatives from the names of parts of
the roof of a house (yii 'eaves* and chou 'ridgepole*), written with graphs
having the 'roof* radical. Yii chou 'the cosmos* and yii 'the cosmos as it
extends* are both common expressions, but chou alone is very rare.
Chuang-tzu ch. 23 (Kuo 800/2) ^HM^m^ » ° Wftrfn**^]* ·
"What has solidity and resides in nothingness is the cosmos-as-it-extends.

What grows older but has no root or tip is the cosmos-as-it-endures."
Huai-nan-tzU ch. 11 (Liu 11, 13B/-2f) » ra*±Tfi±? »
"What goes back into the past and comes down to the present is called the
'cosmos-as-it-endures'. The four directions and above and below are called
the 'cosmos-as-it-extends'. The Way is between them and no one knows
its place." (The Huai-nan-tzii definitions, which remained standard, are
also found in Shih-tzu SPPY, B 5A/-3.)
It may be noticed that the second pair of definitions separates yii and
chou ("the Way is between them"), and therefore goes at least part of the
way towards abstracting them from the cosmos as space and time. In the
Chuang-tzû passage however yii is conceived as solid like particular
objects.
The Mohists use the noun yii 'space', but in stead of the noun chou
they choose the nominalised intransitive verb chiu 'duration' (defined beside
yii in A 40, 41). They use chiu verbally ('endure, continue', A 50, B 14,
46, 64) as well as nominally, and put it beside yii only in order to contrast
the two (B 14). Syntactically the counterpart of chiu is not yii but the
intransitive verb hou № 'dimensioned' (wu hou M№ 'the dimensionless,
the point' : wu chiu 'the durationless, the moment'). The effect is to
destroy the symmetry of time and space (it is no longer possible to think of
events as inside time as they are inside space) and to convert time from
a static to a dynamic concept. One is reminded of Bergson's criticism
of the assimilation of time to space and his choice of the word durée
'duration'.
The target of the Mohist attack is an assumption that principles of
government valid throughout the world at one particular time are valid
for all times. Were there thinkers who inferred to this conservative
conclusion from the interrelationship of time and space? Someone may
well have argued that since the spatial and temporal aspects of the cosmos
interpenetrate like the hardness and whiteness of a stone, and any generali
sation about a stone which is true of it wherever it is white is true of it
wherever it is hard, it follows that any generalisation about men or things
which is true of them throughout space is also true of them throughout time.
But the Mohist may be uncovering a hidden analogy behind a preconcep
tion rather than answering an explicit argument. We can easily see that as
long as the word chou hindered thinkers from fully abstracting the concept
of time from the 'cosmos as it endures', they would be inclined to assume
that principles accepted as valid throughout the cosmos at any one time
ought to be valid for all time. A document of some interest in this connec
tion is the Chou ho EB'R chapter of Kuan-tzu, also written in canon/
explanation form, which expounds the unchanging principles of govern
ment which will guide us through changing situations. The obscure title
chou ho ('the joining of things by the cosmos-as-it-endures', primarily the
joining of heaven and earth, which contain all other things) is explained in
the last of the 13 sections. It seems to refer to a concept more abstract than
chou alone, the cosmic order seen primarily in its temporal aspect, and is
described very much as though as it were the Way itself. It is convenient
2/4 367
to translate it by 'the order of the cosmos-as-it-endures', although this

requires us to render the verb ho by an unwieldy phrase, 'join in one
order*.
Kuan-tzu (ch. 11) BSS 1/47/1, 52/14-53/5 » (=X)
^*i:nt4&»tea r » 4 & £ * J · r J · ± s » ^ £ ± » T t Mtktk
(Proposition 13) "Heeavn and earth are the container of the myriad things,
the order of the cosmos-as-it-endures in its turn contains heaven and
earth."
(Exposition). "Heaven and earth wrap the myriad things and therefore are
called 'the container of the myriad things'. As for the idea of 'the order of
the cosmos-as-it-endures', above it penetrates above heaven, below it
descends^}) below the earth, beyond it emerges beyond the four seas, it
joins in one order and threads together heaven and earth and makes one
envelope. Divide it up, and it reaches where there is no interval and stops
in the unnameable. This at its greatest has nothing outside it, at its smallest
has nothing inside it, and therefore it is said 'In its turn it contains heaven
and earth'. If its examples are not handed down, not one item in a statute
will be fathomed, nothing however trivial will be properly governed."
It may be noticed that Chuang-tzu and Huai-nan-tzu are as aware as
the Mohists that times change and that the ancient methods of government
no longer apply, but they follow a different course. They continue to speak
of yii and chou and the eternal Tao within them; but their Tao is the
principle beyond the principles which can be formulated in words, all of
which are relative.
B 13 ? ^ * # f e » i £ & K ° o
3 1 9
3 2
ft»(^ffiffn^ ^?* 321
319
Huo 'in some direction' (pp. 129f above).
320
Chang (rising tone) 'be prolonged, grow older', used also in the Chuang-
tzu definition of chou Hf 'time' (quoted p. 365 above).
321 =X yu 'again'.
C. "Space travels in a certain direction. Explained by: growing older.

E. Growing older, one travels and again occupies space."
B 14
3 2 2
The Canon was broken at this point by the transposition of B 14b-21
and 22-24a (cf. p. 364 above).
323 x yu « ain',
B ag 13, n. 321.
a s i n B
3 2 4
= S m M 'evening', as in A 40.
C. "Space and duration are not 'as-hard-to-white'. Explained by :·(?).

E . North and South are present at dawn and again at nightfall. The
travelling of space has duration."
B is *t&#s °
Wu chiu 'the durationless, the moment* (cf. A 44, 50).
C. "The durationless is 'as-hard-to-white' with space. Explained by: the

criterion.
E. Because where the hard occupies the white they necessarily fill each
other."
The summing-up ("Explained by: the criterion") is shared with B 3.

The criterion (yin) for some statements is whether they apply to all ('He
loves men', false if there is someone he does not love), for others whether
they apply to some ('He rides horses', true if he has ridden one), cf. § 1 /4/39.
Being 'as-hard-to-white' requires that the mutual pervasion be total. Space
and the moment satisfy the criterion because all space fills one moment and
one moment fills all space, as already shown in B 14 ("North and South are
present at dawn and again at nightfall").
B 16 ffi326^gff^327 , *|£R£|328 o ft&2$ftS» o

(«) · rmmm J > S4«§£i&-ife o a * f t ± 4 · mr j °
326
Tsai causative ('cause to be present, locate'), but still used as in B 13,14.
Cf. p. 163 above.
327
So jan 'where it is so' (§ 1/3/12/3), here referring to a time.
3 2 8
Most editors transpose to wet jan che 'the not yet so' (Liang, Kao, Wu,
Liu). Other proposals arefif7fc#&(Sun), 3!f^c#& (T'an).
2/4 369
329
Shih is used in other elliptical summings-up (B 12 t i ^ "It is specifically
this", B 80 3=f ;Sk "It is as much as this"), the reference being established
by context. Here the reference is to time (yii shih 'at the time in ques
tion', the yii connecting with the earlier chu fit = ;£jft).
C. "Locating it at the time when it is so or during the not yet so. Ex
plained by: at the time in question.
E. 'Yao is good at ruling' is from the standpoint of the present locating
him in the past. If from a standpoint in the past someone located him in
the present, it would be 'Yao is unable to rule'."
2/4/2/3 Problems in the sciences (B 17-31)

The principal surviving document of early Chinese science is the Chou
pet suan ching JSfftlfffi (in its present form not earlier than 200 B.C.),
which calculates the dimensions of heaven and earth and the motions of
the heavenly bodies by applying the geometry of the right-angled triangle
to observations with the gnomon. The gnomon (piao H ) is a post of stand
ard height; by day the astronomer uses it to estimate the direction and
length of the sun's shadow, by night to align with stars a cord extended
from the top of the post to the ground. The early Mohists were already
familiar with this technique. Mo-tzu ch. 35 calls the three tests forjudging
controversial doctrines (ancient authority, empirical observation and
practical effectiveness) the 'three gnomons':
Sun 169/-2—170/4 s f f i M » *mm&±±MiLtl9'#fe °... ftÊ^r
" T o have no model for what one says may be compared to establishing the
directions of sunrise and sunset on a rotating potter's wheel. . . . Therefore
for what one says there must be the three gnomons."
A section on the gnomon at the end of ch. 3 of Huai-nan-tzu (translated
Cullen, op. cit.), out of keeping in several respects with the main tradition
of Chinese astronomy (cf. Maspero (1929) 347-354), may well come from
a Mohist source. (Its successive items strikingly resemble the Explanations
of the scientific Canons.) It begins with a description of 'adjusting sunrise
and sunset' (determining how far they diverge from true East and West):
Huai-nan-tzu ch. 3 (Liu 3, 32A/11-32B/2) J E W * » »M13 » &
»o&mntemtm° 0 (it) * S A » X « M R » J I № *
1
" T o adjust sunrise and sunset, first plant a gnomon on the East (E ), with
draw holding another gnomon ten paces behind the front gnomon (W), and
align them with the North corner( ?) of the sun when you first see it emerge.
2
When the sun is about to go in, plant another gnomon to the East (E ), and
use the West gnomon to align it with the North comer (?) of the sun when
you see it just going in. Then if you fix a position midway between the
3
pair of gnomons (E ), it and the West gnomon will be due East and West."
The Chinese, like the Greeks, noticed that the shadow at noon shortens
as we travel South. Eratosthenes (276-196 B.C.) rightly ascribed the
reduction to the curvature of the
earth, and succeeded in deducing a
remarkably near approximation for
the earth's circumference. The
Chinese ascribed it to the diminish
ing angle as we approach the posi
tion directly under the sun, and tried,
on the assumption that the earth is
flat, to estimate the height of the sun
at noon (which they took to be the
height of the sky):
Huai-nan-tzu ch. 3 (Liu 3, 33A/7-
10) ® c a ^ ± i a f ' » * w - 3 t » j E W j h
j l t I I 0 T i » > Hrfn££ » ffi+M

a · ffjxffi* o £ t t f t n $ £ 9 . mix
2/4 371
"If you wish to know the height of the sky, plant gnomons 10 feet high
1,000 //apart due North and South. Measure their shadows on the same day.
If that of the North gnomon is 2 feet and of the South gnomon 1.9, it
follows that every 1,000 It further South the shadow shortens by 0.1 and
that 20,000 It South there is no shadow, which is being directly under
the sun.
At the place where for a shadow of 2 feet you get a height of 10 feet, the
height is 5 times the distance Southwards; so if you take the number of //
from here to the place under the sun and multiply it by 5, making 100,000
li, you have the height of the sky. Supposing a place where shadow and
gnomon are equal, height and distance will be equal."
The Choupei treats astronomy and the geometry of right-angled triangles
as a single science. There can be little doubt that for the Mohists this was
the basic science. The definitions corresponding to the theses on the
sciences (B 17-31) are the geometrical definitions of A 52-69. These
assume knowledge not only of practical geometry but of astronomy; they
include a definition of noon as the time when the sun is due South (A 56).
We may note also that the alignment (ts'an # ) of gnomons with the sun
is used both to define a straight line (A 57) and to illustrate the pointing
out of a single property in two objects (B 38), and that a term from the
geometry of the right-angled triangle, hsien 3£ 'hypotenuse', appears in
B 27. The Mohist optics, exclusively concerned with shadows, no doubt
originated from questions raised by the shadows cast by the gnomon.
Indeed one can perceive no principle of unity behind the Mohist choice of
scientific topics (geometrical definitions, optics, mechanics) except that all
are on the borders of geometrised astronomy.
Why then do the Canons include no section on the most important of the
sciences ? Probably because the topic of the summa is disputation, and the
purpose of the optical and mechanical sections is to illustrate the kind of
reasoning proper to the sciences, causal explanation. (Cf. § 1/1/2/6.) The
Mohists are not satisfied with explanations in terms of the Y i n and Yang
and the Five Elements, and in the case of astronomy they would have no
causal explanations to fill their place. They must have had some astrono
mical document, and we have noticed the possibility that the Huai-nan-tzu
fragment from which we quoted is of Mohist origin. However, the docu
ment need not have been written inside their school; it might be one of the
lost astronomical treatises recorded in the Han bibliography, or even some
form of the Chou pei itself.
2/4/2/3/1. Optics (B 17-24).

The Mohist optics is primarily the study of shadows. The word ying
jp; = (j!5), 'shadow', appears in every one of the eight relevant Canons, the
word kuang^t, 'light', only in some of the Explanations. The Mohist, as one
would expect of a pioneer of optics, knows nothing about the propagation of
light except what he can learn from the silhouetted shapes and straight edges
of shadows, and how they vary as the relations between object and light
source change. In the present version we translate ying consistently by
'shadow' and resist the temptation to shift to 'image'when the Mohist turns
his attention to mirrors. From his point of view nothing has changed except
that, for reasons he does not or cannot explain, shadows on lustrous surfaces
show up the detail as well as the outline of the object. He can explain the
inversion of the image in a concave mirror only by what he has learned from
the study of true shadows, that is, in terms of light obstructed by the object,
so what he explains is the inversion of the silhouette (B 19, 23). In spite of
this he makes the very great achievement of explaining the inversion of
the shadow by a geometrical figure which is not visible to the eye, a cone
of shadow shrinking to the focal point and then opening out in a second
cone which has rotated (yiin *3I B 19) halfway round, so that the image is
upside down. Shen K u a (A.D. 1031-1095) got no farther some 1,300 years
later, when he compared the figure that he postulated to a drum narrow at
the waist (Meng-ch'i pi-t'an (Hu Tao-ching, item 44), translated Graham
and Sivin 145-147).
The Mohist speaks of the light following straight paths which intersect
at the conjugate focus or at the centre of curvature, two points which he
does not distinguish (B 23). Has he the concept of light rays? The only
straight line in optics which is obvious to the eye is the edge of a shadow,
cast for example by a wall in a strong light. The Mohist's intersecting lines
and Shen Kua's waist drum, conceived rather than perceived, are the
straight edges of cones of shadow. So long as he deals with shadows and
silhouettes, the Mohist can offer not only observations but explanations,
deriving their variations from the figure of the object which obstructs the
light, relative dimensions, distance, inclination, reflected light, curvature
of the reflecting surface, and single and multiple light sources, all depend
ing on the principle laid down in the first optical Explanation, "Where the
light reaches the shadow disappears". T o an extent quite unusual in early
Chinese science we find ourselves in a landscape of geometrically visualized
figures, very far from the metaphysics of Y i n and Yang and Five Elements,
of numerology, time cycles, and resonances. The same is true of the
propositions on mechanics. Indeed the only reference in the Canons to
2/4 373
Yin and Yang or Five Elements is a refutation of the theory that there are
regular ascendancies among the latter (B 43).
The distance from the methods of explanation characteristic of Chinese
science (as of Mediaeval science in the West) can be seen by comparing this
passage from the Ying t'ung MM ('Sameness in response') chapter of the
Lu-shih ch'un-ch'iu:
1 1 9
Ch. 13/2(Hsu503/6-505/3)«HffiS ° J R I R I M £ > ° fft Sifn g»
"It is inherent in things of one kind that they call each other up. If the
energies are the same they act together, when a sound correlates it responds.
When you strum the note shang or chiieh on one instrument the same string
is stirred on another. If you pour water on flat ground it flows towards the
damp, if you put fire to wood of equal quality it tends towards the dry. . . .
Therefore rain is sent down by way of the dragon and the shadow is
pursued by way of the shape."
Anyone content to think in terms of like summoning up like will see
no problem in the reflected image; nothing could be more natural than that
the mirror should respond to the shape by a corresponding picture. But the
Mohist does not think in this way. He refers in B 22 to "what is lustrous( ?)
in the man looking at himself" and "what is lustrous(?) in the shadow",
and in B 23, 24 mentions "everything mirrored" ( № ! l ) in addition to the
shadow (the silhouette), but he does not offer any explanation in terms of
the resonances of things in the same category. What he cannot explain by
interference with the light he does not explain at all.
Several editors have made the mistake of illustrating the optical
Canons with diagrams from modern textbooks of physics. This is very
misleading; the truly scientific approach to the study of B l 8 is to light
three candles in a row, of B22 to lay a teatray on the floor and look down
at your reflection, of B23, 24 to contemplate your face in the front and
back of a soup spoon.
3 3
B 17 S * * & ° i № & S < > °
(*> ° o (5g#«r&a) o
3 3 0
For this phrase, see the Book of Odes (Skih ching) no. 75: ,
te^X&iS^ "How befitting is the black robe. When it is worn out,
I will again make a new one [for you]".
C. " A shadow does not shift. Explained by: remaking.

E. Where the light arrives the shadow disappears. (Just as when being in
it is finished the past ceases.)"
J
"The shadow of a flying bird has never stirred" (№M±Sc%'№W) \&)
was a well-known paradox (Chuang-tzu ch. 33 (Kuo 1106/3)). It reappears
in Lieh-tzu (ch. 4, 88/7, 12), reproduced from the Canon complete with its
summing-up ( S M ^ ^ t & S ' & ' I f e ) . The Mohist accepts the paradox; the
shadow does not shift, its parts appear and disappear as the light comes and
goes. Ssu-ma Piao similarly explains the paradox in Chuang-tzu :
ch. 33, iio9/i5 f m±mit > m^±myk ° ^mmymy^m» »
»trâ > o MTB ' r «^tjfeft J °
"The bird screening off the light is like a fish screening off the water. The
fish moves and screens off the water but the water does not move; when the
bird moves the shadow is born, where the shadow is born the light dis
appears, but disappearing is not going and being born is not coming.
Mo-tzu says ' A shadow does not shift'."
The illustration connects with B 14 Mitfe
"North and South are present at dawn and again at nightfall. The shifting
of space has duration". (At a time when I was freer in transposing illustra
tive glosses I even supposed it to belong to B 14: cf. Graham and Sivin
117.) The tsai is nominalised as in B 41 (for the syntactic mobility of this
word, cf. p. 163 above). T o interpret the analogy let us suppose that the
bird moves to a new position at noon:
When the bird moves,

the light arrives at the bird's the afternoon arrives, the morn
morning position, the shadow ing ceases.
disappears.
It is not that the shadow moves It is not that the morning moves
forward behind the bird. forward behind the present.
The Mohists were consistent defenders of common sense against the

sophists, and it is at first sight surprising to find a well-known paradox of
this school placed at the head of the optical sections. The reason is no doubt
that the Mohist approves of the sophism as a vivid illustration of the idea
that the shadow has the special qualities of an interference phenomenon,
disappearing and reappearing as light enters or is impeded. It will be noticed
that he does not give the argument for the paradox (of which Ssu-ma Piao's
account is a very plausible reconstruction, although probably without
ancient authority). He is not concerned with logic-chopping in this part
of the Canons and no doubt assumes that the argument of the sophists is
2/4 375
common knowledge. He merely lays down what for him is the important
point involved in it, that "where the light reaches the shadow disappears".
This would not be a truism, since in folklore the shadow has an independent
and numinous existence. That the shadow results from obstruction of the
light is essential, for example, to his explanations of shadow inversion in
B 19 and B 23.
B 18 Jb-H ' IftffiS o
3 3 1
The Mohist does not use the particle che casually: yi kuang 'one light',
yi kuang che 'the singly lighted' (§ 1/3/12/2).
C. "Having two shadows. Explained by: redoubling.

E. When two lights flank one light, the singly lighted is in shadow."
The key to understanding this proposition lies in the connection

between the two shadows of the Canon and the 'redoubling' exemplified
in the Explanation.
A Canon in the scientific sections presents either a controversial thesis
or a problem, so the first question to ask is why the Mohist should be
surprised that an object can cast two shadows. The trivial observation that
two shadows imply two lights is certainly not the point being made in the
Explanation. The problem that concerns him evidently arises out of the
conception of light destroying shadow developed in B 17, according to
which lights on opposite sides of an object might be expected to obliterate
each other's shadows rather than both casting shadows. His answer is that
the spaces screened from one of the two lights show up as shadow against
the areas illuminated by both. T o prove this answer he takes the case of
a light with lights on both sides, in which it is obvious that there is no total
shadow and that the intensity of any visible shadow must be relative to
'redoubled' light.
B 19 «:5J 3 3 2
ffi^ 3 3 3
* « » ft** o o
(*) o A**U»;gW o T « ± * A - t M G » № S £ * A * T 3 3 5
° £
KT* > * ^ f t ^ ( i h ) * ± 6 > t r » L f c # . # j £ * s i & T o aasawss >
3 3
Tao (=#J, M . 1950 def. 3), 'turn over', as in B 22; it refers to the process
of turning over, yi 'wrong way round' (B 23) to the inverted state.
3 3 3
Wu 'criss-cross', of intersecting lines ( M . 2703 def. 4).
3 3 4
(K$), cf. § 1/2/1/4/6. Sun accepted hsii in the sense of 'emanate' (but
the word suggests emission of warmth rather than light). Most emend
to chao № 'shine* (T'an, Kao, W u , Sivin).
335 p o r systematic corruption A ( A ) cf. § 1/2/1/3/8. The position of the
word is clearly verbal, since the later Mohist syntax did not allow yeh
after a noun head (p. 154 above).
3 3 6
3 3 7
( · ) , cf. § 1/2/1/3/20. Sun emended to chang * ( = » ) 'block'.
C. "The turning over of the shadow is because the criss-cross has a point
from which it is prolonged with the shadow. Explained by: the point.
E. The light's entry into the curve is like the shooting of arrows from a
bow. The entry of that which comes from below is upward, the entry of
that which comes from high up is downward. The legs cover the light from
below, and therefore form a shadow above; the head covers the light from
above, and therefore forms a shadow below. This is because at a certain
distance there is a point which coincides with the light; therefore the
revolution of the shadow is on the inside."
Corruption and syntactic obscurity account for the very divergent

interpretations which have been offered for this section. Fortunately much
of its terminology appears elsewhere. From other optical sections we may
infer that chang 'prolong' refers to the prolongation of the borders of light
after intersection (B 23), yuan chin 'distance' to distance along the axis of
the concave mirror (B 23, 24), net 'inside' to the space between the centre of
curvature and the mirror surface (B 23). Tuan 'starting/ending/meeting-
point' was defined in A 61. The words we identify as kou 'curve' and yiin
'rotate' (as well as the yi 'changed round' of B 23) occur in two sections
on motion, A 47, 48. There remains the syntactic problem of the two tsai,
neither of which seems to be the ordinary verb tsai 'be in' common in the
dialectical chapters. The later usage as a preposition ('in', common from
the Han onwards) can be ruled out. I take it as tsai 'depends on, is because
o f ; that the succeeding phrase would be verbal and not nominal is con
firmed by the following example:
Mo-tzu ch. 8 (Sun 27/4) ftS^*A«afea**êi«SJK»ft«ift
"This is because kings, dukes, and great men who exercise government in
a state cannot exercise government by promotion of worth and employment
of ability."
2/4 377
Previous editors have believed that this section describes the inversion
of the image in the camera obscura. This is the one point on which I failed
to reach agreement with my collaborator Nathan Sivin in our Systematic
approach to the Mohist optics (where Sivin offers an alternative rendering
of the first sentence of the Explanation: * A *R§3? M "The light enters
and shines (lit. the light's entry is a shining) like the shooting of an arrow").
Wu Yii-chiang even finds references to the dark room (№) and the screen
( * = K ) , although there is no evidence that the Mohists devised special
experimental equipment. (They worked with mirrors, balances, pulleys,
ladders.) The fact that light entering a small hole in the window of a dark
room sometimes casts inverted images on the far wall was noticed in China
from at any rate the 9th century A.D. (cf. Needham 4/1, 97-99). But if the
Mohist is describing this phenomenon it is strange that even on Wu's
interpretation he does not mention the pinhole. Moreover all other Canons
in B 17-21 are about general questions, whether shadows move, why there
can be two of them, why the shadow can be on the sunny side, what
determines its size; while B 22-24 discuss specific operations identified in
the Canons themselves, looking down into a mirror, looking into a concave
or convex mirror. The present Canon introduces the inversion of images as
a general phenomenon requiring explanation, B 23 explains what happens
when they are inverted in the concave mirror. It is conceivable that the
author of the present Explanation preferred to illustrate the phenomenon
by the example of the camera obscura, but if so one would expect a much
more explicit indication of the concrete situation than any commentator
has been able to find in it. The burden of proof rests on anyone who claims
that the Mohist has noticed the inversion anywhere except in the concave
mirror.
One reason why editors thought first of the camera obscura is surely
that the Mohist describes the light as passing the point of intersection
before reaching the reflecting surface (as though passing through a pinhole),
not after being reflected (as in the case of a concave mirror). But why should
we assume that the Mohist knew that the light entering the mirror converges
on the focal point only after reflection, familiar though this fact may be to
his readers two thousand years later? He did not know, as is plain from
B 23, and Shen K u a in the 11th century A.D. did not know either (Graham
and Sivin, 147).
If we are right in identifying a crucial word as kou 'curve', we may
dismiss the camera obscura from consideration. We may also suspect that
the sentence "The light's entry into the curve is like the shooting of arrows
from a bow" means something more than that the light flies the straight
course of an arrow. The graph she 'shoot' in bronze inscriptions is a picture

of a bow and arrow Tuan Wei-yi 212), and the Mohist may be saying
that light approaches a concave surface as the arrow passes across a bow,
along its axis :
It would follow that all light passes through the centre of curvature, and
the Mohist would have a simple explanation of why it converges on it and
diverges from it, as he describes in B 23. This would in turn explain why
the shadows cast by head and feet screening off the light which silhouettes
them are reversed before they reach the mirror (even a small concave
mirror will give a full-length image).
B 20 ftffiE * H 338 o i a f t W 3 3 9
o
(*) o B £ 3 f e R « A ' ffJ*ffi0HA±ia °
3 3 8
Cf. Kuan-tzu ch. 52 ( S P T K 16, 9B/1) & # £ * S £ S § i f 5 $ t f i , ffiBiffi
§t№ "Were you perhaps riding round on a piebald horse and galloping
in the direction of the sun?".
3 3 9
Sun is no doubt right in identifying this graph (used for IB, M . 11137)
as representing chuan 'turn round', written with its orthodox graph
in A 94 and as # in A 95. The phonetic as usual is corrupted in the
Taoist Pathology but preserved in the Mao edition (cf. p. 75 above).
C. "The shadow cast in the direction of the sun. Explained by: turning
round.
E. When the sun's light turns back and illuminates a man, the shadow
will be between the sun and the man."
Although the phenomenon could be observed in more than one kind

of situation, the most obvious place for the Mohist to notice that his image
is on the same side as the sun would be in a mirror, where it would show
2/4 379
up more distinctly with the sun behind the mirror than in the full glare
with the sun behind his back. His explanation is that the light turns back
(fan), presumably reflected from surfaces behind him.
340
B 21 &c&-X'b ° i S : f f i ( i i i i ) ^ № i & ^ S o
Corrected from the Explanation. The latter graph is used regularly for
yi M 'slant' ( M . 14478 def. 2, which however notes the interchange
only in the sense 'split').
Ta 'bigness' or hsiao 'smallness' is what is measured by circumference,
not by length, breadth or thickness (cf. A 55, B 56). Cf. Mo-tzu ch. 53
(Sun 324/7,8) MX [JJ-] # — R , S S S « £ » J 5 £ "The timber is in
size 1 foot square, in length adjusted to the thickness of the city-wall",
335/1 . . . «&jett№RRH^, №-fcrd\ &/\R "The box
is in size 3 | wet round. . . . The shoulders of the 'hooking spurs' are
1 foot 4 inches in breadth, 7 inches in thickness, 6 feet in length".
Emended with T'an, Kao. Sun reads equally plausibly.
The last sentence may be mutilated.
C. "The size of the shadow. Explained by: tilt and distance.

E. When the post slants the shadow is shorter and bigger; when the post
is upright the shadow is longer and smaller. If the flame is smaller than the
post the shadow is bigger than the post. It is not only because it is smaller,
but also because of the distance."
The enlargement of the shadow would be the widening of the top end
of the shortening shadow as the post is tilted towards a light coming from
higher up. The shadow would have to be on a wall or other perpendicular
surface, as noted by Sivin (Graham and Sivin, 128), not on the ground as
supposed by Hung Chen-huan (Hung 20).
B 22 №№M\L » ftS · &ffi7g'J>' Ifc&'&U o

344
(№,) mm™ssâ№iE^^
° I E « « I * * » ° xftfr >ftst*°
' «[4X](W)*InJJb o ^ # Z ( ^ ) * ^
347
^»fr^» 3 4 8
344
Tao, 'turn over', as in B 19.
3 4 5
Cheng chien 'plane mirror', in contrast with the chien wa "when the
mirror is concave" and chien t'uan "when the mirror is convex" of the
next two Canons (proposed by T'an).
346
T'ai ( = « g M . 29454 def. 10), 'posture', written fffi in Mo-tzu ch. 52
(Sun 313/9).
3 4 7
This mutilated sentence appears to be parallel with the preceding, in
which the fact that chii is firmly attested in the dialectical chapters
only in its adverbial use ('both, all') establishes the punctuation. I take
the two sentences as syntactically analogous to Kuan-tzu ch. 55
( S P T K 18, 3B/4) £fMM!l?& ^ t H № " I f name and object fit it
>
will be orderly, if they do not fit it will be unruly". If so, the last two
words should be a pair of opposites parallel with 'approach and
:
recede'; I therefore follow Kao in reading fF0dh(= ff) 'face towards
and turn one's back on', a pair of which several graphic forms are
recorded in Chu Ch'i-feng 1911. It appears in Mo-tzu ch. 46 (Sun
271/9) as M fa . The unknown graph fa (not = M, as printed by
Sun) looks like a mutilated form; the only graph with this bottom half
in the dialectical chapters is ts'an which judging by its corruption
to H in B 38 was originally written in the form ^ (as suggested,
Sun 139/-2). M y emendation to ^ in Graham and Sivin 131 now
seems to me too facile.
348
Ch'ou has been variously emended to ^ (Sun), ^ (T'an), S (Kao), %
(Wu), JR- (Chang Ch'i-huang). But all these suggestions have been
superseded by L i Yu-shu's more recent proposal of 5c, an
obsolete graph for tse W which is actually corrupted to ch'ou in Shih-chi
ch. 23, 1162/1 (Chu Ch'i-feng 1187). Tse is perhaps the most common
early word for the lustre of a reflecting surface. Its usage may be
illustrated from certain Confucian comparisons of various virtues with
qualities of jade, as in Hsiin-tzu ch. 30 (Liang 398/7) SSIiffiiS, tife
"Being lustrous with a warm glossiness is 'benevolence' " : Chia-tzti.
hsin shu S P T K B, 37B/1 "In that a lustrous thing
serves as a mirror, we call it the Way".
3 4 9
Kuo cheng "go beyond the upright" (the mirror plane?), also in B 24.
Cf. the late attested phrase 5tfeSilE ( M . 24015/78), "bend the crooked
beyond the straight, overstraighten".
C. " I f one stands upright looking down in a mirror the shadow turns over,
and the more there is of it the less there seems to be. Explained by: reduced
area.
2/4 381
E. In a plane mirror the shadow is reduced. The appearance, shading

distance and inclination differ from those in the light. When mirror and
shadow are plumb with each other, the man and his shadow approach and
withdraw together; when aligned to keep them plumb(?), the man and h
shadow turn towards and away from each other together(}). The l
parts( ?) of the man looking at himself are mirrored in the mirror without
exception; the lustrous pat ts( ?) in the shadow are numberless and necessarily
recede beyond the mirror plane, therefore they share the same place. Since
this is so of all the parts of his body, it mirrors their portions.''
Several editors (T'an, Hung Chen-huan) find in the Explanation a

description of reflection in multiple mirrors. But this interpretation does
not fit the Canon, and even in the Explanation the Mohist does not make
the specific reference to more than one mirror without which it must remain
implausible. It seems more likely that he is examining the changing
perspective in a plane mirror as he tilts it. When he is face to face with the
mirror he and the image approach and recede from the mirror plane
together; when he tilts it sideways keeping mirror, image and man in
alignment, they simultaneously turn towards or away from each other.
When he looks down into the mirror laid flat on the floor his image is head
downwards; and the nearer he attains to holding himself upright the more
of him appears in the mirror, but compressed into a contracting area. A l l
the lustrous parts of him still show up on the bronze, their number is
countless and they are necessarily held at the same angle to the mirror
plane as his own body, cannot spread out to fill the surface; therefore they
are squeezed into the same place.
If we are right in identifying the word tse 'lustrous', it is evident that
what the Mohist calls the 'shadow' is the silhouette, with the reflected
detail of the figure showing up inside it.
3 5 0 # 3 6 1
B 23 « ( & ) * ? £ »(*) » — /jMfn^ » - A r f n u r » o
(H) o r j o mm&^mm^»° m^wmm^ ·
Chien wa is parallel with the chien t'uan " I f the mirror is convex" at the
head of the next Canon ; since the present section concerns the concave
mirror we should expect the second word to mean 'concave'. Wa is the
word used (in the graphic form M) in Shen Kua's description of the
inversion of the image more than a millennium later (he also uses
cheng for the upright image). This emendation (Chang Ch'un-yi 3R
$6-^ and others) is more plausible than other proposals: 45 (T'an), 91
(Kao), B (Chang Ch'i-huang).
3 5 1
Corrected from the parallel Canon B 24.
3 5 2
The parallelism with the last sentence of the preceding paragraph
suggests the loss of three words, the first of which should be chung.
C. "When the mirror is concave the shadow is at one time smaller and
inverted, at another time larger and upright. Explained by: outside or
inside the centre.
E. Inside the centre. If the man looking at himself is near the centre,
everything mirrored is larger and the shadow too is larger; if he is far from
the centre, everything mirrored is smaller and the shadow too is smaller;
and it is necessarily upright. This is because the light opens out from the
centre, skirts the upright object and prolongs its straight course.
Outside the centre. If the man looking at himself is near the centre, every
thing mirrored is larger and the shadow too is larger; if he is far from the
centre everything mirrored is smaller and the shadow too is smaller; but
it is necessarily inverted. This is because the light converges at the centre,
. . . and prolongs its straight course."
The description of the inversion of the image in a concave mirror is

clear and indisputable. The 'centre' is the centre of curvature, which the
Mohist does not distinguish from the focal point. The 'inside* is the
mirror side, the 'outside* the observer's side. The observer is moving the
object in a straight line towards and away from the mirror. While the object
is between the centre and the mirror the image is cheng 'right way up';
it becomes smaller the further it is advanced from the centre towards the
mirror. While the object is between the observer and the centre the image
is yi 'wrong way round' (the top and bottom exchanged); it enlarges the
nearer it approaches the centre.
The two explanatory sentences with final yeh, the latter with missing
words, are less transparent. They assume that the light converges towards
the centre of curvature before reaching the mirror, not after being reflected
back from it. They make two points:
(1) In one case the light is already 'opening out from the centre' as
it passes the object, in the other it is still 'converging towards the centre'
(where the image will turn over, cf. B 19). This explains why the image is
upright in the former case and inverted in the latter.
2/4 383
(2) In both cases the light continues along straight lines. This
explains the variation in size; the nearer the object is to the centre the wider
the angle and therefore, if the lines are straight, the larger the image.
The Mohist's assumption that the image is inverted before it reaches
the mirror is of course faithful to common sense. It would require a
considerably more advanced optics to recognise that when reflected back
the image is not yet inverted, in spite of the fact that we see it as both
inverted and on or near the mirror surface.
Outside and near the centre Inside and near the centre
Outside and far from the centre Inside and far from the centre
If we have correctly understood the crucial sentence in B 19 ("The light's

entry into the curve is like the shooting of arrows from a bow") the Mohist
has a single principle which explains all these phenomena; light enters a
curve along its axis like an arrow on a drawn bow, and therefore the light
coming from different directions converges at the centre of curvature. This
principle does not imply that he thinks of light as composed of rays. One
could explain the straight edges of a sheer waterfall by its vertical descent
without thinking of rays of water.
The explanation accounts for the inversion and changing dimensions

of the silhouette, but not for the reflected detail of the image. This he
contrasts with the shadow as so chien 'what is mirrored', which I cannot
follow Hung Chen-huan in treating as a technical term for 'magnification'.
Both here and in B 24 there is a clear syntactic contrast between chien che
'the mirrorer' (the man looking at himself) and so chien 'what is mirrored',
further expanded in B 22 £ "The lustrous parts
of the man looking at himself are mirrored in the mirror without
exception".
9 # 3 5 3
B 24 S I B * — < / h > / ( X ) — • * f f n # * ° < t № . . . > °
9 9 9
(j§) o m^mm^J: ° · ffn
3 5 3
The Canon was broken, at the point we mark with a diagonal stroke,
when B 14-21 and B 22-24a were transposed. It was long assumed
that the second half is a separate Canon ; the discovery that the parts
can be spliced by the emendation of — X to —*</h >—*A is one of
Luan T'iao-fu's services to the study of the Canons (Luan (1957) 87).
Comparison with the parallel Canon of B 23, and of each Canon with
its Explanation, confirms the emendation. There seems to have been
further damage here as in the other broken Canon (B 14); although the
second part is immediately followed by the expected summing-up
formula (tftffi#), the latter's wording suggests that it belongs to B 25b.
354
Ch'i, which was presumably at some stage written with the archaic
graph 7T, is corrupted to yi some 80 times in Mo-tzu (Sun 194/2).
3 5 5
Chaoj*T\OG, variously emended to 4ft (Sun), № (Liang), fif (Chang
Ch'un-yi, Wu), fiS (T'an), 85 (Liu). Once elsewhere in Mo-tzu the
y
graph is used for ch'iaol*G \OG «r (cf. M . 11968 def. 2 № = *f): ch. 1
(Sun 3/3) JgTfcfifft "The sweetest water is nearest to
exhaustion, the loftiest tree is nearest to being chopped down". But
ch'iao is not exactly translated as 'lofty'; the Shuo wen ch. 10B
(Tuan 499A/1) defines it as 'high and bent' ( i f , iftifo ftib). It is used
primarily of tall trees with branches bending upward, giving no shade
( M . 3990 def. 2-4), in contrast with chiu P\ ( M . 14436), used of tall
trees with branches bending downward. This seems an appropriate
word to describe the distortion of the image in the convex mirror, the
edges dwindling and receding like the branches of a poplar seen
from below.
2/4 385
C. "When the mirror is convex, the shadow is at one time smaller and at
another time larger, but is necessarily upright. Expained by: . . .
E. If the man looking at himself is near, everything mirrored is larger and
the shadow too is larger, if he is far everything mirrored is smaller and the
shadow too is smaller, but it is necessarily upright. The shadow goes
beyond the plane and therefore recedes at the edges(l)."
The Mohist does not explain why the image in a convex mirror is
'necessarily* upright, but that would follow from the principle applied in
B 23, and possibly the lost summing-up referred back to it. If he conceives
light as following the axis of the curve in a convex as in a concave mirror
the change of size would also be explained as before.
If we are correct in identifying the final word as ch'iao he explains the

recession of the image at the edges by the fact that it goes beyond the plane
(kuo cheng, as in B 22), that is beyond the tangent plane.
2/4/2/3/2. Mechanics (B 25a-29).

There were five Canons on mechanics at Stage 2, one of which has
been lost (B 27). Originally there appear to have been six, the one which
we number B 25b having been mutilated at the same time as the transposi
tion of B 14b-21 and 22-24a at Stage 1.
The six sections begin with horizontal (B 25a, 25b) and proceed to
vertical loading (B 26-29). If a horizontal carrier is attached to supports
at both ends it may or may not bend under a load, depending on two factors,
the weight and its own chi№. This chi, which has sometimes been identified
as the centre of gravity, seems rather to be the full extension of the weight-
bearer (B 25a n.). If the horizontal carrier is pivoted on a fulcrum, which
side will decline again depends on two factors, the weight and the ch'iian W .
The latter word has commonly been mistaken for ch'iian 'counterpoise'
(of the steelyard), but Ch*ien Pao-ts*ung £ $ J I S has shown that it is the
ch'iian used of the power, leverage, positional advantage of a ruler or
minister. In B 25b it refers to the varying pull of the weight according to
its distance from the fulcrum, and could be translated without misgivings
as 'leverage* were it not that in B 26 the Mohist continues to take account
of ch 'uan as well as weight in explaining the vertical weight-bearing of the
pulley. He sees that the pull of the heavier side of the pulley increases with
distance from the pivot, but not that the only factor which he needs to
postulate is the weight of the rope passing from one side to the other. We
therefore translate ch'iian by 'positional advantage*.
In B 26 the Mohist introduces a new pair of concepts which reappear
in B 27 and 29, ch'ieh IS 'suspend, pull upwards* and the more difficult
shou $C, which is something that happens to the weight from down below.
The ordinary uses of shou 'gather in* (for example, a harvest) do not quite
explain what the Mohist means by the word, for which we choose as
English equivalent 'receive from below*. It is certainly implied that the
weight is descending vertically (as it rises vertically when pulled upward);
but is it pulled down or simply arrested as it falls ? B 26 favours the latter
alternative, which is further supported when the Mohist lays down his
broadest general principle in B 27: "Whenever a weight is not pulled up
from above or received from below or forced from the side it descends
vertically; slanting is because something interferes with it**. Clearly it is
arrest of the fall, not downward pulling, which interferes with vertical
descent.
At this stage another pair of concepts emerges, support from the side
(B 28 yi ffi 'leaning on*) and from below (B 29 chu ft 'pillaring*). Since
all weight tends downward the vertical cannot be supported from the side
(B 28), but its descent may be halted even by such a fragile vertical support
as a ruler stood up under the weight which pulls up a wheeled ladder
(B 27). In B 29 the Mohist makes an important reduction in his system of
concepts by showing that chu 'pillaring* is simply a name for a static shou
'receiving from below*.
The two basic terms are chung IK 'weight* and yin 31 'pull* (B 26,
27, 29), which presumably includes ch'ieh 'pulling up* and ch'e IB 'pulling
from the side* (B 26), and is the only word used for pulling down (B 29).
The two concepts are treated as interdependent; not only does a hanging
weight pull its attachment (B 29) but a hand twisting a rope has weight
behind it (B 25a). Yin 'pull* seems to be the only word outside the vertical/
horizontal orientation of the whole vocabulary; no general word for 'push*
appears in the six sections.
2/4 387
The li 'strength* (B 26) which is elsewhere described in terms of

resistance to the downward tendency of weight (A 21) is human bodily
strength, as its definition in A 21 confirms. There are no forces in Mohist
mechanics, only weights and pulls.
356
B 25a (j<)*:ft ffil^Si» K £ i °
(ft) * « * ( * P ) i n 3 5 7 m S f f 5 ^ '
#
o*K358^359m > «Jn
B 25b < « > o . . . o

(«) ojnmjyt-^g 3 6 0
· o · a « * M « s o mm
3 5 6
3 5 7
Corrected from the next sentence.
™ Chiao ( M . 14713 def. 2/1 =fe), 'twist, tighten* (Chang Hui-yen,
Needham v. 4/1, 28). Cf. a Mo-tzu fragment ap. T'ai-p'ing yu-lan
S P T K 336,7B/2 (reproduced Sun 408/4) « » * R A 3 t "Twist (a
rope of) good hemp 80 feet long**.
359 k this chiao to refer to the crossing of the strands in the twisting
t a e
of a rope, the usage from which chiao Ml presumably derives.

360
Ch'ui 31 'hang down* with a deviant radical.
361
Piao 'tip* ( M . 12651 def. 6 =JS, the graph used twice in this same
section). Pen piao 'root and tip* is written yfcSM in Chuang-tzu ch. 23,
quoted p. 365 above.
Textual note. The interpretation of the two parts of B 25 has been
complicated by textual problems involved in the transposition of the
Canons B 14b-21 and 22-24a at Stage 1. Liang Ch*i-ch*ao, assuming
the transposition of whole Canons, identified what seemed to be a
complete but corrupt Canon without an Explanation, in front of B
25: tXM^fiRSH. He took it for the missing Canon of B 25b and
therefore transposed it with B 25a. This hypothesis (followed by
Chang Ch*i-huang, T*an, Needham 4/22, 28, Yabuuchi) became un
tenable when Luan T*iao-fu demonstrated that the last of the breaks
was in the middle of B 24, so that the 7 characters turn out to be the
conclusion of the last of the optical Canons. Most recent editors have
therefore treated B 25a and 25b as a single Canon with a single Ex
planation (Luan 88, Kao, W u , L i u , Ch*ien Pao-ts*ung 66). But this
raises new difficulties; the Canon and the first part of the Explanation
are about the bending of a cross-bar, and have nothing to do with the
section on the beam and fulcrum which follows. We seem therefore to

be driven to the conclusion (already reached by one of the most recent
editors, Ch'en Ch'i-yu BRififtft 36, 39) that Liang Ch'i- ch'ao was right
in distinguishing two Canons, but the second has been lost.
There is the further point that although the first four characters of
the fragment mislocated by Liang Ch'i-ch'ao certainly are the con
clusion of B 24, he had the justification that the last three (the
summing-up, "Explained by: gaining") do look as though they belong
to B 25b. Moreover it is very likely that when the strings broke and the
three strips of B 14b-24a came loose (cf. § 1/2/2/2/1, 2) there would be
some damage to adjacent strips. Indeed we have evidence of this in
the mutilated summing-up of B 14b, and the corruption of the first
character of B 24b, as well as the loss of at least part of B 25b, all of
them close to places where the strings snapped. It is therefore reason
able to postulate that the first strip after the break was fragmented at the
beginning into three scraps which were put together in the wrong order:
B 24b * — * ^ f f 5 ^ n r . . .
25a • f t f f D * i i № f c »
2 5 b . . . m&n
A scribe copying these scraps of the Canons without comparing them
with the Explanations would have no reason to suppose that there were
more than two; it would be very natural for him to take the remains
of B 25b for the missing summing-up of B 24b.
Since our numbering of the Canons is based on their divisions at
Stage 2 we do not give separate numbers to B 25a and 25b.
C. "Carrying a weight without bending. Explained by: which prevails.

E. A cross-bar when you lay a weight on it does not bend, because its
being at full stretch prevails over the weight. If you twine a rope with a
right-hand twist it bends without a weight being laid on it, because being
at full stretch does not prevail over the weight exerted.
C. The beam. . . . Explained by: gaining.
E. The side of it on which you lay a weight will necessarily decline,
because the two sides are equal in weight and positional advantage. If you
level them up, the tip will be longer than the butt; and when you lay equal
weights on both sides the tip will necessarily fall, because the tip has gained
in positional advantage."
The questions answered by B 25a and 25b are "Why does a cross
bar not bend under a weight?" and "Why does the longer arm of a beam
2/4 389
go down when equal weights are placed on both sides ?". The answer to
the first question is that the chi ffi prevails over the weight. Several editors
have identified the chi as the centre of gravity (T'an, L i u , Ch'en Ch'i-yu,
cf. also the translation by 'the centre' in Needham v. 4/1, 28). But it is not
easy to reconcile this hypothesis with the ordinary meaning of chi 'extreme,
to the limit". In any case the common-sense answer to the question is
surely that a cross-bar is held at its full extension by the posts on either
side, so that it can break but cannot bend. It is its extension 'to the limit'
(chi) which prevails over the weight. This will also account for the example
of the rope. If a man attaches a rope at one end and twists it in the direction
which will tighten it (which according to the Mohist is to the right) the
rope bends although he is trying to keep it at full stretch. It is a situation
which vividly illustrates the conflict between holding at full stretch and
exerting effort to rotate the rope, and the man feels in his own wrist the
latter force prevailing. In this case the 'weight' (chung) is that of the
rightward turn. Chung is the weight behind a pull, not necessarily the
weight of an object; cf. B 52, and Lieh-tzu ch. 5 (Yang 108/1), (of an angler's
even pull on the line) iS^fcfc^J, ^ ^ l ^ l S " I n casting the line and sub
merging the hook my hand is never too light or heavy".
The beam (heng $s) of B 25b may be pictured as a beam suspended
from a rope or pivoted on a stone; if it is the beam of the steelyard we must
assume a steelyard of the Tibetan type, with fixed counterpoise and moving
fulcrum, not the type with fixed fulcrum and moving counterpoise which
prevailed in China (cf. Needham v. 4/1, 24-27). The Mohist tells us that
which side goes down depends on two factors, weight and positional
advantage (ch'iian №). If the fulcrum is at the centre the heavier side goes
down. But equilibrium may be recovered by shifting the fulcrum; and
then, if equal weights are added to both sides, the longer goes down,
because it has more ch'iian. This positional advantage which increases with
distance from the centre is of course what we call leverage. The Mohist is
making the point Archimedes was making a little later, although without
the most important part of the Archimedean demonstration, its mathematics.
That the ch'iian of the Mohist mechanics is not, as used to be supposed,
a name for the counterpoise of the steelyard, is the discovery of Ch'ien
Pao-ts'ung (op. cit. 65, 67). Ch'iian and heng are such familiar words for
the counterpoise and beam that it takes an effort of the imagination for a
reader of B 25b to understand them in any other way, whatever nonsense
it may make of the text. Yet in B 26 the pulley is described as having ch'iian
as well as weight on both of its sides. Ch'ien finds no positive evidence of
the counterpoise being called by this name before the end of the Western
Han. In earlier literature he finds only a nominal combination ch'iian heng

'scales', a verb 'weigh* and a noun 'power, authority, positional advantage
in the state*. This historical argument is open to the objection that no other
name for the counterpoise is better attested; and Ch*ien himself derives
the last sense of ch'iian from the power of the counterpoise to lift or lower
the weight according to its position on the beam. However, for the under
standing of the Mohist mechanics all that matters is that of the two nominal
usages, 'counterpoise' and 'positional advantage*, it is the second that fits.
A further merit of the proposal is that once ch'iian is detached from heng
we are no longer committed to picturing a steelyard, the image which has
predisposed scholars to assume a fixed fulcrum.
B 26 mm(&)* W 3 6 2
ota#»363 0
3 6 8
* o s s # r » m m % ± ™ o± # . [Tj T # & t ° m& * m
362
Ch'ieh ( M . 5917 def. 9 =3g, the graph used in the Explanation) 'pull
up from above*. Shou 'gather in, arrest from below* (corrected from
the Explanation) would seem from its corruption to have been written
with the graph 4£ ( M . 13192). Pan is generally recognised to be fan R
'turn over* (cf. the pan-pan of Songs 254/1, M . 14518/61) written with
an unorthodox radical, like the fan i& of B 30, 72. Most indeed follow
Sun in reading the latter graph, but its function in distinguishing a
special logical usage makes this unlikely.
363
Po 'curtain* ( M . 32083 def. 19), later distinguished by the graph IS
( M . 26142 def. 4). There is a strong presumption in favour of this
sense of po, because the Explanation assumes some kind of pulley
without telling us what it is. The mechanics sections always specify
the kind of mechanism (even in B 25b we are in doubt as to whether
the heng 'beam* is that of a steelyard only because the Canon is lost),
if not in the Explanation then in the summing-up (B 28, 29).
3 6 4
A systematic corruption (§ 1/2/1/2/13). The correction seems necessary
here to elucidate the obscure syntax of the opening sentences. The
clause W^Jife, 31 " O n the side where strength is exerted, it pulls**
might be acceptable in some pre-Han texts, but the dialectical chapters
provide no clear instance of yeh after a subjectless initial clause
(p. 154 above).
2/4 391
365
Yi ( M . 13629 def. 5/1, 2 =M) 'slant', variously written Jfo, fl6, * in the
dialectical chapters. For a similar example contrasting with chih
'straight' cf. Huai-nan-tzu ch. 21 ( L i u 21, 3B/4) 3tfel£№ffi3JEW ,
^ f e f f f l S ^ "Its numbers straighten the slanting and correct the
crooked, exclude the selfish and establish the impartial". (Chih yi
'straight or slanting' again ut sup. 4A/4.) But I can offer no supporting
examples of yii yi ('(stop) on the slant' ? '(stopped) by the slant' ?).
366
Ch'e № ( M . 12277) 'pull from the side' with the radical dropped (for a
supporting example, cf. T'an (1958) 160).
3 6 7
Judging by the context hsia and shang should be 'is below' and 'is above',
not 'descends' and 'ascends', in spite of the pi hsia 'will necessarily
descend' of B 25b. It may be doubted whether the Mohist would have
allowed such a syntactic ambiguity; the hsia tf 'descend' of B 27
suggests that there may originally have been a consistent graphic
distinction.
368
Hsia accidentally repeated.
3 6 9
Chut ( M . 38985 def. 16 = » , 5 ) 'drop', as in Sun 14/1 cf. 329/14,
337/14, 377/14.
C. "The changeover from pulling up to receiving as it falls. Explained by:

a curtain.
E. The pull from the side where strength is exerted on the side where no
strength is exerted does not stop. The thing pulled up being stopped on the
slant is because the rope is pulling it sideways. (For example, if you prick
it with an awl.)
Pulling up. The longer and heavier is below, the shorter and lighter is above.
The more the one above gains, the more the one below loses. When, the
rope being straight, the positional advantages and weights of the two have
equalised, they come to a stop.
Receiving. The more the one above loses, the more the one below gains.
When the positional advantage and weight of the one above is spent, it
drops to the ground."
The problem is why, when you lift a weight on a pulley, there is

sometimes a certain point at which you no longer need to pull, because the
weight begins to rise by itself. The solution is based on the same principles
as in B 25b, that the side which rises by itself is the one with less weight
and less ch'iian, the positional advantage which diminishes with decreasing
distance from the pivot. As long as I pull, the weight and ch'iian on the
other side progressively decrease. At a certain point they are equal to the
weight and ch'uan on my own side; then, if I cease to pull, they 'stop'
(chih lb), hang in equilibrium. One more pull, and the weight on the other
side begins to rise of its own accord.
Here the Mohist makes a very interesting mistake; he has not seen
that his explanation requires only the reduction of weight on the other side,
as the rope goes over the pivot, and that the distance of the weight from the
pivot need not be considered as an additional factor. In B 25b he was
considering a beam on a fulcrum, and to explain why one side went down
he had to consider, not only its weight, but its ch'uan, distance from the
fulcrum, leverage. But to deal with the pulley he no longer needs the
concept of ch'uan.
What sort of pulley does he have in mind? He does not say in the
Explanation, which probably implies that the apparatus was identified in
the summing-up of the Canon, as po 'curtain* (cf. n. 363 above). This
would explain why the Mohist seems to be assuming a thing with breadth
as well as length, which will jam if the pull of the cords on either side is
unequal and drags it sideways (as will happen if on one side the curtain
is stuck to the roller by the prick of an awl). There is however some difficulty
in visualising the kind of curtain which is implied. A curtain which goes
down on the far side as it rises in front would of course do as well for his
purposes as any other kind of pulley; but there is no practical point in an
apparatus for lifting a curtain unless it disappears from sight around a
roller. It is hard to resist the suspicion that he is referring to a curtain which
goes up by itself when you tug the cords like Venetian blinds. I know no
evidence of such an apparatus in early China. But M r Ts'ien Tsuen-hsuin
has called my attention to the account of the Imperial library of the Sui
dynasty (A.D. 589-618) in Wen-hsien t'ung-k'ao 3$0ft3^ (ch. 174, W Y W K
1506 C/-3 ff); mechanical fairies, set off by treading on a device in the
floor, rolled up the curtains and opened the doors.
B 2 7 <m . . . * » . . .>370 o
( « ) o mmt > M*&I§*$! 3 7 1
· r# » J
3 7 2
* o »sfttfm 3 7 3
[ K » m * c ] 3 7 4 » s » * s r f n i K * ^ K m » A # ™ * »&mmn °
375
Jim » ± » * » T3Mfc · » HIT* o 377 , o »78$

* W » ? F » it-til o
379
330 jgR j&2pjft 381 , fl^-jr , toft 382 -ft o £
3 7 0
The Canon was lost at Stage 2 (p. 92 above). We restore the first
character from the head character of the Explanation ; for the last two
restored characters, cf. n. 372.
2/4 393
371
Ch'uan ( M . 38427 = & ) , the spokeless wheels of a hearse, half the height
of an ordinary wheel. We take them to be the small wheels (trundles)
at the top end of the ladder.
3 7 2
A n ' X yeK phrase loosely attached at the end of the sentence is likely
to be a quotation from the Canon (§ 1/3/10).
373
Ch'i ch'ien tsai 'the support in front of it*(?). Hsien 'bowstring* is used
in geometry as the standard term for the chord and the hypotenuse
( M . 9754 def. 5). Most follow Sun in emending to yin SI 'pull'; but
the text is supported by a similar verbal use in the K'ao hung chi
of the Chou It MW. (SPPY 42, 5A/-2) & & J t f t "Thereby take
the chord of the inner side of it" (of the curved handle of a mattock).
3 7 4
Four characters accidentally repeated.
375
Ku ( M . 38254), unidentified part of a carriage. The 'trundle axle* of
the translation is a mere guess.
3 7 6
Unknown graph; with some misgivings we follow Sun in taking it for
ft ft 'ladder*.
377
Yi 'slant*, variously written ftfe, flfe, № in the scientific sections.
378
Liu ( M . 17205 = SK), 'stream*, a li-shu form ( K u Ai-chi 2/58B).
3 7 9
Unknown graph, which we take to be hsia ~F 'descend* distinguished
from hsia 'below* by a radical (cf. B 26 n. 368).
380
Chin yeh 'in the present case* probably refers to the situation described
in the lost Canon, cf. B 70, 78. It certainly cannot be passed over as a
vague introductory 'Now . . . *; in the remaining examples it contrasts
with (B 32), (B 33), 'previously*.
381
Fei 'put, stand up* as in B 10, 29. Cf. Kuan-tzu (ch. 69) BSS 3/63/1
£PJ§№#*flfc "The people will be as stable as a square
placed on the ground'*.
3 8 2
Graph unknown except in late binomes ( M . 37738). The formation of
the graph suggests 'step to one side*.
C. "Pulling up from above . . . wheeled ladder.

E. One pair of wheels high, one pair serving as trundles^}), makes a
'wheeled ladder*. If you weight its front, and following the hypotenuse
made by the support in front of it and the trundle axle{}) you hang a
weight in front of it, the result is that the ladder is pulled up and also
when pulled up moves on the ground.
Whenever a weight is not pulled up from above or received from below or
forced from the side it descends vertically: slanting is because something
interferes with it. That the weight which makes the ladder glide fails to
descend is because it hangs vertically. In the present case the fact that
3
394 The 'Canons and 'Explanations'
when you stand up a foot-rule on flat ground the weight will not descend
is because it has no inclination to the side( ?). As for the pull of the rope on
the trundle axleij), it is like a pull on the cross-bar from inside a boat."
The loss of the Canon and the use of otherwise unknown graphs makes
this section very obscure. It is agreed that it describes the erection of some
kind of wheeled ladder pulled up by a weight; but previous editors have
tried to interpret the description without seeking to identify the mechanical
problem which engages the Mohist. A n important clue is the Chin yeh
Tn the present case' introducing the sentence about standing up a ruler on
flat ground, which probably marks it as a direct reference to the lost Canon
(cf. n. 380 above). Sun put his successors on the wrong track by arbitrarily
emending ch'ih R 'ruler* to shih ~B 'stone*. But this sentence surely pro
vides the key to the whole passage. What interests the Mohist is that a
vertically descending weight can be arrested by something as fragile as a
ruler stood up on the ground.
The ladder is erected by a weight suspended from a rope in front.
This rope would have to be thrown across something outside the ladder,
which I take to be what the Mohist calls the ch'ien tsai, the 'support in
front*. This we may imagine indifferently as the top of a wall, branch of
a tree, or part of a carriage on which the wheeled ladder is transported.
Several speculative diagrams have been suggested (Needham 4/21, Kao
146, Ch*ien Pao-ts*ung 70), to which I will add one more:
Support in front
Rope Hypotenuse
- Trundles
Weight
Ladder
High wheels
Ruler
Weighted front
· - . . : : ; v ; r : ; f ; Flat ground
The descending weight pulls up the ladder and as the balance alters the
high wheels shift on the ground. But why is a weight which lifts a heavy
2/4 395
ladder stopped by something as weak as a ruler stood up underneath it?

Because both are perfectly vertical. When the weight rests on the top of the
ruler its pull still holds up the ladder but ceases to move it; it is like a pull
exerted on a cross-bar from inside a boat.
B 2 8 ^ # ^ J E o | S ^ ! j 383 o
(ft) o 4^384 g , ^385 J§t{j386 , ^gfllJ^lE o
383
Ti ladder', written ® , ±# in B 27 (Sun).
384
Pet ( M . 760 def. 1 =ffi), 'turn the back to'.
3 8 5
Sun identified this word as ch'ien S ( M . 12259 def. 4 = '*), 'haul, drag',
as possibly in Sun 335/-6.
386 \ y follow L i u Ts'un-yan in taking this almost unknown graph ( M .
e
38062) as ch'u )S 'bow down* with a deviant radical.
C. "What leans on something cannot be made upright. Explained by:

a ladder.
E. Resting your back against something behind or pushing away some
thing in front, hauling on something in front or bowing under something
behind, when you lean against it you are not upright."
This continues on the theme of support of the vertical, which is

treated from B 27 to 29. Anything which is perfectly vertical is either
supported from below or suspended from above; to be supported from the
side it must slant, like a ladder against a wall.
387 3 3 8 389 0
B 29 (ffi)** £ ^ ( f t ) M i > i&ffiSM
((H)**) ° ^ £ ^ 5 3 9 0
' S ^ ^ » (&) TJSJ 3 9 1
-&°#S** 3 9 2
R » M S f e * T · #1№%±.»
3 9 3
° ft-tfe° B tt 9 394
Chang Ch'un-yi's identification of the character differently corrupted

here and as head character has won general acceptance (T'an, Kao,
Ch'ien Pao-ts'ung).
Sun already proposed this correction; the two graphs are confused
elsewhere in Mo-tzu (Sun 87/2, 3). The word would be the chu
'pillar, support from below' of the Explanation, which is attested with
this radical ( M . 505 def. 5).
Ts'ai is convincingly identified by T'an as the stone base or plinth of a
pillar (shih ts'ai S W M . 24024/324).
Ping Ш 'lay side by side' written with the 'standing' radical (Sun), lei
( M . 27439 = Ж ) 'lay one on top of the other'.
Ch'in ( M . 7210 =$S), a word used of more than one kind of room or
building ( M . 7289 def. 6-10), erh chia 'flank like ears(?)'
The fei Ш quoted from the Canon was probably corrupted to Ш
( M . 18757 = Щ. I propose this correction on the grounds that 'Xyeh*
at the end of a sentence tends to mark a quotation from the Canon
(§ 1/3/10). Y i i Hsing-wu ШI? chose the opposite solution, emending
the fei of the Canon to fa (Yii 3,8A/6).
I follow Kao Heng in taking kuan in the sense of chih Ш 'put' as in
Han Fei tzu ch. 6 (Ch'en 88/1; for examples, cf. his n. 82).
Chiao 'stick' should perhaps be taken as той Ш 'tie' (Kao).
C. "Things piled one on top of another are necessarily supported from

below. Explained by: placing the base of a pillar.
E. Fitting stones horizontally or vertically, flanking a main room with
side rooms, are examples of 'placing'. Let a square stone be one foot from
the ground, put stones underneath it, hang a thread above it, allow the
thread to reach just as far as the square stone. That the square stone does
not fall is because it is supported from below. Attach the thread, get rid
of the stones: that it does not fall is because it is suspended from above.
When the thread snaps it is because of the pull of the square stone. Without
any alteration except the substitution of a name, it is a case of 'receiving
from below'."
'Pillaring' (chu fii), supporting from below, is explained by the

principle laid down in B 27, that all weight tends vertically downward.
'Pillaring' in statics is the counterpart of ch'ieh 'pulling up, suspending',
just as the counterpart in dynamics is shou 'receiving from below' (B 26, 27).
This is proved by removing the support under a square stone; unless
suspended from above it will fall, and the arrest of the fall may be conceived
either as 'pillaring' or as 'receiving from below'. (The stone is imagined as
suspended by a ssu 'thread', not as in B 26, 27 by a sheng 'rope', because
the example requires an attachment which will easily snap.)
At the end of the Explanation most scholars emend ming & 'name'
to shih ~B 'stone' (Liang, Kao, T'an, W u , Liu), and then propose very
forced interpretations of pien 'alter' and yi 'substitute, exchange' in order
to make sense of a sentence which they have themselves corrupted (for
example, "The stone changing in position without having changed in
direction is 'receiving from below' ", L i u 2/41). This is not the only place
2/4 397
where editors' immediate reaction to the Mohist's most interesting

observations has been to emend them out of existence.
2/4/2/3/3. Economics (B 30, 31).

In A 88, among alternatives between which choice is absolute, we find
"When the price is right, 'dear or cheap'" ftffiHb). Why does not the
Mohist class dearness and cheapness among the relative qualities considered
earlier in A 88 ? The two Canons on economics show that for him there
really is a right price, the one fixed by supply and demand. ThefirstCanon
approaches it from the buyer's point of view, the latter from the seller's. A
price cannot be too high if the buyer will pay it (B 30) or too low if the
trader cannot sell for more (B 31). The proof that the price is right is that
you make the deal, which requires that both parties want to make it:
"whether it is right or not decides whether people want to or not".
These sections show the proper capitalist spirit both in the ruthless-
ness of the illustration ("For example, people in a defeated state selling
their houses and marrying off their daughters") and in the plausibility of
the moral justification. Money and grain are each the price of the other,
money is constant in quantity but grain variable, so that their relative values
fluctuate inversely with the harvest; in time of scarcity you pay more,
but in money which will buy less.
B30 nmn* № 9 E < S 3 9 8

RI°
(H) o 7JB4BAH ° nmmm^n»zmaww 96
· ûmm»»w
si»mmmwrnmn ° [ 5 g » ^ j 3 9 7
°
3 9 5
This is the first uncorrupted instance offan 'converse' (§ 1/4/9). It refers
to the argument of the Explanation, that if the coin is the price of the
grain we can look from the other direction and say that the grain is the
price of the coin.
396
Yi 'exchange'. Cf. A 85 Jf JB, ^rffa "Buying and selling are exchanging".
3 9 7
This seems to be a fragment of the illustration to B 31.
C. "There is no such thing as buying too dear. Explained by: taking the
converse of the price.
E. Coin and market grain are each the price of the other. If the coin
demanded is light the grain is not dear, if the coin demanded is heavy the
grain will not be taken in exchange. If the royal coin does not alter but the
supply of grain does alter, when the harvest alters the supply of grain it
alters the coin "
398
B 3 i muujtf >is£^°
MtfaxKifa ™ o s^Si&flc?^ o
4
o
3 9 8
5Aotf ( M . 36124 def. 12 =1S), 'make a sale', as elsewhere in Mo-tz&
(Sun 278/3, 297/7).
3 9 9
Restored from the parallel in the next sentence.
4 0 0
Since the grammar does not allow subjectless initial clauses concluded
by yeh (p. 154 above) we cannot follow Kao and L i u in punctuating
three places earlier.
C. "When the price is right you make the sale. Explained by: all.
E . ' A l l ' refers to the removal of all obstacles to making the sale. When
the obstacles to making the sale are removed, making the sale decides the
price. Whether the price is right or not decides whether people want to
or not. (For example, people in a defeated state selling their houses and
marrying off their daughters.)"
2/4/2/4 Problems in disputation (B 32-82)

The arrangement of these 51 Canons appears to be random, except that
some fall into pairs (B 57 and 58, 63 and 64, 71 and 72, 78 and 79) and a
few into longer sequences (B 37^4-2, 73-75). It is possible that there is some
principle of organisation which has eluded me; if so it would be very useful
to discover it, for in the absence of context many are very obscure. Many
of the theses criticised may be identified as doctrines of thinkers of rival
schools, the Confucians (B 34, 36), Taoists (B 44, 49, 69, 77, 81), sophists
(B 47, 54), the Five Elements school (B 43), Kao-tzu (B 76), Chuang-tzu
(B 35, 48, 68, 71, 72, 79, 82). But the Mohist's preoccupation is less with
refuting his rivals, who are never named, than with sorting out conceptual
difficulties arising in these debates. Thus the Explanations of B 73, 74
refute a couple of objections to universal love, but the issues as presented
in the Canons are quite abstract, "Being limitless is not inconsistent with
something being done to everyone", and "Without knowing their number
we know that something applies to all of them". The scope of the issue also
tends to be narrowly delimited; a Canon will affirm merely that X is not
inconsistent with Y ( ^ S , B 45, 51, 73, 75), or does not necessarily require
Y (sfi'j&ffi, B 49), or that there are the same grounds for it as for Y (l@ Y IRI ,
B 67, 68, 82), or that it is known or unknowable 'a priori' (B 57,58), or is not
establishable on such-and-such grounds (B 78). When a thesis is accepted
or rejected the Canon is often specific as to whether it is logically admissible
2/4 399
or inadmissible (Hf, ^ f ^ I , B 36, 54, 72), fallacious or self-contradictory

B 34, 71, 76, 79), or whether it is mistaken in fact ( ® , B 40), necessarily
does not fit the fact B 35). (The question is never whether it does
fit the fact, which would presumably be outside the scope of disputation.)
On the other hand many Canons neither affirm nor deny, merely present a
problem the nature of which may not be evident even after one consults the
Explanation.
We have identified the fourth Mohist discipline as the explaining of
names, and certainly it is generally true throughout this series that the
problems are conceived as soluble on the plane of names without observa
tion of objects. "The five elements have no constant ascendancies'' ( B 43)
is perhaps the only exception; the evidence offered is factual, but even the
other problems which belong to the sciences ( B 52, 62) are given 'a priori'
solutions. Sometimes conclusions are deduced directly from definitions or
analyses of words in A 1-87 ( B 35, cf. A 74: B 55, cf. A 77: B 76, cf. A 7, 8),
or are reached by establishing the type of name or by distinguishing name
and object ( B 33, 48, 53), or a problem is dismissed in a few words by
calling attention to what something is or may be wet ffl 'called* or to the
principle of deciding between 'X* and 'non-X' in A 73 ( B 34, 47, 54).
The Explanations are often very short, limited to the point which the
Mohist thinks crucial. He assumes that his readers know the current
arguments for and against; in most cases of course we do not, and often we
have nothing to go on but guesswork. Generally speaking we are on firmest
ground in the formally completed demonstrations which become commoner
towards the end of the series.
B 32 tesmm 9
*±*oi 0
4 0 1
(ifr). The graph is invariably in the Canons and Explanations a corruption
of i t or, once only, of IE (A 53). Sun's emendation to the pi of the
Explanation seems plausible only as long as one fails to recognise the
special sense of chih in the Mohist terminology (to fix name on object,
description on fact, cf. § 1 /4/4). The Explanation is quite explicit that
one may be unafraid even when not certain (pi).
C. "When one cannot explain things one is afraid. Explained by: not
having fixed the matter.
E. Having a son in the army, one is not sure whether he is alive or dead.
Hearing about a battle, one is likewise not sure that he is alive. In the
former case one is not afraid, in the latter one is."
This Canon may be read as a general declaration of the importance

of knowing by explanation, not merely by observation and hearsay (the
three sources of knowledge distinguished in A 80: for wu shuo 'have no
grounds to offer', cf. § 1/4/40). As long as one has no properly thought out
principles one is at the mercy of every new fact that comes to one's
attention. But the reasoning man 'fixes' (chih) names on transitory objects
by the art of description and his preferences by the art of considered choice
(cf. A 75 l h 9 f & "fix which he prefers"). The Mohist no doubt decides,
after duly weighing the alternatives, that he prefers his son to be in the
army, for the benefit of the world as a whole, and afterwards is undisturbed
by reports of battles; even if his son does die, it was the right decision.
We noticed under A 51 that the Mohist does not question the certainty
(pi) of knowledge by observation or report; it is only within the sphere of
knowledge by explanation that the certain is the logically necessary. The
man who knows only what he perceives himself or hears from others can
be certain of some things; but his fears do not arise simply from uncertainty,
they come and go illogically. The thinking man will still be uncertain about
judgments which the last two disciplines cannot establish as necessary; but
he is freed from anxiety by his factual and practical decisions according to
the procedures of the first two disciplines.
B33 l"«J > ° t&lEK °

(*) · > W^^^ffiifcffe 402
> «ffluffljfc r m J r it J <>
4 0 2
Shih 'the thing in question, what we judge the object to be' contrasting
with tz'u 'the object/place here', as in B 1, 2 (cf. § 1/3/4/2); the former
belongs to the realm of names, the latter of objects. Yu 3£ 'again' is
as usual written with the graph 'ft*.
4 0 3
Cf. B 10 ftH, & E H & - t f e | S , Mffa "Is it knowing? Or is it supposing
the already ended to be so?: (doubt due to) having passed".
C. "Huo (in one case . . . in the other case . . .) is a name which 'passes
beyond'. Explained by: the object.
E . Knowing that what we judge them to be neither is the place nor is in
the place, none the less we call these places the 'North' and the 'South'.
Having passed beyond them we treat the already ended as so; previously
2/4 401
we called this place 'Southern', therefore now too we call this place
'Southern'."
Previous editors have taken huo here and in A 42, 49, B 13 as yii ^
'region', in which case the theme is some spatial paradox. But we have
concluded that the contrast of huo with mo H 'in no case' in A 42 makes
this claim untenable; the Mohist simply applies the ordinary huo 'in one-or-
other case* in an unfamiliar way to spatial directions (p. 129 above). If so the
Canon is about a peculiarity of the word huo, and the Explanation merely
provides a spatial illustration.
There is nothing in the least surprising about the Mohist discussing
such an abstract word as huo (which is defined in N O 5 as 'not in
all cases'). But what problem would it present ? I would suggest that he is
thinking of the effect of huo on pairs of relative words, and that the nan pei
'the North and the South' of the Explanation is to be understood as huo
nan huo pei 'one South, the other North'. To call one of two minor offences
'heavy' does not forbid us to call it 'light' when we extend the range of
inquiry to major crimes; but if we use huo and say 'one crime is light and
the other heavy' (B 69 W^M^M) the proposition remains true even when
we 'pass beyond' the limits of the situation, as it would in English if we
used the comparative adjective, 'One crime is heavier than the other'. Kuo
'passing beyond' was listed in B 10 as one of the four sources of doubt:
"Is it knowing ? Or is it supposing the already ended to be so ?". But when
we use huo . . . huo . . . we are entitled to "treat the already ended as so";
huo is a 'name which passes beyond'. T o show this, the Explanation takes
a pair of obviously relative words, North and South. The North and the
South neither are these places (as some object may be a stone) nor are in
these places (as hardness and whiteness are in a stone, B 37). As soon as
I pass beyond the limits of the region, the 'South' becomes my North,
but it remains true that it is the Southern part of the region. It is assumed,
as generally in such illustrations (§ 1/5/13), that movement is Southward;
this explains why the Mohist mentions the South without the North
(which is still to the North of the traveller).
4 0 4 4 0 5
B 34 ft * n »tfr£ft$£»fe o
4 0 4
Tsu yung (sufficient to be used), 'applicable in practice'. Cf. ch. 47
(Sun 281/2) ^ a ^ f f i t f e . . . ^ a J S f f l f t " 'Your words can't be
applied in practice'. . . . ' M y words are adequate for being acted on' " .
4 0 5
Pet 'fallacious' (§ 1/4/23).
406 W y{ «i k the distinguishing marks' (§ 1/4/35). Cf. A 73
u ac 2№
" T o lack what distinguishes 'ox' is to be 'non-ox' ". For lun 'sort,
grade' cf. § 1/4/19.
C. "It is fallacious that the knowledge of whether one knows something

or not is sufficient to act on. Explained by: lacking what distinguishes
knowledge.
E. When we sort out one from the other, the non-knowledge lacks what
distinguishes knowledge.''
According to Confucius, "When you know something to recognise

that you know it, when you don't know to recognise that you don't know,
this is knowing" (Analects 2/17 ÛftlZ, ^ f t H S ^ f t , JifcH&). The
Confucian Hsiin-tzu quotes a longer version of the saying which ends
"When one both knows and is benevolent, what more does one need?"
(Hsiin-tzu ch. 29, Liang 396/1 *JB^JB&Sfc).
The Mohist does not object to the saying of Confucius, but sees danger
in the suggestion that knowing one's own ignorance is all that one needs
to know. When one knows X but not Y , knowing one's ignorance of Y
adds nothing to one's knowledge of X . It is still not the case that "having
experienced the thing one is able to describe it", which is what distinguishes
the knowing of Y (A 5).
This section may be compared with B 48, on knowing what one does
not know. The latter illustrates what is described here as 'sorting out' (lun)
the known from the unknown.
B 35 mmm&>jK*vt»№&m °
4 o 8
ssii^»(^)*K â±s*»rnmm»°rijfti'
407
So wet 'the (thing) which a (name) refers to, its meaning'. Cf. A 80
#?EU5£*f]i, @ f " W h a t one uses to refer is the name, that to
which one refers is the object".
4 0 8
Corrected on grounds of parallelism; the corrupted ch'i was no doubt
written with the graph 5f (cf. B 3 n. 275, 24 n. 354).
C. " T o say that there is no winner in disputation necessarily does not fit
the fact. Explained by: disputation.
2/4 403
E. The things that something is called are either the same or different.
In a case where they are the same, one man calling it a 'whelp' and the
other man a 'dog', or where they are different one calling it an 'ox' and the
other a 'horse', and neither winning, is failure to engage in disputation.
In 'disputation', one says it is this and the other that it is not, and the one
who fits the fact is the winner."
Chuang-tzu in his Seeing things as equal argues that there is no winner

in disputation; if two disagree it is useless to appeal to a third party, since
the case for either side is not strengthened by adding to the number of its
supporters (Kuo 107/1-6). The Mohist reply is that there are debates in
which both parties may be right ('whelp' or 'dog') or both may be wrong
('ox' or 'horse'), but these are not disputation as the term is defined in
A 74. In true disputation one side maintains that the object is an ox and
the other that it is not, so that one must be right and the other wrong.
B 36 rn^mw^ ° tsfffe °
C. " 'Defer to the other man in everything' is inadmissible. Explained by:

the start.
E. The time before the man who defers asks the other man to drink first
is 'the start'. It would be impossible for the other to drink first."
Confucians use the word jang 'defer to others' as though deference

had moral value in itself, and sometimes treat it as the essence of li 'manners/
propriety'. Cf. Mencius 2A/6 f S K ^ ' O j I ^ f f l ' t b " A deferential heart is
the starting-point of good manners", Tso-chuan Hsiang 13//« S i , SS^lîfe
"Deferring is the main thing in good manners". We should not expect the
Mohists, for whom good manners is a minor convenience of social inter
course, to take this attitude. The only other examples of jang in Mo-tzii
are in ch. 49, in the phrase jang hsien WkK 'deferring to the worthier'
(Sun 300/1, 4). Deferring to the man one recognises as better than oneself
is the only kind which would make much sense to the Mohist.
The Explanation is brief and rather obscurely phrased, but seems to
be one of the proofs of self-contradiction for which the Mohist shows his
fondness elsewhere ( B 71, 77, 79). No Confucian is known to have explicitly
recommended 'Defer to the other man in everything*. However Confucians
do treat deference as commendable for its own sake, and it is very much in
the Mohist's style to put their position in a rigorous form and prove it to be
self-contradictory. The simplest refutation would be that 'You first',
404 The 'Canons' and 'Expiations'
'No, you first* leads to an infinite regress. But this would be trivial, since
after both have demonstrated their willingness to defer there can be a
convention to decide which does in fact make the first move. The Mohist
attacks from the opposite direction, from the initial state. His point seems
to be that the man who at the start is in possession of the wine cannot
meaningfully defer to a man who is obliged to defer to him.
B 37 5 ^ — i f » HtttoM · » 8 E # o
C. "In one thing you know something and do not know something else.
Explained by: belonging in it.
E. A stone is one, hardness and whiteness are two but are in the stone.
Therefore it is admissible that you know something in it and do not know
something else."
B 37 introduces a series ending at B 42, all concerned with knowing,

pointing out and 'belonging in* (ts'un, cf. § 1/4/30). B 37 pairs with B 38,
both on the knowing of the hardness and whiteness within a stone.
According to the Essay on hard and white, forged between A.D. 300 and 600
in the name of Kung-sun Lung, we can know one without the other because
we perceive them with different senses, sight and touch. Such a criticism
of sense perception is unlikely to be pre-Buddhist (but cf. B 47). The
Mohist's problem is any case different, that we know a priori that a stone
will be hard but not whether it will be white. This is plain from B 38,
where the two qualities are distinguished as 'what I refer to a priori'
(S0?3fc*) and what I do not.
Chien pat ('as hardness to whiteness*) was a general term for mutually
pervasive qualities, of which hardness and whiteness were chosen as
representative (§ 1/4/3). In early Chinese thought such qualities seemed to
present a difficult problem. It appeared that in so far as an object is a stone
it is hard but colourless, and that if it is white it is a white thing as well as
a stone. This problem led to the type of sophistry called 'separation of hard
and white* (HISS), of which Kung-sun Lung*s White horse is the only
surviving document (§ 1/4/3). A n object in so far as it is a horse has shape
but no colour, and therefore no object which is white can be a horse. The
Mohist*s approach to this kind of problem is shown by his analysis of the
idea (yi M) of a pillar in B 57. The idea does not include roundness or
catalpa-wood, yet we are not conceiving the pillar as lighter than the wood
of which it is made. The idea is a scheme known a priori (hsien chih 3feS)
with filled and unfilled places. On the same analogy the scheme of a stone
2/4 405
will have a place in it filled by hardness and another filled by a colour

which may or may not be whiteness, both of them 'belonging' in the stone.
Here the Mohist is attacking the assumption that what one knows and
what one does not know must be different things both of which the object
is (a stone and a white thing). In B 38 he proceeds to refute the assumption
that we can point out instances of the white things which stones and horses
are not.
4 1 0
B 38 »itfc—«*ffS£nJS£ * ! S £ D U 1 ( ^ ) * # *
(*H) o r ^ « ? s 4 1 1
>w@jî3fafc« »fl№»aiffi*»§B?fc
4 1 2
*ft » & - 3 I M r « « »W * I F * § f t o J
mm*. · a i * « " » £ o o r j ft o
ft£ » r * J 4 1 5
f i £ f t o^0 . r*«fit^№* »» * 5 0 f ^ « J »
m < s > 4i« # S 7 № H & ° m 4 1 7
» o
4 0 9
"Pointing out some is in two things", parallel with B 37 ^ftlM
"In one thing, you know some of it". The order is reversed and the
pointing put first because the Mohist is contrasting it with knowing.
4 1 0
Emended from the Explanation.
411
Chih shih 'know what it is' (A 93), contrasting with chih chih 'know them*.
412
Ch'ung 'duplicating', first of the four kinds of sameness, defined as
"there being two names but one object" (A 86).
413
Ch'ang ( = H ) , 'try out'. For other examples of the graphic interchange,
cf. B 61 and Sun 65/5, 146/-2, 155/—1. Ch'ang often has a hortatory
function, as in another example with this graph: Mo-tzu ch. 14
(Sun 65/5) Itf^ftLM 1=1 "Let's examine the source from which dis
order arises". Pre-verbal tang 'ought' is not used in the dialectical
chapters.
414
Heng (=flS, M . 34078 def. 1), 'crossing, athwart'.
4 1 5
A metaphor from the aligning (ts'an) of two gnomons in a straight line
(chih) with the sun, as in A 57.
4 1 6
The identification of the missing word as chih is a conjecture of
Sun.
417
Hsiang (=$S) 'imagine' as in B 57, again in the same context 3&yi 'idea',
cf. § 1/4/37.
C. "Pointing out one of them is inescapably in two things. Explained by:

using the two to align.
406 The 'Canons' and 'Explanations*
E. (Argument). If you know what an X ['white thing'] is, and know in

addition that the X is identical with what I referred to a priori ['hard
thing'], then the fact that you do not by knowing what an X is know that
it is what I referred to a priori shows that even in one of the things the
object is called ['white thing'] you know something and do not know some
thing else.
(Refutation). If you know of Xs, try to point them out; if you know and
inform me I shall know of them. You point at them in combinations
[things both hard and white], which is 'using the two'. You point at them
in crossing directions [things hard but not white and white but not hard],
which is pointing straight to them by 'aligning'. If one says 'Be sure to
point only at what I referred to, and do not refer to anything I did not refer
to', then the man who points is inherently unable to point at it alone. What
he wishes one to imagine is not conveyed, it is as though the ideas were not
yet co-ordinated.
Moreover, if what one does know is X , and what one does not know is X ,
this is knowing the very thing one does not know. In what, constituting one
of the things an object is called, can you know something and not know
something else?"
The Mohist proceeds to defend his claim in B 37 that whiteness is

in the stone, .that an object is not a white thing in addition to being that
colourless thing a stone. He begins by presenting an objection, that 'white
thing' and 'stone' have the same status; with the former as with the latter
one does not, by knowing what the thing is, know everything about
particular objects. The Mohist replies by proving that their status is quite
different, by appealing to the objector to point out white things from other
things, and showing that to do so he has to point out at least two objects
with whiteness in them. In a single object one points at hardness and
whiteness combined (chien chih) :
" I f one says 'Be sure to point only at what I referred to, and do not refer
to anything I did not refer to' " (that is, do not say you are pointing at the
colour, cf. B 70 "The colour of what is in the room is like the colour of
this"), it is impossible to point at the white thing without pointing at the
stone and the hard thing which the stone is. One can only point out white
2/4 407
from hard by comparing with other objects which are one and not the other,
which is "pointing in crossing directions** (heng chih) :
The dialogue form, with a fallacious argument followed by a refutation,

reappears in B 73, 82.
B 39 BffcliF5#№f§ » m£ r # i b J 418 , r ^ E J > ["»0* J > r ( « ) * « 4 1 9
*J >*@;££iii 422
° r»#J ' ^ t g M * 4 2 3
°
'Spring snakes', cf. n. 420 below.
Corrected from the Explanation. There is the same dropping of the
radical in yi in E C 11.
Several graphs are only identifiable by restoration of 'insect* radicals:
$K = W: (both proposed by Kao Heng) and possibly # = H ,
although this would not materially affect the sense (here Kao unneces
sarily emends to tung 'winter'). It is clear that the graph -& cannot
in this context represent the particle yeh ; and its only appearance in
a similar context is in E C App. 13 KSffiÊifa "Its analogy is: dead
snake", where there is actually a variant № . Moreover the phrases
both occur in B 50, which has other indications of being
about snakes. It is remarkable that the text as we have it should use
the graph ifa both for she$fa #£ 'snake' (except in E C App. 5) and
t
for t'o Ah, 12 'other'. Possibly the Mohist distinguished both ixomyeh
by the graph 12, and it was assimilated to yeh by later scribes.
f f
Ch'en 'menial, slave' ( M . 30068 def. 5). For t ao ch en 'runaway slave',
cf. Tso-chuan, Chao 7, fu 1, and also Documents, Pi shih
(Karlgren 80 v. 4) E S c M * ! "when slaves and slave women abscond".
If one knows dogs by the name kou but not by the name ch'iian, one
knows of ch'iian without being able to point them out (cf. B 40).
This sentence is unintelligible without the understanding of several

technical usages ( § 1 / 4 / 1 8 ) . Liang refers to the two sides in disputation,
X and non-X, and is transitive as in A 8 3ffirfn^/fi"Apply it on both
sides, not on one without the other". If a sophist tries to pass off X
and Y as X and non-X we may know that something is 'missed out*
(yi) without necessarily being able to point out what it is. Ch'iao
'skilful* is used of words intended to mislead, sophistries, as in A 10,
94, 95.
C. "Things that one knows but is unable to point out. Explained by:
snakes coming out in spring: a runaway servant: whelp and dog: what is
missed out.
E. 'Snakes coming out in spring*: while they are hibernating it is inherently
impossible for them to be pointed out. ' A runaway servant*: one does not
know where he is. 'Whelp and dog*: when one does not know its name.
'What is missed out*: when sophistry cannot make them into the two
sides."
B 3 9 - 4 2 all take up points involved in the argument that whiteness is
in the stone ( B 3 7 , 38). Although you know what whiteness is, in a single
object you "are unable to point at it alone" ( B 38). This Canon adduces
other examples to show that there is no contradiction in being unable to
point out something that one knows. The last two examples are:
(1) If I do not know that the animals I call kou (whelps) are also
called ch'iian (dogs), I know ch'iian without being able to point them out
(B40).
(2) I know that nothing is left out of 'oxen and non-oxen*; I also
know that things are left out of 'oxen and horses', without necessarily
knowing what they are (cf. B 3 5 , and § 1 / 4 / 1 8 ) .
B 40 *p*ftMiigrf ° m&n °
( » ) o mfmm*»mm ° > «1*» °
C. " I f you know whelps, to say of yourself that you do not know dogs is
a mistake of fact. Explained by: identity of the objects.
E. If the knowing of whelps is of identical objects with the knowing of
dogs, it is a mistake; if the objects are not identical, it is not."
This would be a case of knowing the objects, but not knowing the
name (cf. B 3 9 ) or that name tallies with object, three kinds of knowing
distinguished in A 8 0 . 'Whelp' and 'dog' are the stock examples of two
2/4 409
names for one object, but there seems to have been some doubt as to their
perfect synonymity (§ 1/5/3). The Explanation adds the qualification that
the Canon only applies in case of synonymity. (Cf. also B 54.)
This Canon takes up the objector's claim in B 38 that the hard object
one knows the stone to be is 'identical' (ch'ung, cf. A 86) with the white
object one may not know it to be, in the same way that a whelp is identical
with a dog. The Mohist here points out that in cases of identity one does
know all about the object whether or not one knows both names.
B 41 S f t f t t t »Efrffi^Xit «4 fife o
(a)oFpi#e, r J *m±B' r^Miiftj o « H » n
№ j 426 , ijg£ o
fl ^ M K M I R » »mm °
4 2 4
5faf 'which' (p. 135 above).
4 2 5
Unknown graph. Since it represents a word that we are not expected
to know, we are at liberty to coin an English equivalent.
4 2 6
Judging by the context, this should be a familiar graph corrupted and
fused with a final yeh.
4 2 7
If we have punctuated correctly pi should be an injunctive 'be sure o f ,
as in B 38.
4 2 8
(ft). For the systematic corruption of ping, cf. § 1/2/1/3/21. Tsai, which
the Mohist allows an unusual syntactic mobility (cf. p. 163 above),
may be treated as nominalised as in B 17: "its presence is prolonged
with man".
C. "Understand the idea before replying. Explained by: not knowing

which he means.
E. The questioner says 'Do you know about the blomes?'. You answer
him: 'What do you mean by blomes?'. When he says 'Blomes are such-
and-such', you did know about them. If you do not ask what he means by
blomes, and answer prematurely that you do not know about them, you
have made a mistake of fact.
When about to answer, ensure that the time to answer the question is as
long as the answer. There are profound answers and shallow answers.
Among Heaven's constants its presence is prolonged with man."
The Mohist does not need to say what it is that is eternally present
(tsai) among Heaven's norms. In E C 2 he distinguished the 'a priori'
ethical system of Mohism from the moral judgments made in concrete and
transitory situations, and concluded with another uncharacteristically

dramatic announcement: " I f there were no men at all in the world, what
our master Mo-tzu said would still be present (tsai) ".
This is the one Canon in B 32-82 which is imperative in form (like
some of the Canons laying down norms of description, A 94-97, B 1, 2).
At first sight it does not seem to concern a problem at all. But like B 39
and 40 it takes up a point in B 38, that when a man points out whiteness in
one object the idea (yi) is not conveyed. We cannot accept his claim to have
4
pointed out whiteness because we "do not know which he means ' (hardness
or whiteness). The Explanation illustrates the possibility of misunderstand
ing by a simpler example.
4 2 9 4 3 o
B42§f#»<#> ^'^ f f « $ A № ' <m& . . . o >
r s p j # j ft 432
° ^M>mmr%> -tn ° &-^%\mm
4 2 9
Restored from the Explanation. The Canon is badly mutilated, and
except for B 24 is the only one in Part B to have lost its summing-up;
at Stage 2 it began the bottom row of Part B, in an especially vulnerable
position in the first column in the scroll.
430 ( j g ) < hat?', as in B 57 (§ 1/3/7).
W u = in w
4 3 1
Tsai 'be present' (unlocalised) contrasting with ts'un 'belong in' (§ 1/3/7).
432 \y « h a t ? ' (§ 1/3/7). Since a verb after k'o is passive ts'un should be
u m w
causative: k'o ts'un 'may be regarded as belonging' (p. 159 above).
C. "Where something belongs and that which belongs, 'In which does it
belong?' and 'Which belongs?'. Explained by: . . .
E. The rooms are where they belong, the members of the household are
those who belong. If we start from those present and ask about the rooms,
the question is 'In which may they be considered as belonging?'; if we
take the rooms as topic and ask about those who belong in them, the ques
tion is 'Which belong?'. So in one case we take those who belong as topic
and ask where they belong, in the other we take where they belong as topic
and ask about those who belong."
The last of the series B 37-42 returns to the theme of the first, the
mutually pervasive properties which are in (tsai) the object, which belong
in and constitute it (ts'un), although sophists such as Kung-sun Lung
2/4 411
treat them as 'apart' (li St). In interpreting the analogy with a household
we may notice:
(1) Shih t'ang ('chambers and hall') refers to the rooms composing
a house. It can hardly be taken as a compound word for 'house' (in any
case unattested in pre-Han literature to the best of my knowledge), since
the later Mohist style avoids compounds (p. 164 above), and elsewhere the
word shih 'room, house' always stands alone (EC 6, 12. A 86, 88. B 70).
(2) The key phrase in the Canon is 'In which does it belong?'
(wu ts'un), which is treated in the Explanation differently from the other
three (chit for chit, tsai for ts'un, and an added k'o in the question). The
point at issue seems to be whether properties, which may be conceived in
isolation from all objects, can be treated as belonging to objects. The
Mohist answers that they can, since even if we start from an isolated
property described merely as present (tsai) we can ask 'In which may it
y
be considered as belonging?' (wuk o ts'un), so that as topic of the question
it is something which belongs (ts'un che).
B 43 liftmm ° m&'ti. o
( £ ) o [&7k±'X'xnm * B\i]^xm±»x&& ° ±B%L » °
[^±m-^m-M^mmik^mmmm ° B 11] 434
3 4
The two fragments compose the missing Explanation of B 11. Since
they seem to refer to four of the Five Elements it is natural that they
should have been pushed into the only section where the Five Elements
appear. But the word fu shows that they were being treated as members
of the liu fu 'six storehouses', five of which corresponded to the
Elements.
C. "The five elements do not have constant ascendancies. Explained by:

whichever is appropriate.
E. That the fire melts the metal is because there is much fire; that the
metal uses up the charcoal is because there is much metal."
According to the Lu-shih ch'un-ch'iu (ch. 13/2) the Five Elements

succeed each other in a regular cycle, the order of ascendancy from weakest
to strongest being earth, wood, metal, fire, water. By associating each
dynasty with an element one can explain the rise and fall of dynasties
down to the contemporary decline of the Chou (associated with fire), and
predict the rise of a new dynasty with the element water (Fung (1952)
l/161f). Throughout the history of Chinese civilization varieties of this
system have provided proto-scientific principles of explanation for cyclic

processes in history, astronomy, alchemy, medicine, geomancy (Needham v.
2, 216-268). The Mohist rejection of the principle has an important bear
ing on the question of the nature of Mohist scientific thinking (§ 1/1/2/6).
"The five elements do not have constant ascendancies*' ( E f f &SS8&)
appears among examples of the fluidity of circumstance at the end of ch. 6
of the military classic Sun-tzu ( S P T K 6, 33B/3).
This is one of three Canons with the single word yi 'appropriate' for
the summing-up (cf. B 44, 51). A l l refer to the three kinds of relation
(ho distinguished in A 83, exact, appropriate, necessary. The relation
of two elements is merely a matter of appropriateness; it is in shih $L
'causing' (A 77), where the cause is a 'major cause' (ta ku A # t A 1), that
the relation is necessary.
B 44 fttft&tff S A t t - ifrfiES: o
437 438
as A »*j Agft»ium s»?&-ai -
Most of the later editors (Chang Ch'un-yi, T'an, Kao, Liu) take this
as a reference to Shao-lien (Analects 18/8), who together with his
brother Ta-lien is praised by Confucius for his meticulousness in
observing the three-year period of mourning (Li chi ch. 21, SPPY 12,
13A/4). But this is extremely unlikely, since (1) There is nothing to
connect Shao-lien with injury to health by indulging the passions.
(2) The dialectical chapters mention no one by name except the sages
Yao, Y i i and Mo-tzu; a reference to a minor Confucian hero would be
very surprising. (3) Syntactically the clause can hardly be taken as
"explain by means of Shao-lien", since the grammatical system allows
the instrument only before the verb; it should therefore be taken as
Shut yi . . . 'advise to . . .' (§ 1/3/12/4/6, 7). I take lien as in Shih chi
ch. 113 (2972/-3) # * S a * J E . 5 & S £ , " H i s daugh
ters were all married into branches of the royal house, and he had a
connexion by marriage with the 'Ch'in king' of Ts'ang-wu". Cf.
Mo-tzu ch. 70 (Sun 372/7) understood by Sun as "hold his
kin as hostages", and a doubtful case in ch. 15 (Sun 72/-2) SlSifti
JnLA?#, which may be taken as "Alone of his kin and without
brothers".
For the syntax of huo-che yii pu yu "desire not to have some of it", cf.
p. 134 above.
2/4 413
™Chih(=%§) 'wise', cf. A 6.

4 3 8
Sui (=№) 'although', cf. A 17 n. 87.
C. "Whether lack of a desire or dislike is to be deemed gain or loss.

Explained by: whichever is appropriate.
E. If, because desires and dislikes injure life and involve loss of years you
recommend reducing ties with others, then which person are you loving ?
When one dines on too big a meal, one desires not to have some of it,
because it can be injurious. (Like the effect of wine on a man.)
Moreover the wise man, when benefit to man is for the sake of love, does
not control it however wise he is."
We have more than once insisted on the threat to both Confucians and
Mohists from the apparently unanswerable claim of Yang Chu that man
obeys Heaven by acting according to the nature (hsingft)which he derives
from Heaven, which guides him to live out the full term of his life in good
health. (Cf. E C 1.) The Mohist reaction was to refuse to recognise the
concept of a nature ordained by Heaven, to define sheng ^fe 'life* as "the
body being located with the consciousness" (A 22), and to build an ethical
system on the two terms desire and dislike (§ 1/1/2/5).
Taoism, which inherited Yang Chu's principle of fulfilling one's
nature and avoiding risks to longevity, recommended freedom from all
desire (Lao-tzu 57 ^c^^CffSSS^ "When I myself am without desires the
people will become simple of themselves"). That such a position can be
held at all is a serious matter for the Mohist, whose ethic depends on the
,
assumption that what is desired 'a priori is the one unquestionable value
(§ 1/1/2/5). The issue in B 44 and again in B 45 is whether the ultimate test
of action is:
(1) Gain (yi) or loss (sun, defined A 46, cf. A 47 n. 131), to the body
and the means of nourishing it, in which case it is tenable to reject all desire
and dislike on the grounds that they "injure life and involve loss of
years"; or
(2) Desire and dislike. If one starts from the desire to benefit oneself
and men in general, from the love of self (SB) and the love of men (SA),
one may desire to benefit others even at the cost of loss to one's own body
(cf. A 19 MBffi]S@f % "at loss to oneself and to the gain of those on whose
behalf it is").
The Mohist argues that loss and gain can be valued only in relation
to desire and dislike:
(1) If, because desires and dislikes are dangerous to life, one recom
mends reducing ties with others, then "who/which does one love?" We
might take this simply as a criticism of individualism as immoral. But more
probably the Mohist's point is that the recommendation does assume love,
the love of self rather than others. (Grammatically shut is always translatable
by 'which?', cf. p. 135 above.)
(2) When refraining from over-eating one is not judging desire by
the ultimate test of its effects on health; the reason for not over-eating is
that one does desire something, to avoid injury.
(3) Although the wise man will reject particular desires and dislikes,
there is one desire which there can be no possible reason for controlling,
the ultimate desire to benefit oneself or another which is love.
As in the previous Canon the summing-up refers to the three kinds of
relation as classified in A 83. Here the syntactic pattern is also that of A 83
J#;£j§, Sife "In the case of what Jack is deemed to be, 'appropriate' ".
The relation of lacking a desire or dislike to either gain or loss is not 'exact'
(cheng IE), but a matter of appropriatenes.
4 3 9
B 4 5 Jiffn^:« o i a ^ o
(ffl) o »ms&fift»mtm °( £ « £ « $ 4
ft) °
4 0
4 4 1 4 4 2 4 4 3 4 4 4
Mllffn/gô <B46 ^ ft ^ [ 2 i ] feiliko >
4 3 9
Pu hat 'no inconsistency, no objection' (§ 1/4/12). Cf. the exactly parallel
phrasing of the Canon of B 5. For 'harm, injure' the companion
section B 44 uses shang.
4 4 0
Pi ( M . 29579 def. 2/1 = W), 'haunch', cf. B 54.
4 4 1
Unknown graph identified by Sun as niieh $H 'malaria'.
4 4 2
Ping 'critical' (of illnesses), cf. Analects 9/12 "The Master's
illness was critical", Chuang-tzü ch. 24 (Kuo 844/1) №5£±7?!ffi^
"Your illness is critical".
4 4 3
Delete as accidental repetition.
4 4 4
The whole illustrative gloss entered the text two places two late, at the
beginning of the next section.
C. "It is no objection that there is loss. Explained by: the surplus.

E. When a man who has eaten well refuses the surplus, that he has had
no more than enough is no objection, it could interfere with his having
eaten well. (Like the wounded milu deer lacking a haunch.)
Moreover there are cases where one cannot gain without first losing.
(For example, the malarial crisis in the case of malaria.)"
2/4 415
This continues the theme of B 44. For the school of Yang Chu a 'loss
to life* (sun sheng A ^ , cf. Lu-shih ch'un-ch'iu (ch. 2/3) Hsu 118/—3) is th
ultimate ill, whether or not it serves a desire to benefit others. The Mohist
replies that loss of things which sustain the body, even the loss of flesh in
illness, are not necessarily bad even in relation to bodily survival. The
point as in B 44 is that gain and loss cannot be valued for themselves without
taking account of desire and dislike.
The first illustration refers to one of the more mysterious of the stock
examples, the milu deer (§ 1/5/5). Comparison with B 2 suggests the
tentative hypothesis that it may have been used in disputation to make the
point that removal of a part does not affect what a thing is. A thing which
is recognised as an animal by its four legs is distinguished as a milu deer
by its four eyes; after the loss of a leg it is still an animal because a milu
deer is an animal and it still has the four eyes of the milu deer. (Cf. also
B 54.) The point of the illustration might be that losses to the body do not
affect what distinguishes us as man (jen A ) ; the Mohist does not recognise
a human nature (jen hsing A t S ) which is unrealised if we act at the cost
of ioss to life*.
4 4 5
B 46 « : № 3 I i & » ISffiA o
4 4 6
W% [B 45] (S) JX B S f f i g ElAMfffiA/PM°tjlBU£*& · °
445
Wu lu 'five roads*; the phrase (otherwise unattested) evidently refers to
the five senses. Hsun-tzu in his Right use of names lists the kinds of
sights, sounds, tastes, smells, bodily sensations and emotions dis
tinguished respectively by the eye, ear, mouth, nose, limbs and heart;
he classes the heart separately as the organ of thought, and calls the
rest the 'five offices' ch. 22, Liang 313/-2).
4 4 6
This is the only example of a head character placed too late (§ 1/2/2/5/4);
probably it was mistaken for the end of the illustration written above
it in the margin, and entered the text with it two places too late
(§ 1/2/2/6/2).
C. "When one knows, it is not by means of the 'five roads'. Explained by:
duration.
E. The knower sees by mean of the eye, and the eye sees by means of
fire, but fire does not see. If the only means were the 'five roads', knowing
as it endures would not fit the fact. Seeing by means of the eye is like seeing
by means of fire."
The crucial point about knowing for the Mohist is that it persists even
when the object is no longer in front of your eyes. Seeing is by means of
the eye, knowing is by means of the chih 'intelligence* (A 3). You do not
know unless "by the use of your intelligence having 'passed* (kuo ® ) a
thing, you are able to describe it" (A 5). It is this persistence of knowing
which shows that we do not know by means of the five senses.
It may be noticed that the Mohist agrees with the Second List sophism
"The eye does not see" ( B ^ l . , Chuang-tzu ch. 33, Kuo 1106/2), on the
grounds that the eye is a means to the seer just as light is a means to the eye.
447 1 4 4 8
B 47 (*ft) * i/c i& ° tSS ® °
(X) o HWcMHt» 4 5 0
° ( £ 8 · 0) °
4 4 7
4 4 8
Sun emends the graph (of which this is the only example in the dialec
tical chapters) to tu $1 'observe'; others take it as t'un 'assemble*
(T'an) or ch'un #6 'pure* (Kao).
449
Yi 'deem* ( = yi wet % as in B 71 (§ 1/3/12/4/4).
4 6 0
The pattern is * . . . yeh, fei. . .* (without final yeh, as in B 3), 'It is
that. . ., it is not that ...'(§ 1/3/5/7). For woyu 'my own possession',
cf. Lu-shih ch'un-ch'iu ch. 1/3 (Hsu l , H A / 5 ) ^ S ^ * J W f i f D * I S f c ,
#FA^c "Now my life likewise as my possession benefits me greatly'*,
Chuang-tzu ch. 22 (Kuo 739/3) ^ # ^ 5 W i f e , " I f my own
person is not my possession who possesses it?**.
C. "Fire is hot. Explained by: (?).

E . It is the fire one calls 'hot*, one does not treat the heat of the fire as
belonging to oneself. (For example, looking at the sun.)**
"Fire is not hot*' (X^$&) was one of the Second List sophisms
(Chuang-tzu ch. 33, K u o 1106/1). Hsu Shen ffW (fl. A.D. 100) mentions
it in the form "Coals are not hot" ( a s a sophism of Kung-sun Lung
(on Huai-nan-tzu ch. 14, L i u 14, 7B/-2). Commenting on the former, L u
Te-ming $i$8§B (died A.D. 627) quotes a very interesting interpretation of
the paradox by an unknown authority:
» * * J n » A # « « r · * » f * » A » ffi£*#S«ffe o №X±m ·
"It is just as when a man struck by metal or wood suffers pain; the pain
issues from the man, it is not that the metal or wood is painful. For the
horse which lives in fire, or the insects born from fire, the fire is not hot."
2/4 417
This evidence is post-Buddhist, and may reflect the influence of

Indian philosophy. But this does seem to be the kind of argument to which
the Mohist is offering the common-sense answer. If so, and if it is correct
to date the Essay on hard and white in Kung-sun Lung tzu later than A.D. 30
this is the only Chinese example known to me of a pre-Buddhist criticism
of sense-perception.
Why does the Mohist think that the mere fact that we call fire hot
settles the matter? Presumably because it is hot by definition, would not
be fire unless it were hot, irrespective of whether anyone is warmed by it.
451
B 48 20jfc№ [ I X ] · o
Deleted on the evidence of the Explanation.
C. "Knowing what he does not know. Explained by: picking out by means
of the name.
E. If you mix together what he does know and what he does not know,
and ask about them, he is obliged to say 'This I do know, this I do not
know*. To be capable both of picking out the one and disclaiming the other
is to know them both."
In Chinese as in English 'to know what one does not know* is ambi
guous: (1) 'to know a fact which one does not know', self-contradictory;
(2) 'to know what it is that one does not know*. Chuang-tzu (like Meno in
Plato's Meno 80) did not distinguish these senses and thought the claim
self-contradictory in any case:
chuang-tzuch.2(KuoM)[*Wfo±JSin&fr\ °S' [&M&9a2Ll ·
"Do you know what all things agree in thinking right?—How (literally
'from where?*) should I know it?
Do you know what you do not know?—How should I know it?
If so, does no thing know anything?—How should I know it?"
Each step in the quest for something known involves self-contradiction,
since (1) there can be no standpoint from which to judge whether what
everyone (including oneself) agrees on is right; (2) one cannot know what
one does not know; (3) if no one (including oneself) knows anything, one
cannot know that no one knows anything.
The Mohist answer to the second argument is closer to common sense

than Socrates' answer to Meno. It is based on the difference between
knowing a name and knowing an object (cf. A 80). If a questioner names
what one does not know (for example 'What is the date of the Battle of
Waterloo ?') one can know that it is the name of the object which one does
not know.
B 49 o 1££§T!I °
4 5 2
Pi tat 'necessarily require', cf. § 1/4/27.
453 p grounds for taking wuyen as 'lack some of it', cf. p. 134 above.
o r m e
C. "Being lacked does not necessarily require being had. Explained by:
what it is said of.
E. In the case of lacking some of something, it is lacked only if it is had.
As for the lack of cases of the sky falling down, they are lacking
altogether."
We touched on the concepts of yu 'what there is/what is had/something'

and wu 'what there is not/what is lacking/nothing' under A 88. Taoists
treated them as purely relative:
9
Lao-tzu 2 » S5§f@J£
"Therefore something and nothing engender each other, the difficult and
the easy bring each other about, the long and the short give shape to each
other."
The Mohist too in A 88 classes having and lacking (a rich family or a
sound intelligence) as relative, in contrast with the localising ts'un # 'be
present' and wang t 'be absent', which are absolute. Apparently he thinks
of the universe having particular objects in particular places as absolute,
the objects having properties as relative, a distinction which Taoist generali
sations about yu and wu tended to obscure. Here he points out that if
something is lacking somewhere it does not necessarily follow that the
world has it somewhere else. A collapse of the sky is something that there
never has been.
B so mm^m ° °
Text of Explanation (noting possibilities of systematic corruption) :
(S) o « « i B f t K f t ^ f f i # * » ( ^ / ^ ) ( x / ± ) ^ f t ^ ( a / f i ) » f t f t o
2/4 419
Punctuated text with restored radicals:

( « ) o m№ IB*&^ ° № 4 ^ E r B * № # i : S :
9
454 455 456
» ^№ 4 5 7
^1fi 458
^ 4 5 9 ^ 0
454
Ts'ang 'hide away', written without its radical as in Mo-tzu ch. 46
(Sun 266/8). The word is used of hibernation, sometimes with the
chih 'hibernate* of B 39: cf. Huai-nan-tzu ch. 4 (Liu 4, 10A/1) « f f l
gbiii "The bears hide away hibernating**, Kuan-tzu ch. 52 (BSS
3/4/2) StSflSiK "The insects due to hibernate do not hide away".
455
Ch'un she 'snakes coming out in the spring', as in B 39. Kao Heng, the
first to recognise the second word as she 'snake', did not try out the
same explanation in B 50.
4 5 6
Cf. E C App. 5 3tS#*£3C "Its analogy is: markings on a snake".
4 5 7
Cf. E C App. 13 K S f f i ^ f t (variant №) "Its analogy is: a dead snake".
4 5 8
For tan 'dread' written without the radical, cf. M . 10458 def. 8,
13734/def. 8.
459
Yu shih yeh 'as before' (p. 139 above).
C. "By eradicating occasions for forethought one is free from doubt.

Explained by: whether something exists or not.
E. If in doubt whether they exist, say that the snakes are hiding. If the
snakes are now dead, but the snakes which come out in the spring have got
these markings, we are none the worse for dead snakes being fearful to
us."
Hibernating animals (also mentioned in B 39) almost deserve a place

in the Mohist menagerie of stock illustrations examined in § 1/5. The
interesting point about them is that we have no evidence of their existence
except that each year they reappear in the spring. Here the possibility that
they do not exist during the winter exemplifies the pointlessness of specula
tions which make no practical or observable difference.
B 51 aafcrîE»ffiPF*fflx <«> <> 4

<>
(l)oa<^> 4 6 i
^'fiEffi»fifflxwiBf'№«2
4 6 0
as in A 83.
46i, 462 Restored on grounds of parallelism.
C. "What is about to be so is incorrigible, but this is no objection to

exerting effort. Explained by: the appropriate thing to do.
E. If it is about to be so it is necessary that it be so. If it is about to end
it is necessary that it end. That which cannot end unless you are about to
exert effort necessarily does not end unless you do exert effort."
Rejection of fatalism is one of the ten doctrines of Mo-tzu, defended

at length in Mo-tzu ch. 35-37 (cf. also N O 16). The present passage is
remarkable in touching the issue of fatalism at a much deeper level than
elsewhere in Mo-tzu. Ch'ieh 'will/about to* is defined in the corrupt A 33,
in the sub-section on saying, apparently as the word added when saying
jan ^ 'It is so* before the event. The Mohist does not hesitate to draw
conclusions which might seem to put him at the mercy of fatalists: "What
will be so is incorrigible", " I f it will be so it is necessary that it be so". But
he can do this because he has located ch'ieh in the realm of names and not
of objects; the certainty of the future is simply the logical necessity
(pi, cf. A 51) that "If it will be so it is necessary that it be so". (For the
grammatical point that a verbal phrase before pi is hypothetical, and also
that erh-hou 'only then' implies a necessary condition, cf. § 1/3/11/1.) This
does not require that in the realm of objects it is inherent in the situation
(ku 0) that it should become so. (Contrast Mo-tzu ch. 37 (Sun 179/2)
^ l f r H ^ ± "It is inherent in my destiny that I shall lose it", (Sun 179/4)
^ ^ H T & S ! "It is inherent in my destiny that I shall be poor".) The
Explanation demonstrates very neatly that precisely because of the logical
necessity the outcome of action will depend on choosing the appropriate
course.
The summing-up, like the others with the single word yi 'appropriate*
(B 43, 44), refers to the three kinds of relation (ho ) distinguished in A 83,
the exact, the appropriate and the necessary. The proposer of the fatalist
argument was no doubt using pi 'necessary' in its looser ordinary sense,
'sure, certain*: " I f it will be so it is certain to be so". The Mohist establishes
that the necessity is only in the relation between being about to be so and
being so; the relation between human effort and something being about
to be so is one of appropriateness.
B 52 ifeitMfr
463
>w&m&°
(5) o X & I K . o jgft&fe * . H e o
4 4 4 5
mM^n» <m> « ^-tb
Parallel in Lieh-tzu ch. 5 (Yang 107/3-5), with three additional characters
here enclosed in parentheses:
2/4 421
4 6 3
Fou as in A 93.
4 6 4
Restored from the Lieh-tzu parallel.
4 6 5
For the syntax, cf. A 26, 27 S S i f e . . . 3 t ? № "the one of them which
is harmful", "the one of them which is beneficial".
C. "Whether in equal conditions they snap or not. Explained by: what is

equalised.
E. If hairs support equal weights, and although the weights are light a
hair snaps, it is because the hairs are unequal. If you had equalised the
one of them which snapped, none would snap."
This and B 62 are the only problems of B 34-82 which connect with
the mechanics sections (B 25-29). Of hairs supporting equal weights some
may snap and some not. But this does not prove that equal weights exert
unequal pulls. If the weights are equal the hairs must be unequal in
strength.
The author of Lieh-tzu (c. A.D. 300) quotes the whole Explanation in a
corrupt form which can hardly be punctuated with confidence, and
mistaking the head character for part of the text. He uses it to illustrate the
story of a fisherman who could pull in the heaviest fish with the weakest
line because "he moved his hand evenly" (Kf#^f&), or, in the fisherman's
words, because "the pull of my hand is never too heavy or light" (-^Ê
M L ) . Elsewhere he embodies what he takes to be the Mohist's point in a
sophism which he ascribes to Kung-sun Lung:
Lieh-tzu ch. 4 (Yang 88/-1 cf. 88/8) T^BI^^J J » H M ^ t i i °
" ' A hair will pull a weight of a thousand chun*. . . when the shih (the
totality of other pulls and pressures) is perfectly equal."
B 53 4 6 6
»£ 4 6 7
^4rf5jS^*ffnSB# o Iftfflff*— o
4 6 8 4 6 9 4 7 0 4 7 1
(§) ° < ? > » « » ^K^SS A'^«iSAo**:tt®-tb »
* K * » A * ° ftftfc* 4 7 2
» A K K R A * o m±m&» A S 4 7 3
*
466 "Yao's being an example", not "Yao's example" (which would not
account for the yeh in the phrase, both here and in the Explanation
cf. p. 155 above). Yi ( M . 28504 def. 22 =B) 'example. Cf. Mo-tzU
ch. 35-37, where in parallel passages yi is written both with the
radical (Sun 169/-3-2, 177/-2,-l) and without it (Sun 174/-3-2).
4 6 7
"Is born in the present." Cf. E C 2 ^#£jtf№40±
Aife "The love which benefits is born from thinking. Yesterday's
thinking is not today's thinking".
4 6 8
The parallelism of the first four clauses requires a missing word to which
the third refers back as the fourth refers back to huo 'meat soup*.
469
Huo 'meat soup* (=11): the radical is written in a few sentences later.
The graph without the radical elsewhere represents hao 'crane', but
always as an example of loan-naming (§ 1/5/4).
470
Shih ( M . 34836 def. 3 = * ) , 'show to'. The pre-verbal placing of the
instrument shows that the meaning is 'show to others by means of the
name', not 'show to others the name' (§ 1/3/12/4/7).
47i, 472 Embedded 'Yyeh' clauses, "refer to a friend as being a rich merchant",
"point out this as being meat soup" (p. 155 above).
473
Shih sheng yeh 'this sound'. Shih is required to resume the exposed
element (p. 115 above) and itself requires yeh after the noun (p. 155
above). Sheng 'sound' (a word which in other texts might refer merely
to Yao's fame) is the Mohist's term for the vocal enunciation of names
(§ 1/4/25).
4 7 4
Restored from the Canon.
4 7 5
"Like 'Surely in the city gates?' and in Jack." Tsang ('Jack') is literally
'hideaway', abusive name for a runaway slave (§ 1/5/12). Cf. Lii-shih
ch'un-ch'iu ch. 24/3 (Hsii 24, 8A/3), about a minister suspected of
absconding: ^^föÄföF? "Tso I am sure is still within the gates".
Of the two directive phrases introduced by yii 'in' the first is put on
the plane of names by the particle tat, which expresses a fear or
timidly ventured opinion. (An obscure reference to tai in A 33 seems
actually to mention this effect.) The object talked about however is in
Jack, wherever he may be. Cf. E C 12 iJ$£Äfcftfö!J$ "Jack as he is in
himself resides in Jack". Cf. also B 39 "'Runaway servant': one
does not know where he is".
C. "Yao's being an example to us is engendered in the present, yet he

resides in the past, and they are different times. Explained by: the things
taken as example being two.
E. Of . . . and meat soup, the former we show to others by means of the
name, the latter we show to others by means of the object. T o mention
one's friend as a rich merchant is to show to others by means of the name,
to point out something as meat soup is to show to others by means of the
object. In the case of Yao being an example, the vocal sound is engendered
2/4 423
in the present, the object taken as example resides in the past. (Like
'Perhaps within the city gates?' and in the runaway Jack.)"
Among the ancient sages accepted by both the moralistic schools,

Yao §H§ and Shun were especially revered by the Confucians, Y u S by
the Mohists. It is remarkable that although the sages named in Expounding
the canons are, as we would expect, Y i i (EC 5) and Mo-tzu himself (EC 2),
the only sage mentioned in the Canons and Explanations is Yao (B 16, 53).
But on closer inspection we see that in both Yao represents the sage whose
example is not necessarily relevant to the present, since times have changed.
The present section clears up a confusion involved in describing Yao
as ayi 'example, model'. If Yao was an example, in the same sense that he
was a ruler or a man, it would seem that being an example is a fact about
him which obliges all later rulers to model themselves on him. The Mohist
dispels the confusion by one of his standard methods, distinguishing name
from object. Yao becomes an example only by the vocal enunciation of
words in the present referring to him as an object in the past.
The illustration takes the example of a runaway slave (cf. B 39). When
we say 'Perhaps he is in the city gates?', perhaps being still in the city is
on the plane of names, the ku $&. (the thing we are talking about as it is
in itself, cf. n. 475 above) being in the runaway himself is on the plane
of objects.
B 54 >ffn r mz^Kfo j nj o t s s m o
(So) o o r J pj o (^M(II)*№) « o47
4 7 6
The emendation to pi 'haunch' ( M . 29579 def. 2/1 = W ) is supported
by examples of pi in other Illustrations (B 45, 69) and by Han Fei tzu
ch. 34 (Ch'en 754/1), where advising a ruler to get rid of a favourite
is compared to "advising the right buttock to get rid of the left"
C. " A whelp is a dog, but 'whelp-killing is not dog-killing' is admissible.

Explained by: the identity of the object.
E. A whelp is a dog. It is admissible to call it 'killing a dog'. (Like the two
haunches.)"
" A whelp is not a dog" ( $ J ^ : £ ) appears among the sophisms in the

Second List in Chuang-tzu ch. 33 (Kuo 1106/3f). Judging by this Canon,
one argument for it was that "Whelp-killing is not dog-killing", a proposi-
tion which the Mohist concedes but regards as compatible with " A whelp
is a dog". The Explanation is very brief; but there is a detailed argument
in N O 15 that ' X is Y , but doing something to X is not doing it to Y* is
admissible for one class of proposition, for example "Robbers are people,
but killing robbers is not killing people (murder)". In each case a com
bination of words has assumed an idiomatic significance, and presumably
the same had happened with 'ôw-killing' and'ch'uan-K\\\mg. It is possible
for example that the former suggested slaughter of a domestic dog for food
and the latter the crime of killing a noble's hunting hound (cf. § 1/5/3),
even after the general distinction between the words had been forgotten.
The Exphnation seems at first sight to be saying the opposite of the
Canon, and most editors have resorted to the facile expedient of emending
k'o ('admissible') to puk'o ('inadmissible') either in the Canon (Kao) or in
the Explanation (Sun). But we can see the Mohist's point if we manufacture
an English parallel:
" A shank is a leg, but it is admissible that pulling someone's shank is
not pulling his leg (teasing him)".
A n English sophist denies that this is consistent, and concludes that a
shank is not a leg. T o show that he is wrong it is sufficient to answer:
"One can call it pulling his leg".
If tugging at a man's lower limb could not be called 'pulling his leg'
the sophist would be right; but since it can, 'pulling his leg' (teasing) is
obviously a special idiom which has nothing to do with the matter.
The crucial difference between Canon and Explanation is the wet 'call'
of the latter. On reflection we see that the Mohist is simply applying a
principle laid down in B 3, that since names when linked (It M) change
their meaning we may have to reject one of two apparently parallel state
ments (p'ien ch'i MM)> but that "inherently the thing is what we called it"
(SPfflnHAfe).
The point of the illustration ("Like a pair of haunches") may be that
the dropping of one or other of propositions about X no more affects what
X is than does the loss of a part (the loss of a haunch which leaves the milu
deer still a four-legged animal, according to our speculative interpretation
of B 45).
B 55 « ( « ) · tftfttt o
Emended from the Explanation.

2/4 425
478
Shih tien met 'cause a hall to be beautiful' is so unidiomatic that no
commentator known to me has ever taken the sentence in this sense,
but a consideration of the Mohist's special style shows it to be
acceptable (p. 163 above). Most editors who do not emend the tien
follow Chang Hui-yen in giving it the sense of 'march in the rear*
(Kao, Chang Ch'un-yi). But the parallelism of the Explanation, again
in defiance of ordinary idiom, requires it to be a noun corresponding
to wo T, me'.
479
Shih applied to persons ('employ') is sometimes intransitive, 'do as one
is employed to do' (cf. B 69). Cf. Lii-shih ch'un-ch'iu ch. 3/5 (Hsu 3,
18B/-2—19A/3) A £ # ^ « I 3 & , S f f i f t ± f f e , f S * J i M f t f f e . Jfiiffi
fl№«H«*«£. A E # « f . M*#ffD«£. tf£
ffi^ft, ^JEMft "In the case of the body and four limbs which man
possesses, that he can employ them is because when something acts
on them he is certain to know. If he does not know when something
acts on them, the body and four limbs are no longer under his control.
The same applies to ministers; if commands do not activate them they
can no longer be kept under control. Better not have them at all than
have them not under one's control".
4 8 0
Restored on grounds of parallelism with the next sentence.
4 8 1
Emended on grounds of parallelism with the preceding sentence.
C. "Beautifying a hall. Explained by: shih (cause/intend an effect).

E. To command is to intend an effect. Whether I obey or not, you intend
the effect on me; whether the hall becomes beautiful or not, you intend
the effect on it."
Before identifying the Mohist's problem, one sees that it centres on

the word shih, which is wider in scope than the English verb 'to cause' in
that it can be used of employing someone to do something, intending an
effect which may not be realised. In A 77 the Mohist distinguished two
kinds of shih, one necessarily effective (ku $C, 'a cause'), the other not
(wet IB 'telling', for example ling ^ 'commanding'). Here he points out
that the shih in question is of the second kind, allowing the possibility that
the beautified hall may not be beautiful.
The problem perhaps arose from the causative use of met 'beautiful'
allowable in certain contexts:
Kuan-tzu ch. 53 (BSS 3, 6/3) * B J § £ W № S » # » / J v f e ° 7 K M f t ·
"That enlightened kings did not beautify their palaces was not because
they preferred them meaner, that they did not listen to bells and drums
9
was not because they disliked music/ (Cf. also Han Fei tzu ch. 9 (Ch'en
152/2) K'^MW^'MMM "Rulers delight in beautifying palaces, terraces
and pools".)
If met 'beautiful* is a relative concept, as stated at the end of B 3, how
is it that we accept as an objective fact that a king beautified his palace ?
T o discuss this question the Mohist cannot detach met tien 'beautify a hall'
from its syntactic contexts, which would transform it into an adjunct/head
construction ('beautiful hall*); he therefore has to expand it to shih
tien met.
B 56 ffi£*482
» Xft 483
gift o o
Ta 'big* is used of measurement by circumference, in contrast with

measurement along one dimension (in this case depth). Cf. B 21 n. 341.
4 8 4
• It is remarkable that ch'en 'sink* should be written with different
graphs in adjoining sentences, in all the Ming editions consulted (Tao,
T*ang, Mao); the assimilation of the graphs in modern editions
obscures what may be a significant distinction. The other examples in
Mo-tzu raise a suspicion that the graphs distinguish falling-tone
(transitive) and level-tone (intransitive) readings: ^££ffiftl£fcf f H
:
"We have submerged and soaked Chou of Y i n in drunkenness"

(Sun 99/7), but ftifcfe©^ " A t home they were submerged in wine
and music", SSfe^JfcftJi-iJl " H s i Shin's drowning was because of
her beauty" (Sun 176/6, 3/8 cf. also 312/8 S&BH, an obscurely named
engine of war).
The Taoist Patrology and the cognate W u M S have an idiosyncratic
reading, The reading of the other editions (Lu, Mao, Mien,
Horyaku) is confirmed by the Canon; there is little danger that it is a
correction after collation with the Canon, which would have been
unidentifiable until late Ch'ing scholarship began to link the Canons
and Explanations.
C. "The extension of the bramble is because its submergence is shallow.

Explained by: what we use.
E. In the case of the thing we use to submerge the bramble, that its
submergence is shallow is not because the bramble is shallow."
2/4 427
On the present interpretation B 56, like B 55, deals with a problem of

causation. According to common sense the spread of the bramble {ching,
regarded as the lightest of woods, M . 30940 def. 3) over the surface of the
water is caused by its shallow submergence. But there is a problem here
which a sophist might well point out; the cause precedes the effect (A 1),
but it might be argued that if the bramble is pressed deep underwater and
then released, the lower parts cannot rise until the upper parts have spread
out to make room for them. The Mohist answers that the depth of the
stick with which we push the bramble is obviously not caused by the depth
(and therefore on this hypothesis the spread) of the bramble.
B 56 is almost unique in that it has no apparent relation with other
sections and does not contain a single important word found elsewhere in
the corpus. It is of interest as an especially vivid reminder of the extent to
which the understanding of Classical Chinese depends on context. M y own
interpretation has the advantage over most that it involves no emendation
to the body of the Canon or Explanation (although I do transfer the illus
trative gloss at the end to B 57). But one other proposed interpretation
avoids emendations, and also tries to cover the illustration; it is supported
by W u Yii-chiang, L i u Ts*un-yan, and also Kao Heng (who however
cannot resist gilding the lily by emending chii ^ to yu W 'possession'):
C. "The greatness of Ching is the weakness of Shen which belongs
to it. Explained by: being the tool.
E. If Shen is the tool of Ching, then Shen being weak is not Ching
being weak. It is as though it were exchangeable for one fifth of it
Kao, W u and L i u all offer evidence for the political relations between
Shen and Ching (Ch*u /f), and for ch'ien 'shallow* used of the weakness
of a state.
4
B 57 K ( f f i ) * 1 & 8 « f § ] » * 8 7 » fe KS(^) 5fe^ft
488 # 489
o o
( ^ £ ± - ) < > (JX) o tt£HHb '
4 9 4 9 1
» 5fe@«ffi
ft 492
oâgJJUc498 » ^ ^ - t f e # ^ 4 9 4 o
Corrected from the Explanation.

T'uan, 'round* (=BB B 24), three-dimensional in contrast with the
two-dimensional yuan H , H . Both here and in B 62 it is corrupted
in the Taoist Patrology (cf. p. 75 above) to W in the Canons and Î9 in
the Explanations.
«8 W u ( H B ) <i what?', as in B 42 (§ 1/3/7). For the syntax of * 8 f i l f l l ,

= n
5teftHiL "wherein one deems is known 'a priori' ", cf. B 41 ^P№S*!I
Ilife "not know which he refers to". For the wet, cf. p. 118 above.
4 8 9
Corrected from the Explanation ; the confusion (which implies that zvu
was written with the graph 3c) is found again in B 73. Previous editors
follow Sun in emending the Explanation from the Canon; but this
overlooks the contrast of hsien chih 'foreknow' with the wet k'o chih
'not foreknowable' of the next Canon (cf. § 1/5/8).
4 9 0
"Like replacing the ones in five". Cf. B 12 f&EffnE— "Thefingersare
five and the five are one", B 59 M "Five has one in it". The
wordy/ 'substitute, exchange' (cf. § 1/4/38) suggests that the illustra
tion belongs to this rather than the preceding section.
4 9 1
Cf. Chuang-tzu ch. 33 (Kuo 1111/1) %h±M "(The
sophists) replaced men's ideas", Hsiin-tzu ch. 21 (Liang 296/14)
"Ittrfnffi^r^ "The mind cannot be made by force to replace its
ideas".
4 9 2
For yi hsiang ( = 38), 'idea and image' (the two are presented parallel
in B 38), cf. § 1/4/37.
4 9 3
Ch'iu ( = « , $E, $C M . 31333 def. 1/2, 3), 'catalpa'. Cf. N O 3
il/fc "having the idea of the wood of this pillar".
494
Yang-yang ( M . 17363/143) 'vast, flooding, overwhelming' is a common
binome, but I have not noticed other cases of yang-jan. But there are
cases of jan after yang-yang in the sense (in which it is also written
with the graph SI, M . 17363 def. 10) of 'not knowing where to
go/having nothing to depend on' (M. 17363/143/7, 44144/167). Cf.
Lu-shih ch'un-ch'iu ch. 23/1 (Hsu 23, 2B/4) 5 4 J l f i ± ? $ } £ * & K * i f f f
^P£n§fj8 "Now I see the people fleeing east yang-yang-ly and not
knowing where to settle". Song 44, Mao ^ commentary:
^$P@f!S* "Yang-yang-ly anxious, not knowing anythingfixed".T h e
Mohist would probably write the word as yang-jan since he seems to
avoid repeating syllables (p. 165 above).
C. "When we deem a pillar round, the place wherein we deem is known

a priori. Explained by: the idea.
E. In the case of the pillar's roundness, when we see it, the place for it
within the idea remains unchanged, because we know a priori the idea and
image. If the pillar were lighter than its catalpa-wood, the place [for
lightness] within the idea would be undecided. (Like changing the ones
in five.)"
2/4 429
B 57, 58 both discuss problems raised by yi 'ideas, mental pictures'

(§ 1/4/37) and by hsien chih, 'knowing a priori' (§ 1/4/13). In B 57 the
example is a pillar, the stock example of something we recognise by a mental
picture (§ 1/5/9). Our idea of a thing can contain only what we know about
it a priori, for example the hardness of a stone but not its colour (B 37, 38).
Is our idea then an attenuated mental picture divested of inessentials, in
which case the essential pillar will lack the weight of the wood or stone of
which it is made ? The Mohist's answer is that the scheme of the idea has
places to be filled by roundness or squareness, catalpa-wood or other
materials, and the substitution of one for another does not affect their
places in the scheme. Similarly, as the illustration notes, the replacement
of one thing by another does not affect a set of five; indeed we can substitute
at will the five colours of A 93, the five fingers of B 12, the five elements of
B 43, the five roads of B 46, without ceasing to have the idea of five things.
The scheme within which one deems would be what the Interpretation
of Lao-tzu in Han Fei tzu calls the li 'organisation, structure':
(Ch'en 365/1) mm » JWfcfcfc* °
"The li is the pattern of the completed thing."
(Ch'en 369/1) »# B M f t K « № £ # f t o
" A l l li is an apportioning of squareness or roundness, shortness or length,
coarseness or fineness, hardness or softness."
We shall inspect the idea of the pillar more closely under B 58.
4 9 5 4 9 6 4 9 7
B58te ±;«*^ »tftffirÎfflJ » r S i f f J *
f§ (*)(it) *m &mi&m »r *im jft°fcmmmm· m&№
498 499 500
1
mmmm™ ' rm(#)*frJ 5 0 2
ft°
4 9 5
Sun and most other editors start the Canon two places later, at the cost
of inflating the summing-up of B 16 to the unintelligible phrase tk&
But once it is recognised that the presence of a head character
at first or second place in the Explanation is a firm rule (§ 1/2/2/5/4),
there can be no doubt that W u Ju-lun and W u Yu-chiang are right in
starting the Canon at ch'ui 'hammer'. The graph is written in its
standard form ($£) throughout the Explanation except in the Horyaku
edition, which in the last two cases reads the ft of the Canon (a variant
overlooked in Wu Yu-chiang's collation). For examples of this inter
change, cf. Chu Ch'i-feng 0085 bis, 0251, 0072.
496
Wei k'o chih 'not knowable a priori' (§ 1/4/13).
497
Kuo wu, written 5SE(=ffi) in A 98, 'exceed the matching' (one image
containing more than the other). Cf. § 1/4/34.
498 This graph (also written S, M . 3174) is in practice often treated as

equivalent to Wt and from P i Yiian's onward all editions I have
y
consulted have in fact printed the latter without comment. In any

case there is general agreement that Sun is right in identifying the
word as tuan Wt $k y y WL (the last, traditionally read hsia, appears
to be only a miswritten form, M . 24341), 'hammering block*. Cf. Han
Fei tzii ch. 35 (Ch*en 783/-4) "Hammer and
block are the means of levelling the uneven".
4 9 9
Corrected from the other examples of ch'ui 'hammer*.
5 0 0
For shihyu 'employ on*, cf. Han Fei tzu ch. 20 (Ch*en 359/-2) J i ^ V
MW:WMMit "The ruler does not employ his horses on fighting
battles and hunting down the fleeing**.
5 0 1
For the pattern 'Xyii Y t'ung', 'It is the same with X as it is with Y*,
cf. § 1/4/36.
5 0 2
C. "The idea of a hammer is not knowable a priori. Explained by:

usability, including more than what matches.
E . Hammer and block both being employed on the shoes is their 'usability*.
That to complete [the idea of] decorating shoes you include more than the
hammer, just as to complete [the idea of] the hammer you include more
than the decorating of shoes, is 'including more than what matches*.**
What exactly does the Mohist mean by saying that we know 'a priori*
the idea or image of a pillar (B 57) but not of a hammer ? Presumably that
the visualisable shape is implicit in the definition of the former but not of
the latter. T o understand him we must look at early definitions of these
two things; in both cases they are in terms of function. We have a definition
of the hammer only a little later than the Canons, from the remains of the
the lexicon Ts'ang chieh p'ien #SBJB by L i Ssu ^JW (died 208 B . C . ) :
(TSCC) 41/-3 [" S U » ffl№tft o
" A 'hammer* is what one uses to beat things." (Cf. also Shuo wen ch. 6A
(emended Tuan 266A/9f) f" 86 J » " A 'hammer* is what
one hits with**.)
A hammer then is defined by its 'usability*, is indeterminate in shape,
and cannot be visualised in use without imagining in addition the thing
being hammered ("including more than what matches**).
The Shuo wen defines ying 'pillar* by a synonym, chu ft (Shuo wen
ch. 6A, Tuan 256A/-1). Chu is the word used verbally in B 29 of supporting
from below, and this is the aspect brought out in the Shih ming (c. A.D. 200):
2/4 431
Shih ming (TSCC) 88/7-9 r tt J » ° > ^«§f^* °

II(KB r m J ° W M l » » fB»ff±*-tli o
"Ying (pillar) is it stands by itself t'ing-t'ing-ly (sticking up
vertically), without any side support. In C h ' i and L u they call a pillar
ch'ing. Ch'ing is 'adequate to'; it occupies a place alone standing by itself,
and is able to bear adequately the load of a weight above."
The fanciful etymologies are characteristic of Han thought, but the
substance of the definition may very well descend from the time when the
pillar was a stock example in disputation (mentioned also by Chuang-tzu
and Hsiin-tzu, cf. § 1/5/9); it connects very neatly with the descriptions of
side support and vertical weight-bearing in B 28, 29. Now the function of
a pillar, unlike that of a hammer, imposes a particular shape as the most
efficient, vertically elongated, although as noticed in B 57 it is not necessarily
round or square. Moreover it can be visualised as performing its function
in isolation; by definition it needs the support of nothing external except
the ground it stands on, and we do not have to imagine a load above it
because the upper part is itself a load supported by the lower.
,
If we have guessed right, the idea of the pillar is known 'a priori in
the same way that the circle is (A 93). The definition of the circle is built
up from the abstract concepts jo 3a 'like* and cheng IE 'straight' (p. 57 above);
the visualisable figure is not described, yet is known 'a priori' from the
definition. In the case of the pillar we are at the disadvantage that the
Mohist defined only abstract concepts and assumed the definitions of ox,
horse and pillar as common knowledge. However, we can see that a defini
tion in terms of the Mohist statics (B 29), which is substantially that of the
Shih ming, would impose 'a priori* the geometrical figure of the pillar.
B 59 — ^ f t — f f D ^ f e E » ( £ ) · # « » o
(-) o £ 3 M R . - W i L B » + r g o
503
Chu in Mo-tzu is always corrupt (Sun 87/2, 3 *ft 153/-2 *1£), so that
Sun's emendation to wei 'position' is very probable. Cf. B 23 (44)* ?iL
C. "One is less than two but more than five. Explained by: establishing
the level.
E. Five has one in it; one has five in it, twelve in it."
That at different levels of classification the units grouped together

(ho fe) count as one is affirmed in B 11,12. This Canon forbids the counting
of units not on the same level (wei 'seat, station'), units which, if we
have rightly emended A 83, "stand together*' ( * f e £ ) . Five has one in it

(yen M) but one has five in it (cf. B 12, "The fingers are five but the five
are one", according to whether you are counting the fingers or the parts of
the body) and twelve in it (twelve months in one year ?).
Such sophisms as "Jack has three ears" ascribed to Kung-sun Lung
(Lii-shih ch'un-ch'iu ch. 18/5, Hsu 842/4-843/6), and " A chicken has three
feet" (ISHJE) in the Second List in Chuang-tzu ch. 33 (Kuo 1105/-1),
were probably based on the confusing of levels:
Chicken
Foot Wing
A
Foot Foot
The forger of Kung-sun Lung tzu (between A.D. 300 and 600) under
stood the 'chicken* sophism in this way:
ch. 4 (Ch*en 136) mmm o mm&—» »îfn-*H °
"Chickens have wings. You refer to chicken's feet as one, but count the
feet as two. They are two and one, therefore three."
The decimal place system (used in reckoning with counting rods, cf.
Needham 3/5-17, 83) would be a vivid illustration of the Mohist's point.
But attempts to treat this section as a direct description of the decimal
system break down in the last sentence, which Sun emends ( + , — < £ >M
" . . . and ten, two fives in it"), while others are content with forcing the
syntax (T'an, W u : " . . . and ten, two in it").
5 0 4 5 0 5
B 60 2 N M I f f ' mvn ° f&ffi^ °
504
Cho #f, ff? 'hoe, mattock', verbally also in the more general
sense of 'chop up'.
505
Tuan 'starting-point' (not 'point', A 61).
506
Ch'ien when verbal is 'move forward to the next position' (cf. A 42).
507
Wu 'in no case', the negation covering all stages of the progress
(p. 131 above).
5 0 8
The choice of graph for wu and the absence of yeh after fei pan shows
that the meaning is "There is nothing with it which is not half", not
"What has nothing with it is not half" (pp. 125, 133 above).
C. " I f you hoe only a half at a time, you do not make progress. Explained
by: the starting-point.
2/4 433
E. T o go forward hoeing a half is to take it from what is ahead. Whenever

you move a stage ahead the centre does not make a half, you are as though
at the starting-point.
If you take half of what is both ahead and behind, your starting-point is
the centre. Whatever you hoe is necessarily half, there is always a remainder
which is the other half, and it cannot be hoed."
As Sun already recognised, this expounds a form of one of the para

doxes in the Second List:
Chuang-tzu ch. 33 (Kuo 1106/4) — R £ S » H # 3 M * > MWF>M °
" A stick one foot long, if you take away a half every day, will not be
exhausted for a myriad generations."
The Mohist version refers to hoeing a field, unless we give cho the
general sense of 'chopping u p \ T o hoe half, the man either
(1) hoes from the starting-point to the centre, which at each stage
becomes a new starting-point from which he hoes half the remainder, so
that he never reaches the end; or
(2) hoes from the centre to the end, and then from the centre of the
remainder to the end, so that he never gets back to the starting-point.
In either case he does not move (tung IS), does not make progress,
since he never gets away from the starting-point in the former case, from
the centre in the latter.
Both the sophist and the Mohist use the word pan 'half merely for
convenience of description; the argument does not of course depend on the
two parts at each successive cut being equal in length.
B 6i Ri«nb»w^rfn^î* o ·
( 5 ) · <...»RI o »!i]m *& · 3^r«H&
5io
· *
A lacuna may be suspected (due to the scribe's eye slipping from one
k'o to the next), because (1) The last sentence seems to imply a missing
proposition about something of limited duration, parallel with the
intact proposition about something of limitless duration; (2) K'o wu
yeh cannot be three head characters, since the number never exceeds
two (§ 1/2/2/5/5).
Ch'ang (=Hr, the graph used in the Canon), as in B 38.
C. "It is possible for there not to be; but once there is, it is impossible
for it to be dismissed. Explained by: it having been so.
434 The 'Canons* and 'Explanations'
E. The fact that . . ., it is possible for there not to be. The fact that if
full provision was made full provision has been made it is impossible
for there not to be. The duration is both limited and limitless."
Except for the necessary, which is the unending (^PB A 51), the
fitting of name to object 'stays' ( i t A 50) only as long as there is the object.
Previously it was 'not yet so' cf. B 16), afterwards it is ended (yi E ) ,
and we are in danger of 'supposing what is ended to be so' ( B S $ * B 10).
However, even at a time when there is no object, we can refer forward to
it by ch'ieh J=L 'about to* or back to it by yi B 'already' (A 33). This yi,
although it is yi 'to end' used adverbially, does not abolish the event from
history. Its duration was limited, but the duration of its 'having been so'
(H?^) is limitless.
This section connects with a sophism of the Second List (Chuang-tzu
ch. 33, Kuo 1106/4) " A n orphan colt has never had a mother" ( W S ^ H f
W S ) , on which the lost Six Dynasties commentator L i ^ noted:
(Kuo 1111 n. 22) » "B T № J flfl&S ° \ ft J > №\ T « J
-tfa o S t r i f e s o
"The colt at birth has a mother, but when you say 'orphan' it has no mother.
When the appellation 'orphan' comes into effect the name 'mother' is
dismissed. The mother has been the mother of the colt."
Cf. Lieh-tzu ch. 4 (Yang 89/1 cf. 88/-4) T M S * # W © J » ·
" ' A n orphan calf has never had a mother', because it would not be an
orphan calf." (Sophism ascribed to Kung-sun Lung.)
The Mohist does not seem to be directly attacking the sophism, which
itself contains the temporal particle ch'ang. The sophist was not denying
in general that past events ever happened; his point was apparently that
the orphan colt has never had a mother during the time it has been an
orphan colt. One might make a similar sophism out of the English sentence
" N o Pope has ever been a young man".
B 62 T f i f f i O T ( » ) * % 5 1 1
> o
(IE) o ( A ) * A 5 1 2
*i§f Jftrfn^P^R 513
» °
5 1 1
This correction by Sun is widely accepted (Liang, Chang Ch'un-yi,
Liu). Other suggestions (Kao, \$ 'pull sideways'; Wu, IS 'firm')
overlook the frequency of this confusion, due to yao 'quiver' having a
li-shu variant graph J t ( M . 12583). (Cf. K u Ai-chi 2/13B). Several
examples of the confusion are quoted in Tai Wang's WM. collation for
2/4 435
Kuan-tzu ch. 6 (BSS 4/26/-5) ( * ) * ^ W $ ; S № * "Wishing to

immobilise the tip while shaking the fishing-rod".
512
Wan 'ball* (tabooed under the Sung as a homophone of Huan S ,
personal name of Ch'in-tsung ife^) is variously corrupted to %
(Tao, Wu), / L (Lu, T'ang-ts'e) /1 (Mao, Mien, Horyaku). W u
Yu-chiang takes the second of these (fan) as the original graph for
wan in its Shuo-wen form, although in theory at least the latter is a
distinguishable character ( M . 96 H»).
513
Chung hsiian 'coincide with the plumbline*, cf. Chou li (ch. 39) SPPY
39, 15B/5 ft^t+K, ft*** "The upright coincides with the
plumbline, the horizontal with the waterlevel".
C. "Upright, but it cannot be made to waver. Explained by: being

spherical.
E. A ball whatever its position coincides with the plumbline, because it
is spherical."
A line from the highest to the lowest point on a ball will always be
perpendicular to the ground, "because it is spherical". The Mohist can
see that this follows from the definition of a circle (A 58, "having the same
lengths from one centre"), but evidently he does not appreciate the value
of trying to fill in the intermediate steps of the demonstration. This is
interesting because it suggests that although he certainly has the idea of
the geometrically demonstrable (A 80, 98) he probably did not develop
true proofs in the manner of Euclid (cf. p. 57 above).
B 63 TMMfc»*
(ft) · <n >™%ft%<&ft&mm -S[tt] 5 i ?

s » « i · 5fe«» IK
CA'w IS 'mark off from*, written with the characteristically Mohist 'man*
radical (§ 1/2/1/2/7). Since it implies something from which X is
marked off, p'ien chii should mean "refer to without referring to that
from which it is marked off".
The two graphs are similarly confused in A 78.
This seems to be one of the cases where the identical graph following
the head character has dropped out (§ 1/2/2/5/4).
436 The 'Canons' and 'Expfonations'
6 1 7
Delete on grounds of parallelism.
C. "Going forward in space there is no getting nearer. Explained by:

spreading out.
E . That which being demarcated cannot be referred to without referring
to that from which it is demarcated is 'space*. Travellers as they go forward
spread out, at first nearer together, afterwards further apart.
C. Travelling over a distance implies duration. Explained by: at first and
afterwards.
E . The travellers are necessarily nearer at first and only afterwards farther
apart. Being near or far is distance, being before or after is duration. For
people to travel over a distance necessarily requires duration."
B 63, 64, seem to imply as their targets two sophisms about space:
(1) The slower of two travellers, falling farther and farther behind
the other, will eventually be getting nearer, because when the faster reaches
the ultimate limit of Space he will begin to catch up.
Answer: We cannot delimit a space without postulating space beyond
the limit, from which we are dividing it off. If two travellers advance at
different speeds, the one in the rear falls farther and farther behind.
(2) A finite distance can be travelled in a moment.
Answer: This would imply being both far and near at the same
moment.
B 64 is unlikely to be an answer to H u i ShhVs paradox "I go to Yueh
today but came yesterday" ( ^ 0 ffiiiif?iS Chuang-tzu ch. 33, Kuo
1102/-2). The Mohist's target is the idea of starting and finishing a journey
simultaneously, not of finishing it before you start. H u i Shin's idea was
perhaps that a paradox arises from combining the concepts of point and
moment; if I crossed the border at midnight, I can equally well be said
to have left the other state today or to have arrived in Yueh yesterday.
B 63 adds to our knowledge of the Mohist conception of space. The
question whether it is finite or infinite seems to be left open in B 73
(perhaps only because the first debater does not want his case against
universal love to depend on its resolution). It is infinite if, however far we
advance, there is always room for a further measurement (A 42). Any area
which we mark off by a boundary cannot be 'referred to without referring
to that from which it is demarcated'. Consequently, when we advance in
space, there is always further space ahead.
2/4 437
5 2 2
mm» ( » # t i i ) » o
518
Hsiang yii 'being with each other', being together (ho 'p) in a class
(cf. B 11, note).
519
520
The use of mao is sufficiently far from ordinary usage to seem to require
emendation. Most follow Wang Yin-chih 3:5I;£. in emending to lei sS
in both places. But T'an accepts the text as it stands, and an examina
tion of mao 'figure, characteristics' as a technical term (§ 1 /4/20) shows
him to be right.
521
522
The yu phrase is parenthetic as in A 32, perhaps a gloss (§ 1/2/2/6/4).
523
Cf. A 71 r#Ufk^> "Being 'so' is the characteristics
being like the standard", B 22 3 t f l t t $ t "It is so of all its parts".
C. "The belonging together of things with one standard is complete, for

example the tallying of squares. Explained by: the square.
E. If the characteristics of the square are complete, that both things have
the standard but are different, one of them wood and the other stone, is
not inconsistent with their tallying in being square. Anything of which they
complete the characteristics (as in being square), is so of both the things."
524
B 66 ffi^ ^*m*nJI »t&£# °
5 2 e
e> r ( ^ ) * ^ ffi^*j >ft*<m>™ft>mmft>&m*m
fe528 o %m*#ft » mm ft» &&&mz*m&»âs-tn»
524
K'uang chii 'referring arbitrarily' (in offering as criteria). The term
reappears in B 76.
525
Sui (=88) 'although' (§ 1/2/1/2/1).
526
Both T'an and Kao note that the seal forms of the graphs are easily
confused (Hf, ^r).
527
Restored from the parallel in the next sentence.
528
The radical emendation proposed in G (2) 162 n. 2 (ft^f <^№rfn >M
PJf&) doubtfully supported by an elusive parallel in Kung-sun Lung
tzfi ch. 4, Ch'en 128/—1) now seems to me unnecessary.
C. "By referring arbitrarily one cannot know differences. Explained by:

what they have.
E. Although oxen are different from horses, it is inadmissible to use oxen
having incisors and horses having tails as proof that oxen are not horses;
these are things which they both have, not things which one has while the
other has none. If you say 'Oxen and horses are not of a kind', and appeal
to oxen having horns while horses have not, in this the two kinds of thing
are dissimilar; but if, when you refer to oxen having horns and horses not,
you treat this as the dissimilarity between the kinds, this is referring
arbitrarily, as in the case of oxen having incisors and horses having tails."
To be 'of a kind* (lei) is to 'have respects in which they are the same*
(WJ^fi^l, A 86). Lei are of varying generality (B 2 'animal', 'living thing').
However, 'horse' is the typical example of a name for a lei (A 78), and the
present Explanation implies that 'Oxen are different from horses' and 'Oxen
and horses are not of a kind' are both true, although wrongly justified. The
shift to the second formula shows merely that the debater has now chosen
to specify which of the four senses of yi 'different' he intends, namely
pu lei 'not of a kind' (A 87).
The defining characteristics of ox and horse must have been common
knowledge, and the Mohist never specifies them. But Taoists who give
examples of characteristics derived from Heaven and not from man seem
to be taking them from the 'ox and horse' disputation:
Chuang-tzu ch. 17 (Kuo 590/-1) ^MSHJE.» &MX ° fH( = '
"That oxen and horses have four feet is ascribed to Heaven, haltering
horses' heads and piercing oxen's noses is ascribed to man."
Huai-nan-tzii ch. 1 (Liu 1/12A/6) * ^ J E ^ i f o * # » H#Mlif6£JE* ' %
o mmzu o ,A *o
"Therefore that oxen have cloven hooves and bear horns, and horses have
spreading manes and undivided feet, is from Heaven; haltering horses'
mouths and piercing oxen's noses is from man."
"Oxen and horses have four feet" appears in B 12 (cf. also N O 18
"Horses have four feet"), and the Huai-nan-tzu differentiation may well be
the "dissimilarity between the kinds" mentioned here. If so, the Mohist
accepts the conjunction of horns with cloven hooves as distinguishing oxen
from horses, but not the horns alone. There can be hornless oxen, perhaps
also horned horses. Cf. Lu-shih ch'un-ch'iu ch. 6/5 (Hsu 6/18B) J H W ^ ^ i
"There are cases of horses growing horns" (in a list of prodigies).
2/4 439
529 530
B 67 * * J ^S;> «I^[pJ ' o
* J * "I» r *frJS§*Ni J * "J » I O * « r S C f "I» ffiB T ^ J H ^ f t * *j J
5 2 9
The proper place of division between B 66 and 67 (both in the Canons
and in the Explanations) was for a long time controversial. The forger
of Kung-sun Lung tzu modelled his confused argument about oxen and
sheep (ch. 4, Ch'en 126-139) on the Explanations of both, which he
must have read as continuous. Sun and Liang divided the Canons
two places too late, giving B 66 the summing-up "explained by:
something inadmissible" (W^Î), and T'an and Kao have continued
this tradition. But B 66 is about 'having* (W) horns and tails, not about
'something inadmissible*, and the present Canon requires 'pu k'o X*
(deny X ) to balance 'k'o chih' ('admit it*). For the causative use of pu
k'o, cf. p. 159 above. The true places of the divisions were discovered
by Chang Ch'un-yi (followed by L i u and, for the Canons, by Wu).
530 p j f
o r t l <x yji y t'ung', 'It is the same (there are the same
l c o r m u a
grounds) with X as with Y ' , cf. § 1/4/36.

5 3 1
The final confirmation of Chang Ch'un-yi's divisions are that they are the
only ones which provide a head character. Chang's claim that pu is the
head character did not impress his successors; but once it is recognised
that no head characters have been lost and that they are frequently dis
placed to the second position (§ 1/2/2/5/2), it can be seen that he was
right.
6 3 2
"Oxen and horses are non-oxen and non-horses** (not " . . . are not oxen
and are not horses", which would require a final yeh *tfa, § 1/3/5/8).
C. "There are the same grounds for denying that oxen-and-horses are
not oxen as for admitting it. Explained by: the total of the two.
E. If it is admissible that since some are not oxen they are not oxen, then,
since some are oxen though some are not, it is equally admissible that they
are oxen. Therefore, if it is inadmissible to say either 'Oxen-and-horses
are not oxen* or 'Oxen-and-horses are oxen*, then, it being admissible of
some but not the others, it is likewise inadmissible to say ' "Oxen-and-
horses are oxen" is inadmissible'.
Moreover, if neither oxen nor horses are two, but oxen and horses
are two, then, without the oxen being the non-oxen or the horses being
the non-horses, there is no difficulty about 'Oxen and horses are non-oxen
and non-horses'."
In this section (which connects with B 12) 'Oxen-and-horses are not

oxen* looks at first sight like some mysterious sophistry. A corrupt list of
sophisms in Hsiin-tzu's Right use of names does in fact contain an item
which may be read as ^fclf ^ H i f e (Liang 316/6 + 12), although one should
probably punctuate one place later ('A horse is not a horse'). But is not the
proposition, odd as it is, too close to common sense for a sophism ? Given a
choice between 'Oxen-and-horses are oxen' and 'Oxen-and-horses are not
oxen', one's first reaction, before recognising them as false alternatives,
would be to plump for the latter. On closer inspection it seems more likely
that the Mohist is exposing an illicit use of ' X + Y is not Y ' in disputation;
he points out that neither alternative is acceptable, and that the proper
analysis is ' X + Y is non-X + non-Y'. A n example of misuse of the
formula would be one of Kung-sun Lung's arguments for ' A white horse
is not a horse', that " A white horse is a horse and something white"
(ch'en 6 i &m%mm&&).
The context of the argument may have been the Second List sophism
" A yellow horse and a black ox are three" (cf. B 12, note).
533 534 535

B 68 ( * ) * » ilt(«)*tfe
ilfciilftj«:^ · i&£H ·
(t&) · IE*#»jfc«]lfc-Br o fft«ih»&a»jitJH:ihjfeJH;»«JHSP"^ · & & * 3 6
533, 534 Corrected from the Explanation. Cf. the systematic corruption
(fit)* IS (§1/2/1/3/5). For the syntax of the causative constructions in
this section, cf. § 1/3/13.
5 3 5
' X yii Y t'ung', ' X is required by Y ' (§ 1/4/36).
5 3 6
Both here and in the similar construction in B 82 ch'ieh is presumably
the temporal particle, since no other usage is attested (apart from
ch'ieh 'moreover', which precedes the subject). There are no cases of
the adversitive ch'ieh 'even' which follows the subject in such texts as
Mencius. The effect of inserting a pre-verbal particle into this type of
clause will be to transform the construction from 'verb—object' to
'subject—passive verb' (§ 1/3/13).
537 We follow the reading of the W u manuscript. The Taoist Patrology and
all other editions repeat the tz'u.
C. " Y o u cannot use 'that' for this without using both 'that' for this and
'this' for that. Explained by: their being different.
E . It is admissible for the man who uses names rightly to use 'that' for
this and 'this' for that. As long as his use of 'that' for that stays confined to
2/4 441
that, and his use of 'this* for this stays confined to this, it is inadmissible
to use 'that* for this. When 'this' is about to be used for that, it is likewise
admissible to use 'that* for this. If 'that* and 'this* stay confined to that
and this, and accepting this condition you use 'that* for this, then 'this'
is likewise about to be used for that."
The problem of the relativity of naming is discussed here and in B 72,

82. For the Mohist, as for Chuang-tzu, the crucial example is the varying
use of the demonstratives according to standpoint:
Chuang-tzu ch. 2 (Kuo 66/3-5) » ° itft— 9
"This is likewise That, That is likewise This. There they use 'is-this' and
'is-not' from one point of view, here we use 'is-this* and 'is-not* from another
point of view. Are there really a That and a This ? Or really no That and
This ? Where neither That nor This finds its opposite is called the axis of
the Tao."
The whole art of disputation is discredited if, as Chuang-tzu maintains,
the distinctions marked by 'that* and 'this* are unreal. The Mohist answer
is that the relativity of the demonstratives has no bearing on the reality of
the distinctions. X and Y remain different whichever way we apply the
demonstratives (or even, if the Mohist agrees with Chuang-tzu that all
naming is relative, whichever one we choose to name 'horse*). You can
call 'that* what I call 'this*, but only if you also call 'this' what I call 'that'.
B 69 P i W H i S 9ffi&b°
5 3 8
(Pi) o <D| > 5 3 9 ^ , ^ § f J t § 540

o ;=gfli)541 o , $-tfe542 » ^ %
E · Piffn^W » A ^ * i b o « ^ j f n * * # £ o SljTJPfCPg » * * * * · m
9 9
$ A « A # nU№&& * ftAÂS <# ·
638
Ch'uan ( M . 10691 def. 6 = * ) , 'threadtogether*. Sincech'uan^KWAN
is both graphically and phonologically the same word as kfl«/*KWAN
H (Karlgren (1957) No. 159), t'ung ch'uan may be taken as synonymous
with —*M ( M . 1/431) 'strung on one thread*, implying that X and Y
require each other as the formula *Xyu Y t'ung' (§ 1/4/36) implies that
X is required by Y .
5 3 9
For the dropping of the identical word after a head character, cf. §
1/2/2/5/4.
8 4 0
Cf. Huai-nan-tzO, ch. 9 (Liu 9, 12A/7) %-HBT'XM , i f i i l № 0 f * £ - f e
"Because their bodies have limitations and their abilities have things
beyond their scope'' (of the deaf and the dumb, who do some jobs
but not others).
The gloss must have entered 8 places too early, if as is probable the
second word is the pi #$ 'buttocks' also corrupted in B 54, a standard
example of two inseparable things.
For intransitive shih 'do as you are told', cf. B 55 n. 479.
5 4 5
The parallelism gives no place for hung in the first position, requires
it in the second (as the opposite of tsui 'crime', as in A 35-38).
C. "Taking the lead and following others' leads involve each other.
Explained by: the credit.
E. If you always confine yourself to taking the lead, there will be nothing
which you comprehend from all sides; if you always confine yourself to
following others' leads, which is doing what you are told, you will have no
possibility of choice. (Like the buttocks.)
To take the lead and not follow leads is refusal to learn; if when your
knowledge is less you refuse to learn, it will necessarily be little. To follow
leads and not take the lead is refusal to teach: if when your knowledge is
more you refuse to teach, it will merely be lost.
If you make someone steal someone else's coat, the blame is heavier in
your case than in his; but if you make someone give someone else wine,
the credit is greater in your case than in his."
Ch'ang ho 'lead and accompaniment' (in singing) is a common

metaphor for initiating and following another's initiative. It belongs to the
vocabulary of Taoist and Taoist-influenced thinkers, who recommend the
latter:
Chuang-tzu ch. 5 (Kuo 206/2f) * H t W M ^ # ' V&fiJAifuB °
"He has never been heard to take the lead, all he does is sing the accom
paniment".
Lu-shih ch'un-ch'iu ch. 18/1 (Hsu 18, 1A/7) ' A5teft« °
"Others take the lead, I accompany. Others go ahead, I follow."
These Taoist terms do not appear elsewhere in Mo-tzu, but the idea
is attacked in its Confucian form "The gentleman complies and does not
initiate" (Mo-tzu ch. 39, Sun 186/11 S*«ff5"*№, cf. A 15).
The Mohist thinks one cannot do without either (like the two but
tocks). The last paragraph attacks the argument that it is better to follow
than to lead because you escape blame. The Mohist answers that if you
2/4 443
expose yourself to the risk of greater blame by taking the initiative you also
have the opportunity of winning greater credit. Judging by the summing-up
of the Canon ('the credit') this paragraph is the heart of the Explanation.
B 70 \m^%\mvn\»mm%\±- ° °
3gSKfe J * 4 7
o Agffi»£0?#t& o »548 f S^549^^K|550J^ o
5 4 6
This restoration (Liang, Chang Ch'un-yi, Kao, Liu) is necessary because
it is implied throughout that the unknown thing is inside the room.
The scribe's eye will have jumped from one so to the next.
547
Shih ch'i se 'the colour of this'. For the effect of ch'i in emphasising the
demonstrative, cf. § 1/3/4/9. The choice of shih rather than tz'u jtb
shows that the debater is referring to a kind of thing which is white,
not pointing out a white object (§ 1/3/4/2).
548
Yu 'It is just as with . . . ' ( § 1/3/8). It is the same with inferring that an
unknown colour is white as with disputation over whether a known
colour is white or black (sheng 'win' is winning in disputation, cf.
A 74). The Mohist is referring to the very first example of judgment
between mutually exclusive alternatives in A 88, and in similar words:
FBfi8#, SH-tfe " O n both sides prevailing decisively: white or black".
549
Jo 'or' (§ 1/3/11/8).
550
Shut 'which ?', probably adverbial, 'with which of them ?' (p. 135 above).
551
Chin yeh 'in the present case', contrasting with yu 'it is as with' as in
B 27, 78 (pp. 138f above).
552
Ch'in chih 'knowing by experience' and shuo chih 'knowing by explana
tion' are distinguished in A 80. , ;
C. "When you hear that something you do not know is like something you
do know, you know them both. Explained by: informing.
E. A thing outside you do know, the thing in the room you do not know.
Someone says: 'The colour of the thing in the room is like the colour of
this'. Then the thing you do not know is like the thing you do know. It is
as with 'White or black, with which does one win ? This is like its colour,
and what is like a white thing is necessarily white'. In the present case too
you know that its colour is like the white, therefore you know that it is
white.
A name by means of something you are clear about determines something

you do not know, it does not bring in the thing you do not know to cast
doubt on the thing you are clear about. (Like measuring an unknown length
by means of a footrule.) What is outside, you know by experience; what is
in the room, you know by explanation.''
The example may be taken as a white object out of sight in another

room. But there are some grounds for suspecting that 'in the room' is the
Mohist's way of saying 'in the object' (§ 1/4/26). If so, the example is
probably once again the stone of B 37, the whiteness of which is compared
with that of another thing in B 38. The argument is unaffected in any case.
The Mohist seems at first glance to be labouring an obvious point.
But in the last part one finds him speaking as though his argument about
being like a white thing, in which the question of names has never been
raised, leads to conclusions about names in general (fu ming). The whole
section turns out to be an answer to an objection to the radical nominalism
of A 31,32, 78. According to A 78 we name an object and extend the name
to all similar objects, so that a common name is equivalent to 'like the
object' ( T J ffe^f). It might be objected that on this account a common
name does not enable us to know an unknown object, merely tells us that
it is like a known object (in a Western terminology, tells us what it resembles
but not what it is). The Mohist replies that it is exactly as with disputation
over a known colour, which is proved to be white by comparison with a
standard for 'white' (cf. A 96).
This section connects in theme and terminology with a story about the
sophist H u i Shih:
Shuo yuan (ch. 11) S P T K 11, 4A/7-4B/1. ^ I B ^ I B » r f i ^ ± a * f t
. fiipptmô J i s 9
r « « IWEMM+'B* raft
^smwmbs 9
° j » ^ s »r ^ A t k a m ^ ^ m 9
B»3s±»
M « ? « B » »£tt»3»oflMi№? JIB 9
!"*!§№<> J r i * f i J 6 « 0 9
3 ! ± № £ n ^ » rf5RWf§3£ · flO*№ ? J I B » T ^ 1 » ^ ° j ^ B >T

m H K X № * 0 » * R B f ^ » · ffii«A»£ ° ^ I B f t S S »
9
o j £
B> T#o J
" A client said to the K i n g of Liang:
'When H u i Shih talks about anything he is prone to use illustrative com
parisons. If you forbid him illustrative comparisons he won't be able to
speak.'
The K i n g agreed. A t the audience next day he said to H u i Shih:
'When you speak about something I wish you would simply speak directly,
without illustrative comparisons.'
2/4 445
'Let's suppose we have a man who does not know what a tan is', H u i Shih
said. 'If he says "What are the characteristics of a tan like ?" and you answer
"Like a tan", will it be conveyed to him?'
'It will not.'
'If you proceed to answer in stead " A tan in its characteristics is like a bow,
but with a string made of bamboo**, will he know?'
'It will be known.'
'It is inherent in explanation that by using something he does know to
convey what he does not know one causes the other man to know. T o give
up illustrative comparisons as you are telling me to do is inadmissible.'
'Well said', said the K i n g . "
This evidence is valuable because the Mohist defines only abstract
terms; one would like to know how Chinese dialecticians defined their stock
examples such as ox, horse and milu deer. The H u i Shih story suggests
that, in agreement with the nominalist principle that the basis of naming is
similarity, they conceived the typical definition as being by analogue and
differentia rather than like Aristotle by genus and differentia.
5 5 3
B 71 K s f § * f # »»o t & g X a o
5
:f«Hb o ( f f i ) * £ 5 5 4
*A£a^J · (JEW*)" »fl'JAW^Jffe o
553 prf 'confused, fallacious' in later Mohist usage seems always to imply
self-contradiction ( § 1/4/23). The reference is presumably to Chuang-
tzu's thesis that of anything said we may equally well say the opposite.
6 5 4
Corrected from the parallel in the next sentence. Chih was no doubt
written with the graph ^H.
6 5 5
This parenthetic phrase looks like a gloss; the sequence of the argument
suggests that it should be at the end of the sentence.
5 5 6
Yi for yi wet % 'deem' (§ 1/3/12/4/4).
C. " T o claim that all saying contradicts itself is self-contradictory. E x

plained by: what he says himself.
E . T o be self-contradictory is to be inadmissible. If what this man says
is admissible, there is saying which he recognises as admissible (and so
not self-contradictory). If what this man says is inadmissible, to suppose
that it fits the fact is necessarily ill-considered."
In this argument it is to be noticed that being inadmissible (puk'o)

is distinct from although a corollary of being logically confused or self-
contradictory (pet), and not 'fitting the fact* (tang) is distinct from although
a corollary of being logically inadmissible (pu k'o). Cf. p. 39 above.
B 71 and 79 are essentially refutations of the propositions " A l l pro
positions are false** and " A l l propositions are true*', similar to those of
Aristotle in Metaphysics 1063b/30-35. But they are not presented symmetri
cally, because they are designed to deal with an anti-rational philosophy of
the kind that Chuang-tzu presents in Seeing things as equal. According to
Chuang-tzii everything said is admissible from one point of view and
inadmissible from another; but he is not directly vulnerable to the Aristo
telian refutation, since he never suggests that everything may be affirmed
or everything denied from a single standpoint. He does however treat all
'saying* (yen) as vitiated, in that whatever we say is equivalent to the
opposite said from the other standpoint, and even if we try to escape from
alternatives by affirming "The myriad things and I are one** the affirmation
is added to the myriad things and me, and "one and the saying make two**
( · — T h e sage can only guide us, by words which " i n referring
to something refer to nothing and in referring to nothing refer to something*'
(SftHBWnB, ^pBlffinB), towards the wordless illumination in which every
thing becomes shih, 'This'. The Mohist is making a brave effort to refute
by reason an explicit rejection of reason. Chuang-tzu does not reject all
saying as inadmissible, but does seem to imply that it says the opposite of
itself, is self-contradictory. He does not ask us to say 'This' of everything,
but he does 'reject denial' B 79).
5 5 7 5 5 8
B 72 1 1 S M » oI M S o
5 5 9 562 5 6 3
(jf) o MM r s J " J » o ffii№ r ft J *
657
Wei 'is specifically' (pp. 118f above). For the syntactic combinations of
wet 'call', cf. § 1/3/12/1.
658
Fan 'converse', the graph uncorrupted (§ 1/4/9).
559
Hao 'crane', the stock example of a name which can be loaned for
another thing, typically a dog (§ 1/5/4).
560
Yu chih 'still' (p. 139 above).
561
Fu before a common name marks it as referring to the general concept
(§ 1/3/4/10).
562
Pi shih 'that and this' (of kinds of thing, not like pi tzu jlfc of objects, cf.
§ 1/3/4/2), as in the Chuang-tzii ch. 2 passage quoted in B 68, note.
2/4 447
The final yeh probably marks a distinction between wei shih hao "call
this thing 'crane' " and wei pi shih shih yeh "say that that and this
are this." (§ 1/3/12/1/3).
The graph $J and the preposition hu serve to distinguish a sentence
(
pattern from the quantifying construction wu M . . . yii № . .
(pp. 133f above), which would have made a sentence meaning "The
caller is not specifically either of the things he is called." I take the
wet as nominalised and equivalent to the wei shih ffUlk of B 12, 'what
is specifically X ' ; the Mohist cannot supply shih here because he is
talking about shih 'this' and pi 'that*. For the use of copulae without
complements and their nominalisation, cf. p. 118 above.
The parallelism with the next sentence, the phrasing of the Canon, and
the Mohist's avoidance of hu except when distinguishing from quan
tifying constructions, support this deletion. (But cf. p. 165 above.)
Hsing 'proceed' (what is so of one object being so of all of the same
kind), cf. § 1/4/4.
C. "That something is specifically what I call it is inadmissible unless that

is its name. Explained by: the converse.
E. T o call this thing 'crane' is admissible, but still it is not the crane itself.
To say that both that thing and this are this thing would not be admissible;
the sayer would have nothing which is specifically the thing it is called.
If that thing is still specifically the thing it is called, what I call this will not
'proceed'; if that thing is not specifically what it is called, what it is called
will not 'proceed'."
The examples of 'being this thing specifically' {wei shih '№;§·) in B 12

are objects fitting the names 'ox' and 'horse'. These are the primary names
of kinds of things; the Mohist presumably would not object to coining more
general names, since in his technical terminology he coins so many himself.
The fact that dogs are called by two doubtfully synonymous names,
kou $U and ch'iian A , and that a dog maybe given the personal name Hao
'Crane' (B 8), provided the stock proofs that all naming is conventional,
and the justification for the Second List sophism " A dog may be judged
to be a sheep" (§ 1/5/3, 4). The assumption that any name might be given
to anything, and that by 'calling' (wei M) we can divide and classify in any
way we please, is essential to Chuang-tzu's denial of the reality of all
distinctions in Seeing things as equal. As we noticed in discussing B 68
the point on which the Mohist insists in this controversy is that the
difference between X and Y is not abolished by a change of naming. You

can call dogs 'cranes', but the difference between the creatures remains.
If they were the same the converse (fan) would also apply; we could also
say that both dogs and cranes are 'dogs'. The essential point is that a com
mon name is given to objects which are like each other (A 78), so that we
can 'proceed' (hsing) from what is so of objects to what is so of the thing
they are ( B 1). As stated in the summing-up of B 8, the difficulty in calling
dogs 'cranes' is in the 'not being so' (^#&). No change in the conventions
of naming can alter the fact that either
(1) the birds are still specifically 'cranes', in which case to give the
name to dogs will not entitle me to describe cranes as having teeth or having
four legs; or
(2) 'crane' has become a general name for two kinds of thing, in
which case calling the birds 'cranes' will not entitle me to describe cranes
as having beaks and wings.
B 73 mm^mm » I B S H S °
(H) · r mm&m » so Ri« · mmwi^m · ^mmm^w»wmm
#
A«:Rft(ft) &™* »»JA**ibo«**«Ht«ft*«!ffl*lJ!«ft ·
Accidentally repeated.
The phrase has intruded from the next clause.
Cf. the other example of this corruption in B 57.
The graphs when confused will have been written in the forms (~K)*Jc.
(cf. B 3, n. 275: B 57, n. 489). Ch'i is the reading of the Mao, Mien
and Horyaku editions; the Taoist Patrology and the other early
editions further corrupt it to 'ft".
C. "Their being limitless is not inconsistent with doing something to

every one. Explained by: whether it is filled or not.
E . (Objection). The south if limited is exhaustible, if limitless is inexhaus
tible. If whether it is limited or limitless is unknowable a priori, then whether
it is exhaustible or not, whether men fill it or not, and whether men are
exhaustible or not, are likewise unknowable a priori, and it is fallacious to
treat it as necessary that men can be exhaustively loved.
(Answer). If men do not fill the limitless, men are limited, and there is no
difficulty about exhausting the limited. If they do fill the limitless, the
2/4 449
limitless has been exhausted, and there is no difficulty about exhausting

the limitless.''
The few Explanations which give both objection and answer are
especially interesting since they show how the later Mohist disputation was
conducted in practice. Of the three examples (B 38, 73, 82) this is the most
fully and at the same time most economically developed. Philosophical
dialogues in Mencius, Chuang-tzu and Kung-sun Lung tzu, although ofte
more interesting in content, offer nothing to match the extraordinary
formal elegance of this argument, which becomes more apparent the more
closely one examines the senses of the key words in Mohist usage (Hat
y
'be inconsistent with', wei k o chih 'unknowable a priori', pi 'necessary',
pei 'fallacious, confused, self-contradictory'). The Mohist has considered
the meaning of every important word that he uses. The following words
have already been defined in the senses understood: ch'iung 'limit' (A 42)^
ying 'fill' (A 65), chin 'exhaust' (A 43), chih 'know' (B 5), pi 'necessary*
(A 51), not to mention chien 'total' and at 'love' (definitions lost, cf. § 2/2) #
The objector is fully aware that his argument cannot establish more
than that since the number of men may be infinite the possibility of loving
them all "cannot be treated as necessary". But this is enough to require an
answer, since it is the dearest conviction of the later Mohists that their
ethical system is logically necessary. The objector no doubt knows of the
arguments (unfortunately lost to us) for the two sides of H u i Shih's paradox
"The south is limitless yet is limited" ( f t ^ ^ ^ J f S W H ) ; he sees however
that for the purposes of his argument he need not postulate that space is
infinite, only that "Whether it is limited or limitless is unknowable
a priori". The Mohist in his turn narrows his answer to the crucial point
that "there is no difficulty about exhausting the limitless", which he
demonstrates very neatly in two sentences. One is especially impressed by
the absence from the Canon itself of any reference either to men or to love;
it abstracts the logical issue involved in defending universal love against
attack from this direction, "Their being limitless is not inconsistent with
doing something to every one".
B 74 * » K # f f i » K « r t l l » i & £ ( 0 J 3 ) * F p a ^ ·
(2) °-^ 5 7 3
» rffl№Xfi£* £* J #
> I£# 5 7 4
« 5 7 5 :
?XPP^
-tfe · * | H A » BQftXftffrlH ·
B 75 o mm^m °
5 7 2
A n emendation of Sun well supported by the Explanation.
5 7 3
For another example of an isolated erh at the head of the sentence
("If they are two . . ."), cf. A 60.
574
Huo che 'some* (pp. 133f above).
675
Yi 'leave out', used elsewhere of failing to establish a pair of comple
ments (§ 1/4/18).
676 \ Y h t we take as the Explanation of B 75 is commonly taken as the
a
conclusion of B 74 (where however it hardly makes sense), leaving

B 75 with B 11 as the only Canons without Explanations among the
propositions.
C. "Without knowing their number we know that all are included. Ex
plained by: the questioner.
E. If they are two men, we do know their number. In 'How do we know
that your love of the people includes them both ?', some men are left out
of his question. If he asks about all men, then one loves all whom he asks
about.
C. Not knowing their location is not inconsistent with loving them. Ex
plained by: a man who has lost his son.
E. Like knowing that love includes all of them when you do know their
number; there is no difficulty.*'
Suppose that the questioner did ask about a fixed number of men (as
usual, the Mohist takes two as the typical number greater than one); he
would then not be asking about all men. But an objector asks how one
loves all men; and if one can ask about all without knowing the number,
one can also love all.
The Mohist's analogy between 'asking' and 'loving' is reasonable,
since he holds that one loves all if one loves individuals solely as men, just
as one would be asking about all if one were to ask about any characteris
tics of a person which are essential to his humanity.
One can love a lost child without knowing where he is. The Explana
tion of B 75 does not elaborate this obvious point, merely notes that that
even the objection in B 74 assumes that one can know one loves all men if
one knows their number, irrespective of whether one knows their location.
5 7 7 6 7 8 6 7 9
B 76 t « ± S f t ^ '&(^)*» » i&ffiffS *
2/4 451
5 7 7
For the three cases of net wai we follow the readings of the W u manu
script. The Taoist Patrology and all other editions invert the two words
except in the first example in the Explanation. It would be natural for
a scribe to slip into writing them in the more usual order wai net
(invariable elsewhere in Mo-tzu).
5 7 8
Emended following Chang Ch'i-huang (3,23A). The corrupt character
is in exactly the sentence position of the pei of B 34, cf. also the four
other examples of pei at the end of Canons in B 71-79.
5 7 9
Sun conjectured that the phrase is a corruption of the binome SIPH*
(Chu Ch'i-feng 1796), expressive of disorder. But the Mohist uses wu
fif, E ( = f f i ) of the matching of one figure with another (§ 1/4/34), for
example a circle with its standard (A 98) or the partly matching ideas
of using a hammer and of making shoes (B 58). I propose therefore
to take it as a reference to matching the two eyes with the face (either
both are in it or both outside it).
5 8 0
Cf. n. 577.
581
K'uang chu 'referring arbitrarily* as in B 66.
C. "It is a fallacy to suppose that benevolence is within and morality

outside. Explained by: matching with the face.
E. T o be benevolent is to love, to be moral is to benefit. Loving and benefit
ing are on this side, the loved and the benefited are on that side. Neither
loving and benefiting nor the loved and the benefited are within or outside
each other. T o suppose that benevolence is within but morality outside is
to refer to loving and to the benefited, which is referring arbitrarily. (Like
the left eye being excluded from the head and the right eye included.)'*
The doctrine that benevolence (jen) is internal but morality (yi)

external is that of Kao-tzu H r f , criticised in the first episodes of the
Kao-tzu chapter of Mencius. We also find it, together with Kao-tzu's
theory of the moral neutrality of human nature, in the ChiehfflLchapter of
Kuan-tzu (cf. G(9) 227-231). The Mohist, who cares nothing for the
subjective side of experience, does not share Mencius' concern with proving
that morality as well as benevolence is internal (cf. Lau (1970) 258-261).
But he does wish to give the two major virtues the same status. His
refutation is deduced directly from his definitions of jen and yi (A 7, 8).
The illustration offers as parallel the two eyes. It does not matter
whether you say that they are both in the head or both outside it; what
you cannot say is that one is inside and the other outside.
B 77 m±&& » m&mm ·
T'an and W u take this word as t'o ffij 'another', which is however
unattested as subject anywhere in Mo-tzu. It seems necessary either
to delete the yeh (Kao) or assume a lacuna.
C. "That it is useful to learn. Explained by: the objector.

E. Thinking that he does not know that it is useless to learn, he therefore
informs him. This is causing him to know that it is useless to learn, which
is teaching. If he thinks that it is useless to learn, to teach is to contradict
himself.""
The Taoist rejection of learning is criticised as self-contradictory by

the procedure used in B 71, 79.
583
B 78 tt ±ÎS^K** o o
5 8 3
Although fei 'blame* is defined among the ethical terms (A 30) we find
it in use only in logical contexts, always treated as equivalent to a
causative fei W 'reject*, the added radical serving to give it full
syntactic mobility (B 11-19, cf. p. 160 above). Cf. Lii-shih ch'un-ch'iu
ch. 18/7 (Hsu 18, 23A/-3) M^MSx^M "This is the rejecter being
the same as what he rejects**.
584
Lun 'to sort, grade* (§ 1/4/19).
5 8 5
Dropping the radical on grounds of parallelism.
C. "Whether denials are admissible or not does not depend on how many
or few they are. Explained by: deserving rejection.
E . In the sorting out of admissible and inadmissible denials, if his grounds
are that in principle the claim deserves rejection, however much is denied
his denial is the right choice; as for claims which in principle do not deserve
rejection, however little is denied his denial is the wrong choice. To say
in the present case that to deny too much is inadmissible, is as though one
were sorting out the shorter in comparison with the longer.**
2/4 453
Here the Mohist appeals to the basic difference, examined in A 88,

between comparing to decide which is more beautiful (B 3) or longer or
shorter (cf. B 80), and judging between contradictories, shih and fei.
Sorting out the longer and shorter in a pile of sticks implies equal numbers
of both, and it would be reasonable to object that too many sticks are being
classed as short. But the ox that I declare X to be, it either is or is not.
If X is this (shih) and you say it is not (fei), then my denial (fei) of your
claim will be the choice of what it indeed is, the right choice (shih).
B79
#B£SMil» ^m\- · ' *«I2Nl» ft^»ife ·
5 8 6
The head character has been wrongly provided with the radical.
C. " T o reject denial is self-contradictory. Explained by: he does not

reject it.
E. If he does not reject his own denial, he does not reject denial. Whether
his rejection is to be rejected or not, it amounts to not rejecting denial."
B 78 reverses the argument of B 71 that it is self-contradictory to deny

all propositions. At first sight there seems to be no corresponding contradic
tion in affirming all propositions; but the Mohist shows that there is, by
presenting the affirmation as a double negation, 'rejecting denial'. The
double negation is in any case better adapted to dealing with Chuang-tzu
(cf. B 71, note).
5 8 7
B80^(«)*ffi ^S'tSft^rSJ ·
( % o B S S M ' *ft»ft»*5gsB&* o ft±d6te*Ste#»H*ifeft ·
Corrected following Sun.
C. "Whether or not something is extreme in degree. Explained by: to the

degree of being 'this'.
E. If it is the longest or shortest, nothing is longer than this, nothing is
shorter than this. As for this being or not being this, nothing is this to a
more extreme degree than this is."
Y may be longer than X and shorter than Z ; but as far as being Y is

concerned, something either is Y or is not. Relative concepts are discussed
also in the first part of A 88, in B 55, and at the conclusion of B 3 and of B 78.
B 8i &TO#_t-tk » mjm «
588 Variant 11 (Mao, Mien, Hôryaku). But ch'ing ('what a thing is in itself,
cf. § 1/4/6) is acceptable, since technical nouns can be used causatively
after so(% 1/3/13).
C. "Choosing to be below is for the sake of seeking to be above. Explained

by: the marsh.
E. Higher and lower are being measured in terms of better and worse
(they are not like mountain and marsh). If dwelling below is better that
dwelling above, what is taken as the essential character of being below is
being above."
A preference for the lower, softer, weaker, over the higher, harder,
stronger, is a Taoist theme especially prominent in Lao-tzu. In the case
of refusing the higher, there is an ambiguity as to whether we mean social
status, prudential advantage or moral superiority. The Mohist declares
that if higher and lower are being measured in terms of value, it is impos
sible to choose the lower except as a means to the higher (which indeed it
is for Lao-tzu, if you follow one strand in his systematically ambivalent
thinking); if one chooses the lower as an end in itself, one is by definition
regarding it as the higher.
B 82 S&H&IRI ' t a f f i ^ ^ H 589

o
r^(S)di » S&MLBAM O ^&(X)*±№&\fn^ <± >™№& · *
«**ih ' fl№rfn**ih» * 4^*ihftJiiJff *±ik& o& r *± J 13
589
Chou (— Ml) 'all round, everywhere', as in N O 17 MStA 'love men
without exception'. The interchange is common in Mo-tzu. Cf. ch. 52
(Sun 325/9) 'the road around it' (written MM 339/7, 359/13),
ch. 69 (Sun 358/7) TfC^MI 'islets (chou =iffl) in the water'.
5 9 0
Restore on grounds of parallelism.
The many other problems require discussion at length: (1) The Ex
planation presents an argument and refutes it; its organisation may
be compared with B 38, 73. (2) For the syntax of the causative
constructions, cf. § 1/3/13. Shih shih (verb—object) "treat this as
'this' ". Shih ch'ieh shih yen (subject—passive verb) "This will be
2/4 455
treated as 'this' among them" (not 'treated as this by hint', since the
directive with the agent in a passive construction is unattested,
p. 116 above). (3) The Canon is closely similar to that of B 68, the
Explanation of which also has tSB-jtbife " I f that is going to be treated
as 'this' " (cf. "This will be treated as this among them").
Comparison with B 68 also accounts for the mysterious graph %
which appears eight times, and which elsewhere is a systematic
corruption of (§ 1/2/1/2/13. Here Sun, Liang, W u and L i u make
the same emendation, and try in various ways to account for the fact
that it is negated by pu and is therefore a verb; T'an emends to
and Kao to A . )
B68tftjlfctftjfc;Htfejfcia
82 S A H Ji U
6 8 t t « «
82 &(X.)*±tk&
The original word was therefore chih i t (used of a name staying fixed
in an object, § 1/4/4), a word graphically and phonologically related
to chih and confused with it in B 2 (cf. B 2, n. 266). (4) B 68 is a
criticism of the relativism of Chuang-tzu. B 82 connects with another
passage in Chuang-tzu's Seeing things as equal: (Kuo 108/lf) S^PTI: ,
"Treat even what is not this as 'this', even what

is not so as 'so\ If what is this is ultimately this and what is so is
ultimately so, there is no difference for disputation between this and
not this, so and not so". (5) The formula 'Xyu Y t'ung' implies that
X is required by Y (§ 1/4/36). The Canon will then be a direct answer
to Chuang-tzu: " Y o u cannot treat as 'this* without treating only this
as 'this* " . That it is X that is required by Y is especially obvious in
the concluding sentence: "Therefore you cannot show that 'this' does
not stay without showing that it does stay". (6) Neither Chuang-tzii
nor the Mohist is using the terminology of disputation over right and
wrong alternatives. Both negate shih not by fei but by pu shih which
we take to be the opposite of naishih TbJk (p. 118 above) 'is/is not this'
(among indefinite possibilities). The summing-up of the Canon denies
that shih applies 'to all cases' (chou), not 'to both sides' (liangW). The
topic seems to be the transference of shih from one thing to another
as discourse proceeds, of which there was a striking example in B 80
-%L±ML&#M&%, X f i K S "As for this being this or not being this,
nothing is more it than this is". This explains why Chuang-tzu thinks
that if we are really saying anything about A , B , C . . . when we refer
to them successively as 'this', there cannot be any ultimate difference
between this and not this. (It would make no sense for him to say that
if the right alternative is ultimately right it cannot be distinguished
from the wrong one.) It also explains how the Mohist can use shih
13 times in a short Explanation without sorting out its references,
which would be hopelessly confused if he were switching backward
and forward between alternatives (a point overlooked in the earlier
account of this Canon in G(3) 96). For an example of a syntactically
similar construction actually applied to right and wrong alternatives,
cf. Hsun-tzu ch. 2 (Liang 16/6f) &&##m±.90,
"To recognise right as right and wrong as wrong is called wisdom, to
take right as wrong and wrong as right is called foolishness". (7) The
argument and refutation are in the same form: Sentence 1 . . .,
tse . . . yen. Sentence 2 Chin . . . yii shih, erh . . . yii shih. Sentence 3
Ku . . . Since yen 'in it' contrasts with yii shih 'in this', its reference
is to the object which may or may not be indicated as 'this'. The chin
'now* which immediately follows marks a temporary choice of some
thing as 'this*. Although Chinese chin like English 'Now . . .' some
times merely introduces a new stage in discourse, and does so once in
Names and objects ( N O 10, but in an exposition with a much more
leisurely development), here it definitely indicates a point in time.
There is no other example of chin at the head of the sentence in the
Canons and Explanations except in the combination chin yeh 'In the
present case' (cf. B 70 n. 551).
C. " Y o u cannot treat as 'this' without treating only this as 'this'. Ex
plained by: not embracing everything.
E. (Sophism). As for the ones which are not this one, this one among
them is about to be treated as 'this'. M y present 'this', when it does stay
in this one, will not be staying in this one. Therefore 'this' does not stay.
(Refutation). If 'this' does not stay, it does treat one as 'this* but does not
stay in it. M y present 'this', when it does not stay in this one, will stay in
this one. Therefore you cannot show that 'this' does not stay without
showing that it does."
The point at issue is whether the demonstrative shih 'this' (the one
being talked about, the one in question) can be said like 'ox* and 'horse* to
'stay* in a thing, distinguish it from other things for a duration of time
2/4 457
(cf. A 50). Since in discourse it constantly flits from one thing to another,
it may seem that it does not. If so, the validity of all disputation is under
mined, since alternatives are distinguished as right or wrong by shih 'is-this*
a n d / « 'is-not*.
The Mohist shows that one cannot, as Chuang-tzu pretends in Seeing
things as equal, "treat even what is not this as 'this* (shih pu shih)", exten
the demonstrative to everything. It does not 'embrace everything*; that at
any moment it may shift away from X does not alter the fact that it is
temporarily staying in it, distinguishing X from everything else for a
duration of time. The argument is like that of B 68, that 'that* can be
substituted for 'this* only if 'this* is also substituted for 'that*.
2/4/2/5 Appendix: Kung-sun Lung's 'Essay on pointing things out'

interpreted in the light of B 32-82
The Chih wu lun fafelra is (with the 'White horse* essay) one of the
two chapters of Kung-sun Lung tzii which are genuine pre-Han documents.
Of all Chinese philosophical writings it is perhaps the one which most
fascinates readers with the taste for solving difficult riddles, but no two
commentators have ever agreed in interpreting it. The solution I proposed
in 1955 has not satisfied more recent inquirers (Chmielewski(l), Cheng and
Swain), and now seems even to me to depend too much on arbitrary textual
emendation. Whether or not my new proposal is found acceptable, I hope
to show that in future the essay must be studied, not in isolation, but in the
light of the problems in disputation of B 32-82.
Inquirers have never agreed even on the significance of the crucial
word chih. But this seems to me an unreal difficulty; as I showed in my
earlier study, we have plenty of evidence as to how pre-Han philosophers
used the word chih,bo\h as a verb ('point out*) and as a noun ('what is being
pointed out'); they applied it not only to the gesture of pointing but to
the meaning of discourse and the meanings of words. T o ignore the philo
logical evidence and start by asking which of the concepts of Western
philosophy fits the Chinese argument ('universal', 'quality', 'logical class*)
is wrong-headed in principle. The real difficulty arises from the fact that,
until we notice parallels in the Mohist Explanations, the literary form
appears to be unique; we seem to have no clue to the overall organisation
of the argument.
In my earlier study I noticed that the essay falls into two halves, of
which the second resumes and develops points raised in the first. But this
account now seems to me altogether to underrate the contrast between the
two halves. The first starts by proclaiming that "chih are not chih"(JB#MB ),
3
458 The 'Canons' and 'Explanations
and concludes with a chain of arguments leading up to the same paradox.

But the paradox does not reappear in the latter part; on the contrary we are
told explicitly that "it is not that chih are not chih, it is chih in conjunction
with things which are not chih" (t&**J&-fe, J&JSfe#J&1b), and the final
sentence is "Moreover if chih inherently and of themselves are deemed not
chih, why is it that they are deemed chih only in conjunction with the things
on which they depend?" ( f i . * J & I I S « # » , ^tefeifn7!rllftj&). The
'White horse" essay is an example of disputation in practice, presented in
dialogue form; is the present essay another, consisting of an argument
followed by its refutation?
This suggestion can be wholly convincing only if we find other
evidence that it was a convention in disputation to write down an argument
followed by its refutation (possibly separated by an interval ignored by
transcribers after the tradition of disputation was forgotten). Now several
of the problems of disputation in B 32-82 have precisely this form. The
one which has been recognised by all editors is B 73, but I have noticed
two more (B 38, 82). The structural parallelism is especially striking in
B 38; there too the refutation ends on a rhetorical question introduced by
ch'ieh 'Moreover . . .'. But this new insight into its structure must alter
our whole approach to the essay. We can no longer treat it as a meandering
discussion of the relations between chih and things; it is the defence of the
sophism "Chih are not chih" followed by its refutation.
The word chih has three main functions:
(1) Noun, 'finger'.
(2) Verb, 'point out one from another'.
(3) Noun (also written i f , 1B) 'meaning', the direction in which
discourse points, its meaning or drift, the main point in contrast with
details or side issues.
The Canons and Explanations use the verb chih of pointing out an
object by gesture, in contrast with chii 9 'referring', picking out an object
by name (B 38, 53. Cf. also B 39). Names and objects has one example of
the noun chih 'meaning' in an obscure context (NO 9). Chih 'finger' appears
only as a part of the body, without reference to its function of pointing
(A 75, B 12, E C 8, N O 1, 3, 7).
Hsiin-tzu in the Right use of names differs from the Mohist in extending
the verb to the pointing out of objects by means of names:
(Liang 312/8) ft»*B±£3(I » «0*Blf&» o
"Therefore wise men made divisions and distinctions on behalf of them,
and instituted names in order to point out objects."
2/4 459
(Liang 320/6) » SffiKSS » o

"Therefore when the name is sufficient to point out the object, and the
sentence sufficient to display it to the utmost, he goes no further."
Similarly Chuang-tzu uses the nominalised chih of the meaning of
a name:
Chuang-tzu ch. 22 (Kuo 750/10) r J S j r f f l j r ^ J H ^ * &%№H » X
"These three, chou,pien, and hsien, are different names for the same object,
their meaning is one."
Nominalised chih appears also in two versions of the programme of
disputation quoted on p. 20 above; here there is little point in asking whether
it refers to the meanings of names or to the general drift of discourse:
0
Shih chi chi chieh j£!2Jfc» ch. 76 (Shih chi 2370 n. 2) ff«51J& » J «
mm*
"(The disputant) . . . dredges out his ideas and makes his meanings
intelligible, and makes it plain what he is talking about."
Han shih wax chuan fttlfSPTK 6, 3B/6f J E f & f f i » sB£R °
"Unconsciously failing to sort out one's meaning is called 'obscurity'."
The nominalised chih 'meaning/what is pointed out' derives from the
verb chih 'point out' by a semantic shift similar to that from 'mean' to
'meaning' (what is meant) in English:
Han shu (ch. 53) 2411/2 °
"His style is concise and his meaning plain."
Comment of Yen Shih-ku (581-645): H B J S B * l £ 0 r » »5g
" 'Meaning' refers to what the sense runs towards, like a man pointing out
a thing with his hand." That is, the chih is 'what is pointed out'.
This detail is important in connection with explanations, criticised
in my previous study, which treat chih as both the 'pointer' and 'what is
pointed out'. Since the primary meaning of the word is 'finger', it seems
at first sight easy to accept the suggestion that it could be used for a
pointer, less easy to suppose that it could mean the object of pointing. But
as a matter of fact chih 'pointing' detached itself sufficiently from chih
'finger' to be written sometimes with other graphs ( p , '№), and there seems
to be no evidence that it was ever used for 'pointer' or 'sign'; on the other
hand it definitely tended to merge into the object of pointing, as in the
case of the English word 'meaning'.
In my older translation I was content to represent chih by the
word 'meaning'. But although 'meaning' appears to be unavoidable in
the translation itself, such an equivalent cannot of course be more

than an approximation. 'Meaning* does not bring out a crucial aspect
of chih whether it is used of gesture or of words, that it always implies
pointing out one item from others. This aspect might well be fundamental
to the Chinese argument. If so, my choice of the word would tempt me to
impose Western problems concerning meaning on to the thought of
Kung-sun Lung. (It will not have escaped historically-minded readers that
at the time when Neo-Realism was fashionable Fung Yu-lan recognised
the chih as universals, and that it was during the ascendancy of linguistic
philosophy in England that I identified them as meanings.)
Now other early references to a sophism about chih strongly suggest
that it affirmed the impossibility of comprehending the cosmos by pointing
out from the whole. The reference in Chuang-tzii ch. 2 to 'showing that
chih is not chih" ( ^ B ^ ^ H - t f e ) and "showing that the horse is not a
horse" ( P i S ^ ^ S - f e ) ends with the declaration:
(Kuo 66/7) x * — » s t e o
"Heaven and earth are the one item-pointed-out, the myriad things are
the one horse."
Among the Second List sophisms we find:
Chuang-tzu ch. 33 (Kuo 1106/2) t B ^ S » M ^ f f i °
"When we point out we do not arrive, when we arrive we do not detach."
Here it may seem more natural to take the word chiieh in the sense
'come to an end*; but the only philosophical use of chueh in the later
Mohist corpus (A 88) is of detaching something from its interrelations.
(For chih 'arrive', cf. A 32 n. 108.)
Lieh-tzu (c. A . D . 300) ascribes the sophism to Kung-sun Lung and
gives a longer account of it, which may or may not have older authority:
Lieh-tzu ch. 4 (Yang 88/4f, l l f ) #M^(>b)* ± » ^ M H ^ M > HVffcm . . .
"When there are ideas, it is not determined(J). When something is pointed

out, it does not arrive. When there are things, it is not exhausted. . . . When
there are no ideas, determinations^ ?) are the same. When nothing is pointed
out, everything arrives. That which exhausts things exists as constant."
(Our suggestion that 'fr is a corruption of i t as in A 96, 97: B 26, 32, and
is to be understood in the Mohist sense, 'fix what a thing is', is only tenable
if Lieh-tzu is borrowing from a pre-Han source.)
Something in the essay which has distracted attention from the theme
of pointing out is the recurrent phrase 5 ? T W R "The world does not have
chih", commonly understood as "Chih do not exist". Until the essay is
2/4 461
recognised as the defence of a sophism followed by its refutation it appears

to be a discussion of the relation between chih, which do not exist, and things,
which do exist. I myself assumed that Kung-sun Lung conceived meanings
as hypostatised entities without concrete existence which he was trying to
relate to the concrete things which names mean. But this kind of hypostati-
sation, characteristic of Western thought and encouraged by Indo-European
language structure, now seems to me foreign to Chinese disputation (cf.
§ 1/1/2/1/7). On closer inspection one notices something suspicious about
T'ien-hsia wu chih, "The world does not have chih". The two words wu ch
would be sufficient to convey "Chih do not exist"; why does Kung-sun
Lung consistently supply t'ien-hsia 'world', a word which appears no fewer
than 14 times in this short essay ? This suggests that 'world' may itself be
an overlooked concept in Kung-sun Lung's argument, the cosmos as a
whole in contrast with the things which compose it. We find this contrast
in Chuang-tzu:
Chuang-tzu ch. 21 (Kuo 714/-3) 5fc T %T J »* $ £ 0 r — f e ·
"As for 'the world', it is that in which the myriad things are one."
On this hypothesis we could take T'ien-hsia wu chih either as "In the
case of 'world' there is no item-pointed-out", or as "The world does not
have the item-pointed-out/Nothing within the world is the item-pointed-
out" (by 'world': on this interpretation the reference should be established
by the context). Both readings give the same sense, although syntactically
it is the second which agrees with other examples of the formula T'ien-hsia
wu X throughout the essay. This suggestion makes sense of an especially
difficult sentence which my former interpretation forced me to emend:
X T M B # , ' W « J f l » & ' t i l (Of 'world') "That nothing within the world
is what we are pointing out is in the case of things it being inadmissible to
say that nothing is what we are pointing out".
In my earlier study I took chien i£ in the sense of 'annex to itself,
common in non-philosophical usage. But comparison with other texts
shows that problems of pointing were discussed in terms of the concepts
t'i 'unit/individual/part' and chien 'total/collection/whole' which we
examined under A 2:
Chuang-tzu ch. 25 (Kuo 909/-4) 4 J & S i : W | l i f D ^ # J R » rfnJR»ftW# ·
"Now the fact that when we point out from each other the hundred parts
(t'i) of the horse we do not find the horse, yet there the horse is, tied up
in front of us, is because we place the hundred parts on another level and
call them 'horse'." (For 'levels' cf. B 59).
Explanations B 38 » [ K Z l J ·& o
"Pointing at them collectively (chien) is 'using the two*."
Kung-sun Lung: fifH#^T±§T^ °
"Moreover items pointed out from each other are collected together
(chien) by the world."
The last example confirms that for Kung-sun Lung the world is the
whole from which we point things out. Applying our interpretations of
'world* and of chien, we find that the opening sentences of the refutation
assume an unexpected coherence and have the same theme as the Chuang-
tzii ch. 25 passage:
"That nothing within the world is the meaning originates from each thing
having its own name and not being deemed the meaning (the item pointed
out by 'world*). When the items not deemed to be the meaning are pro
nounced to be the meaning, we are speaking collectively of the items not
deemed to be the meaning. It is inadmissible to go from that of which the
components are not deemed to be the meaning to something of which every
component is deemed to be the meaning.
Moreover meanings are collected together by the world. That nothing
within the world is the meaning is in the case of things it being inadmissible
to pronounce that nothing is the meaning. . . . "
Here we may look first at the phrase pu wet chih. If we take the essay
as a metaphysical discussion of the relation between meanings (or univer
s a l , or logical classes) and things, it is possible to take pu wet chih as 'is not
deemed to be a meaning" (is deemed to be something else, such as a
thing). But if the theme is a paradox arising from the relation between the
item pointed out as world and items pointed out from it, between the
meanings of 'world* and of 'ox* or 'horse*, we can take pu wet chih only
as 'is not deemed to be the meaning* (of 'world*, the word introduced at
the head of the sentence). Similarly "IIlls will be 'may be pronounced
to be the meaning* and ^ J B 'is not the meaning*, as I already appreciated
in my earlier study. But why is it that the refutation introduces the new
formula pu wei chih, while the argument used fei chih alone ? We noticed
in § 1/3/3, 1/4/32 that 'wei X* implies satisfying the conditions for being
deemed an X , and is not equivalent to ' X yeh' ('is an X*). In B 3 we are
told that a yung fu I I ^ 'bold fellow* is not being deemed to be a fu
'husband* ( ^ S ^ ) , although it is not denied that he may well be a husband.
2/4 463
Things are deemed to be what is meant by such names as 'ox* or 'horse*,

not 'world*; nevertheless "no thing is not the meaning'* ( ^ ^ b ^ t i t ), a state
ment which recurs in the refutation as well as in the argument, and which is
not synonymous with " N o thing is not deemed to be the meaning". The re
futation is insisting on a distinction in the part/whole relationship which we
may suspect that the defender of the sophism overlooked, that although
for example a finger is not deemed to be a man, there is a sense in which
we can say of a man's finger that it is the man. (Cf. N O 1 "Therefore one
man's finger is not one man, but this one man's finger is this one man".)
The affinities between Pointing things out and the Mohist Explanations
raise the question whether the document may actually come from the
Mohist school. (Except in the case of the White horse, we cannot be sure
that the forger of Kung-sun Lung tzii had authority for ascribing his
fragments of ancient disputation to Kung-sun Lung.) But it is more plau
sible to take the affinities as evidence that the earlier sophists had more in
common with the Mohist dialecticians than we should guess from the other
surviving materials, the White horse and the lists of paradoxes. Pointing
things out shares some linguistic features with the Canons (such as shut 16
'which?'), but it does not observe the grammatical limitations which we
examined in § 1/3; short as it is, it uses erh rf5 after the subject, ho chih
'go to', shih $i 'supposing', all unattested in the language of the summa.
Fung Yu-lan noticed two trends in early disputation (Fung (1958) l/214f),
associated with the names of H u i Shih and Kung-sun Lung respectively,
the former arguing that all distinctions are illusory, the latter that even the
hardness and whiteness of a stone are separable. The sophism that it is
self-contradictory to claim that names point things out from each other
would belong to the school of H u i Shih, and Kung-sun Lung would be the
proper person to refute it.
A l l previous translations and expositions of the essay, whatever their
interest as philosophical exercises, can be rejected on purely philological
grounds. Criticising five earlier versions in my earlier paper I noticed "the
forced and inconsistent manner in which the key terms and constructions
of the original are reproduced", and added that "These inconsistencies are
not due to carelessness; each of the five scholars would obey the ordinary
rules of translation if he could, but his interpretation of the thought forces
him to do violence to the language. From one point of view, such weaknesses
are encouraging. If several conflicting interpretations yielded equally
acceptable renderings, we should be driven to the conclusion that the
problem is insoluble. But the failure of previous attempts suggests that a
translation which is linguistically acceptable throughout in itself guarantees
that the interpretation of the thought on which it depends is essentially

sound". But my own earlier version, although not I think linguistically
unacceptable, must also be rejected on philological grounds; it overlooks
the affinity with Mohist Explanations divided between argument and
refutation, and it resorts to too many conjectural emendations. The more
recent version of Cheng Chung-ying and Richard H . Swain is open to the
same two objections, and in addition slips back into the old assumption that
Kung-sun Lung did not know how to write grammatical Chinese:
9 9
Cheng and Swain 140 %T(\fn)*M% ^Mi%¥ ?
"If (acts of reference) are not objects of reference, and there are no concrete
things, how can one speak of objects of reference?"
Ut sup. 141 X T « S » « * » « f ^ # * * ^ » J & * ·
"If there were no acts of reference, then the fact that each concrete thing
has a name would not constitute an act of reference."
In the translation we shall bind ourselves by the following rules:
(1) Each key word or phrase must be given the same sense through
out. Except when chih 'meanings' are described as chien 'collected together*
or are expressly related to wu 'things', we take chih throughout as the mean
ing of 'world*. Fei chih 2£JB 'is not the meaning* (of 'world'): pu wet chih
^f§f!f 'is not deemed the meaning' (of 'world'): k'o wet chih ^TIBJif 'may be
said to be the meaning' (of 'world'). We take wu M throughout as 'not have
within it/not have as a component' (always referring to the world). We
take chien ^ in both occurrences as a transitive verb, 'collect together'.
(2) Each syntactic construction must be reproduced consistently:
X m Y •& ' X is Y ' : XffDY 'When X , Y \
(3) We follow the Taoist Patrology text without any textual emenda
tions.
(4) We take as the ultimate test of a translation its success or failure
in organising the essay as a coherent argument followed by a coherent
refutation. Once its structural affinity to Explanations B 38, 73, 82 is
recognised, there can be no question of treating it as a more or less haphazard
collection of dicta about chih and things.
9 692
XT«st^ 9
mîmvtfmm · mm^ ?
r » j -tfe# » X T ± W « 4 o r ^ j t t t ' xTzmmfo ·
59s
KiXTzm
xymm •ffiVff^mmteo ^mnm% r 9

J* · 9
r mm
2/4 465
591 Yiieh's emendation of fffi to M (followed G ( l ) 300) is unnecessary

if we take the erh as emphasising the subject as in Analects 3/22 HfR
lTn$flft, fft/F$nit " I f even Kuan knows the rites, who does not know
the rites?" (although the usage is admittedly unattested in the later
Mohist corpus and in other fragments of the sophists), and wu as
verbal, 'treat as a thing*, as in Chuang-tzu ch. 20 (Kuo 668/5) ^J^ffiff
"Treat things as things and not be made into a thing by
things". In this construction the wu would be passive, "be treated as
a thing* (p. 160 above).
The two sentences are traditionally punctuated X T M H , %M*S№L
5 9 2
m%. * f B # , XTtm&, n I 3 ! f § ¥ . In G ( l ) 300 I rejected this punc

tuation on the grounds of parallelism with the later sentences № ^ ~ F
. . . $.%~TM% . . . (a parallelism which now seems to m
deceptive, as to Cheng and Swain), and for two syntactic reasons:
(1) Throughout pre-Han literature k'o-yi precedes an active and k'o
a passive verb (cf. § 1/3/12/4/2), so that we cannot take the second
clause as 'there is nothing by which a thing may be called a thing".
(2) The phrase ^ f a ^ f is left unattached to either sentence. Cheng
and Swain 151 defend the traditional punctuation, but without
answering either of the syntactic objections.
593 Y en quotation device (§ 1/3/10), taking a word as used in its pre
c n 6 f
ceding context (as in A 64, B 31). On the present interpretation it

marks the chih as that of the proceding sentences, 'what we are pointing
out* (as the world), not chih in general ('items we point out from each
other').
594
Fei yu 'It is not the case that there is*. Cf. Kung-sun Lung tzu ch. 2
(Ch'en 61/1) XTf&&te&Z&& "It is not the case that there are
colourless horses in the world".
5 9 5
One is tempted to delete the erh as syntactically inoperative; but the
point is perhaps that the quotation from the initial proposition of the
essay is being completed.
Sophism. When no thing is not the meaning, the meaning is not the
meaning.
Defence. When there is no meaning one thing rather than another within
the world, no one can call a thing not his meaning. By treating the world
itself as a thing, might one say that it is one's meaning ?
As for the 'meaning', it is nowhere within the world: as for a 'thing',

it is somewhere within the world. It is inadmissible to deem what is some
where within the world to be that which is nowhere within the world.
When the world does not have the meaning within it, a thing may not
be said to be the meaning. That it may not be said to be the meaning
is its 'not being the meaning'. Its not being the meaning is 'no thing not
being the meaning'.
That, when the world does not have the meaning within it, a thing
may not be said to be the meaning, is nothing not being the meaning.
Nothing not being the meaning is 'no thing not being the meaning'. 'No
thing not being the meaning' is 'the meaning not being the meaning'.
The argument is very abstract, and we can best come to grips with it
by taking ox and horse as the characteristic 'things', as is usual in Chinese
disputation. By names such as 'ox' and 'horse' we point things out from
each other. When, ceasing to point things out, we speak of the 'world', we
cannot say of any thing that it is not what the name points to. But how
can we treat the whole world as one more thing which a name points out ?
The things pointed out from each other are components of the world, but
there is no component of the world pointed out by 'world'.
This involves a paradox: each thing is not what 'world' points to
(an ox is not the world, neither is a horse, neither is a . . .), yet the con
verse of this is that no thing is not what 'world' points to (oxen and horses
and . . . are the world).
For the sophist it is contradictory that if each is not none is not, and
if none is not each is not; he is overlooking the distinction between
distributive and collective which engages the Mohist (as discussed under
A 2, B 12) and will be explored in the Refutation. For him it seems to
follow that oxen, horses and the rest both are and are not what 'world'
points to; "the meaning is not the meaning".
r XT mm J m»tk™i&m±m% > · ûmmn±m >

»%Tzm& r %rmm J m m^mmm& ° ^mmmn ·
o
9
r J •& · mm%%» r mmm j · r #«t j & » m m
mm»wm r^wtj mm r m®m J · 9

2/4 467
For sheng yii 'is born from* in logical contexts, cf. E C 2, B 53.
Unless we emend to H with Y i i Yueh, chih must be taken as the verb
'go to* (not used in the later Mohist corpus).
Shut 'which ?', not necessarily of persons (p. 135 above). For an example
cf. B 41, which also has an example of ching 'directly, prematurely*.
Ch'ieh-fu is not a compound in pre-Han Chinese; the topic-marker fu
before common nouns generalises them, so that the reference is to
'items pointed out from each other* in general.
Refutation. That the world does not have the meaning within it derives
from each thing having its own name and not being deemed the meaning.
When though they are not deemed the meaning we say they are the
meaning, we treat as a collection the items not deemed the meaning. It is
inadmissible to make a jump from that which has items within it not
deemed to be the meaning, to something within which every item is
deemed to be the meaning.
Moreover, meanings are what 'world* treats as a collection. The world
not having the meaning within it is in the case of the things it being
inadmissible to say that it does not have the meanings within it. It being
inadmissible to say that it does not have the meanings within it, is 'nothing
not being the meaning', and nothing not being the meaning is 'no thing
not being the meaning'. It is not that 'the meaning is not the meaning',
it is that a combination of meaning and thing is not the meaning.
Supposing that the world had within it no meanings which are things,
what would you be thoughtlessly calling 'not the meaning'? Supposing
that the world had within it no things, what would you be thoughtlessly
saying 'is the meaning' ? Supposing that the world did have the meaning
within it, but no meanings which are things, what would you be thought
lessly calling 'not the meaning', thoughtlessly saying 'has no thing within
it which is not the meaning' ?
Moreover, if the meaning inherently and of itself is deemed 'not the
meaning', how is it that it is deemed the meaning only in combination
with things on which it depends?
The objector replies by introducing the concept of chien 'collection'

and the distinction between fei 'is not' and pu wet 'is not deemed to be',
tools which he shares with the Mohist dialecticians. He shows that the
argument mistakes the collective function of 'world' for a distributive
function. Each thing has its own name and is deemed to be 'ox', or 'horse',
but not 'world*; what we call the 'world* is the collection of all of them.
The fallacy was to suppose that if the collection is deemed to be the 'world*
each of its members should likewise be deemed to be the 'world*.
Consequently the paradox disappears. Each thing is not deemed to be
what 'world* points to (pu wet chih) ; an ox is deemed an ox, not the world;
a horse is deemed a horse, not the world. But it remains true that no thing
is not what 'world* points to (fei chih) ; oxen, horses and the rest do
compose the world.
In the sophism "the meaning (oxen, horses . . .) is not the meaning
(the world)**, his objection is therefore to the first 'meaning*, which he
corrects to 'combination of meaning and thing* ( JeiSi'^J) or 'meaning which
is a thing* (%fe). The name 'world' simply points, the names 'ox' and
'horse* combine pointings with things. He asks three searching questions
about the first 'meaning', showing that the sophist cannot maintain that it
is simply what 'world' points to:
(A) Is it denied that the things are being pointed out from each
other as 'ox' or 'horse' ? Unless they are, the sophist has no right to say
that they are not what 'world' points to.
(B) Is it denied that the world is composed of things? Then the
sophist has no right to hold that they are what 'world' points to.
(B) Does he suppose that, without any 'ox' or 'horse' to point them
out, the name 'world' itself points them out? This raises again the diffic
ulty of A, that he ought not to say that they are not what 'world' points
to, but also an additional difficulty, that he can no longer claim that no
thing in the world is not what 'world' points to.
Finally, if the first 'meaning' (oxen, horses . . .) is, like the second,
simply what 'world' points to, how is it that it is differentiated from the
second (the world) precisely by its dependence on the things with which
it combines?
Both the argument and the refutation could be presented throughout
in terms of the gesture of pointing. It would seem however that the
debaters are thinking primarily of the meanings of names, because the
beginning of the refutation refers specifically to names, because one can
hardly 'point out' the world as a whole except by the name 'world', and
because the essay refers throughout to wu 'things' (general categories of
things, cf. § 1/4/33), not to shih 'objects' (particulars). The Mohist, who
confines chih to the gesture of pointing, refers only to the pointing out of
objects (B 53). Even in the two passages from Hsiin-tzu's Right use of names
which we quoted above, which do refer to names, we are said to use names
to point out objects.
2/5
NAMES AND OBJECTS
Names and objects is the one document in the corpus which is not
organised as a series of canons and explanations and which cannot be seen
as part of a single summa. It is a consecutive treatise on the art of naming
the similar similarly and the different differently, and re-examines the pro
cedures of consistent description (A 88-B 12) in the light of a new discovery,
that the tz'u i $ 'sentence/proposition' is something more than a string of
names. The first half of the document as we reconstruct it is as dislocated
and mutilated as Expounding the canons, patched together from fragments
in the Ta-ch'u and the first quarter of the Hsiao-ch'ii. For the rest we are
on firmer ground, in the long homogeneous passage which survives intact
as the main body of the Hsiao-ch'ii (NO 12-18).
Although the fragments of the damaged half cannot be rearranged
with full confidence, they can be seen to fall into three groups, one of which
follows immediately on to the title (NO 1-5) while another leads directly
into the long intact passage (NO 9-11). The argument follows the same
direction as the Canons on consistent description, which explained in
A 88-98 the grounds for deeming something so of 'the object here' (tz'u jlfc),
and from B 1 how one 'proceeds' (hsing ff) to the 'kind' (lei S ) , from what
is so of a particular object to what is so of the thing it is judged to be
(shih jk). The one change in terminology is that Names and objects no longe
distinguishes the particular by the demonstrative tz'u. Its treatment is also
much less abstract; it gives a variety of concrete objects described as shih X
'this X*, yi — X 'one X*, shih yi X 'this one X ' , and also supplements the
generalising shih with mou 'such-and-such*.
We may identify the themes of the three sequences as follows:
N O 1-5. What is so of particulars.
N O 6-8. Sameness and difference in being this and being so.
N O 9-18. Whether what is so of a particular is so of 'this', what it is
judged to be.
The term pien 'disputation* is used to cover several kinds of argu
mentation (NO 6), of which only the first is the disputation proper of the
Canons. This, the fourth of the Mohist disciplines, does not appear at all
470 Names and Objects
in the treatise. The theme throughout is the art of description, the relating
(ho ·£·) of names to objects; indeed the very first sentence is "Names and
objects do not necessarily go together (ho)".
NO 1
T C 4A/2-4 ° i g * ^ ( £ ) * £ «"> ° » fc^SftSi&l
e^«^S&* ^:'îclll 5A/10-5B/2 8 0 1 o
f£?m$Slhtr
N0 2
T C 3A/3-4... »#frfejw» 2
> °
4A/4-9 mifcrnnfo o «« 8 £ & °
rmj r » j r a u # · ° msmm^ »^rtuJ r£j
N0 3
sn*i%» T C 3B/2-4... ât-fess-ai ° ' #ata±*
-fe ° * f ^ A * ^ « A - f e ° jft8№755ft*-fe«>« -
N0 4
T C 4B/2-3 — B7WtifD& ' —B75Jtifn*& · H B S » E3B
NO 5
H C 6B/3-5 T $ J ° ° &#£S£?£til
6 0 0
Emended following Chang Ch'i-huang 5B/5. Cf. A 80 %MIS, I" £ J ft
"The mating of name and object is ho".
6 0 1
5«« (=81), 'although*.
602 Y u "y u x> (§ ^3/6).
y i i x h a v e s o m e
6 0 3
Yen 'only then', cf. § 1/3/11/5.
8 0 4
Huo 'catch (of game)', M . 20758 def. 1/3. For the syntax of shih huang
yeh "this half-disc", yi chih chihjen yeh "visualise the finger as being
the man", yi huo yeh "visualise as being game", cf. pp. 155f above.
4 0 8
For wei 'be deemed' without a complement, cf. p. 118 above.
N O 1 "Names and Objects. Name and object do not necessarily go together.

If this stone is white, when you break up this stone all of it is the same as the
white thing; but although this stone is big, it is not the same as the big
2/5 471
thing. In all cases of naming otherwise than by reference to number or

measure, when you break up the object all of it is the thing in question.
Therefore one man's finger is not one man, but this one man's finger is this
one man; one side of a cube is not a cube, but one side of a wooden cube
is the wooden cube.
N O 2 . . . To have a certain Ch'in horse is to have a certain horse and to

know that it is the horse of someone coming from there. In this ['Ch'in
horse'] there is a term of convenience ['Ch'in']. In cases of naming on the
basis of shape and characteristics ['horse'] we necessarily know that this
thing is ' X ' , only then do we know ' X ' ; in cases where naming cannot be on
the basis of shape and characteristics ['Ch'in'] we may know ' X ' even if we
do not know that this thing is ' X ' . In all cases of naming on the basis of
residence and migration, if they are things which have entered its confines
all are of this place; if they have left this place, by this criterion they are
not. Of naming on the basis of residence or migration, district and village
names, 'Ch'i', 'Ching' and the like are all instances; of naming on the basis
of shape and characteristics, 'mountain', 'hill', 'house', 'shrine' and the like
are all instances.
N O 3 Knowing is different from having a pictorial idea . . . but this half-

disc is this jade. Picturing a pillar is not picturing wood, it is picturing the
wood of this pillar. Picturing a finger as being the man is not picturing a
man; picturing as being a catch of game is however picturing the birds.
N O 4 In the first case we say that if the instanced is this ['the stone'] some
thing is so of it ['white'], in the second that though the instanced is this
something is not so of it ['big'], in the third that it has changed in place
['Ch'in horse'], in the fourth that it. . . .
N O 5 'Some' is not all ['stone' when broken up]. The loan-named ['Ch'in
horse'] is not now so. A n example ['pillar'] is a standard for being deemed
such-and-such ['wood']; the thing exemplified is the standard by which the
example is deemed such-and-such. Therefore if something coincides with an
example, it is this thing, and if it does not it is not; this is exemplifying."
We reconstruct this sequence on the basis of the resumptive passages

N O 4, 5. The first two items of N O 4 are close in wording to N O 13 ^ c % ,
sKTijiktTn^, ^ S S f f n ^ ^ " O f a thing in general, something may be so if
the instanced is this thing, or not so even though the instanced is this
thing . . .". But the latter is distinguished by the fu wu 'thing in general'
exposed in front of it; here the instanced 'being this' is part of a particular
object being the object. For nai shih 'is this (of an indefinite number of
things to bey, cf. p. 118 above.
The Mohist distinguishes three ways of naming (there may of course
have been more in the missing parts of the treatise):
(1) "Naming on the basis of shape and characteristics (hsing mao)."
The explicit examples are all of naming by shape ('mountain', 'hill',
'house', 'shrine', and in N O 7 'sword'), the configuration of the body; the
surface characteristics called mao (cf. § 1/4/20) would presumably be named
by such words as 'white'. With this kind of word we necessarily (pi) know
what the object is if we know the meaning of the name; it is only if we
recognise mountains when we see them that we know mountains. (In the
case of 'Ch'in horse' on the other hand, I can know all about the land of
Ch'in without knowing that the horse in front of my eyes is a Ch'in horse.)
(2) "Naming by reference to measure or number", for example 'big'.
We are explicitly told that this is the only category in which what is so of
the whole is not so of the parts; the parts of a big white stone are white but
not big. Since he has examples elsewhere which superficially contradict this
generalisation (cf. the blind horse and yellow ox of N O 18), he is presum
ably thinking of ideal objects which fit their names exactly, not merely in
the appropriate respect (to use the terminology of A 96, 98). The distinction
between so of a part and so of the whole leads him to supplement the old
definition of 'all' (A 43) by a definition of huo 'some', presumably in reference
to the t'i 'parts'. (Cf. A 46 R W ^ ^ ^ F "of its parts some are removed
and some remain".)
(3) "Naming on the basis of residence or migration", for example
"Ch'in horse". The Mohist takes the position that one is not actually 'of
C h ' i ' unless resident in C h ' i ; after leaving the state one may still be called
a man of Ch'i, but this is merely a "term of convenience" (pien wet), an
example of the chia 'loan-naming' of B 8. The point that distinguishes
loan-naming is that what is so of the thing properly so called is 'not so'
(pujan, the summing-up of B 8); and in the case of a migrant from a state
it is "now not so" (chin pujan). By this usage he seems to abandon the
unqualified rejection of loan-naming as self-refuting (pet W) in B 8.
We find an adverbial use of yin 'by this criterion' in N O 2 and again
in N O 17 (it is also used verbally in N O 9). The syntactic mobility does not
imply any change in the concept of yin, defined in A 71 (if our emendation
is correct) as $ ? " w h e r e it is so". The criterion for being a black man is
the part of the body where you should be black, the criterion for being a
man of Ch'i is the place where you should be residing; if you have left the
state, "by this criterion you are not" (B^tfe).
2/5 473
In the case of a Ch'in horse I see that it is a horse and in addition

"know that it is the horse of someone coming from there". This leads the
Mohist to observations about the difference between knowing and having
a yi, idea or mental picture. B 57, 58 already noted that we have an 'a
priori' idea of the look of a pillar but not of a hammer, which is recognised
not by its shape but by its function. Here the Mohist makes a much more
fundamental point, that a mental picture is interprétable in more than one
way, as being pillar or wood, birds or a catch of game. This complicates
the older claim that an idea can serve as the/a 'standard' of what something
is (A 70) ; and he supplements the fa with the new concept of objects as hsiao
che 'exemplifies', which exemplify the standard by which something is X
and can themselves serve as standards for X . {Hsiao is similarly combined
with fa in the Great Appendix of the Changes, both as parallel with verbal
fa 'take as model' and in hsiao fa 'imitate a model'.) The recognition that
mental pictures are variously interprétable is perhaps the Mohist's crucial
step in his discovery of the proposition. The authors of the Canons and
Explanations never quite broke away from the assumption that ' X yeK is
simply the name ' X ' with an insignificant appendix evoking the idea of an
X . The author of Names and objects understands that a proposition affirms
that something is a catch of game, or is a pile of dead birds; it does not
present a mental picture which might be interpreted as either.
It may be noticed that in N O 1 the phrase ffiSîâl^ is translatable
either as "all is the same as the white thing" or as "all composes the same
thing with the whiteness" (the ho t'ung 'sameness in being together'
of A 86). But here the ambiguity of nominalised pat (cf. § 1/1/2/1/4) does
not affect the argument.
N0 6
H C 6A/9-6B/1 ' №\>№-]m£fi » #fâ«L£*5 > W\m±m >
6B/9 r &mm J ° r sëiB J °
T C 4A/9-4B/2 m u a » mm» mmzm » m&±m ° &m>Mim >
> t&zm » m&zm™° n&teM » * « £ f i ° t R U t
NO 7
T C 5A/6-10/2 S A ; i ( M - ) * # 6 1 0
M A Z I ^ » » « f ô # t i î » #cfô ° fë£
6 1 1 612
A * J i r g * ± A t e H » A £ i l # - « t i ï ' Wk ° fë&№fé&U
PI" . . .
NO 8
T C 3A/10-3B/2 fr№±mn*u-Z№n ° imR±^^W^^(M)
6 0 6
In this long sentence three rhymes ($2 chi^KpG, M /*/*LJ>G, M
#
jy// NGl3G) mark off three pairs of phrases before the conjunction
yen 'only then' (also used in N O 2. Cf. § 1/3/11/5). The pair of phrases
following the yen is unrhymed (Hi pi/*PIaR). I formerly took the yen
as final (G(5) 2), but the placing of the rhyme confirms that Sun
Yi-jang and T'an Chieh-fu were right in identifying it as the con
junction. The rhyme-word chi, which I translated 'successions* in
G(5) 2, is used of the crucial principles and institutions which bind
organised society together: cf. Mo-tzû ch. 11 (Sun 49/-4, on the five
punishments), 12 (Sun 54/-5, on officiais): Sg^l
"UM "They may be compared to the main thread binding a skein of
silk (cf. Sun 49/-4 n.), the main rope controlling a net". For mu
'describe*, cf. § 1/4/20.
6 0 7
For y eh che as a quotation device, cf. § 1/3/10. Shih yu 'This is as
with . . .* introduces a parallel example in A 75, B 27, 78. But here it
may refer to synonymity, as in Mencius 4A/1 T WM J 38 ï!rï?f tfe " Yi-yi
means t'a-t'a". Wu ch'i 'How can I . . .?* is unattested elsewhere in
the corpus.
6 0 8
At first sight this seems to be a list of 10 kinds of sameness. But on
closer inspection one sees that it consists of two sets of four (Nos. 1-4,
6-9), summed up in Nos. 5 and 10 respectively. The first four are the
types of sameness distinguished in A 86, summed up as 'sameness
with the same name'. Originally no doubt the arrangement was shown
by the layout of the text (as with the layout of the Canons at Stage 2,
the head characters of the Explanations, the titles of Expounding the
canons and of Names and objects itself).
6 0 9
6 1 0
Emended following Y u Yueh (Yu 206/-3). Y u (followed by Sun, 259/2)
also emends the second chih *L to yeh it*. ; but I agree with T'an
(T*an 230) in finding this unnecessary.
611
Chih chih jen yeh "The finger's being the man", cf. p. 155 above.
6 1 2
Chang Ch'i-huang (op. cit. 7A/1) is probably right in taking chiang
chien and t'ing chien as types of swords different in shape. But the
ordinary meaning of t'ing chien is "draw a sword" ( M . 12106/13), and
Sun understands chiang chien as "support oneself on a sword" (Sun
259/4).
2/5 475
613
Yeh must either be omitted or else supplied in the preceding phrase
also; the latter solution ("Willow-wood being wood and peach-wood
being wood are the same thing"; cf. the observations on the use of
yeh p. 155 above) is hard to reconcile with the earlier "The finger being
the man and the head being the man are different things".
6 1 4
Emended following Sun 256/8.
N O 6 "The purpose of disputation is

(1) by clarifying the portions of 'is-this* and 'is-not\ to inquire into
the principles of order and misrule:
(2) by clarifying points of sameness and difference, to discern the pat
terns of names and of objects:
(3) by settling the beneficial and the harmful, to resolve confusions and
doubts:
only after that may one by describing summarise what is so of the
myriad things, by assorting seek out comparables in the multitude
of sayings.
'This is as with calling it . . . * implies sameness, 'How can I call it
. . . V implies difference.
The sameness of the identical,
the sameness of constituents,
the sameness of the connected,
the sameness of the same in kind
— sameness with the same name.
Sameness in being separated off,
the sameness of the accessory,
sameness in being this thing,
sameness in being so
— sameness with the same 'root*.
There is the difference of not being this thing, there is the difference of not
being so. Sometimes the different is to be deemed the same and the same
is to be deemed the different.
N O 7 The sameness of a tall man and a short man is their characteristics

being the same, therefore they are the same thing. A finger as being the man
and a head as being the man are different things; a man's members are not
things which are as one in their characteristics, therefore they are different
things. A chiang sword and a t'ing sword are different things; a sword is
something named according to shape and characteristics, their shapes are
not as one, therefore they are different things. The wood of willow-wood
and the word of peach-wood are the same thing . . .
N O 8 . . . the circularity of a big circle and the circularity of a small circle

are the same. What a measurement which reaches just one foot does not
reach, and what a measurement which does not reach 1000 miles does
reach, are not different; as for their being the same in not reaching, it is to
the longer and shorter distances that it refers.''
The fragment on the purpose of disputation comes from the beginning

of the Hsiao-ch'u; we have no positive evidence of its location in Names
and objects. Its significance emerges only when we perceive that it is adapted
to the fourfold classification of knowledge by explanation which underlies
the organisation of the Canons (§ 1/6/1). It implies a wider conception
of disputation than that of the Canons, embracing three of the four
disciplines:
(1) "Clarifying the portions of 'is-this' and 'is-not' " ('Explaining
names'), the argumentation from the definitions of names as to whether X
is or is not an ox, which is the only kind of disputation recognised in A 74,
B 35. Its purpose, the Mohist insists, is "to comprehend the chi of order
and misrule", the principles and institutions which hold society together
as the chi (the main thread) binds a skein of silk. We are again reminded
that the ultimate purpose of disputation is practical, to equip us to approach
the problems of government with clear heads.
(2) "Clarifying points of sameness and difference" ('Explaining how
to relate names and objects'), the art of consistent description by which
we learn to call the similar by the same name, the dissimilar by different
names. The author, who is primarily interested in the sentence rather than
the name, formulates its purpose in a new way: "to discern the patterns
of names and of objects", the organisation of names in sentences (cf. N O 10)
and of objects in the world.
(3) "Settling the beneficial and the harmful" ('Explaining how to
act'), which is "to resolve confusions and doubts". This refers to the ethical
system of the summa, in which the basic defined terms are the 'benefit' and
'harm' of A 26, 27.
The remaining discipline, 'explaining objects', is still excluded from
the scope of disputation, which embraces only the theoretical preliminaries
to the sciences which "by describing summarise what is so of the myriad
things", and to the particularised argumentation which "by assorting seeks
out comparables in the multitude of sayings".
2/5 477
The next fragment begins by resuming the four types of sameness

y
distinguished in A 86. Sameness 'as units'(t i H ) , the opposite of which was
'not connected or attached' (A 87 ^ S I R ) , is renamed 'sameness in being
connected' (lien t'ung). Sameness in being together (ho ) is renamed chii
jl» t'ung, which may be taken as 'sameness in composing (the complete set)'.
This is intelligible if, as we tentatively suggested under A 86, 87, ho was
the sameness of properties such as hardness and whiteness composing one
stone. (Cf. N O 1 S H H S I R I "all of it is the same as the white thing".)
But Sun Yi-jang may well be right in identifying the word as verbal chii {ft
'be together' (although this chii is firmly attested in the corpus only as
pre-verbal 'all, both').
These four are classed as 'sameness with the same name'. But the
author of Names and objects has perceived the difference between a complex
of names and the sentence, which he will introduce in N O 10. He proceeds
to distinguish four more types of sameness, under the mysterious heading
of 'sameness with the same root'. The two more important ones (the only
ones for which he mentions corresponding types of difference) are 'sameness
in being this' and 'sameness in being so'. We may identify them in the light
of the examples of "being so if the one instanced is this" (TbM\f\]$s) in
N O 4 and N O 13:
(Being this) (1) N O 14 " A white horse is a horse"
(2) N O 14 mmm-fo " A black horse is a horse"
(Being so) (3) NO 11 ?5;J№tfe£ " I f this stone is white"
(4) NO 1SS-tMA "Although this stone is big"
(5) NO 14 m&m "Rides a white horse"
(6) NO 14 mmm "Rides a black horse".
Then 'sameness in being this' will be the sameness of white horse and
black horse in both being horses (Nos. 1, 2), and 'sameness in being so'
their sameness in that it is so of both that one rides them (Nos. 5, 6). The
'root' which is the same will be
Nos. 1, 2 JSife "is a horse"
5, 6 HI "rides"
and the roots of the sentences without parallels will be
No. 3 ft "is white"
4 A "is big".
The root of the sentence then is the complement of the nominal or
the main verb of the verbal sentence. We shall find this hypothesis further
supported when we come to the description of the proposition in N O 10.
Of the other two types, ch'iu is the ch'ii E of the Canons (§ 1/4/7), as
Sun already recognised. Ch'u is to mark off from other things, to group on
1
grounds not of similarity but of difference from the rest: cf. A 73 / L 4 ,
1
flis^^ , W*tfe " A l l oxen, and non-oxen separately grouped, are the two
sides". Ch'iu t'ung should therefore be sameness in not being this or not
being so. Sun also identified the fu 1ft offu t'ung as fu ffli
'attach/, and found
a case of a proper name F u written with both graphs (Sun 257/-4). Neither
graph appears elsewhere in the dialectical chapters. But it may be noticed
that the only point of similarity or difference in Nos. 1-6 which is not a
matter of being this or so is the difference between 'white horse* and 'black
horse'. In "The horse is white" and "The stone is white" horse and stone
would be the same in that being white is so of both of them; and the
sameness would remain if we cut off the root of the sentences and attached
it to 'horse' and 'stone' to make 'white horse' and 'white stone'. We may
guess that this was the significance of fu t'ung.
The fragments of N O 7, 8 are difficult to sort out. N O 7 seems to be
concerned with sameness in being this. In N O 8 yuan 'circular', like chih
'reach', is verbal (§1/5/10), so the theme is presumably sameness in being so.
N0 9
C 10/2A/7-10 H S ^ W H b O J | g f H i f n t * * « ' ' k ( t e ) * £ £ - l b · g P
6 1 6 # 6 1 7 6 1 8 6 1 9
T C 4B/4-5 < @ ^ J ? > ^(04) « ' i£H^ «£ MX *^
62o
mm»£Si§ «iE62io
N O 10
5feSr T C 5B/2-6 №k*k » » &Mu|*j#l 622
o ft»iffi*HJJiB&t
< « > 6 2 3
» o ^Kwmmwu o m^mvm^mm»*
« c .is. . the fragment which was incorporated in the Canons as A 89-92

This
(p. 109 above). We emend yeh to chih on grounds of parallelism. The
last sentence contains the only example in the corpus of pien as a
quality of mind, 'subtle in making distinctions, logical in ordering
ideas', as in Lu-shih ch'un-ch'iu ch. 18/5 (Hsu 842/5) < ^ ^ t t B " K ± H 5
T£$S "Kung-sun Lung spoke on Jack having three ears with extreme
subtlety".
These are the two words written as a gloss in the Canon A 94, which we
take as the broken off end of the strip incorporated as A 89-92
479
(p. 110 above). They fit very neatly at this position, where the
parallelism requires two words. Li 'unobstructed' ( M . 1932 def. 2. Cf.
A 26, n.) is used as in the k'ou chih li 'fluency with the mouth' of
the longer fragment (yin li "the sound is uninterrupted").
617
Tz'u 'the next' and tuan 'starting-point' (the graph here partly obliter
ated, but surviving in the next sentence) are geometrical terms
defined in A 61, 69. Pi 'compare' is similarly the counterpart of the
geometrical pi 'put side by side' defined in A 68. I propose, as a
hypothesis which makes sense of this obscure passage, that the Mohist
has adapted tz'u and tuan for the sequence-positions and pauses of
the sentence.
6 1 8
Yin chih "by the criterion of its reach", cf. N O 12 Vf'ZW^SiSffnIE
"The parallelism of propositions is exact only as far as it reaches".
The yin 'criterion' is sojan "where it is so" (A 71); the criterion for a
comparison is a place where the two propositions do or do not
correspond. For chih used nominally as a technical term, cf. L i u
Hsiang 9m (79-8 B . C . ) , Pie lu Sm ap. Shih chi chi-chieh A f H * »
ch. 76 (Shih chi 2370 n. 2) ffi^T£J№firE»H£, ffflSteUT "The
disputation recognised throughout the world has 'five wins and three
arrivals', and correctness in phrasing is the least". (Unfortunately the
categories of chih 'arrival' are not enumerated. For the complete
passage, cf. p. 20 above). Yu chih "amplify the meaning"(?). For chih,
cf. Chuang-tzu ch. 22 (Kuo 750/9) r JUJ T J T ^ J H # , H £ ^ K ,
W it—ffe "Chou, pien and hsien, these three, are different names for
the same object, their meaning is one", and pp. 458f above.
6 1 9
Fu 'add name to name' as in B 11 (§ 1/4/11).
6 2 0
Yin ch'ing " B y the criterion of the ch'ing (what the thing is in itself, as
exhibited in its definition, cf. § 1/4/6)".
6 2 1
Fu cheng 'compound or simple'. The familiar expression cheng ming
J E £ "giving things their right names" (of which there is an example
in B 68) would make cheng a suitable word for referring to the single
name which is the right one, although I do not know other examples.
Hsun-tzu also distinguishes between compound and single names, but
with a different terminology: chien %. 'double', tan 'single' (ch. 22,
Liang 314/4).
6 2 2
Invert yeh che, following Sun (259/8), on grounds of parallelism with
the next sentence beginning fu tz'u "The proposition . . .".
6 2 3
So sheng <chang> "that from which it is engendered and becomes
full-grown"; the restoration is required by the two phrases to which
it refers, yi ku sheng, yi li chang "is engendered according to a reason,
becomes full-grown according to a pattern". Sheng chang "be born

and grow up" is a common phrase ( M . 21670/275).
N O 9 "Hearing is the sensibility of the ear; to attend to what you hear so

that you grasp the idea is discernment by the mind. Saying is the flow of the
tongue; to make a case for what you say so that the idea can be seen is the
subtlety of the mind. <When the sound flows consecutively > , discern the
sequence-positions and starting-points; comparisons, with the degree of
correspondence as the criterion, amplify the meaning. When sequence-
positions are compounded, discern the starting-points of the voice; names,
with what is essential to the thing as criterion, are compound or simple.
N O 10 The proposition is something which is engendered in accordance

with the thing as it inherently is, becomes full-grown according to a pattern,
and 'proceeds' according to the kind. It is irresponsible to set up a proposi
tion without being clear about what it is engendered <and grows up > from.
Now a man cannot proceed without a road; even if he has strong thighs
and arms, if he is not clear about the road it will not be long before he gets
into trouble. The proposition is something which 'proceeds' according to
the kind; if in setting up a proposition you are not clear about the kind, you
are certain to get into trouble. Therefore . . . . "
Names and objects has now arrived at the point which the corresponding
series of Canons reached at B 1; having 'separated the roads' (A 97) we
'make the man proceed' (B 1), from what is so of particulars to what is so
of the kind. The same key words reappear, tao 'road', jen 'man', hsing
'proceed', lei 'kind', as well as the/)/ 'comparing' of B 6. What is quite new
is that the author of Names and objects sees clearly that what 'proceeds' and
is compared is not a name but a proposition. The word tz'u 'proposition'
appeared once in Expounding the canons (EC 1), but in its ordinary sense
of language deliberately phrased for effect; this is its first introduction as a
technical term (cf. § 1/4/31).
The rest of Names and objects is devoted to the detailed comparison
of parallel sentences. It appears from one fragment in N O 9 that to analyse
their correspondences the author uses terms which (like English 'parallel')
derive from geometry. Although the context is obscure it cannot be a
coincidence that the three words pi 'compare', tuan 'starting-point' and
tz'u 'sequence-position' all occur both in two successive sentences of N O 9
and in two adjacent sections on geometry, A 68, 69. We place sentences
side by side and compare them word for word, beginning at the pause, the
2/5 481
'starting-point of the voice', just as with the comparison of measurements,

which "is possible only with a starting-point for both" (A 68 MWSffiS
"Pi). The test of parallelism is how far the sequence-positions (tz'u) tally,
how far the correspondence reaches (chih, as in N O 8 "What a measurement
which reaches just one foot does not reach").
The proposition is said to be engendered (sheng), become full-grown
(chang) and proceed (hsing), like a man proceeding on a road, knowledge
of which is indispensable "even if he has strong arms and thighs". This
suggests an implicit comparison with the engendering and full growth of
the human body. But how are we to understand this odd organic metaphor ?
If we are to take it with full seriousness, it would seem that a proposition
has two parts:
(1) The part first engendered, " i n accordance with the ku", a word
which covers reasons in general but would here refer primarily to the
described thing as it is in itself; this is the part which would be true or false.
(2) A part which completes it "according to the li", the pattern of
pauses and phrase-positions (comparable, within the image of the man
setting out on his route, to the articulation of his thighs, arms and other
members). That this is the significance of li can be seen by collating the
four examples of ch'a 'discern, scrutinise':
NO 6 "discern the li of names and objects"
NO 9 "Attending to what you hear so that you grasp the idea is
discernment by the mind"
"discern the sequence-positions and starting-points"
"discern the starting-points of the voice".
Then the part first engendered is precisely the part called the root
(ken Wt) in N O 6, the complement of a nominal or main verb of a verbal
sentence. We can easily understand how the uneliminable core of the
Chinese sentence, expandable by the addition of dispensable units such as
the subject, would be seen by the Mohist as the 'root' from which it grows.
This would be as natural for Chinese as it was for Greeks to think of the
assertion-making part of the sentence as 'predicated' of the subject (from
which the Indo-European verb takes its person and number). Returning
to the first of the six sentences we analysed under N O 6, the root will be
Hife "It is a horse",
which is judged true or false by comparison with the ku, the horse itself.
#
This connects with the phrase in A 39 1S "The pillar's engendering
of the ti ('complement', not of course in the grammatical sense)", where
we took the ti to be the name 'pillar' applied to the object. The sentence
grows from this root to its full length as
SSS-tfe " A white horse is a horse",
"according to a pattern", which is that of compounded names before a

pause, mentioned in N O 9 'ffcfc, S F S S "When sequence-positions are
compounded, discern the starting-points of the voice" (cf. Kung-sun
Lung's White horse: , t £ J S J "Putting together 'white' and
'horse', one compounds the names as 'white horse' ").
For the Mohist the //, the pattern or structure of the sentence, would
reveal itself only by comparing parallels and noting the same or different
words at corresponding phrase-positions (even a modern grammatical
analysis of the Chinese sentence depends heavily on this method). Then
the four types of sameness distinguished in N O 6 would show up at
different positions:
1 2 3 4
W H I T E H O R S E H O R S E yeh RIDE WHITE HORSE
B L A C K H O R S E H O R S E yeh RIDE B L A C K HORSE
The sameness at Position 3 is in being this, at Position 4 in being so;
the difference at Position 1 is perhaps that of accessories (fu fitt). We can
now appreciate the full significance of N O 6 ?f3fp]pj;£Ji, %&%ZM
"By clarifying the points (/places) of sameness and difference, discern the
patterns of names and objects". The recognition of sameness or difference
at corresponding positions reveals relations which are both between the
words (as grammatical or as logical relations, which we should not expect
the Mohist to distinguish) and between the objects described.
N O 11
H C 6B/1-3 si^mn»&sm«'Ki&tti#t»mm»· * » a
6B/5-8 r » J 6 2 4
*ib 9
VmWSi2& ° [" # J
6 2 5
ife^ 6 2 6 9
ttlffiffiiRfTte ° r « J fe* · s r ^ ^ S ^ R ^ J i f e o
6 2 7
9
r* J ^m^zn^^m9 =fZM °
N O 12
H C 6B/9-9A/2 ^m^mW^mmU ° » £ # * # 0 f £ i i n IE O Jfcfctt
6 2 8 9
° < » & > ^ KwaafcfofciRi ° K * ^ * *
6 2 9
< §f > B№£ - > &m\№tt*S№ · fifths
mzmnmm nm&»m\fn$t m^m^m^m^k * « r
9 9 9
2/5 483
624 p v ( ^ , M . 38642 def. 9), 'offer an illustrative comparison*. Cf. Ch'ien

=
fu lun mj&k (ch. 29) BSS 191/5 £&lB£;£W8,

^ à l ^ S i ^ ^ è . "Illustrative comparisons derive from something not
being clear when you tell it directly; you therefore borrow another
thing of which it is so or not so in order to illuminate it".
«25 T > 0 ( ffe) 'another', as in B 3 and N O 15, 16 below.
=
626 Mou 'equal', adapted for the 'equalising' or parallelising of propositions.

Cf. Chou li mm ch. 42 (SPPY 42, 12A/-3) "The distances
apart must be equalised". The Mohist's usage seems to be otherwise
unattested.
627
Yuan 'adduce', as in A 93. It is to adduce X in support of Y , while t'ui
'infer' is to draw Y as a conclusion from X . In their primary senses
yuan 'pull' and t'uei 'push' are opposites, paired in Lii-shih ch'un-ch'iu
ch. 7/2 (Hsu 298/1). I am not aware of other pre-Han examples of
yuan 'adduce', but it is common from the Han onwards in such
phrases as yuan-yin lie31 'adduce' ( M . 12407/2), yuan li 'adduce
a precedent' ( M . 12407/49).
628, 629 Restored on grounds of parallelism.
N O 11 "One (A) uses names to refer to objects, (B) uses propositions to

dredge out ideas, (C) uses explanations to bring out reasons, and (D) accepts
according to the kind, proposes according to the kind. What is present in
one's own case is not to be rejected in the other man's, what is absent from
one's own case is not to be demanded of the other man's.
(A) 'Illustrating' is referring to other things in order to clarify one's case.
(B) 'Parallelising' is comparing propositions and letting all 'proceed'. (C)
'Adducing' is saying: 'If it is so in your case, why may it not be so in mine
too ?'. (D) 'Inferring' is using what is the same in that which he refuses to
accept and that which he does accept in order to propose the former.
N O 12 (A) Of things in general, if there are respects in which they are the
same, it does not follow that they are altogether the same. (B) The paralle
lism of propositions is valid only as far as it reaches. (C) If something is so
of them there are reasons why it is so; but though its being so of them is
the same, the reasons why it is so are not necessarily the same. (D) If we
accept a claim we have reasons for accepting it; but though we are the same
in accepting it, the reasons why we accept it are not necessarily the same.
Therefore propositions which illustrate, parallelise, adduce and infer
become different as they 'proceed', become dangerous when they change
484 Names and Object
direction, fail when carried too far, become detached from their base when
we let them drift, so that we must on no account be careless with them, and
must not use them too rigidly. Hence saying has many methods, separate
,,
kinds, different reasons, which must not be looked at only from one side.
From N O 12 we are no longer putting together patterns out of

mutilated scraps; the text is mercifully intact until what it is unnecessary
to doubt is the end of the treatise. It concludes with the last of five resump
tive formulae (NO 18), all of which were listed in N O 13, so that there is
no question of another accidental break.
N O 11 lays down a procedure in four stages for testing inferences from
what is so of an object to what is so of the thing it is. Since the four stages
are anticipated in the introductory paragraph and resumed in the series of
warnings which follows, we mark them by letters to facilitate comparison.
The procedure is applied in N O 14-16; we take our examples from the
argument of N O 15:
Problem: Is it legitimate to describe the execution of robbers as
'killing people* ?
(A) Illustrating. One offers other things for comparison with the
robber: Jill's younger brother, a boat.
(B) Parallelising. One "compares propositions and lets them all
'proceed' " (from something being so of objects to it being or not being
so of the thing they are):
"Her younger brother is a handsome man, but loving her younger
brother is not loving handsome men. . . .
A boat is wood, but entering a boat is not entering wood.
Robbers are people, but abounding in robbers is not abounding in
people, being without robbers is not being without people."
(C) Adducing. If challenged, one "uses explanation to bring out
reasons" (cf. A 72 IS:, Bxfi".Wife " 'Explanation' is the means by which one
clarifies"), adduces examples which he does accept, and asks " I f it is so in
your case, why may it not be so in mine too?":
"How shall we make this clear? (*^C£?£;£.). Disliking the abundance
of robbers is not disliking the abundance of people, desiring to be without
robbers is not desiring to be without people."
(D) Inferring. "Accepting according to the kind, proposing according
to the kind", one "uses what is the same in that which he refuses to accept
and that which he does accept in order to propose the latter":
"The whole world agrees that these are right; but if such is the case,
there is no longer any difficulty in allowing that, although robbers are people,
2/5 485
loving robbers is not loving people,

not loving robbers is not not loving people.
killing robbers is not killing people.
The latter claims are the same in kind as the former. . . ."
For the phrase shuyi 'dredging ideas' (purifying them, presenting them
unsullied by verbiage), cf. § 1/4/37.
N O 13
3 1
*JS » * — f t i f n — [ ^ A ] [ ]« 2№ ·
N O 14
630 This clause is restored from its recapitulation in N O 16. There it lacks
the pu but the examples which lead up to it make it clear that the
shih M must be negated; in any case the unemended clause is un
acceptable since it would be a mere repetition of the first clause of the
present passage. Moreover the displaced characters ^fcjfe two clauses
later seem to be a fragment of the missing clause. (Hu Shih, followed
by Chang Ch'un-yi, T'an, W u Yù-chiang, T'ang Chun-yi.)
631 Two fragments have crept in at this point, breaking a clause the
continuity of which is guaranteed by its recapitulation in N O 18:
(i) The two characters from the missing clause (cf. n. 630 above), (ii)
Twenty characters, beginning in mid-sentence, mistakenly repeated
from N O 12 above 1W№ . . .
N O 13 " O f the thing in general, there are cases where

(1) something is so if the instanced is this thing,
or (2) is not so though the instanced is this thing,
or (3) is so though the instanced is not this thing,
or (4) applies without exception in one case but not in the other,
or (5) the instanced in one case is this and in the other is not.
N O 14 A white horse is a horse. T o ride a white horse is to ride horses.

A black horse is a horse. T o ride a black horse is to ride horses.
Jack is a person. T o love Jack is to love people.
Jill is a person. T o love Jill is to love people.
, ,>
These are cases where 'something is so if the instanced is this thing .
486 Names and Object
In previous translations (G(5) 6, G(12) 177) I rendered all nouns in

the four inferences as singular ("To ride a white horse is to ride a horse",
"To love Jack is to love a person"). It is only when the passage is seen in
its context in the reconstructed Names and objects and compared with the
earlier account of the art of consistent description in A 88-B 12 that it
becomes clear that, odd as it may seem at first sight, the Mohist is proceed
ing from the love of Jack and Jill to the love of all men. And this is not odd
at all when we appreciate that Jack and Jill (Tsang and Huo, abusive names
for bondsman and bondswoman) are the stock examples of the humblest
people, the people with absolutely nothing lovable about them except that
they are human (§ 1/5/12). T o love Jack is sufficient proof that one loves
a man simply as a man, and therefore that one loves all men, just as to ride
a white horse is sufficient proof that one rides horses. There is of course
an important difference between the two cases; to ride horses does not
imply that one has ridden every horse in the world, while love of men does
imply that one loves every man. The Mohist will be making precisely this
point in N O 17.
There would of course have been no possibility of misunderstanding
if the Mohist had been writing in a language with grammatical number.
It is not however that the difference between singular and plural is in itself
relevant to the Mohist's case. Throughout N O 14-16 phrases such aspai ma
'white horse* refer to particular objects, without regard to number; it is a
matter of indifference whether one translates the first sentence of N O 14 as
" A white horse is a horse" or "White horses are horses". Phrases such as at
jen 'love of men* carry their ordinary sense as word-combinations, which may
be general, as in at jen, or in sha jen 'killing people* (murder, massacre),
or may be particular as in shih V jen 'serve a man* (be in the service of a
lord, have a husband). In the case of at jen 'love of men* and ch'eng ma
'riding horses', this becomes quite evident when they recur in N O 17.
What matters for the Mohist is to decide when it is legitimate or illegitimate
to describe what is so of objects as so of shih Jk 'what they are', a problem
confused by these elusive changes of idiom which he brings to light by
sorting propositions into different classes.
The Canon on "fixing the kind in order to 'make the man proceed'
(B 1 i h S K f f A ) described this sort of argument as "explaining, on the
grounds that it is so of the instance here, that it is so of what it is judged
to be" ( m f t ^ - f e i a ^ ^ t i i ) . That we are starting from what "is so of
the instance here" is the implicit assumption behind the elliptical N O 13
"Of the thing in general, something is so if (the instanced)
is this thing" (distinguished from the similar phrase in N O 4 by the
2/5 487
presence of the generalising fu wu "the thing in general"). In each pair of

sentences the first affirms that X is the thing, the second 'proceeds' from
'It is so of X ' to 'It is so of the thing* by linking them with a final yeh.
N O 15
632
m± (») * sa A * » A ft o * ^ M A - & »
6 3 3
Si Aft o 6 3 5
» SS##il&*«*-& * « * f t . * A f t # * A * f t ° ft
A A f t » &&#&№m&#MA&*&WJi2. ? ffi£S£#ffi£A-fe »
m » ^ * m « A f t o ittffin^di^ o , flijsi&AAft»as^s
Aft > ^ * f i l * > P * A - f e » » J £ A # « A - f e ' &Sli[&«ll]« £ o jifcn

36
638o 639
BflflJJl ? '№^*ft] i!fc71;&ffi^(S)*^ ^ft -
6 3 2
Corrected from the parallel following.
633
Shihjen 'be in service to a lord, be married to a husband* ( M . 241/71).
634
Ch'eng mu 'ride wood* should have some idiomatic meaning I have
failed to trace.
635
Ju mu 'enter wood*; there are later examples of the phrase used of an
arrow piercing or ink soaking into wood ( M . 1415/198).
6 3 6
Delete as accidental repetition.
637 f> (=flb) ku 'other reason* as in N O 16, here accidentally inverted.
0
6 3 8
That this is a gloss is a plausible suggestion of T*an.
6 3 9
Corrected from N O 13.
N O 15 "Jill's parents arejen (people), but Jill's serving her parents is not
serving jen (serving a husband).
Her younger brother is a handsome man, but loving her younger brother
is not loving handsome men.
A carriage is wood, but riding a carriage is not riding wood.
A boat is wood, but entering a boat is not entering wood.
Robbers are people, but abounding in robbers is not abounding in people,
being without robbers is not being without people.
How shall we make this clear ? Disliking the abundance of robbers is not
disliking the abundance of people, desiring to be without robbers is not
desiring to be without people. The whole world agrees that these are right;
but if such is the case, there is no longer any difficulty in allowing that,
although robbers are people,
loving robbers is not loving people,
not loving robbers is not not loving people,
killing robbers is not killing people.
The latter claims are the same in kind as the former; the world does not
think itself wrong to hold the former, yet thinks the Mohists wrong to hold
the latter. Is there any reason for it but being, as the saying goes, 'clogged
within and closed without' ? (Gloss: 'Having no empty space in the heart:
it is indissolubly clogged').
These are cases where 'something is not so though the instanced is this
thing'."
The thesis "Killing robbers is not killing people" does not appear
elsewhere in Mo-tzii; but it is mentioned as a sophistry in Hsun-tzu ch. 22
(Liang 315/-2) and as a principle originating in the time of Y i i ?B (favoured
sage of the Mohists) in Chuang-tzu ch. 14 (Kuo 527/6). Evidently it was
not a doctrine of the school but a position assumed in debate on some
controversial issue, a position which seemed to their opponents to defy
common sense.
There is no contradiction between the Mohist conception of universal
love and the execution of criminals. The equal desire for the benefit of all,
when applied with the Mohist's ruthless consistency, requires the sacrifice
of some for the benefit of the rest, criminals of course, perhaps the innocent
as well. (Cf. the mutilated exposition of E C 6 "Loving them equally, we
choose one among them and kill him".) But we can well imagine their
opponents deriding them for professing an equal love of all, yet executing
robbers: "Robbers are people, killing robbers is killing people". It might
not be easy to come to grips with this accusation because of the manner in
which words change their sense in combination. Sha tao 'killing robbers'
implies just execution. (Cf. E C 6 M>W&3kW&& "Killing robbers on
your own authority is not killing robbers".) Sha jen like English 'killing
t
people', suggests wanton slaughter or murder, not only for Mohists

(cf. Lu-shih ch'un-ch'iu ch. 1/5 (Hsu 95/-3) "By the
law of the Mohists one who kills a man dies") but in ordinary usage, as
can be seen from the many examples in Mencius. The Mohist does 'kill
people' in the neutral sense of the words, but not in the pejorative sense
which implies that he does not 'love people'. It was noticed in B 54 that
although (in some idiomatic sense which I have failed to trace) "Killing a
whelp is not killing a dog", it remains true that a whelp is a dog because in
the ordinary sense of the words "one can call it killing a dog". The author
of Names and objects approaches a similar problem by applying the
procedure laid down in N O 11.
The problem is one of description, of how to relate names to objects,
and does not admit of the strict proofs of the puzzles in disputation proper,
2/5 489
solved by analysing the definitions of names. Throughout N O 14-16 there

are not even logical connectives, simply parallel propositions laid side by
side. The single implicit principle is that like must be described alike.
Enemies of Mohism assume that there are the same sort of conditions for
describing someone as 'loving people* or as 'killing people*. But as soon as
propositions are sorted in different classes it can be seen that the consistent
way of talking is to say "Robbers are people, but killing robbers is not
killing people** (like "Her brother is a handsome man, but loving her brother
is not loving handsome men**). There is no logical compulsion to put the
proposition in the category the Mohist has chosen for it, but if you classify
it otherwise 'killing people* will have lost its pejorative significance and
become a neutral description of executing robbers to which the Mohist
would not object. In terms of modern Western categories, it is a matter
of semantics rather than of logic. The only appeal is to consistency: " I f it is
so in your case, why may it not be so in mine too?** (NO 11), "The world
does not think itself wrong to hold the former, yet thinks the Mohists
wrong to hold the latter'*.
N O 16
#2NHfe ' * * * * * * * > * iksmm · 1 t t t f t t i f n * S # * ·

M##iifcrfnf№± ofts-ifesm· mmm^mm ? · mm*
Wife] * itTj < ^ > M 4 * I » t f t °
The restoration of these characters on grounds of parallelism is a
proposal of H u Shih (followed by Wu, T*ang Chiin-yi). Less plausible
solutions have been offered by Sun (followed by Chang Ch*un-yi and
by T*an.)
The omission of this character, although not previously suggested,
seems a necessary corollary of Hu*s emendation (n. 640). It cannot be
understood either as 'moreover* or as 'about to*, and must have
entered the text by assimilation to the neighbouring clauses.
T'ang Chün-yi has proposed this ingenious emendation, which takes
# as a mistake for four characters written close on top of each
other. W u Yii-chiang reads <WKX>#^!"tfe. Earlier solutions did
not respect the parallelism with the preceding sentences.
Wu adds H at the beginning of the sentence in the interests of parallel
ism. However Mohists refer to fatalists indiscriminately as * W
or simply ^fa^ (cf. for example Mo-tzu ch. 36, 37 (Sun 175/—5,-3,
178/-4).
6 4 4
Cf. N O 13, n. 630.
N O 16 "Moreover reading a book is not a book, but to like reading books

is to like books.
A cockfight is not a cock, but to like cockfights is to like cocks.
Being about to fall into a well is not falling into a well, but to stop someone
being about to fall into a well is to stop him falling into the well.
Being about to go out of doors is not going out of doors, but to stop someone
being about to go out of doors is to stop him going out of doors.
If such is the case, there is no difficulty in allowing that:
Being about to die prematurely is not dying prematurely, but to stop
someone being about to die prematurely is to stop him dying prematurely.
Fatalism is not fate, but to reject the doctrine of fatalism is to reject fate.
The latter claims are the same as the former; the world does not think
itself wrong to hold the former, yet condemns the Mohists as wrong for
holding the latter. Is there any reason for it but being, as the saying goes,
'clogged within and closed without'? (Gloss as in N O 15.)
These are cases in which 'something is so though the instanced is not
this thing'."
This series is designed to answer .two fatalist arguments, both of them

undocumented. The first seems to be the one attacked from another
direction in B 51. We cannot stop an event from happening; we can stop
some event from being about to happen, but being about to happen is not
happening; what was prevented from being about to happen cannot have
been the destined event. The Mohist replies that the proposition belongs
to a class in which doing something to X does imply doing it to Y , in spite
of the fact that X is not Y .
In the second argument one would like to know what the fatalist
understood by fei ming 'rejecting fate'. The Mohist doctrine of feiming is
the denial that length and conditions of life are destined by Heaven,
and it seems strange that anyone should have questioned that to reject the
doctrine of fatalism is to reject the existence of fate. But one may guess that
the fatalist was saying something different, that one can refuse to believe in
fatalism but cannot refuse to die when one's time is up.
The fatalism for which Mohists attack Confucians is not determinism
and has nothing to do with the problem of free will. Fatalists do not insist
that your decision to go out of doors is predestined, only that whether you
2/5 491
go out or not you will not miss your appointment in Samarra. For Con
fucians, whether you act well or ill is decided by yourself, but whether you
live or die, succeed or fail, is decreed by Heaven. The Mohist's list of
parallels is designed to lead by degrees from the innocuous " T o like reading
books is to like books" up to the anti-fatalist conclusion. A Confucian
would be in a quandary as to where to raise his first objection; is he to deny
that you can stop a child falling into a well, and make nonsense of Mencius's
most telling illustration of the disinterested impulse to sympathy without
which we should not be human (Mencius 2A/6)?
N O 17
g A £ o mm<^>^^mmmmk^m^^mw:mm%mm^ °
#ft°
645, 646 Commentators agree that ^ must be supplied in the second

position; but in the first they correct ^ to ^ (Wang Yin-chih, T'an),
or retain it and omit ^ (Yii Yiieh, Sun, Chang Ch'un-yi, T'ang), or
correct the latter to 5fe (Wu). But except for H u Shih (who drastically
emends the whole passage) all agree on the general sense.
N O 17 " 'He loves people' requires him to love all people without exception,
only then is he deemed to love people. 'He does not love people' does not
require that he loves no people at all; he does not love all without exception,
and by this criterion is deemed not to love people.
'He rides horses' does not require him to ride all horses without exception
before being deemed to ride horses; he rides some horses, and by this
criterion is deemed to ride horses. On the other hand 'He does not ride
horses' does require that he rides no horses at all ; only then is he deemed
not to ride horses.
These are cases in which something 'applies without exception in one
case but not in the other'."
Throughout the three classes of N O 14-16, to describe something as so

(or not so), of what the instanced is (or is not), may or may not universalise.
The Mohist takes the two examples of 'proceeding' from the instanced to
the thing it is (NO 14); that one loves men simply as men implies that one
loves all men, that one rides horses does not imply that one rides every
single horse. In the other classes the examples which universalise are in
fact rather few: loving handsome men, being without robbers, rejecting
fate, perhaps liking books and liking cocks.
It is apparent that in dividing his propositions into three main classes
the Mohist has opened up indefinite possibilities of further subdivision.
But this is the one further distinction which it is imperative that he should
make; otherwise his parallels would obscure the fundamental Mohist claim
that love of men implies the love of every man.
The question whether 'loving some men' or 'not loving some men' is
the criterion (yin, cf. A 97) for love of man had already been raised in A 96.
N O 18
M A i t f i l H I A * » « A ± # i * » A * · A±m#A-tfe » o«
6 4 7 648
< A > £ & * f t A - t i l » ft Jt£*7!rftJiUb ° ±M± S ( » ) * H> »
MrfnE3J£-tb o M o [—JSJR-fe] 6 5 0
o J f g d t ( g ) * & 65i^f -
6 4 7
6 4 8
This is the reading of parallels in Huai-nan-tzu S P T K 16/15A/3, 4
and a comment of Ssu-ma Piao on Chuang-tzu S P T K 10/42A/6. The
same graph is corrupted to 81 in B 3.
6 4 9
6 5 0
Delete as an accidental repetition.
6 5 1
N O 18 " I f you inhabit somewhere in the state you are deemed to inhabit
the state: when you own one house in the state you are not deemed to own
the state. The fruit of the peach is a peach; the fruit of the bramble is not a
bramble. Asking about the man's illness is asking about the man: disliking
the man's illness is not disliking the man.
The ghost of a man is not a man; the ghost of your elder brother is your
elder brother. Sacrificing to a man's ghost is not sacrificing to a man;
sacrificing to your elder brother's ghost is sacrificing to your elder brother.
If this horse's eyes are blind we say that this horse is blind; though this
horse's eyes are big, we do not say that this horse is big. If these oxen's
hairs are yellow we say that these oxen are yellow; though these oxen's hairs
are many, we do not say that these oxen are many.
2/5 493
Both one horse and two horses are 'horses'. 'Horses have four feet' implies
four feet to one horse, not to a pair of horses. 'One-or-other of the horses
is white* implies one-or-other not of one but of two horses being white.
These are cases where 'the instanced in one case is this and in the other
is not'."
N O 14-17 examined the pitfalls of proceeding from 'so of X* to 'so of

what X is*. But even the preliminary decision as to what X is presents
similar difficulties, exposed by offering parallel sentences in which a
criterion is adequate in one case but not in another.
The previous lists all began with simple examples leading up to the
controversial examples which they illuminate (loving people, killing people,
stopping someone dying prematurely, rejecting fate). The present list leads
up to examples connected with problems in A 96 and B 12.
(1) A 96 raised the question of whether the black parts or the parts
which are not black should serve as the criterion for being a black man.
Under A 94 we noticed the sophism " A white dog is black", possibly based
on the argument that if its blind eyes prove it blind its black eyes prove it
black (cf. B 3 "Most of a white horse is white, most of a blind horse is not
blind"). Here the Mohist seems to be attacking a similar sophism, and we
can make a guess as to the direction of its argument. Which is the criterion
for the colour of animals, their eyes or their hairs ? Not their hairs (because
the hairs can be many without the animals being many), but their eyes
(because if the eyes are blind the animals are blind). The Mohist replies
that the eyes can no more serve as a constant criterion for describing the
thing than the hairs can. 'Many* would be a word which "names by re
ference to measure or number" like the 'big* of N O 1; but the eyes can be
big without the animal being big.
(2) It is difficult to translate the last example literally because of the
necessity of fitting the noun into the Indo-European category of number.
But the problem concerns the distinction, not between singular and plural,
but between distributive and collective, which is not exhibited by Indo-
European number (we say indifferently ' A horse has four legs' or 'Horses
have four legs'). Under B 12 we noticed a sophism " A yellow horse and a
black ox are three", possibly based on the argument that (a) In tfrM?3ftL
"Niu ma have four feet" the name niu ma must apply to each of the oxen
and horses with four feet and therefore calls each of them an 'ox-horse'.
(b) Since some ox-horses are not oxen and the rest not horses it follows that
ox-horses are not oxen and are not horses (a claim refuted in B 67).
(c) The two objects are therefore three things; the yellow one is a horse.
the black one is an ox, and both being four-footed are ox-horses. Although
we can have little confidence in this reconstruction, it is likely that some
thing of the kind stimulated the Mohist to his examination of distributive
and collective uses in B 3, 7, 11, 12. He sees that whether we are speaking
collectively or distributively has nothing to do with whether a pair of names
refers to one thing or two things. In B 12 he explained why 'having four
feet' can be said of two things ("Oxen and horses have four feet"), here he
points out that if we can speak collectively of niu ma one of which is an ox
we can equally well speak of horses one of which is white.
Part III
APPENDICES
3/1
ABBREVIATIONS
C Canon
E Explanation
TC Ta-ch'ii A ^ C (Taoist Patrology text)
HC Hsiao-ch'ii / h i * (Taoist Patrology text)
EC Expounding the canons
NO Names and objects
BSS Basic sinological series (Kuo-hsiieh chi-pen ts'ung-shu ISIS2£2

»»)
G Graham
HSYK Hsueh-shuyueh-k'anmftMn
HYHP Hsin Ya hsiieh-pao $rffi*$R
KMJP Kuang-ming jih-pao 0 %
KSP Ku shih pien *
M Morohashi Tetsuji I t J M t e k , Dai Kan-Wa jiten AStftlJ&ft
(Tokyo 1955-1960)
MTCC Yen Ling-feng ISM*, Mo-tetf chi-ch'eng
SPPY ^-^flo raw « s
SPTK Ssw-/>w ts'ung-k'an Qffifltftl
TSCC 7*'wwÂw C A I - C A ^ H # J D S
W Y W K l^m-jyw weit-Afc fl№$CI»
3/2
T H E TAOIST PATROLOGY TEXT

O F MO-TZÜ C H . 40-45
14 13 11 58 7 5 3 1 A
63 62 60 10 56 54 52 50
15 12 59 8 6 4 2
64 61 57 55 53 51
9
^ >
*θ ^ #C S
ft -lb SSM*
t
4fr
t u
If ^7 ^ Hb
% i ^T*
Ш
A t
500 The Taoist Patrology Text
t
3/2 501
lW£H4Wä4«
«£4#g*£ g ЛИ*
1ОДяЫ >4 « К
* « 4 # 4 « ί # I **>#4$
•I £ ^ ^^4 Й44ЙК*
The Taoist Patrology Text
» 1ttЩ £ Ц t ^ ^ ^ 4 s ^ 4 l 4 ^ ^
»·1
4 € ^ £ ^ ^ И ^ 4 ^ * ¥ ! *
4s*es«H K&^$4%*K+W
ni
^ $ 4 4 4 ; < 4 № й & $ 4 ® ЩЩ
« Н Ц ^ # 4 * ^ 1 *v£'l £ W M $ 4
H ^ ^ 4 * * V r 4 < ^ 4 < ^ V ^
W ^ M £ ^ # 4 M M N & # 4 * |
3/2 503
Ht
9
•Я
4a
3/2 505
3/2 507
V f ^ 4.^ jrtfe £ £Д-s? * Н * * &
*** i
3/2 509
3/2 511
+
+
3/2 513
WW H№ tu&steéК
u* tf№t ^ ^ ^ ^
3/2 515
4
3/2
HK-^vÔ <04^W&- 04**^W

3/2
4¥ ^ α V ^ 4
Vψ ^»•WP» VfcF^
The Tacist Patrology Text
3/2 521
\
3/2
3/2 525
3/3
T H E ARRANGEMENT OF T H E 'CANONS'
AT STAGE 2
22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1
щят&ъштщm * & ν η m«£&*Ρ**ΙΛΑ
m mш±* чΨ кшж m m ж mm mш m m # «· #m
Ш^:£ШМ#Ш & m Ш Ш Ш № Ш m ib Ш Ш Ш № Ш
явтВги§т ft ftft ft -Ш, m m
Ä ïffi fF & Ш ft Ш
ft m №. ffe ft ft 3£ JA
ft Ш. № « ft
M ft
71 70 69 68 67 66 65 64 63 62 61 60 59 58 57 56 55 54 53 52 51 50
Я Й * И 1 Й l i t m я Φ ш 13 m ш в m Φ m ψ & ±
m ïïi m ш m ЙЖ na * m « η ш - « Ψ ^ л m * и
Ü ГнШ Κ Ψ ± - ÏPS Φ -Й (fr № Й и Ä в λ
ft ïïo ïïn ffl ft ffl if ft 5f Hi lui ft RH Ш Ш * ft iTr ft ft ft
^ ft ft Ii* I I ft ft ffl
ft m ш ft ffn ft ft *
* "/f ft ft
ft m w
m m
ft ft
528 The Arrangement of the 'Canons' at Stage 2
47 46 45 44 43 42 41 40 39 38 37 36 35 34 33 32 31 30 29 28 27 26 25 24 23
affl*«rH*aiSIII±3E±*JBS.ffiSIIBI8#9fST»l»»l
w -tb tii -tb ^ M № Kr « T -tb T -tb s ^ tb -tb it, iiî -tb M M m »
-tÈ ^ -tb -tb :^ -Z H « SB 15 1b
-tb -tb -tb &
R — -tk -tM -tb
-tb -tb
95 94 93 92 91 90 89 88 87 86 85 84 83 82 81 80 79 78 77 76 75 74 73 72
m m m a m m m m m m m s m & m m %\ m & m E 3 m ta
I f t l M f l î P W ï f l i i H - K f f {fr ffi « tf ^ 1% i i fîR Sic $§ 9 K №
id t& - s• £ w £ M £ « t: m M m. m ta m m
r
c %\ \t ^
m ?·? m m %u m m # # M ^ % m ·& mmtL » * -tb
« m&m m
№ ^1
& * te
-tb -tb
V
3/3
24а 13
16 15 14b 23 22 14а 12 11 10 9 8 7 6 5 4 3 2 1
ж m m жт m --^.тшж^ш>тщтрх
m ψ — um. шшш^шш^^^^шп
/ h * : ff * Ш « « 4 > 1 й * Ш - « Л
* ö SM?« Ä m я — nana
0«?fflS5©Mff*ff!aBÎIIÈff
t a s ff -iE » 1 * 5
ff ffi 4> W i t
л m
Mi 1Ä ff m - â /h
ff Ä II II
-ν л
ft
13
58 57 56 55 54 53 52 51 50 «й 48 47 46 45 44 Щ 43 42
^ ^ ? * Й « ^ 1 Й 1 Е | й 4 9 ^ 1 © Ы й й 1 Ш ^
Л К & А %К ffiiE M f*ffiâ#K#
# № * ff * ift 1ЙЙА*ЙИ
IE «I № «t ïïo £ S & & ΰ ff
Л iE Л Ж ffl П&. Л fil #
fflftÄ -fefô Χ sg ffH
fr ff № ffi ff ff
Ä ff И Ж ST
ff
530 The Arrangement of the 'Canons' at Stage 2
24b
41 40 39 38 37 36 35 34 33 32 31 30 29 28 27 26 25a 25b 21 20 19 18 17
l i t A S *
* «j » }& — * » » » » s
M M W 18 « £ £ rffi 11«m. «es. s ; s * & * B # ta %
ta & RT ; urn At /h ta ^ # ta
tatiMit t f i f i f i # ta JE : tatatata^^KS
R T * fi £ # £ &
* *n ta ^ ta ^ mm mmw « a ; * * "
ft^ffiës&atate -i>
!t tete^ta m ta v fi ta
ib ÏE Ê r . # m
82 81 80 79 78 77 76 75 74 73 72 71 70 69 68 67 66 65 64 63 62 61 60 59
ffi# -
* m M
F * Si lit •5J « m Mm m
S # K1$ S te n №ÏÏD *«
nff ^
«
lit
t a t a » £ № rfn Sf —
* » ffi
te f$ * lit £
^±tata^^î« « n
tateff^ K Ï E w * » ta m # ÏE ta ïïn
g ta 3g 3 m m te s ai] ta * m te
fl'J is ta * ta s
TffcM#m ta m ft
ta£« te 5
mm
^ IS IS ? fi p # ta
ta «
m * te
3/4
BIBLIOGRAPHY
T H E following items are cited by the author's surname alone : K A O Heng

(1958), L I Yü-shu, L I A N G Ch'i-ch'ao (1922), L I U Ts'un-yan (1965),
L U A N T'iao-fu (1957), PI Yuan, S U N Yi-jang, T A N Chieh-fu (1958),
W U Yii-chiang.
Editions of classical texts identified by a surname may be traced under the
editor's name (for example, 'Han Fei-tzu (Ch'en)' under 'Ch'en Ch'i-yu').
3/4/1 WESTERN LANGUAGES
Altchinesische Logiker : Bibliographie (cyclostyled). Seminar für Ostasia

tische Kultur und Sprachwissenschaft der Universität München, 8
München 23, Leopoldstrasse 10. (No date: latest entry 1968).
CHAN WING-TSIT, A source book in Chinese philosophy (Princeton, 1963)
——. Chinese philosophy, 1949-1963 : an annotated bibliography of Ma
China publications (East-West Center Press, Honolulu, Hawaii, 1967).
C H A O , Y . R . , "Notes on Chinese grammar and logic", in Philosophy East
and West, 5/1 (1955), 31-41.
C H E N G C H U N G - Y I N G and S W A I N , R I C H A R D H . , "Logic and ontology in the
Chih Wu L u n of Kung-sun Lung T z u " , in Philosophy East and West,
20/2 (1970), 137-154.
"The Optics of the Mo-ching", in
C H ' I E N L I N - C H A O (TSIEN LING-CHAO),
Actes du Ville Congrès International de l'Histoire des Sciences, p. 293-2
(Florence, 1956).
CHMIELEWSKI, J A N U S Z , "Notes on Early Chinese Logic", in Rocznik
Orientalistyczny : (1) 26/1 (1962), 7-22; (2) 26/2 (1963), 91-105; (3) 27/1
(1963), 103-121; (4) 28/2 (1965), 87-111; (5) 29/2 (1965), 117-138;
(6) 30/1 (1966), 31^52; (7) 31/1 (1968), 117-136; (8) 32/2 (1969), 83-103.
C H O W TSE-TSUNG," A new theory on the origins of Mohism", in forth
coming Wen lin %Vt> v. 2.
CREEL, H . G., Confucius and the Chinese Way (New York, 1960) (1st edition
entitled Confucius : the man and the myth, 1949).
532 Bibliography
. What is Taoism? (Chicago, 1970).

. Shen Pu-hai (Chicago, 1974).
C U L L E N , C , " A Chinese Eratosthenes of the flat earth: a study of a frag
ment of cosmology in Huai Nan tzü", in Bulletin of the School of Oriental
and African Studies, 39/1 (1976), 105-127.
FORKE, ALFRED, Me Ti des Sozialethikers und seiner Schüler philosophisc
Werke (Berlin, 1922).
. "The Chinese sophists", in Journal of the China Branch of the Royal
Asiatic Society, 34 (1901-1902).
. Geschichte der alten chinesischen Philosophie (Hamburg, 1927).
FUNG YU-LAN, History of Chinese Philosophy, translated by Derk Bodde
(Princeton, 1952).
. A short history of Chinese philosophy, translated by Derk Bod
(New York, 1958).
G E I S S E R , F R A N Z , MO Ti: der Künder der allgemeinen Menschenliebe (Bern
1947).
G R A H A M , A . C.
G(l) "Kung-sun Lung's Essay on Meanings and Things", in the
Journal of Oriental Studies, 2/2 (1955), 282-301.
G(2) "The composition of the Gongsuen Long tzyy", in Asia Major,
N S 5/2 (1957), 147-183.
G(3) " * Being* in Western Philosophy compared with shihjfei and
yu/wu in Chinese Philosophy", in Asia Major, N S 7/1-2 (1959),
79-112.
G(4) "The dialogue between Yang Ju and Chyntzyy", in Bulletin of
the School of Oriental and African Studies, 22/2 (1959), 291-299
G(5) "The logic of the Mohist Hsiao-cWu", in Toung Pao, 51/1 (1964),
1-54.
G(6) "Two dialogues in the Kung-sun Lung tzü", in Asia Major, N S
11/2 (1965), 128-152.
G(7) "The 'Hard and White* disputations of the Chinese sophists", in
Bulletin of the School of Oriental and African Studies, 30/2 (1967
358-368.
G(8) "'Being* in Classical Chinese", in " T h e verb 'be* and its
synonyms" edited by John W . M . Verhaar, Foundations of language,
Supplementary series vol. 1, 1-39 (Dordrecht, Holland, 1967).
G(9) "The background of the Mencian theory of human nature", in the
Tsing Hua Journal of Chinese Studies, 6/1, 2 (1967), 215-274.
3/4 533
G(10) "Chuang-tzu's essay on seeing things as equal", in the History

of Religions, 9 (1969-70), 137-159.
G ( l 1) "The grammar of the Mohist dialectical chapters", in A Sympo
sium on Chinese Grammar, edited by Inga-Lill Hansson (Lund
1971).
G(12) "Later Mohist treatises on ethics and logic reconstructed from
the Ta-ch'u chapter of Mo-tzu", in Asia Major, N S 17/2 (1972),
137-189.
G(13) "The concepts of necessity and the 'a priori' in later Mohist
disputation", in Asia Major, N S 19/2 (1975), 163-190.
G(14) "The organization of the Mohist canons", in Ancient China:
studies in early civilization, edited by David T . Roy and Tsuen-
hsuin Tsien (The Chinese University Press, 1978), 167-179.
G R A H A M , A. C. and S I V I N , N A T H A N , " A systematic approach to the Mohist
optics", in Chinese Science, edited by Shigeru Nakayama and Nathan

Sivin, M I T East Asian Science Series, vol. 2, 105-152 (Cambridge,
Mass., 1973).
HEATH, Sir T H O M A S L . , The thirteen hooks of Euclid's Elements (New Yo
1956).
H u S H I H , Development of the Logical Method in Ancient China (Commerc
Press, Shanghai, 1922).
HUGHES,E . R., Chinese philosophy in classical times (Everyman's Libra
London, 1942).
KAO K U N G - Y I and O B E N C H A I N , D I A N E B . , "Kung-sun Lung's Chih wu
lun and semantics of reference and predication", in Journal of Chinese
philosophy, 2/3 (June 1975), 285-324.
KARLGREN, BERNHARD, Book of Odes (Stockholm, 1950).
. Book of Documents (Stockholm, 1950).
. "Grammata serica recensa", in the Bulletin of the Museum of Far
Eastern Antiquities, 29 (1957), 1-332.
L A U , D . C , "Some Logical Problems in Ancient China", in the Proceed
ings of the Aristotelian Society, N S 53 (1952/3), 189-204.
. Mencius (London, 1970).
. Review of Creel (1970) in Asia Major, N S 18/1 (1973), 121-123.
LEE, CYRUS, "Mo-tzu: on time and space", in Chinese Culture, 6/1 (1964),
68-78.
LEGGE, JAMES, The Chinese classics (Hong Kong University Press, 1961).
534 Bibliography
LESLIE, DONALD, Argument by contradiction in pre-Buddhist Chinese rea

ing (Australian National University, Canberra, 1964).
Liu, JAMES J. Y . , The Chinese Knight-Errant (London, 1967).
L i u T S ' U N - Y A N , "One item from the Canon of the Mohists", in Jao

Tsung-yi chiao-shou nan-yu tseng-pie lun-wen chi MS^Kfi^CÎSffiSêB
XM, 1-11 (Hong Kong, 1970).
MASPERO, HENRI, "Notes sur la logique de Mo-tseu et de son école", in
Toung Pao, 25 (1928), 1-64.
. "L'astronomie chinoise avant les Han", in T'oung Pao, 2 6 (1929),
267-356.
M E I , Y . P., The Ethical and Political Works of Mo-tse (London, 1929).

. Mo-tse the neglected rival of Confucius (London, 1934).
N A K A Y A M A SHIGERU and SIVIN, N A T H A N , Chinese Science, M I T East Asian
Science Series, vol. 2 (Cambridge, Mass., 1973).
NEEDHAM, JOSEPH, Science and civilization in China (Cambridge,
1954- ).
. The grand titration ; science and society in East and West (Lon
1969).
T
I
PULLEYBLANK, E . G . , "Fei ^ wet f£i and certain related words", in Studia
serica Bernhard Karlgren dedicata (1959).
RUBIN, V . A., "Tzù-ch'an and the city-state of ancient China", in Toung
Pao, 5 2 (1965), 8-34.
SCHAFER, EDWARD H . , "Cultural history of the elaphure", in Sinologica,

4 (1956), 2 5 1 - 2 7 4 .
SOLOMON, BERNARD S., "The assumptions of Hui-tzû", in Monumenta

serica, 2 8 (1969), 1-40.
TSIEN TSUEN-HSUIN, Written on Bamboo and Silk (Chicago, 1962).
M . (editor), "The verb 'be' and its synonyms", in the

VERHAAR, JOHN W .
Foundations of language, Supplementary Series (Dordrecht, Holland,
1967- ).
WALEY, ARTHUR, The Way and its Power (London, 1934).
. The Analects of Confucius (London, 1938).

-. Three ways of thought in ancient China (London, 1939).
WATSON, BURTON, Mo Tzu : Basic Writings (Columbia University Press,
1963).
WILLIAMSON, H. R . , MO Ti: a Chinese heretic (University Press, Tsinan,

1927).
3/4 535
YATES, ROBIN D .S., Towards a reconstruction of the tactical chapters of

Mo-tzü (chüan 14), University of California (Berkeley) thesis, 1975.
3/4/2 CHINESE A N D JAPANESE
3/4/2/1 Classical texts

Chuang-tzü jfl:?. See Kuo Ch'ing-fan.
Han Fei tzü See Ch'en Ch'i-yu (1958).
Hsün-tzü ^ - f . See Liang Ch'i-hsiung.
Huai-nan-tzü ffiShP. See L i u Wen-tien (1933).
Kuan-tzü ® i \ BSS.
Kung-sun Lung tzü See Ch'en Chu (1937).
K'ung ts'ung tzü 7 L 3 B \ S P T K .
Lieh-tzü . See Yang Po-chiin.
Lü-shih ch'un-ch'iu gBc#Sc. See Hsü Wei-yu (1955).
Lun heng MTOW. See Huang H u i .
Meng-ch'ipi-Van See H u Tao-ching (1956).
Mo-tzü JS-1\ See Sun Yi-jang.
Shuo-wen I S X . See Tuan Yü-ts'ai.
Tai-pHngyü-lan SPTK.
Yen Vieh lun SftSUm. See Wang Li-ch'i.
References to dynastic histories are to the Pai-na Htfi editions,
except for
Shih cht Ä I 2 , Peking 1959.
Han shu Öttf, Peking 1962.
3/4/2/2 Sinological works

(For items contained in the Mo-tzü chi-ch'eng 8 · ? Ä Ö £ of Yen
Ling-feng which was not available until proof stage, the abbrevia
tion M T C C and volume number have been added after the edition used.
Items for which this is the only edition mentioned may be assumed to have
been previously inaccessible to me.)
CHAN CHIEN-FENCX Mo-chia ti hsing-shih lo-chi g ^ K J ^ Ä S H
(Hupei, 1956). ( M T C C v. 46.)
. Kuan-yü Mo-chia ho Mo-pien-che ti p'i-p'an wen-t'i KlfölliCft] j|J|$
#fö#t4*J!fflS, H S Y K (1957) 4/14-17.
C H A N G C H ' I - H U A N G S R Ä I I , Mo-ching Vung-chieh MUMM (Taipei, 1960).
( M T C C v. 29.)
536 Bibliography
CHANG CHIH-CH'UN 3ft±J6, P'ing-chuMo-tzti ching-hua lu W8c.*&^W0Mk

( M T C C v. 21).
CHANG C H ' U N - Y I 3ft#6—, Mo-tzu chien-ku chien S ^ P f l l A ^ (Taipei,
1962) (First published 1922). ( M T C C v. 22.)
. Mo-hsueh fen-k'o S * # f 4 (Shanghai, 1923). ( M T C C v. 22.)
. Mo-tzu chi-chieh M^MM (World Book Company, 1936). ( M T C C
v. 23-26.)
CHANG HSIN-CH'ENG 3fi'i>a, Wei-shu t'ung-k'ao {SUffi^f (Shanghai,
1957).
CHANG HSING-YEN * f r ! R , Ming Mo lun-chi £ S l f t & ( M T C C v. 31).
C H A N G H S U A N 36'JIL, Mo-tzu ching-shuo hsin-chieh S ^ S I S S f W , Kuo-ku
yueh-k'an MtkJifl (1919) vols. 2, 3. ( M T C C v. 21.)
. Mo-tzu ching-shuo tso-che k'ao K S P 4/238-245.
CHANG HUI-YEN 3fi*K~B, Mo-tzu ching-shuo chieh rS^StftS? Hsiian-t'ung
fi$c chiyu B M edition, 1909. ( M T C C v. 9.)
C H A N G P I N G - L I N ^ f f i l S , Kuo-ku lun-heng SftStfra^r, Chang-shih ts'ung-shu
vol. 10.
CHANG SHIH-CHAO * ± £ Ü , Tung-fang tsa-chih MïïBM 17 (1920),
20/43-52; 20 (1923), 21/75-78; 21 (1924), 2/19-26 (three papers on the
Canons).
CHANG TAI-NIEN 3fitS№, Mo-tzü ti chieh-chi li-ch'ang yii chung-hsin
ssü-hsiang täBiRÄÎI+^SSl, K M J P (1954), 3/24.
CHANG T'IEH-CHÜN 3ftü&®, Mo-ching ssü-hsiang-lü ch'üan-chen MI&SJS
#»Ä, № Ä j i t o * » , 2/10 (1967), 8-10.
CHAO CHI-PIN S i f f i g , Mo-tzü tut K'ung-men lo-chi ssü-hsiang ti p'i-p'an
chi-ch'enghofa-chan » - f f f Ä P 9 a i » S « « » : « « l * « I » M . In his JC'rni-
Ä /w B & f t (Peking, 1963).
CH'EN CH'I-YU M t f f « , Han Fei tzü chi-shih WÏÏTMn (Peking, 1958).
. Mo-tzü ti k'o-hsüeh № f 4 ^ , Chung-hua wen-shih lun-ts'ung
ÄjfeÄ*, 4 (1963), 33-65. ( M T C C v. 46.)
CH'EN C H U E l f t , Mo-hsüeh shih-lun H S + i r a (Commercial Press, 1934).
( M T C C v. 33.)
. Kung-sun Lung tzü chi-chieh ^MM^MM (Commercial Press, 1937).
y
. Mo-tzü k'an wu k an wu ll^^Jl&yÜJSi (appendix to Su Shih-hsüeh,
op. cit.). ( M T C C v. 33.)
CH'EN KUNG Bfctfc, Mo-hsüeh yen-chiu M W 2 £ (Taichung, 1964).
CH'EN Kuo-FuE*Hfà, Tao-tsangyüan-liu k'ao M№M^ (Peking, 1963).
3/4 537
C H ' E N L I « « , Tung-shu tu-shu chi 3KSIMNB, S P P Y .

C H ' E N P ' I N - C H ' I N G K â K , Lun Pie-Mo Ira BUM, Yu-shih hsueh-chih
• t t , 5 (1966), 1.
CH'EN WU-CHIU , Mo-ching hsuan-chieh W№№№ (Chung-kuo kuei-
tu hsiieh-she 1935). ( M T C C v. 38.)
C H ' E N Y U A N Sfcki, Stt* te* rte' ft* i f e S W J , Yenching Journal 4 (1928),
537-651.
C H E N G C H E N - C H ' I U £Mfi&, Law Mo-tzu chih pien-hsiieh f r a l l - J ^ ^ f S ,
Wen-shih hsiieh-pao XifeSfS, 3 (1966), 79-94.
CHIANG CH'UAN tCJSfe, 7M tew chih-yen fH^J&B (Wen-hai ch'u-pan she
no date), (First published 1917).
CHIANG PO-CH'IEN J U G S , Chu-tzu t'ung-k'ao St^ffl^f (Cheng-chung
IE 4» Book Company, 1948).
C H ' I E N L I N - C H A O $8№M, Yang-sui , Wen-wu ts'an-k'ao tzu-liao
S53*#*, 95 (July 1958), 28-30.
CH'IEN M U HSU Hsing wet Mo-tzu tsai-ch'uan ti-tzu k'ao PfffMM^
fffcSM^S, K S P 4/300-303.
. Hut Shih Kung-sun Lung MM^Mi (Shanghai, 1931).
. Mo-tzii M^r (Commercial Press, 1931).
. Chuang-tzu tsuan-chien, Hi:-? £SH (Hong Kong, 1951).
. T'uei chih p'ien t S i h U (On the ancient Chinese logical methods Vui
and chih), H Y H P , 6/1 (1964), 1-43.
C H ' I E N PAO-TS'UNG Mo-ching li-hsueh chin-shih mWM^ 9
K'o-hsiieh shih chi-k'an ? 4 f e № J , 8 (1965), 65-72.

CHIH WEI-CH'ENG Mo-tzu tsung-shih St^ffiM (Shanghai, 1925).
( M T C C v. 32.)
CHOU FA-KAO JS&BJ, Chung-kuo ku-tai yii-fafyW^KSkfe,Academia
Sinica, Special publication No. 39 (Taipei, 1959- ).
C H O U F U - M E I Mill it, Mo-tzu chia-chieh tzu chi chetig M ^ i i H a ^ ^ i K
(Taiwan University, 1963).
. Mo-tzu hsû-tzù yen-chiu M^JEL^fflju, Wen-shih-che hsiieh-pa
15 (1966), 1-111.
C H U CH'I-FENG Tz'u t'ung S*M (Taipei, 1960).
CHU CH'IEN-CHIH /fcitil, Lao-tzu chiao-shih ^r^^M (Shanghai, 1958).
C H U HSI-TSU £M№, Mo-tzu Pei-ch'eng-menyi-hsia erh-shih-pïen hsiHan
sen wei-shu shuo « ^ « l « F 5 a T — + » « 8 I A « * l f t , KSP4/261-271.
Chung-kuo li-tai che-hsueh wen-hsuan <£HJBft£f*#*iS (Peking, 1962
538 Bibliography
Chung-kuo che-hsüeh-shih tzü-liao hsüan-chi ^W&^-^MMWßSt (Peking,

1964).
Chung-kuo che-hsüeh-shih chiao-hsüeh tzü-liao hui-pien » M I j S j f c Ü ^ Ä f t
« I i (Peking, 1 9 6 4 ) .
C H U N G L E I Ulf, Mo-tzü 'Fei-kung' 'Chien-ai' ti fan-tung shih-chih IB?
Ä i i W R l f f , K M J P (1961), 1 2 / 1 .
F A N K E N G - Y E N ftiif W, Mo-pien su-cheng £S№$k№, Kuo-hsüeh hsiao ts'ung-
shu mm^MW, 1 9 3 4 . ( M T C C v. 2 8 . )
F A N G S H O U - C H ' U jj¥£%ty Mo-hsüeh yüan-liu (Taipei, 1 9 5 7 ) (First
published 1 9 3 7 ) . ( M T C C v. 3 9 . )
y
Fu S H A N ff lij, Mo-tzü Ta-ch'üp'ien shih Shuang hung k an
cht m%mm ch. 35. ( M T C C . 6 . )
V
F U K U I SHIGEMASA iiS-№ffi3ft, "The resuscitation of Mo-chia (the Mohist

school) in the former Han dynasty", in Töhögaku 3 9 / 1 - 1 8 (1970).
FUNG YU-LAN M~#M> Chung-kuo che-hsüeh shih »frBS^W.bfc (Commercia

Press, Shanghai, 1 9 4 7 ) (First published 1 9 3 4 ) .
. Kuan-yii Hut Shih shih-shih ti yi-hsiang tzü-liao B S f ö / K t t + ^ t ä — 5 1
Ä « , K M J P (1962), 4 / 7 .
. Chung-kuo che-hsüeh-shih hsin-pien ^ f f l Ö ^ J & R J B (Peking, 1 9 6 4 ) .
H A N LING-HSIAO Mo-ching chung tichih-shih che-hsüeh SlS+föÄl
Chung-hua wen-hua fu-hsing yüeh-k'an »NÛMÄ!R>J ?'J, v. 5 / 6
(1972), 51-53.
H o L U N - C I I I N G MfölM, Mo-teß c/»ao-ÄM (Taipei, 1 9 6 8 ) . (This

is an incomplete reprint of Kao Heng's Mo-ching chiao-ch'iian under a
different title and author's name.)
Hou WAI-LU ^ ^ f f i , Chung-kuo ku-tai ssü-hsiang hsüeh-shuo shih ^^TÖ*
ftS&¥iäJft (Shanghai, 1946).
Hou W A I - L U ^ f f i , CHAO CHI-PIN and Tu KUO-HSIANG tt№,
Chung-kuo ssü-hsiang t'ung-shih ^ H Ä S . S Ä , vol. 1 (Peking, 1957).
HSIEH H S I A N G Ürffl, Mo-tzü hsüeh-shuo yen-chiu sS^^tftSff'Ä (Hong

Kong, 1 9 6 7 ) .
Hsu F U - K U A N IfctKü, Kung-sun Lung tzü chiang-su ^ r ^ f f i ^ ü l S c
(Taichung, 1 9 6 6 ) .
Hsü HSIAO-T'IEN ffftf^ , M O - ^ M k'ao-cheng §5^ 3&I8 (Shanghai, 1 9 2 6 ) .
Hsü WEI-YÜ № « , Lü-shih ch'un-ch'iu chi-shih g f t # f t J № (Peking,

1955).
Hu CHI-CH'UANG Chung-kuo ching-chi ssü-hsiang shih ^ I S ^ S
(Shanghai, 1962).
3/4 539
Hu S H I H S Ü S S , Chung-kuo che-hsüeh shih ta-kang ^W£&%J<M (Com

mercial Press, Shanghai, 1947) (First published 1919).
. Hsiao-ch'ü p'ien hsin-ku /h^M$Tß'i', Hu Shih wen-ts'un
1st collection 2/35-74. ( M T C C v. 21.)
Hu T A O - C H I N G ÄfläSÄ?, Kung-sun Lung-tzü k'ao Äf^ffi-?'^ (Commercial
Press, 1934).
. Meng-cWi pi-Van chiao-cheng (Shanghai, 1956).
Hu Y Ü N - Y Ü t№K3i, Mo-tzü ching-shuo ch'ien-shih S ^ S I Ö Ä S , S _ h ,
Kuo-hsüeh hui-pien vol. 2. ( M T C C v. 28.)
H U A N G H U I J i № , Lun-heng chiao-shih mMkW (Commercial Press, 1938).
HUANG KIEN-CHUNG Mohists*philosophy and their logical doctrines,
Shih-ta hsüeh-pao W ^ f f i , 4 (June 1959), 1-17.
HUNG CHEN-HUAN #k88'££, Mo-ching kuang-hsüeh pa-tHao li-shuo M I M Ä S
AfäÄtft, K'o-hsüeh shih chi-k'an S * * * ? ] ' , 4 (1962), 1-40.
. Mo-ching li-hsüeh tsung-shu ^%'Jj^^!iM y K'o-hsüeh shih chi-k'an
^ S A Ä f O , 7 (1963), 28-44.
1
JEN CHI-YÜ Mo-tzü S ^ (Shanghai, 1956).
J U A N Y Ü A N E/ÜTG, Ching-chi tsuan-ku %Wfklk (World Book Company,
Shanghai, 1936).
?
K A O H E N G Ä >\ Mo-ching chung yi-ko lo-chi kuei-lü S « * — .
, 1
In Chung-kuo ku-tai che-hsueh lun-ts'ung -r ffli&fä©#Mraiä8, compiled
by Wen shih che ÄJfc^f editorial board (Peking, 1957).
. Mo-ching chiao-cWüan 1 « S (Peking, 1958). ( M T C C v. 41.)
. Mo-tzü hsin-chien ( M T C C v. 41.)
KAO PAO-KUANG Mo-hsüeh kai-lun ffi^JKän (Taipei, 1956).
K A W A S A K I K Ö J I Mfc^Tn, Bokushi kenkyü bunken mokuroku SS^Sff3^3Sc8R
Q » (Hiroshima, 1973). (Not seen.)

K I T A M U R A K A I T S U JhWffiäS, Bokushi kaisetsu S^üfftft (Tokyo, 1935).
K u A I - C H I I M I § , Li pien (World Book Company, Taipei, 1970).

K u S H I H S Ä , Mo-tzüpien-ching chiang su ^MnffiB. ( M T C C v. 37.)
K U A N F E N G I I B I S , LunMo-tzüyü Yang Chu hsüeh-p'ai fraM-?№**M,
Chung-hua wen-shih lun-ts'ung ^ ^ Ä Ä I W J S , 6 (1965), 1-56.
Kuo CH'ING-FAN ^ K T S , Chuang-tzü chi-shih Jffi?*^ (Peking, 1961).
Kuo Mo-jo SÄIÄ />'*-/> W +Sfc#J*f (Shanghai, 1950).
L A N G CH'ING-HSIAO &m^ y Mo-tzü che-hsüeh ( M T C C v. 41.)
L A O SSÜ-KUANG ^ S 3 t , Mo-tzü yü Mo-pien M^MMM, Ta-hsüeh sheng-
huo XmStfS (1962), 123/4-13, 124/14-19, 125/16-21.
540 Bibliography
LEE, CYRUS Mo-tzû yen-chiu (Taipei, 1968).

Li CHING-CH'UN $ j l b # , Mo-ching pien-lun-shu ti yi-ko wen-t'i >Ê&-Êm
* № — f l f f l f l , Che-hsueh yen-chiu 8 (1958), 46-50.
Li L I Ting-pen Mo-tzû chien-ku chiao-pu J&fcS^tffllÉtttôï (Com
mercial Press, 1925). ( M T C C v. 27.)
L I SHIH-FAN Tan-Van Mo-pien kuan yu pien ti li-lun WC$MLM
mi&mmm^, K M J P (1964), 2/28/4,3/8/4.
Li YÛ-SHU ^r&M, Mo-pien hsin-chu MISfrfè (Taipei, 1968).
LIANG CH'I-CH'AO S6gFi®, Mo-tzû hsueh-an (Commercial Press,
1921). ( M T C C v. 18.)
. Mo-ching chiao-shih S S t t f f (Commercial Press, 1922). ( M T C C
v. 19.)
. Tu Mo-ching yu-chi SRSSfi&iB, K S P 4/253-261.
. Tzu-Mo-tzu hsueh-shuo - ? S ^ * t f t (Taipei, 1956). ( M T C C v. 18.)
LIANG CH'I-HSIUNG Ift&Si, Hsun-tzu chien-shih ^^WM (Peking, 1956).
LIANG HENG-HSIN 3£flï'£>, Wo kuo ku-tai k'o-hsueh-chia tut she-ying chi-pen
chih-shih ti kung-hsien ^ H * ^ i f 4 * ^ » « 3 K * * № a k № R ( R , Ta chung
she-ying JK^mï (1960) 3/17-19.
LIU CH'ANG MM, Hsii Mo-tzu chien-ku I S M ^ P f l l i (Yi-wen yin-shu kuan
no date). ( M T C C v. 30.)
L i u C H ' I 3 0 ^ , Lun-li ku-li m&£№ (Shanghai, 1943).
L i u S H I H - P ' E I SUSPS, Mo-tzû shih-pu fèlS (Yi-wen yin-shu-kuan
mXWmt, no date). ( M T C C v. 19.)
L i u T ' A O 8330, Ya-shih li-tse-hsiieh yu Mo-tzu pien-hsueh chih pi-chiao
yen-chiu &&mW№m?MmZ№£№'ft, Chung-hua wen-hua fu-hsing
yueh-k'an *mXW№k%n, v. 4/10(1971), 36-45; v. 4/11 (1971),66-74.
Liu TS'UN-YAN Mo-ching chien-yi M I S ^ H , H Y H P , 6/1 (1964)
45-140; 7/1 (1965), 1-134.
L i u W E N - T I E N SOlfcft, Huai-nan hung-lieh chi-chieh №W&Wk№ (Com
mercial Press, 1933).
. Chuang-tzu pu-cheng ffi^WlE (Commercial Press, 1947).
L o K E N - T S E f l t K S and K ' A N G K U A N G - C H I E N MJtS., Mo-tzû
(Chungking, 1945).
f
. Chu-tzu k ao-suo WiT^'M (Peking, 1958).
Lu TA-TUNG Mo-pien hsin-chu 33IS$n£ (Chunghua Book Com
pany, 1926). ( M T C C v. 36.)
3/4 541
L U A N T ' I A O - F U ÎS"ffi", Mo-hsueh chiang-yi W M I S J (Ch'i L u ta-hsueh

1925).
. Mo-pien t'ao-lun mMtim, Tzu-hsueh-she ts'ung-shu -?»f±*«
(1926).
. Mo-tzu yen-chiu lun-wen chi S ^ W ^ f r a X * (Peking, 1957). ( M T C C
v. 33.)
MA H S U - L U N H&lift, Chuang-tzu Tien-hsia p'ien shu-yi %T^££it
(Shanghai, 1958).
MAKINO KENJIRO Bokushi kokujikai M^m^M (Tokyo,
1911).
Mo tzu I №
(1) Tao tsang MM edition, 1445 (reproduced Commercial Press, 1923).
( M T C C v. 1.)
(2) T'ang m edition, 1553 (reproduced S P T K ) . ( M T C C v. 2.)
(3) Mao K ' u n ^ t t edition, 1581. ( M T C C v. 3.)
(4) Horyaku MM edition, 1757. ( M T C C v. 4.)
Mo-tzu son chiian, Ming ch'ao-pen, Huang PH-lieh pa H ^ f H ^ ,
SSÎJIK (The Y u tfl manuscript of Mo-tzu ch. 1-13). National Peking
Library (microfilm at the library of the School of Oriental and African
Studies, London).
Mo-tzu yin-te i S ^ l j l ^ , Harvard-Yenching Institute sinological index
series, Supplement No. 21 (Peking, 1948).
OHAMA AKIRA zfc^Gn, Met to jitsu £ £ If, Nihon chiigoku gakkai ho
^mm^n, 3 (1951), 21-35.
O K A S A K A T A K E O O R S i : $ i , "On the consciousness of language in Mo-jing",
in the Kyoto kyoiku daigaku kiyb ^ # ^ » * B S , 29 (1966), 1-24.
P'EI HSUEH-HAI &m№ Ku-shu hsu-tzu chi-shih SmjA^Mn
9 (Peking,
1954).
Pi YUAN #S5, Mo-tzu S ^ , SPPY. ( M T C C v. 7, 8.)
SHEN YU-TING ifrff J№, Mo-ching lun shu S№ttftfe, K M J P (1962), 10/5.
SHIMA, H . ft—*, On the thought of early Mo-pien 81 Shukan toyogaku
30 (Dec. 1973), 1-22.
Su S H I H - H S U E H 5§£f$», Mo-tzu k'an-wu S - ? ? J R (Chunghua Book
Company, 1928). ( M T C C v. 10.)
SUN CHUNG-YUAN S ^ H , K M J P November 1, 1963, Mo-chia 'Sha tao
fei sha jen* ti ming-t'i pu-shih kuei-pien HHt [ W&WWtK J S ^ J ^ J S ^ S
WM\ February 21 Mo-chia ti yi-chung fan-po fang-shih, 'chih' H ^ f f t
542 Bibliography
—SR®#Ä» T i h J . (Answers: 1964 January 17, 24; February 28;

March 8.)
S U N H S I N G - Y E N SJl/T/J, Ch'ing Sun Hsing-yen shou chiao, Shen Tseng-chih
shou pa, chiu ch'ao-pen Mo-tzü » » S f ö ¥ t t , fctÄW^tt,
National Central Library, Tai-pei.
S U N S S Ü - F A N G S l B f ö , Mo-hsüeh t'ung-lun T ^ i l Ä . ( M T C C v. 34.)
S U N T Z ' Ü - C H O U H^CÄ, Hsü Hsing shih-fou wei Mo-chia ti wen-Vi IffffS??
3<Sä?täßSa, K S P 4/189-191.
——. Mo-tzü Pei-ch'eng-men yi-hsia shu-pHen chih chen-wei wen-Vi ^•J'fll
« P ^ ß l T l k Ä i Ä f i l f f l « , K S P 6/188, 189.
Y I - J A N G MfäM Mo-tzü chien-ku <>R-?WU? (Peking, 1954) (First
f
SUN y
published 1894) (This reproduces the BSS edition and has the same
page numbers). ( M T C C v. 12-15)
TAI CHÜN-JEN JRSIfc, Ming-chia yü Hsi Han li-chih £^Jl?SÖtfE?&,
Wen-shih-che hsüeh-pao 3C5№«?-8, 17 (1968), 69-85.
TAKATA ATSUSHI iÄffi}?, The Chinese sophism, Töyö gakuhö Ä # H ? B ,
45 (1962) 1/69-92, 2/102-118.
-—-. Bokkei no shisö S $ S * > S $ I , Gakushüin daigaku bungakubu kenkyü
nempö m^±mXm^m%^n, 10 (1964), l-47; Tokyo joshi daigaku
ronshü M&icttmmn, 15/1 (1964), 1-46.
-—. Bokushi (Tokyo, 1967).
T'AN CHIEH-FU IfW^f, Mo-ching kuang-hsüeh «TsffÄ^, Tung-fang tsa-
chih 30/13 (1933), 154-168.
. Mo-tzü Hsiao-ch'ü ti-ssü-chang chiao-shih ^ ^ / h l f c ^ P H ^ R S , Wen
che cht k'anX^n, 5/1 (1935), 153-172.
. Mo-ching yi-chieh Wt%WM (Commercial Press, Shanghai, 1935).
( M T C C v. 34.)
. Mo-pien fa-wei (Peking, 1958). (Revised edition, Peking,
1964). ( M T C C v. 35.)
T ' A N G C H Ü N - Y I JjifSJS, "The six different interpretations of ' L i * (31) in
the history of Chinese philosophy", in the H Y H P , 1/1 (1955), 45-98.

. " A n Interpretation of 'Argument* (pien) in the 'Hsiao Ch'ii' chapter
of the Mo-tzü", in the H Y H P , 4/2 (1960), 65-99.
T ' A N G Y Ü E H I S M , Lun hsien ChHn wu so wei Pie-Mo Sra5te*«^ll£UIS,
K S P 4/245-248.
T ' A O H U N G - C H ' I N G P £ № K , Tu chu-tzü cha cht I R I S ^ f l J H (Peking, 1919).
3/4 543
TENG K A O - C H I N G Spoilt, Mo-ching hsin-shih 1#M$rf¥ (Commercial

Press, 1931). ( M T C C v. 31.)
0
TÖDÖ AKIYASU Ä S J 3 ' S , "Why did the argument of 'Chien po i t'ung'
SÖÎRI arise?", in the Tokyo shinagaku-hö %&%№&M, 8 (1962),
1-10.
T'SAO YAO-HSIANG *«rffl, Mo-tzü chien - ß - ? ^ . ( M T C C v. 17.)
TS'EN CHUNG-MIEN Mo-tzu ch'eng-shou ko-p ien chien-chu
y
ffi?i$
fftlfifffiÈ (Peking, 1958). ( M T C C v. 45.)

Tu KUO-HSIANG ttWïs Kai tsen-yang k'an-tai Mo-chia lo-chi î£;Sft№?#
BMM№, Che-hsüeh yen-chiu © ¥ # f 3 E , 10 (1959), 49-56.
. Tu Kuo-hsiang zven-chi tt№XÄ (Peking, 1962).
TUAN WEI-YI a * f t » , Ku chuan-wen ta tzü-tien l ï â f c X ^ ^ A (Taichung,
1965).
T U A N YÜ-TS'AI f£5${, Shuo-wen chieh-tzü chu Î È X J W ^ J E É (Taipei, 1955).
T U N G TSO-PIN tffËft and W A N G H E N G - Y Ü ï f ï f è j "Reforms of some
specific scripts under the Empress W u ( A . D . 684-704)", in the Bulletin
of the Institute of History and Philology, Academia Sinica, 34/2 (1963),
447-476.
W A N G CH'I-HSIANG Chou ChHn ming-chia san-tzü chiao-ch'uan
mmzw.^Wcn (Peking, 1957). .
W A N G CHIH-HSIN Mo-tzü che-hsüeh S ^ f (Nanking, 1925).
W A N G C H I N G - H S I B i f t H , M o s t e ^ i ^ . ( M T C C v. 17.)
W A N G K ' A I - Y Ü N 3i№]31, Mo-tzü chu Wk^Q, Hsiang-ch'i-lou ch'üan-shu
( M T C C v. 16.)
W A N G L I - C H ' I ZEÎSS, Yen-tHeh lun chiao-chu Ü Ä f i l f t t E (Shanghai,
1958).
W A N G M I N G EE J3, Ts'ung Mo-tzü tao T'ai-p'ing ching ti ssü-hsiang yen-pien
0
i&m?mx<miY№Mmn, K M J P (i96i), 12/1.

W A N G NIEN-SUN S & f ö , TVsÄ« *UI-<:«Ä WYWK.
WANG SHU-MIN Textual Criticism of Mo-tzü, Academia Sinica
30/1 (1959), 71-102.
WANG SHU-NAN EEföffi, Mo-tzü chiao-chu pu-cheng ffi-?M9:ffilE, Tao-lu
U'mg-k'oMUmm No. 7. Kuang-hsü Ä « 13 (1887). ( M T C C v. 11.)
y
W A N G T I E N - C H I S I S , Chung-kuo lo-chi ssü-hsiang shih-k o fen-hsi #89
m№ŒM&.mftVr (Peking, 1961).

W A T À N A B E T A K A S H I A i â - ^ , Bokka no hêigikôsho ni tsuiterS^ÂSWïfl
K O ^ - C , TöÄyö shinagaku hö Ä H I Ü , 3 (1957), 1-19.
544 Bibliography
. Bokka no shudan to sono shiso HiC*>^BH £ * *>Siffi, Shigaku zasshi

A W K I S , 70 (1961), 1198-1231, 1351-1385.
. Bokushi shohen no chosaku nendai M^tii^&^Sfctfrfc, Toy5 gakuho
Mffmn, 45/3 (1962), 1-38; 45/4 (1962), 20-38.
. Bokushi , Chugoku no shisoka ^M^SMU (Tokyo, 1963),
1/120-132.
. " O n towns and villages protected by Mo-chia (Mo-school)", in the
Tohbgaku 5 C # » , 27 (1964), 33-48.
. Bokka shiso 3&£&№ f in Toyd shiso JfclftSM, edited by Uno Seiichi
and others, 4/3-96 (Tokyo, 1967).
. Kodai Chugoku shiso no kenkyu #ft+BSfil«>Sff3fc (Tokyo, 1973).
Wu FEI-PO ffi*W, Mo-pien chieh-ku mffiMtft. (Peking, 1923).
. Kuo-ku lun-ts'ung iS&iraSI, Hsiieh-yi hui-k'an »S*^J 13, 79-140
(four papers on the Canons) (Commercial Press, 1926). ( M T C C v. 21.)
Wu JU-LUN gil&m, K'ao-ting Mo-tzu Ching-hsia pHen ^ J g S ^ I S T *
(Appendix to Wang Shu-nan, op. cit.).
. Chu-tzu chi-p'ing f f ? J I № (Taipei, 1970).
Wu K U A N WR, Mo-tzu yii Wo-hsueh m^M^W. In Hsii Hsiao-t'ien
(1926).
Wu T S E - Y U %Vm Mo-tzti yi-chu
t ( M T C C v. 46.)
Wu Y U - C H I A N G JSttEC, Mo-tzu chiao-chu H ^ S f f i (Tu-li ch'u-pan she
1944). ( M T C C v. 43, 44.)
Wu-wei Han chien jftiKSIffi, edited by the Archaeological Institute of the
Chinese Academy of Sciences and the Kansu Provincial Museum
(Peking, 1964).
YABUUCHI KIYOSHI Kftfff, Bokushi (Tokyo, 1968).
Y A N G HSIANG-K'UEI $§[R]1S, Mo-tzu ti ssil-hsiang yii Mo-che chi-t'uan Wi*
№gMMm%mm f wm shih che x&n (wssj, 3/11-18.
. Chung-kuo ku-tai she-hui yii ku-tai ssu-hsiang yen-chiu ^Bl^f^ttltf
$S-£ft&filffift (Shanghai, 1962).
Y A N G K A N G - C H I E N SMS'lIt, Tan-Van Mo-tzu ti chieh-chi ch'eng-fen Ifclfc
a - ^ K l B P f t ^ , K M J P (1961), 11/10.
YANG K'UAN Vk%, Mo-ching che-hsiieh H f f i t f S (Cheng-chung J E ^ Book
Company, 1942). ( M T C C v. 42.)
Y A N G PO-CHUN Ifei&ML, Lieh-tzu chi-shih (Shanghai, 1958).
YEH MANG P t ^ , Mo-ching chung ti kuang-hsiieh chih-shih S f f i t f & t f t *
»», K M J P (1962), 9/4.
3/4 545
YEN LING-FENG j S c S ^ , Mo-tzü chih-chien shu-mu l № f t L l Ä @ (Taipei,

1969). (From Kuo-li chung-yang t'u-shu-kuan kuan-k'an S Ä ^ Ä H Ä t
tin, 1/2 (1967), 15-38; 1/3 (1968), 21-41.)
. Hsien-tsai Mo-tzü chu-p'ien nei-jung chih fen-shi cht ch'i tso-c
chien-ting S f f » ^ » « f t « ^ Ä « ? R Ä № # » « Ä f Yu-shih hsüeh-chih
mmm, 6 (1967), 3.
. Mo-tzü chien-pien S-JIflMS, Jen-jen wen-k'u \KXW> (Taipe
1968).
. (Wu ch'iu pei chat) Mo-tzü chi-ch'eng f№Htf2, 46 v.
(Taipei, 1975).
YIN T'UNG-YANG F « , Mo-tzü hsin-shih ü ^ f f S . ( M T C C v. 20.)
Yü C H ' A N G T H , Hsiang-ts'ao hsü chiao-shu (Peking, 1963).
( M T C C v. 18.)
Yü H S I N G - W U - T Ä i F , Mo-tzü hsin-chengfi-T^ffiftin Shuang chien ch'i
chu-tzü hsin cheng » « l » » ^ * ? « (Peking, 1962). ( M T C C v. 42.)
Yü Yü Chung-kuo ming-hsüeh 4 » ® ^ * (Taipei, 1959).
Yü YÜEH # 1 « , Chu-tzü p'ing-yi (Peking, 1954).
3/5
INDEXES
3/5/1 CHINESE GLOSSARY
THE following is a full glossary of words in the dialectical chapters

(Mo-tzii ch. 40-45). The listing of occurrences is selective, but, unless
otherwise stated, is complete for the meaning illustrated. For full listings
the reader is referred to the Mo-tzu concordance in the Harvard- Yenching
sinological index series (with the warning that it is based on the text a
punctuated and emended by Sun Yi-jang).
II X X is an antonym
=X X is a synonym
*X X is a emended character
fX X is a corrupt character
548 Indexes
RADICALS
—.
549 554 560 563 36 567
1 549 555 560 564 568
549 555 560 564 568
; 549 /h 555 it 560 564 a 568
549 X 555 560 564 568
j 549 P 555 ft 560 ffiî 564 568
549 ih 555 560 564 m 568
550 «< 555 560 564 .f- 568
A 550 X 555 'X 561 m 564 n 568
)L 551 a 555 /R 561 RI 564 M 569
A 551 rtJ 556 Jl 561 a 565 569
A 551 556 561 m. 565 569
551 & 556 561 m 565 569
Jl 551 r 556 562 565 569
u 551 556 562 565 tr 69
5
71 551 556 562 m 565 10
m 569
~h 552 ï 556 562 565 # 569
552 \ 556 /s 562 M" 565 ES- 569
b 552 556 H. 562 565 569
c 552 X 557 ÄE 562 565 H 569
552 P 557 r 562 ft 566 m 569
+ 552 557 Ô 562 566 & 569
P 552 558 M 562 m 566 u
& 569
r 552 X 558 @ 562 7
Ml 566 570
552 jf 558 563 n 566 m 570
X
*
552 558 563 s 566 570
P 552 558 563 9 567 M 570
• 553 H 558 563 % 567 570
± 553 B 559 563 567 570
± 554
*
M 559 563 567 14 |j| 570
554 559 M. 563 567
554 560 563 567
-k 554 ± 560 563 567
3/5 549
(2) (particle) of. (fife) E C 8, App.

2: B 26: NO 9.
— yi, one. #
t( ik) B 2. t^(*ih) B 82.
Z. saw, three. (3) See -g,
X shang, above ( || ruler, supe hu, (1) interrogative particle, B 41.
rior, EC 9: A 36, 38. ((ft) E C 5. (2) preposition replacing in
T (1) hsia below ( II _h): subject, constructions confusable with the
quantifying ^f /10$ . . . jfe pattern
t
inferior, A 36, 38.

(2) See ft. (But cf. p. 165), B 72, 74: NO
15, 16.
(1) pu pre-verbal negative.
4s Graph (also written ]g) substituted
t
(2) See -g. for ] E by the Empress Wu jjSj

shih, generation. (§ 1/4/2), written ft in the Taoist
JL ch'ieh, (1) be about to, defined Patrology and ft or 2§ in the Mao
A 33: A 84: B 41, 51, 68, 82: and Horyaku editions: A 53, 56,
NO 16. 83, 84, 98: B 21, 23, 24, 31, 62.
(2) moreover, B 27, 38, 44, 45: A ch'eng, ride, NO 14, 15, 17.
NO 16.
f See P..
XL ^
jc. ch'iu;(l) mound, NO 2.
(2) See flg. itL A 94: B 62.
*. (l)*'o ( =flfe),another (§1/3/4/8),
| ^ A 17: B 3: NO 15, 16.
·+. chung, (1) centre, middle, defined

(2) she( = m. See Jg.
(3) yeh, particle (§ 1/3/12/8). Type
A 54: passim. 1: final particle in the pattern *(X)
(2) hit the mark, coincide with, Y yeK (X is Y). For its frequent
A 83: B 62: NO 5. (t^) A 85, 88. use in embedded phrases (A 39 j | ,
JlkfS'fii "see that this is a pillar"),
cf. § 1/3/4/8. Type 2: in the nomi-
nalised verbal or directive pattern
*H wan ball. (t^L) B 62.
y 'X chih Yyeh'. Type 3: contras-
ii. C/IM, master: put in dominant tive after temporal words (*n% j|ff,
position, B 42. and proper names. Type 4:
after subject with adjunct shih ^.
/ *t II
#
For t ( ± ) , see § 1/2/1/13.
luan, disorderly ( ||
7) nai, particle contrasting the com
plements of 'X Y sentences
J ^
(§ 1 /3/3), EC 2, 5 : NO 1, 3, 14-16,
18. J*jjik is this (answering -f yu, give to, B 69: NO 11.
question 'What is it?'), NO 4, 13. f shih, (1) serve, E C 3: A 39: B 58:
A. chiu, endure, persist: duration NO 15.
(§ 2/4/2/2); defined A 40: 44, 50: (2) service, business, affairs, EC 7,
B 14, 15, 46, 61, 64. the 12: A 28.
durationless, the moment, A 44,
50:B15.
chih (1) (pronoun) him, her, it,
them (§ 1/2/1/2/13). A 75: ~- erh two.t
B 74, 75. (Replacing jjfc as adjunct # ching, well.

before noun) this (§ 1/3/4/6). JL wu, (1) five.
A 31, 78: B 3, 50. (ftfj) B 71. (2) See fr.
550 Indexes
ft ho, what? (§ 1/3/7), B 41.

(*#:) A 43, 60, 67 (§ 1/2/1/3/4).
TZ (1) wang, be or become absent,
A 76, 85, 88: B 3, 17, 26. tf* A 12.
(2) wang (= Jg), forget, EC 9. №• zvei, seat, position, f f i B 59.
£ ?, A 12. t(*?£), B 23.
£ C/IKJO, interchange, interplay, A 88:
& lai, come, NO 2.
B25a. ft J/M'A, to cause, to commission,
# yi, also (backward reference), analysed A 77: A 11, 14: B 9, 29,
(§ 1/3/1), A 33: B 23, 24, 32, 55, 55, 69, 77. (Used intransitively) do
67, 68, 73, 76. after all not, as you are told, B 55, 69.
EC 10, 11. fR t'ung, agree, conform, defined A
39. (U) A 39.
ft mow, equalise, parallelise, denned
A £p NO 11: NO 12.
A. jen man. f(* A) A 98, B 19: NO 2, № SeeS,gg.
15;(fA)B71 (§1/2/1/3/8).
1%. pien, convenient, NO 2.
4- chin, now, EC 2,5: A 40: B 16,50,
fl. hsien, appear, (tft) B 4. (^,) A 1,
53, 82: NO 5,10. îfe now on the 11, 59: B 4: NO 9. f g B present
contrary (§ 1/3/12/8), B 27, 32, day, EC 2. (§ 1/2/1/2/11)
33, 70, 78. f A 40.
it hsin, trustworthy, defined A 14.
jen, benevolence, defined A 7: EC
2, 7:B76. fft lun, modes of behaviour graded
according to relative social posi
tfc A 83.
tion. (f&) A 9. {f^lj the arranging
4* ling, command (§ 1/4/22), defined in sequence of lun, defined EC 9:
A 18: A 77, B 55. EC 10.
W yi, preposition marking nominal % chu, all, both (§ 1 /3/6), A 9, 39, 67,
unit as means: conjunction be 70,74,78, 86, 88: B 2, 3,7,12, 22,
tween verbal units, marking former 35, 48, 58, 65, 66: NO 11. (+{B)
as means and latter as end A 43, 60, 67. f B 4, 12, 22.
(§ 1/3/12/4).
Representing ,Q£f§ *deem\ B 47, f£ pei, (1) double. Defined A 60: EC
71. 4, B 5.
J
(2)(= fif). See ft.
ft wu match point by point (§ 1 /4/34),
y
B 58, 76. ( £ = ffi), A 98. (tfr) ft yi, lean on, B 28.

B 58. f* hu(= jfft), tiger(?), A 32.
jen, bear the weight of a respon tik p'ien (§ 1 /4/18), in one case but not
sibility, defined A 19: B 5. (fffi) the other, A 46: B 3, 4, 7, 63, 66:
A 12. NO 12. (Verbally) do in one case
№ pi, put side by side to measure but not the other, A 83.
against each other (for radical, cf. (2) pien (= S), all the way round,
§ 1/2/1/2), defined A 68. (fffiO A 49.
A 68. (Hi) A 88. \ik chia, loan-name, B 8: NO 5.
ik fan, converse (§ 1/4/9), B 30, 72. fl£ ch'u, to divide off from the rest
(•Mft), defined A 73: A 74. (R) (§ 1/4/7), B 63. (fjg) A 73. (R)
A 83. (fft) A 83. ( № ) B 3. B12. ( £ ) N 0 6.(tS)Bll.(1«)
1ft B58. B 12.
& See ft. f B 59. ft ch'uan, transmit, A 80, 81: B 38.
ft tso, originate, A 15, 16, 17, 18. ( t » ) A 81. fA 95.
A 68. Pk shang, wound.
3/5 551
ffc man, impolite, A 9. (Used adverbially) love

f|l hsiian (= JJ), to circle round, everyone, EC 5, 13: point
denned A 47. at them both, B 38.
f* yu, ample(?), NO 9.
iflt hsiian, let the scales settle (§
1/2/1/2/10), A 75.
*W- meng (= H|), cover over, (tg£)
A 41.
X- hsiung, elder brother. Jl ^

fo/awg, light.
A fan, generalising pre-nominal par
£ foien, beforehand ( || |&), B 63, 64. ticle, E C 6: A 73: B 27. f E C
t( 3E = *l). B 73. ( f « | ) , B 57.
#
App. 7. Cf. B 62 n. 512.
Used of knowing, referring or
desiring 'a priori* (§ 1/4/13), EC 2:
A 93: B 38, 57. W
& mien, escape, enable to escape, EC
8. fA 49, 88. & hsiung^ malign: (of a bad year),
· * I'll, hare(?)(§ 1/5/6), (1&) A 49, famine, EC 3.
y
88. A ch u, come out ( || A)> A 78: B 76:
NO 16: bring out, utter, express,
A 32: NO 11. | B 71.
^ ju, enter ( || &), A 12: B 76: NO

16. (tA) A 98: B 19: NO 2, 15.
t(*A) B 71. 71 tao, coin (the knife-shaped coins of
1*1 inside ( || ^(v). fB 76. the states of the North East), B 30.
^ a pair of, A 64: B 27, 54: ( · № ) A 75.
NO 18. (Used adverbially, § (1) fen, division, A 2: B 22: NO 6.
1/4/18), in both cases, A 68, 73, (2) See g .
88: B 25, 48, 70. (Verbally) do in m Seej£.
both cases, A 83: B 39. £'1 lieh, arrange in sequence, EC 9,10.
%\ pieh, separate, A 97.
*«I li, benefit ( || defined A 26:
EC 2, 3, 6-8, 10, 11: A 8, 12, 13,
hung, mode of address to noble, 26, 27, 35, 75, 84, 93: B 44, 76:
A 9. NO 6.
^ hung, together, NO 15. Run smoothly, B 11: NO 9.
t & Seejg. #»J Me (= jjil), pull from the side,
Mi, his, her, its, their (at some B 26.
stage written "JV, B 3 n. 275). $] tao ( = $]), turn over, B 19, 22.
Resumptive after demonstratives
f«J tz'u, prick, B 26.
(St jtt» IS) renders them posses
sive and contrastive (§ 1/3/4/9). m See^.
#
t # , B 24. m B 3. t( $$) B 73. •ft ch'ien, before ( || f£). (Used ver
& chii, be fully provided, EC 2: NO bally) move to a position ahead,
6 (?): implement, B 56. A 42: B 60.
Hts Men, a total of countable 'units' M hsiao, scrape, pare, B 10.
(H), A 2, 46, 86: B 67. (Used IiJ tse, (if . . . ) then (§ 1/3/11/1-4).
verbally) do to every one, B 73. M Men, sword.
552 Indexes
1) *P pan, half.
# tsu, die, EC 3.
ft It, strength, denned A 21: B 5,
*) nan, South (§ 1/5/13).
10, 26. fA 75.
jfo chia, put on top of something, add
to, impose on, EC 3, 5,10: A 79: P *
B 25a, 25b. (t#0) B 25a. (Adver It wei, dangerous, EC 2: NO 12.
bially) more, EC 4: B 7.
tit hung, achievement, denned A 35:
EC 3: A 36: B 69. (X) A 83, r ^
B 51. & hou (1) have bulk, have dimen
y
•b chieh, do violence to, B 27. sion, defined A 55. fA 69 n. 182.

% yung, courage, denned A 20: B 3. dimensionless, A 65, 69.
wu, (1) busy oneself at, A 94. (2) have more of, do more for,
treat as more important, EC 3, 4,
( 2 ) ( = f | ?), mist(?), BIO. 5, 7, 9, 10: B 69. (|| ft).
th tung, move, stir, defined A 49:
EC 2: B 60.
№ sheng, win, prevail (§ 1/4/24). Of
winning in disputation, A 74, 89: ·£; ch'ii, leave, depart from ( || gfc) A
B 35, 70. Of one force prevailing, 60, 88: B 22, 29: NO 2: get rid of,
B 25, 43. $f$jg£ win in disputation, EC 2: A 46, 93: B 3, 4, 7, 29, 31,
defined A 74. 45, 48, 61.
ts'an, align (primarily of three
*] ^ things in a row, for example, two
to wu, negative imperative equivalent posts with the rising or setting
sun, A 57 n. 158), at one stage
to m±. (§ 1/3/5/2), A 83. written ^ (B 22 n. 347), A 57,
S. poo, wrap, B 3. B 38. B 38. (-Hfc) B 22.
it hua, transform (§ 1/4/38), defined

)X (1) fan, turn back, B 20. (ft?) B 26.
A 45: A 85. (2) See №.
it (1) pet, North. i i yu, friend, B 53.
(2) pet (= #), turn back on, B 22.
№ B 28. A chi, reach, A 62, 63.
shou, receive, A 80.
C ^ Jpl ch'ii, take one rather than another,
choose, E C 8: A 95, 96, 48, 81:
*g k'uang (Sung taboo, § 1/2/1/4/2). NO 11, 12: take away, B 60.
(1) (tffi) square frame, A 53.
(2) (ttt) square off, A 59.
k'ou, mouth (as organ of speech),

A 32, 78: NO 9.
01 ch% delimited area (§ 1 /4/7), A 48, *T k'o, admissible, allowable. A suc
63: B 22.
ceeding verb is passive after oj,
active after "^DX- inadmis
sible (used causatively B 67,
+ shih, ten. § 1/3/13). ^ o j not admissible
*+ ch'ien, thousand, NO 8. without further evidence, B 67.
3/5 553
£ yu, right ( II £ ) , B 25a, 76. f E C 2. $ (1) shang, merchant, B 53.

g B 65. (2) See ft.
6] kou (= curve. Mistaken by a • i t ti, complement (§ 1/4/29).
post-Sung editor for a substitute vulgar graph) A 22. (aig) A 89.
graph for the tabooed H^J, RnJ < t » A 51, 85. (fg) B 5.
(§ 1/2/1/4/6), A 47, B 19. ch'ang, take the lead in singing,
cA'tA, hoot, A 79. B 69.
ancient times, the past, A 40: M wen, ask about.
B 16, 17, 53. | A 83. * See
£ A 16, B 65. -4- hsi, be pleased with, A 26.
£ See?£. 4- shan, good, A 10: B 16, 81.
ho, be together, analysed A 83: -ft sang, lose, B 26, 75.
(ffi") A 83: (of tallying) A 14, 80: y
ch ien, deficient, A 16.
( t S ) B 65: (f*&) B 65: (of com y
posing one and the same thing), Us ch ang, taste, B 44. Pre-verbal
A 86, 87: B 11: (of light con particle of indefinite past, B 61.
verging), B 23. \%) B 61. Pre-verbal particle in
injunctions ('Try to ...'); CUf)
ft hsiang, face towards, (tffl) B 22. B 38.
$ J't/ttg, (1) same ( || %), analysed
A 86, NO 6: passim. same P ^
in length, defined A 53: A 54, 58.
& ssu, four. t(*ffl) B 2.
(2) See flfl.
0 yin (originally writtenflg?), criter
# B9. ion (§ 1/4/39), A 97: B 3,15. (f
& rning, name (§ 1/4/21), analysed defined A 71. (Used verbally)
A 78: A 9, 31,32, 80, 86 :B 3,29, depend on, have as grounds, E C
33, 39, 48, 53, 68, 70, 72: NO 1, 1, NO 9. (Used adverbially) by
6, 9, 11. ( t £ ) A 78. call by this criterion, NO 2, 17.
a morally good name, A 10. IS k'un, get into difficulties, NO 10.
3? fou, negative verb equivalent to @ ku, inherently (§ 1/4/14), B 3,
preceding verbal unit negated by 38, 39.
^ (§ 1/3/5/3), B 12, 34, 73, 78. M kuo, state, NO 18.
CfC) A 94: B 52. SI yuan. See jlit[§|.
wu, I, my.
IS t'uan, spherical, cylindrical, B 24.
* pi, compare (§ 1/2/1/2/5), B 6.(J£) № ) B 57, 62.
A 88: NO 9, 11.
U yuan, circular, circle (§ 1/5/10).
% kao, tell, inform, A 32, 81: B 38, Defined A 58: NO 8. (M) A 70,
70, 77. ( t § ) B 9. 93, 98.
M chiin, ruler, defined A 34: EC 9,
A 39.
• ming, (1) to name (§ 1/4/22), A 20,
78, 79: NO 1, 2, 7. t*u, earth.
(2) decree (of Heaven), destiny, & tsai (§ 1/4/30): (with noun of
NO 16. place) be in, EC 6, 10, 12: A 37,
B 14, 20, 32, 33, 37, 70: (before
# chou, all the way round, of all cases
verbal phrase) depend on, B 19:
(§ 1/4/18), B 69: NO 17. (jfl) B 82. (intransitive) be present, exist,
(t^) B 5, NO 13. EC 2: B 3, 42. (Nominalised)
fa ho, accompany in song, B 69. presence, existence, B 17, 41:
« (=H).SeeE|. (causative) locate, B 16 (§ 1/3/4).
554 Indexes
it ti, the ground, B 27, 29. fB 21. 1/3/4/10), B 70, 72: NO 9, 10, 12,
equal, B 52. 13, 16, | A 50.
$ ctoag, city-wall, A 14, 93. 3^3fe
:
with regard to, B 27.
A t'ieny heaven, sky, EC 1, 7: B 41,
» See$.
& (1) chih, hold (§ 1/4/5). Of uphold 49. fB 24.
ing one of alternative claims, E C A yaoy die young, NO 16.
1: NO 9, 16. (Nominalised) the k shihy lose, fail, NO 12. fNO 17.
claim upheld, EC 1: A 93, 94. Of yang. centre, A 88.
controlling conditions, E C 2, 8. & chia, toflankon both sides, A 62,
(Nominalised) conditions, EC 3.
(2) chih gjfc). hibernate, B 39. 63: B 18, 29.
£ Vang, hall, B 42. tse ( = §g) lustrous (f^), B 22.
f
£• (1) Men, hard ( || A 88: B 15, # See^g.

37. Verbal phrase is-as-hard- # hsiy how?: NO 11, 15.
to-white (distinct but mutually & to, snatch from, rob, B 69.
pervasive, § 1/4/3), defined A 66: # fen, exert oneself, A 21. (tUD
B 4, 14, 15. A 21.
(2) ch'ien (= Ig, $ ) , drag, haul,
B 28.
tui, pile up. ( t № ) B 29. (ff|) B 29.
f*vA 11, B 25b (§ 1/2/1/3/7, 1/3/1).
pao requite, A 36, 38. fB 3.
y
ifhao, to like, NO 16.
& Yao, ancient sage revered by all
# kit, (injunctive particle) let us for
schools, B 16, 53.
the present . . . ( ? ) , A 33.
5 Mo, Mohist, NO 15, 16.
jfe
shihy commence, defined A 44:
See ^p.
B 33, 36.
it hsing, surname, A 78, 88.
* £P
t& hsien, confusable, NO 6.
-ir shih, knight, A 19, B 10. chia, marry off daughter, B 31.
# shou longevity, B 44, fNO 16.
y
JL hsia summer, B 10.

y
f & ~f tzu, (1) child.

(2) polite substitute for 2nd person
£h wat, outside ( || j^J). pronoun, -f'^t-f" honorific form
£ £o, relatively much, many ( || ^?). of the name Mo-tzu used by the
A 88: B 3, 6, 22, 43, 44, 59, 69, Mohist school, EC 2.
78: NO 12, 15. fA 78. # ts'un (§ 1/4/30), belong in, B 37,
#L yehy night, B 6. 42: (be where it belongs) remain,
stay as it is, A 46, 85, 88. (Used
mengy dream, defined A 24.
causatively) keep where it belongs,
preserve, EC 6, 8.
¥ tziiy style-name, ( f ? ) A 78.
-fc ta big (measured by circum t(*?) B 63.
y
ference, B 21 n. 341) ( || /J\). # hsiao,filial,defined A 13: EC 10.

A (l)/w, husband, B 3: in brave It (1) shu, which? (§ 1/3/6), A 96:
fellow, commoner,EC 3, B 3. B 6, 42.
(2) fiiy topic-marker, which before (2) See ffc.
common nouns generalises (§ I£ hsiieh, learn.
3/5 555
#fr chiang, intend to, A 12(?).

?, NO 7. X the purpose of
£ chat, house, NO 18. X is to, NO 6.
^ yii, space, A 41: B 13, 14, 63. #• tsun, honour, EC. 9.
(t^) defined A 41. (f^) B 63.
*f tui, reply, B 41.
t(*^) A 78.
A 41.
% yi, appropriate, one mode of rela
tion (^) contrasting with the •b hsiao, small ( || ^ ) .
exact (IE) and the necessary y (1) shao, relatively few, little
A 83, 88, 96: B 31, 43, 44, 51. ( || ^), A 88: B 7, 22, 44, 59,
t ?, ECApp. 12. 69, 78.
£ h'o, (guest, stranger) the tempor (2) shao, younger, A 88.
M See _h.
*. *
arily present, EC 2.
% shih, room, house (§ 1 /4/26), EC 6,
12: A 86, 88: B 31, 42, 70.
A 88.
£ hai, harm ( || flj), defined A 27:
EC 3, 4, 8: A 26, 27, 37, 75, 84: №. chiu, approach ( || £ ) , A 88: B 22.
NO 6. Interfere with, be inconsis
tent with (§ 1/4/12), A 20: B 5, 27,
45, 51, 65, 73, 75. t(*JS) B 5,
r
NO 13. K ch'ih, measured foot, foot-measure
jung, allow room for, A 42 (?), A12. (§ 2/4/1/4), A 2, 42, 60, 63, 65, 67:
B 27, 29, 70: NO 8.
£ ch'in ( = Jg), chamber, B 29.
& wet, tail, B 66.
£ chia, family, A 88. | A 41.
& chii, dwell, live in, NO 2, 18.
?s su, (stay overnight) put off to the
<& li, shoes, B 3, 58.
next day( ?), A 88.
yft chii, shoes, B 3, 58.
& han, cold, B 10.
JSi shu, attached, dependent on, A 87.
t fu, rich, enrich, EC 11: B 53. (jjfg)
A 88.
% kua, a few ( || EC 5: B 69, 78:
reduce to only a little, B 22. iU shan, mountain, B 81: NO 2.
£ ch'a, scrutinise, discern, NO 6, 9. fNO 9.
It (1) shut, (solid) particular and
concrete object (§ 1/4/21, 1/4/26), «( £p
A 31, 32, 78, 80, 86, 88: B 3, 33, JH See JH.
53: NO 1, 6, 11.
(2) shih, fruit, NO 18. X IP
(3) See M. X See Si.
3| hsieh, to draw, A 58. £ «50, left ( || £ ) , B 76.
$ shen, make a thorough inquiry *5 cA'uio, skill, sophistry, A 10: B 10,
into, B 71: NO 6, 12. 39. S^flJ sophistical turns (§
1/5/14), A 94, 95.
It she, shoot with bow, B 19. tl C/M, oneself.

chuan, alone, on one's own respon C yf, come to an end, analysed A 76:
sibility, EC 6. A 51: B 10, 33, 51: (adverbially)
556 Indexes
already, A 33, B 61. ;fvf#B have

no choice, EC 8, B 69. tfj hsingy shape, NO 2, 7. Of the
human shape or body, (Jfl]) A
21,22.
ifc ch'ang, constant, unchanging, A \ ^

48: B 41, 43: NO 12.
is. wang, go off, go away, A 54, 65, 88.
| B 29.
piy the other (of alternatives,
+ P'ingy (1) flat, level, defined A 52: § 1/3/4/7). (Used causatively, §
B 27. 1/3/13), B 68. fA 73 n. 187.
(2) calm, serene, defined A 25. ft hotly after ( || gif, # · ) . ffijfa only
then (§1/3/11/1,4), A1:B3, 8,51,
64: NO 17. (ffijg) A 68: B 45, 49.
* £p (îfiHJ?) A 69. only then (§
to yu, young, EC 9. 1/3/11/4), EC 2: NO 17.
taiy await the fulfilment of a con
dition (§ 1/4/27), A 35, 77, 78:
r ^ B 49: NO 10, 17.
t# A 61 n. 165. $L (1) ts'ung, follow from, proceed
M fu storehouse, B 11.
y
with, EC 3: A 93: B 10.
#
0L tUy measure, length, A 88: B 70, 81. (2) t ( № ) . see fê.

ft tej get, (pre-verbally) succeed.
t# = for 38). A 48, B 19.
Used of occupying the same
J& lien, honest, defined A 17. spatial position, A 65, 67: B 15.
t# A 80. ^fs} E have no choice, EC 8, B 69.
#j miaOy shrine to ancestors, NO 2. hsiy shift in place, B 13, 14. ( f f é )
fit fei, to put down, to place on the A 49: B 13, 17 (§ 1/2/1/3/9).
ground, B 10, 27, 29. (frS) B 29. fll fu, ( = $1)* doubled, compounded
fik kuangy broad, B 4. (§ 1/4/11), B 11, NO 9.
* Seejffi. f/S hsiin, follow, accord with, B 10,
NO 9. (flg) A 15. ( f i ® B 10.
#
t(*ffi) B 4, 64. t ( » ) B 68.
It te, merit putting another in one's
& lu shed, booth, B 10.
y
debt, EC 9.
$H cheng, distinguishing marks, A 45.
}t chierty set up, B 59.

hsin, heart (organ of thought), NO
9, 15, 16. f(*lh) A 96, 97:,B 26,
?I ;ym, pull. 32. t(*!E) A 53.
016 pi, certain, necessary, defined A 51:
% fu, negative equivalent to ^fC'i, one kind of relation (-§·), A 83:
(§ 1/3/5), EC 2: A 16,18, 75: B 27, A 3, 14: B 32, 38, 41. Used of
32, 39, 41, 44, 60, 79. logical or causal necessity, EC 2:
tiy younger brother. A 1, 4, 8, 39, 74, 77, 78, 87: B 2,
fa hsierty chord, hypotenuse, B 27. 3,8, 9,15, 22, 23,24, 25b, 29,35,
48, 49, 51, 60, 64, 69, 70, 71, 73:
f$ jo, weak (|| 55), E C 2: A 12. NO 1, 2, 10, 12. fB 47.
& ch'iang, strong ( || | § ) , NO 10; put
strength
?$ miy into,
spread ECA9:40,A41.
over, 12.
3/5 557
t * B22. 4K WO, I, my, me.

& chih, intent, EC 3: A 8, 13, 20, 80, ^ huo, some, one or other of two
83,88. %•£ (= -g) Will of Heaven, (§ 1/3/6), defined NO 5, examined
EC 1. B 33. (Used of place) somewhere,
& chung, do one's utmost, be whole A 42, 49: B 13. gj5g- some, B 44,
hearted in service, defined A 12. 74.
*f7f hsin, open, A 14. $. ch'i, distant kin, EC 9.
chi urgent, A 19.
t
JR chan, battle, B 32.
f£ hsing, nature, EC 1.
t * OS?) A 6, 88 (§ 1 /2/1/3/7). f *p
& /MI, cease, B 17, 69. f hit, door, A 49.
.% ch'uan { — III), thread together, ff\ so, relative pronoun substituting
B 69. for object or directive (§ 1/3/12/3).
& (1) ow, dislike ( || $0, analysed
A 84: passim.
(2) o, ugly ( || A 30.
(3) wu, (equivalent to J^fnJ) in ik See tttj-
what? from what? (§ 1/3/7), B 38,
42. (J*) B 42, 57.
If shit, dredge, clean up (§ 1/4/37),
NO 11.
f£ t'an, tranquil, A 25. #T C/JC, bend, snap, A 88.
& chih ( = © ) , wisdom ( || /&), de m fit, soothe, cherish(?), EC 2.
fined A 6: A 75: B 44. (f&l) A
U (1) C/KIO, summon, B 3. fJo B 3.
6, 88.
(2) ch'iao ( = ^ ) , receding at the
\% (1) wei, is only, is specifically
edges( ?), B 24.
(§ 1/3/3), A 55: B 11, 12, 46, 72.
(2) See fif|. tfc B 3.
4fc C/JM, push away, B 28. fB 11.
№ See%i
& yii, the more . . . the more . . ., ^ ch'ieh, pull up from above, B 26,
B 26. 27, 29. (%) B 26.
^ at, love. $6§£ l °v e a s
individual, -H 5
support, EC 4.
A 7. love everyone, EC 5, 13. 4ft C/HV* (§ 2/4/2/5), (1) finger, EC 8:
& yw, foolish ( || g*), A 75. fB 10. A 75: B 12: NO 1, 3, 7.
* yi, idea (§ 1/4/37), A 14, 70: B 38, (2) point out, B 38, 39, 53.
41, 57, 58: NO 9, 10: (verbally) (3) meaning(?), NO 9.
conceive, have idea of, NO 3. t E C App. 3.
tW A 81. Cf. B 20 n. 339, B 57, n. 487. t'ing> pull, draw. gJ$J (?), NO 7.
\k chieh, make contact, be in touch,
t& shen, take pains over, A 75.
A 5.
I* Seef|. ttt. A 53 n. 152.
it •/«, think, defined A 4: EC 2, B 50. C ^ ' M I ( = 3E), hang down, droop,
JUL yingy answer, B 41. B 25b.
tg chu> afraid, B 32. #L (1) f'm (push, || $|), infer, defined
NO 11: B 2, NO 12.
(2) See ft.
(3) See
t* B 55. W B5.
ch'eng, come about, bring about, $1 ywaw (pull, || ff£), adduce as
A 1, 19, 76, 77: B 19, 58. Of evidence, defined NO 11: A 93,
establishing a claim(?), A 88, 94. NO 12.
558 Indexes
tft sun, reduce, lose ( II defined

A 46: A 19: B 44, 45.
X wen, markings, EC App. 5. B 50.
*4& yao, shake, waver, (t^) B 62. #
t( ^.)» so corrupted only when
4* See H. chin is pronoun (§ 1/2/1/2/13),
£ mu, describe (§ 1/4/20), NO 6. A 31, 75, 78:B 3, 50, 74, 75: (for
& wao, bend, B 25a. ( ± ) * ± ) B 82.
ta?, pick out, select, EC 6: A 96.
t& B62. ft ^
$$. c/w, depend on, B 42. *T ssu, cut (§ 1/3/1), A 46.
Ifc m, show what X is like by offering 9f cho, hoe, mattock : chop up, B 60.
something which is like it, A 31. IfiJt tuan, cut off, EC 8.
fll cho pluck out, B 50.
t
№ ying, coincide in position, defined

A 67: A 68, 69.
* /**£(§ 1/4/10). (1) square, defined
A 59: A 80: B 29, 65: NO 1.
(2) just now, A 33: NO 8. (t#)
A 88.
A p'u ( = ^?), a thing in its rough (3) direction, B 33 : guideline,
shape(?), A 58, 59. NO 12.
shou, receive from below, arrest as (1) yûy directive preposition (§
it descends (§ 2/4/2/3/2), B 62, 1/3/12/6).
27, 29. (f/R) B 26. (2) wu. See §§.
ft* See<fg. & (1) See tft.
$t &«> renew, reform, B 17. (2) ?, B41.
& (§ 1/4/10), depend on, be %L lii, travellers, see $fJÊc-
relative to, A 88. % p'ang, side ( || ^0-
& ku, reason, cause, defined A 1:
A 77, 94, 96: NO 10-12: the thing
as it is in itself (§ 1/4/14), EC 12:
A 88: B 7. (Initially), therefore, VL chi, already, EC 1, 8.
A 32: B 19, 22, 24, 33, 37, 70, 77,
82: NO 1, 5, 7, 10, 12. for
this reason, NO 12. jfejbffcM
there is no other reason for it, « ./VA, sun, day, EC 2, 4: B 20. (fB)
NO 15, 16. B 10, 20, 47. 0 pfi sun at noon,
defined A 56: B 10.
#t hsiao, exemplify, NO 5.
J> (l)taw, morning, ( f a §1/2/1/3/4),
4t chiao, teach. A 40, B 14.
& pat, smash to pieces, defeat, B 31, (2)taw( = ti), dread. (fjL) B 50.
NO 1. t-f (•ff») A 85, 88.
=
pi ( ^)» collapse from exhaus j£ hsi, formerly.
tion, B 10. h yi, substitute, exchange (§ 1/4/38),
# kan, dare. A 45, 85 : B 29, 30. (Used of inver
$t chingy respect, A 9. sion) the two sides exchanging
places(?), A 48, B 23.
it shu, number, B 11, 12, 22, 74.
*fl ming> (1) clearly understand, make
lit (1) /w, spread out, B 63. clearly comprehensible, A 6, 29,
(2) fu (=|R), blossom, B 3. 30, 72: B 70: NO 6, 10, 11, 15.
3/5 559
(2) eyesight, A 3: clear-sighted,

A 6.
ftfj See ^ mu, wood, piece of wood.
& ch'un ( = teeming of insects pen butt, basis, B 25b, NO 12.
y
coming out of hibernation in wet, not yet. Expressing negative

spring, B 39, 50. alternative after the verb (§
1/3/5/4), A 89, B 3. Tôjftj not
A shihy this, the said thing (what a
knowable without further informa
name means and an object is
tion, not knowable 'a priori'
judged to be), § 1/3/4/2-7. When
adjunct the head is followed by -fji, (§ 1/4/13), A 75: B 58, 73.
§ 1/3/12/8. Used between verbal # yi ( = slant ( || IE), B 21, 22:
units in the 'X Y sentence, C№)B26:(Jfe)B27:(t*)B21.
§ 1/3/3. Used of the right one of H (1) ts'ai, resources, A 3.
two alternatives, § 1/3/4/4. (2) ts'ai, plinth of pillar, B 29.
l t s
it is/it is not, * this/it
$L See K..
is not this, § 1/3/3. know
what it is, A 93, B 38. jfc tungy East,
9$ shih, a time, A 35, 40, 44: B 41, 53. ttt. B26.
fA 82. i& chuy pillar: to support from below,
chih, know (systematically replaced B 29. (£Ë)*|i B 29. f A 59.
by in the Canons and Part A mou, 'X' (substitute for any name),
Explanations, § 1/2/1/2/2), defined NO 2.
A 5, analysed A 80: passim. (t#P) 4£ rouy soft, yielding ( || ig), A 88.
A 11.
#C (1) chiao, compare and make the
* ying ( = g£), shadow, t i l B 23. same, B 38.
jjk poo, criminal, violent, EC 1: B 3. (2) chiao ( = gc), twist, B 25a.
y
#t t aOy peach.
$1 t'ung. See fflS-
yiiehy say (before direct speech), fll ken, root, NO 6.
A 33: B 38, 41, 48, 66, 67, 70: £ ?, ECApp. 1.
NO 4, 11. # Si, ladder, B 27. fc$) B 27. ($|J)
t ( * R ) B 10, 20, 47. B 28.
# tsuiy most, A 1(?), 61. /tang, bridge, A 51.
C/J'Î, abandon, reject, E C 8, B 3.
ft # c « , bramble, NO 18.
fit c/rW, hammer, B 58. (f|) B 58.
% (1) yu, have, there is ( || For
relation to existential 'to be', cf. § №) B
-
5 8
1/2/1/2, A 88 comment, B 49 com $i chi, culminating, extreme in

ment. For its quantifying function, degree, B 25a.
cf. § 1 /3/6. Of possession of proper fà. chien, bar of door, A 53.
ties, B 47, 66. ^ £1 have what dis f# ytfwg, willow.
tinguishes it as . . . (§ 1/3/12/4/3). № ying, pillar (§ 1/5/9), A 39: B 57:
(2) yu ( = 50, also, in addition NO 3. (fjg) B 57.
(forward reference) (§ 1/3/1), A 28,
42: B 13, 14, 33, 38. tW BIO.
fu, (1) wear. 3£HK jade pendants Jung, to flower, display itself
on cap, A l l . spontaneously, A l l .
(2) submit, devote oneself to l é (1) s/m, hinge, pivot, A 49.
task(?), A 94. (2) See flg.
560 Indexes
# ytieh, music, EC 7.
ft piao, tip, B 25b. (JK) B 25b. ft ssu, die.
f$ c/ru, orange, B 3. Hi tai, pre-verbal particle expressing
flt /teM£, cross-bar, B 27. a fear or venturing an opinion
t * B57. (A 33, comment), A 33, B 53.
fjt ch'iian (§ 1/4/8). (1) weigh one ffc shu> separate, distinct, NO 12.
consideration against another, de
fined EC 8: EC 7: A 84.
(2) leverage, positional advantage,
B 25b, 26. *H t fS tuan (= $$), hammering block,
B 58.
t # B 58.
ifl B 55.
tz'u, next, defined A 69. Used 3* sha, kill.
nominally of the successive stops # tien palace hall(?),B 55. fJR B 55.
t
in a sequence( ?), NO 9.
3fc yii, desire, analysed A 84: passim
# £P
(II SB).
-ft Ao, sing, EC 1. # WIM, mother.
ft. See {B. # WM, (1) negative imperative, B 37
(§ 1/3/5/2).
(2) used for wu 4H£, to avoid
confusion with the latter's quanti
fying function (§ 1/3/6), EC 10:
jh cAft, stay (§ 1/4/4), defined A 50. B 43, 60, 72: NO 15, 16.
Of physical motion, A 12(?): NO
16. § 1/2/1/3/10) B 26. Of
name temporarily staying with
changing object, A 43, 50, 78, 93: # See ft, Bit.
B 68. <t#) A 51. (tjfc) B 82.
(Used transitively) cause to stay,
fix, A 75, 96, 97: B 1. (fifr) A 96, 4. ^
97: B 32. ( t £ ) B 2. AL mao, body-hair.
it ctewg (see also ffr)* exact, dead on
(§ 1/4/2), one mode of relation
( £ ) , A 83: EC 7, 8: A 56, 84, 98: A. IP
NO 9, 12. (Of bodies), upright, f\ shih, a type of surname, B 8.
B 21-24, 28, 62. (Of edges),
& (1) people, A 35, B 64, 74.
straight, A 53. A 53. (Of
surfaces), plane, B 22, 24. (Used (2) See | g .
transitively) adjust, put right,
EC 1, 2: B 28, 31, 51, 68, 70.
*t tz% this (§ 1/3/4/2), independent
pronoun not used as adjunct & cJrV, energy, zeal, A l l .
(§ 1/3/4/6). Tz'u (the object or
place here) contrasting with shih
Jl? (what it is judged to be), B 1, * *p
2, 33. Used causatively, 'treat as 7^ shut, water.
this' (§ 1/3/13), B 68. $L ch'iu, seek, defined EC 7: EC 1, 7:
jJt suit year, B 30. A 4, 28, 94: B 81: NO 6, 11.
3/5 561
3f hsia, descend, the radical perhaps lack the distinguishing charac

distinguishing from hsia ""p in teristics (§ 1/4/35), A 73, B 34.
sense of 'be below', B 27. $ jan, so (equivalent to 'like this',
B 25b, 27, 29. § 1/1/2/6), A 1, 24, 33, 43, 51, 70,
p=- Chiang, Yangtse river, EC App. 9. 71, 97, 98: B 1, 2, 8-10, 16, 22,
ft See J | . 33, 51, 61, 65: NO 4-6, 11-16.
<• № ( = №)» shave wood, B 10. After descriptive word, A 25, B 57.
f&M* see ?£. fB 11.
ch'en, sink, B 56. (For the two
n See ^J.
graphs, cf. B 56 n. 483, 484.)
& je, hot, B47.
j£ fa, standard, model, defined A 70:
ft shou,ripe,mature, EC 3. (J^) A 51.
EC 9: A 71, 95, 96: B 65: NO 5.
tB 29. M i/ii<> torch, illumine, B 20.
chih, set in order ( || j|L), defined
A 28. Of ruling the state, EC 11:
A 28, 85: B 16: NO 6. Of ordering
one's own life, A 28: B 44. Of # cheng, contend over, A 73, 74.
curing illness, A 76. & (1) wei (§ 1/4/32), analysed A 85,
if yang. (§ 1/3/14), undecided passim: do, make: is to be deemed
(?) B 58. as, counts as (§ 1/3/3). Before verb
tiL /«/, flow, NO 12. B 27. rendering it passive, E C 5. X
ch'in, soak into, encroach on. g Y, deemX to beY. deem,
suppose.
(§ 1/4/31), encroach on and vitiate,
EC 1. (2) wei (§ 1/2/2/5), for the sake of,
defined A 75: EC 1, 2, 5, 7, 9-12:
ch'ien, shallow ( ||
A 7, 75.
% shen, deep ( || gg), EC 9: B 41. fk chiieh, noble rank, B 6.
& yin, vicious. See jgS?.
2$ chieh ( = $!), exhaust, EC 9.
:
A lou, leak, EC App. 8.
Uk ?, E C 1 . № ch'iang, wall, E C 2. (J^) A 75.
;'t Unknown graph, perhaps yii
( = ff), foster, EC 2.
jf marsh, B 81. **f" ww, ox: stock example of some
:
M shih, damp, A 77. thing which an object either is or
i l Seejg. is not (§ 1/5/2).
* wu, thing (§ 1 /4/33), EC 2: A 5, 6,
78: B 2, 9,12, 65, 80: NO 12,13.
A /wo, fire, B 11, 43, 46, 47. (f^) t'e, one of a pair(?). (tBf) A 82.
B47. (f^C)B21.
*'«w, charcoal, B 43. A IP
Mf yen, (1) directive pronoun equiva ^ ch'iian, dog (§ 1/5/3).
lent to yii chih >.
(2) (conjunction) only then (§ It /aw, offend against, A 37.
1/3/11/5), NO 2, 6. k*uang, mad. <j:E$ refer arbitrarily,
& wt/j have not, there is not ( II ^f: B 66, 76. t A 53.
for quantifying function contrast ft Aow, whelp, dog (§ 1/5/3), A 79:
ing with zvu cf. § 1/3/6). B 8, 35, 39, 40, 54.
Originally written ^ (§ 1 /2/1 /2/14)
#
t ( & ) B 57, 73. ( · . £ ) A 67. » (§ 1/3/8). (1) (With X) it is as
(with Y).
562 Indexes
(2) (pre-verbal) still, even so.

(pre-verbal) still, B 72. ^tjgr-til is
* ?, EC App. 7.
as it was, B 50.
yi, different ( || f^|), analysed A 87,
ffi tuy alone, B 21, 38: NO 11.
NO 6: passim.
fl[ (1) Huo, 'JilP, abusive name for
bondswoman ('captive* ?, § 1/5/12), «§· liiehy summarise, NO 6.
EC 2: NO 14,15. % hua draw, delineate, A 32.
y
(2) huo, catch of game, NO 3. 1£ (1) tang, be plumb with, dead on

( II 3 § pass, exceed, §1/4/28), A 44:
ffc shouy animal,
B 22. Used of name fitting object
flt liehy hunt, EC App. 10. or statementfittingfact, A 14, 50,
74: B 12, 35, 46, 71.
(2) See Hf.
shuaiy altogether, throughout, NO

12.
*fc yiy doubt, A 75, 83: B 1, 10, 50:
NO 6.
3L vw, jade, A l l .
i wa/i£, king, B 30.
r ^
M pingy illness, A 76, 85: B 9: NO 18.
ft /i, pattern, arrangement (§ 1/4/17),
Of illness becoming critical, EC 2:
A 75: B 78: NO 9, 10.
B45.
« se, (1) zither, B 6 (§ 1/5/6).
& chiy illness, EC 2.
(2) glistening(?), B 6. | A 49.
M ?. B 45.
$ huang, semi-circular jade orna
ment, NO 3.
& />ai, white. (f@) NO 18.

"5" paiy hundred.
& sfott, extreme in degree, B 80. • (1) chiehy all (confined to 'X is Y*
sentences, § 1/3/6), EC 9, NO 2.
(+*) B 80.
(2) chieh ( = #£), together, A 98.
4 IP
shengy live, be born: (used causa- & yingyfill,defined A 65: A 66, B 4,
tively) engender, generate ( || ^g). 15, 73.
Defined A 22: EC 4, 6: A 88:
B 2, 32, 44. Of name stemming JL yiy increase, gain ( || J||), EC 3, 9:
from object (A 39, B 53, NO 10), A 19, 47: B 44, 45, 77.
action from thought (EC 2), $L taoy rob, robber.
desires from situations (EC 2). & chin, exhaust: (adverbially) all
(§ 1/3/6), defined A 43: passim.
m ip a ip
#J yungy use. Of employment in office,
A 8. Of utilising as evidence for 3 mw, eye.
a claim, A 98: B 66. | B 22. J. chihy straight, defined A 57: A 98:
^fl^, unidentified binome(?), A B 23, 26, 27, 38.
88. t# NO 18.
3/5 563
*fl>' miao, blind in one eye. (tSM) B 3. ^» Ch*in, name of a state, NO 2.

(TO) NO 18. # yi shift, transfer, A 79.
y
39 (1) hsiang, pre-verbal particle of

№ chii. fJIJS, describe a curve(?),
reciprocal action. A 47.
(2) hsiang ( = image, imagine.
ffi ch'eng, cite as authority, EC 9.
B 38, 57.
chung, multitude, many ( II U), EC
5: A 15: B 78: NO 18.
^ hsiieh, cave, tunnel. [M/^ demar
flfc ni, look askance, A 4.
cated empty space (§ 2/4/1/4), A
# tu, inspect in order to verify, A 14. 48, 63.
St &'MW£, hollow, NO 15, 16.
* lp % ch'iung, limit, defined A 42: A 75:
shih, arrow, (tpfe) A 50 (§ 1/2/1/3). B 61, 73.
final particle in verbal sentences
marking a change of state (§ A IP
1/3/12/7), A 28, 98: B 26: NO 10, JL li, to stand, make to stand, A 83,
15, 17. B 22: NO 10.
fa (1) chih, the intelligence, consci #f ping ( = lay side by side, B 29.
ousness, defined A 3: EC 2: A 4, tuan, starting-point of a measure
5, 6, 22, 23, 25, 98.(g)A 75: B 5. ment, defined A 61: A 1, 2, 60, 63,
(2) See 67, 68, 69: B 19, 60. Used of
chii, carpenter's square, A 59. speech-pause( ?), NO 9. (flL|)
^ft tuan, short ( || Jj|). NO 9.
# IP
^ IP teng> grade, rank, A 9.
^ shin, stone (§ 1/5/7). t& See
** yu\ arsenolite. (f#) EC App. 3. cfow, needle, B 5.
* IP * IP
(1) chi, sacrifice to, NO 18. i£ SM, millet, B 6, 44.
(2) chi ( = $£), border between, SeeB$.
A 49. #1 *i, marketed grain, B 30.
31 chin, forbid.
*S See %.
(1) lit rites, propriety, defined A 9: if) yiieh, tie together, bind together
(It) EC 9. by a treaty, A 34.
(2) See ft.
tZ chi, main thread binding skein of
silk, the institutions binding
society together, NO 6.
Yii, the sage especially revered by & chi, provide with the full amount,
the Mohists, EC 5. B 61.
e o n
ft ch'in> bird, NO 3. & lei ( = JJ|), Pi^ e above another,
B 29. fB 38.
& ssiiy thread, B 29.
> IP
chiieh, snap, B 29, 52. Of snapping
tfc ( = ftD, catalpa, B 57. the relations between things, treat
ft tfo'A. See $WE. ing them in isolation, A 88.
564 Indexes
№ chingy canon, doctrine accepted as

authoritative, EC 1. ^ erhy ear, B 5, 29: NO 9.
H yuan, proceed along an edge, B 23. 5t shengy sagely. sage, EC 2, 7,
H hsuan ( = Jg), hang, B 27, 29, 52: 9, 12: A 97. that which is
plumbline, B 62. sagely, A 83.
Ht hut, embroider in various colours, Wl wen, hear, analysed A 81: A 80:
B 58. B 32, 70: NO 9. f(*Pe1) A 62, 63.
0. shengy rope. (tBB) A 80.
№ ?, defined A 64. Unknown graph, perhaps w£
( = =J), be in charge of, A 17. <Jfi)
defined A 15.
I& ts'ung, sense of hearing, NO 9.
5 crime, defined A 37: A 38, $ shengy voice, sound (§ 1/4/25), A
75: B 69: NO 16. ( || 11 (sound of metal). Used of the
•£5 fa, punishment, defined A 38. vocal enunciation of names, A 78,
B 53, NO 9.
1*1 ^
h mei, beautiful ( || gg), B 3, 55: NO ftf kan, liver, B 3.
15; of fair repute, A 29. kung, forearm, NO 10.
$ ch'un, multitudinous, NO 6. Jft fei, lungs, B 3.
6 (1) yi, righteousness, duty, defined
A 8: EC 3, 7, 9: B 76. JSL ku, thigh, NO 10.
m it (1) neng, be able. (Nominalised)
(2) ( = 411), ° d e l , example, B 53.
ability, A 8, 13.
(2) t'ai ( = possibly written as
« *F § 1/2/1/2/12), posture, B 23.
fifr hstUy long (of distance, or contrast
ing length with breadth), B 64.
( t « ) B 4, 64 (§ 1/2/1/3/5).
ft yii, feathers, B 10. M fu, dried meat, A 75.
W pi ( = f№), buttock, haunch, B 45.
(tlj|) B 54. (f№) B 69.
laoy old. Jlfc wan, arm, EC 8. EC 8.
Jfc See fl$.
$t che, post-nominal particle, which 0 c/wao, stick together, B 29 (?):
after nominalised verbal phrases NO 15, 16.
appears always to mark the agent
(§ 1/3/12/2). particle com huo, meat soup, B 53. (ft) B 53.
bination equivalent to 'unquote';
after a single word in an Explana & *
tion marks it as quoted from the Hi cA'eftj minister, A 34. jcSfS runaway
Canon (§1/3/10). slave, B 39.
ffi? erh, conjunction of subordination № wo, sleep, defined A 23: A 24.
and co-ordination in verbal com (1) Tsang, 'Jack', abusive name
binations (§ 1/3/12/5). for bondsman ('hideaway' ?), stock
example of the man who can be
loved only as a man (§ 1/5/12),
EC 2, 4, 6, 7, 13: A 78, 83:B53:
ffl 0Uy mate, A 80. NO 14. ({$) EC 4.
565
(2) ts'ang (= hide away, B 50.

tin, look down at, B 22. &
ft *
colour, A 93: B 70.
6 (1) oneself, A 15, 98: B 40: /wf (= ft), allotted portion in life,
NO 15, 16. A 8, 13. (#), EC 9.
(2) from, A 33, 54: B 16, 27. & 70, (1) be like: (of measurements)
fNO 18. (t£L) A 33. be equal (§ 1/3/8, 1/3/14). (Before
1* B22. verb) seems, A 98: B 22, 38.
not like, A 7,11, 74: B 3, 81.
£ *P with regard to, B 27.
iL chih, arrive, reach, EC 9, 10, 11: (2) if (§1/3/11/6).
B 17, 29: NO 8, 17. #0fM have (3) or (§ 1/3/11/8), B 10, 70.
a degree which they reach, NO 12. # kou, if (§1/3/11/6), EC 6: NO 1, 2.
HM using as criterion the degree f E C App. 12.
which they reach, NO 9. 3tf||
f- mao, reed, B 3.
slight to the utmost degree, EC 9.
$>C chih, cause to arrive, A 32. #'] (1) ching, bramble(?), B 56.
t * («*(), A 51, 85 (§ 1/4/29). (2) Ching, a name of the state of
Ch'u ^ , NO 2.
% (1) mo, none, no one ( || nJ5).
& £p
(2) ww( = f ) , evening, A 40, B 14.
& (§ 1/2/1/2/14). (1) chii(=m> lift,
$ meng, populace, A 34.
A 21, B 5: (g£), B 5, 10.
(2) yii, together with (originally £ chu, become manifest, A 6, 7.
written ^ ?), A 22, 96: B 2, 3 and •$j wan, ten thousand,
passim: (f-R) A 83, B 41: (1 jj$) it p'eng, dense(?), B 10. | B 10.
A 67: (fJ4) NO 7. IL pi, cover, B 19.
(3) yii, interrogative equivalent of
final yeh B 2, 10. t$ tang, dissolve, A 85.
}% po, have less of, do less for, treat
I? fcmg, raise up, EC 8, 10.
as less important ( || JJ[), EC 5, 7,
(1) chit, refer, pick out object by
name, defined A 31: A 32, 79, 97: 9, 10: B 69.
B 38, 53, 63, 66, 76: NO 1, 11; 41 chieh, suppose that, EC 4.
pick out for favour, EC 10. i f f & See ^ .
A 32.
it *
(2) See 151 (1).
iS. ch'u (§ 1/4/26), occupy spatial
% tA21.
position, A 22, 86, 88: B 13, 53,
>t IP 62, 75, 81: dispose in their proper
#* ww, dance, EC 1. places, NO 6: a spatial position,
A 66: B 22, 39: NO 6.
hsii, empty, A 64.
C/*OM, boat, B 27.

№ ch'uan, boat, NO 15.
1*0 (*$) shih, louse (§ 1/5/6), A 88,
B 6. ( f S ) A 49.
* IP she, snake, E C App. 5. (•{& = $&)
jl /*att£, good of its kind, adequate EC App. 13: B 39, 50.
to its function, A 88. tt See Jjfti.
566 Indexes
f t IP t IP
ft hsing (§ 1/4/4). (1) to go, B 27, 63, cT yen, say, a saying, defined A 32:
64; NO 10. Of proceeding from EC 2: A 14, 33; B 71: NO 6, 9,12.
what is so of an instance to what 18 cAtfott ( = JpJ), scrupulous, sque
is so of the kind, B 1, 72: NO 10,
amish, defined A 16.
11, 12.
mao, describe (§ 1 /4/20), A 94. (£8),
(2) conduct by which one is A 5.
morally judged, defined A 10:
EC 3, 7, 9: A 14, 18, 80, 89. Vt shuo, explain, defined A 72: EC 1:
A 80, 98: B 1, 2, 32, 44, 66, 70,
(3) iEff» the five elements, B 43. 82: NO 10. X explained by
X, B 1-82. #№l& have/lack
#f hengy horizontal, crosswise, B 25a,
25b, 38. grounds to offer (§ 1/4/40), A 99,
B 32.
n ytiy expound, EC 1 :(}&?) EC 1.
^ lp
t$ pei> self-falsifying, illogical (§
yi> cloth coat, A 35, 76: B 3, 69. 1/4/23), B 8, 71, 73, 77. (fg$) B
34, 79. Of ft) B 76.
4C Miu, fur coat, A 35.
Meng (graph tabooed under Sung
and replaced by flf, § 1/2/1/4),
® IP sincerity: defined A l l .
© hsi, West. •ft* Seef<j>.
(1) lun, to sort things out in dis
JL IP course, putting them in their
proper grades and positions (§
JL (1) C/MCH, see, analysed A 82: A 5, 1/4/19), A 6, 89: B 34, 78: NO 6.
39: B 4, 9, 46, 57. (2) See {ft.
(2) See fS.
* shut, who ? which ? (§ 1 /3/6), B 41,
$L kuei, compasses. 44, 70. fB 29.
ik Ming (= what is essential to
fJL S/M'A (at one stage written ^ , B 3
being an X, that without which
n. 272). (1) look at, A 14: B 47. something may not be named 'X'
(2) cause to look at, show (= 7$), (§ 1/4/6), EC 2: B 81: NO 9.
B 53. t(*№), B 3. (*$J) NO 15.
ifc fei, blame, defined A 30. Used for
Ul cA'm, (1) parent, close kin. (fiSA) causative ^ (§ 1/3/13), reject, B
NO 15. 77, 78, 79.
(2) (adverbially) by direct personal i$ mou, to scheme, devise, A 75.
contact, A 80, 81: B 70. ifc chu, (1) collective pre-nominal
#| kuan, (followed by object) observe, particle, EC 1, 2: A 32: NO 1, 2.
A 75, 95, 96: NO 12: (followed by (2) = B 16: NO 11.
directive) travel around in observ If wei, call, say of (§ 1 /3/12), analysed
ing, A 80, 81.
A 79: passim. X >M Y. It is X
which is meant by Y, EC 7-9.
^ ^lIBik. It is of X that it is said,
A IP A 21, NO 8.
ft Mao, horn, B 66. |g no, assent (phonetically related to
jo y§ 'like', § 1/3/4), A 93, 98.
If Meh, loosen, dissolve, NO 15, 16. Mh ( = S6)> remember, record,
I* f?.B5. B 11.
3/5 567
:
& yu, praise ( || ij| ), defined A 29:
t after JEJ^, EC 2, passive after Jg.,
EC 11. B 34).
fEC App. 3, A 32. lu. :£JLJ£& five roads (five senses),
ft tUy read, NO 16. B46.
£ pietiy alter (§ 1/4/38), B 7, 29, 30. № ?, B 27.
/
shou ( = f^), make a sale, close the
deal, B 31. k ^P
i% jang, defer. % shen, one's person, EC 8: A 18,19,
11$ A 59. 81, 88.
№ ?, B 28.
X IP
4 cA'f, how? (rhetorical expressing * IP
inadmissibility), NO 6.
$- Me, carriage, B 27: NO 15.
jf chun, army, B 32.
# unidentified part of a carriage,
tlL mao (written ^ in Wu manus B 27.
cript). (1) looks, visible character
j& fwif, carry load( ?), B 27.
istics (§ 1/4/20), A 48: B 22, 65:
NO 2, 7. ( f S ) A 32, 47, 71. # cA'i'ngr, light ( || Jf).
(2) See m- Ift lun, wheel.
|& Milan, small spokeless wheel of
a ip hearse, B 27.
t* B56n. 485. chuan, turn round, A 94: NO 12.
({#) A 95. ({$) B 20.
B25a.
H / M , carry on one's back, B 25a.
(t jS). B 25a. Jh *
it (1) kueiy noble, socially or morally £f (=jf), offer an illustrative
or financially valued ( || JljS), E C comparison, defined NO 12.
3: A 9: B 6. Of price (expensive), ift te'w (§ 1/4/31), phrasing, verbal
A 88: B 6, 30. formulation (EC 1), used of
(2) See » . sentence or proposition (NO
% maiy buy ( || A85: B 3, 30. 10-12).
tA 85. pien, the distinguishing of right
If May price, B 6, 30, 31. and wrong alternatives, disputa
tion, defined A 74: B 35: NO 6:
1t shang, reward, defined A 36: E C
to make a distinction, E C 3:
i i (ii 13). subtlety in making distinctions,
% hsien, worthy, EC 10. NO 9.
* ^P
& *'K ( = №)· E C App. 10 №fe L IP
hunter. & ying, go towards to meet, B 20.
& Mao, jump over, A 93. sitchin, near ( || ^S).
H MUy head in the direction of, A 75. il chut, pursue, EC App. 2.
& *'ao, escape, B 38, 39.
s. % i& m, go to meet. j$i$f£ inn (welcome
JL tsu, (1) foot, leg. to travellers), E C App. 11.
(2) sufficient (the verb is active i& mi, lose one's way, EC App. 2.
568 Indexes
& feng> meet with (circumstances),

B 10. (M), B 10.
x ip
JL li, village, N O 2. Chinese mile,
x£L Sung, (be unobstructed through
(ll)*=P*S, N O 8.
out) current everywhere, A 34:
be fully conversant with, B 41. 4 (1) chung, heavy ( || f®.
(2) ch'ung, duplicate, redouble,
& lien, connect, A 87, N O 6. Used of B 18. Used of identity of object
personal ties( ?), B 44. called by different names, A 86:
i$ tai, reach. in the case of, B 38, 40, 54: N O 6.
N O 17. •jt Hang, measure, B 6; N O 1. +B 23.
jft chin, advance, go forward, B 60, 63.
3& kuo, pass, exceed, err ( || 'gj be 46- i p
plumb with, dead on, § 1/4/15),
& chin, metal, money, A 11,14: B 43.
A 5, 50, 98: B 10, 22, 24, 33, 40,
41, 58, 69. (tjft) B 10. $% chui, awl, B 26. | B 58.
i t tao, road (§ 1/5/14), A 97, N O 11. t i t N O 8.
i& yii, happen on, chance on, EC 3, № chietiy mirror, B 22, 23, 24. (jg)
8: A 15. Used of one event hap B 22, 23, 24.
pening to be at the same time as m See§g.
another, B 10. (fft) B 10.
it.ta, get through all obstructions,
A 78. (1) cfc'awg, long ( || $g).
(2) chang, prolong, li 19, 23: grow
i& (1) sui, then, consequently, N O 12. older, B 13, 41, N O 10: elder,
(2) chut ( = H), drop, B 26. EC 9, A 88: lead, A 85.
i£ yun, (originally written Jip without
radical ?, § 1 /2/1 /3/19), rotate. (1 f£) PI *P
defined A 48: B 19. (fit) A 80. fl men, gate.
g32 residence or migration, N O 2. ffl />*, close, N O 15, 16.
i i yuan, far ( || 2£). fA (1) C/MCW, interval, A 62, 63,64, 69:
#
i t (1) shih, merely, barely, B 29, 45, B 20. (tM) A 62, 63. t( IW) A 80.
^TB] have an interval between,
69.
defined A 62: $T||g] have no in
(2) ti. See ]g. terval, A 69.
if. tsao, meet with, EC 2. (2) chien, intervene, be between,
it. ch'ien, change residence, N O 4. defined A 63.
i t yiy overlook, leave out of con M ktian, put, B 29 (n. 393).
sideration (§ 1/4/18), A 75: B 39,
74. (^) EC 11, B 39.
i£ huan. U S ( = fgifjg) uneven, Jf- IP
irregular, A 88. P/C k'eng, pit, hole, EC App. 6.
№ ch'u, get rid of, EC 8.
1$. ch'en, propose, lay out, EC 1.
£ IP f# yii, corner, A 59.
•ft pang, state, B 31.
#P hsiangy district, N O 2.
yung ( = 38)» P » ot E C
App. 8.
H*. sui, although, EC 6: B 78: N O 15.
(If) A 17, 35, 37: B 44, 66. (Df)
58 chiuy wine. N O 1, 2, 10.
3/5 569
H chi, cock.
$L (1) tsa, mix up, B 48.
IS" shou, head.
(2) tsa (= ffj), make a circuit,
(tritt) A 59.
№ (1) //, part from, apart, B 4, 11: J* *s
NO 12. ,8/ ma, horse. Stock example of a
(2) li (= $g), meet with (a mishap), common name (§ 1/5/2).
A 75.
J$ chii, colt, EC 1.
& wa«, difficulties, EC 2: B 2; object
to, raise difficulties, A 88, 94. MM Jf» ssii team of four horses, B 2.
t
there are no difficulties, B 67, 73, $c sao stinking, putrid, A 75.

75: NO 15. $L huan ( = #C)» enjoy, EC 2.
II ?,A75. JR ?, B41.
ft /i, sable (of horse), NO 14.
m ip
H" /m'ao, sleet, A 85.
* ( 1 ) * « » ( = 0 , jR), crane (§1/5/4), # IP
A 88: B 6, 8, II. fl£ (1) limb, member, EC 2: NO
*
(2) See f l . 7: countable unit (part of whole
or member or sub-class of class),
defined A 2: A 1, 7, 61, 67, 82, 86,
fei negative copula (§ 1/3/3): the
y
87: B 12, 22. Used causatively
wrong one of alternatives (§ (§ 1/3/13), EC 8. (|f) A 46.
1/3/4/4): with or without final (2) See |S.
(§ 1/3/5/7-10): causatively (§
1/3/13). (t.DlO B 79. % IP
% mi, consume, B 43.
% kao high ( || ~f).
t
-S" yin, sound, NO 9. U IP

%L fa, hair of head, B 52.
ir ip
№ faftt (= g|), instruct( ?), A 85. n *
?, B 47.
№ tou,fight,B 3, 10: NO 16.
$1 yew, face, B 76.
H lei, sort, kind, EC App. 1-13: A 78,
86; B 1,2, 6, 66: NO 6, 10-12,15.
not of a kind, A 87: B 66. f yii, sell ( || M), A 85: B 30, 31.
68 ku, look round at, pay attention to,
EC 9. EC App. 9.
^ IP &t/«, spirits of the dead. EC 5,

S/MV*, eat.
NO 18.
ifc ;ym, drink.
& />ao, eat one's fill, B 45. & IP
It yang, to nurture, EC 4. Jk yii, fish.
yii, surplus food, leftovers, B 45. № fu(=Wi ?)» attach(?), NO 6.
570 Indexes
IP
Mj niao, bird, A 88: B 2. black.
A
& cKun quail, A 45. (fH)
y 85(?).
& IP
ft ?, EC App. 1.
& mi, milu deer (§ 1/5/5), B 2, 6,
11, 45.
M. li, to link (§ 1/4/16: at some stage 9L shtiy rat, EC App. 6.
written ?, B 3 n. 275), B 3.
(M) A 79. fB 3.
# (1) cA'f, come out level with each
# IP other, A 63.
huang, yellow. (2) Ch'i, name of a state, NO 2.
571
3/5/2 PROPER N A M E S
Abe Ryüichi 70n Chi Ch'ih aft {c. 300 B . C . ) 22

Akiyama Tadashi ik\Ui& (1702-Chiang Ch'uan ï£$k 6n
1763) 74 Chiao Hung mt (1541-1620) 68n
Archimedes 54, 389 Ch'iao Ch'ui 11
Aristotle 44, 445, 446 Ch'ien M u mm 3n, 6n, 52, 196
Ch'ien Pao-tsung 386, 387,
Bergson, Henri 366 389, 394, 395
Bodman, Richard 227 Ch'in K u - l i (early 4th
century B.C.) 16, 19, 52, 246
Chang Chan 3ft ?Ä (4th century) 76 Ch'in-tsung ^ (1126-1127) 86,
Chang Ch'i-huang 3ÄÄM 101, 435
102n, 249, 258, 293, 321, 349, Ching, King of Chou J S f t ï (544-
382, 387, 451, 470 520 B . C . ) 9
Chang Ch'un-yi 3ft№-^ 349, 382, Ch'ing H u H $ (died 549 B.C.) 9
395, 412, 425, 434, 439, 443, 485, Ch'ing Y i n KJK (died 549 B.C.) 9
489, 491 Chomsky, Noam 169
Chang Hui-yen 3ftÄB (1761-1802) Chou Fa-kao J S & ^ 124
70, 219n, 321, 387, 425
Chu if
11
Chang Tai-nien 3ft tS№ 6n Chu Ch'i-feng 182n, 204,
Chao Ch'i jütfe (died 201) 207, 208 21 In, 297, 349, 380, 429, 451
Chao Chi-pin ffißfö 66n Chu Hsi-tsu fctfffl. (1879- ) 4n
Chao Chiin-ch'ing №M№ (c. 200) Ch'uan M i 9, 10
308 Chuang Po ffifâ 208
Ch'ao, Prince ϊ ί ΐ β 9 Chuang-tzù ï t ë i (?369-?286 B.C.)
Ch'ao Kung-wu M & Ä (died 1171) 8, 15, 18, 21, 25, 39, 40, 59, 62,
69 67, 70, 173, 180, 181, 183, 193,
Che-tsung (1086-1100) 87 195, 200, 218, 224, 225, 262, 319,
Ch'en Chen-sun UUU 69 334, 340, 398, 403, 417, 431, 441,
Ch'en Ch'i-yu WSfifc 388, 389 445, 446, 447, 453, 455, 456, 457,
Ch'en L i fi|« (1810-1882) 71 459
Ch'en Yuan W i t (1880-1971) 85 Confucius ï L i (551-479 B.C.) 3, 6n,
Cheng Ch'iao M№ (1104-1162) 68n 10, 11, 14, 17, 18, 20n, 24, 25,
Cheng Chung-ying ) Α Φ ^ 464 179, 198, 261, 358, 402, 412
Ch'eng Hsüan-ying Γ&£?& (fl. 631- Creel, H . C. 198
650) 66, 76
Ch'eng-tzü S i (late 5th century Enoki, K . S — S É 70n
B.C.) 25 Eratosthenes 370
Ch'eng Y i mm (1033-1107) 69 Euclid 44, 71, 307, 328
572 Indexes
Fang Shou-ch'u 3n, 6n Huan T'an UU (c. 43 B . C . - A . D . 28)

Forke, Alfred 54n 174
Fu Shan f#[J4 (1607-1684) 70 Huang Chen K)M (1213-1280) 69
Fukui, Shigemasa JftWlSffi 64n Hui, King of Ch'u « * 3 £ (488-432
Fung Yu-lan №#m (1890- ) B.C.) 6
6n, 33, 172, 198n, 460, 463 Hui Shih MWi (late 4th century B.C.)
19, 21, 25, 58, 61, 62, 66, 172,
Han Fei tzu (died 233 B.C.) 173, 175, 180, 181, 195,224, 225,
8, 21, 22, 23, 24, 59, 192, 261, 261, 266, 289, 302, 306, 334, 341,
262, 272, 286 436, 444, 449, 463
Han Y i i (768-824) 68, 69 Hui-wcn, King of Ch'in
Hatta Ryukei AM№81 (1692-1755) (337-311 B . C . ) 5n
70n Hume, David 301
Hayashi Shuichi Wfi—* 70n Hung Chen-huan 379, 381,
Hou Chi B№. 226 384
Hou Wai-lu fefVa 66n
Hsi chung m$ 11 Kao Heng M¥ (1900- ) 73,
Hsi Shih 426 258, 352, 396, 407, 419, 427
Hsiang-fu (c. 300 B . C . ) 23 Kao Ssu-sun iSHJKJ^ (c. 1160-c.
Hsiang-li ffiSR 23 1220) 69
Hsiang-li Ch'in HlMiJ (c. 300 B.C.) Kao-tsung W£ (1127-1162) 87
22, 23 Kao-tzu S ^ f (late 4th century B.C.)
Hsu Hsing ufff (4th century B . C . )
25, 61, 398, 451
52
Karlgrcn, B. 221, 352, 357
Hsu Fan №3G (4th century B.C.) 52
Hsu Fu-kuanfefflSL66n K u Ai-chi 82, 195n, 216n,
Hsu Shen fttK (fl. 100) 173, 174, 434
175, 416 K u Chich-kang KSBPM (1893- )
Hsiin Hsu ^W) (died 289) 89 69n
Hsiin-tzu tu-f* (?298-?238 B . C . ) 7, K ' u Huo 7M (c. 300 B . C . ) 22
18, 20n, 21, 32, 37, 39, 43, 52, Kuan Chung 1S№ (died 645 B.C.)
59, 60, 63, 64, 65n, 67, 70, 179, 17
187,191,195,196,197n, 199, 209, K'ung Ch'uan T L ^ (4th century
213, 228, 233, 234, 262, 268, 269, B . C . ) 20n
270, 281, 301, 322, 325, 350, 355, Kung-meng (late 5th century
402, 415, 431, 440, 468, 479 B . C . ) 14
H u Shih ffl.® (1891-1962) 22n, 72, Kung-shu Pan (late 5th
95, 243, 293, 485, 489, 491 century B.C.) 6, 7, 11
H u Tao-ching S K I $ 19n, 69n, 172, Kung-sun Hung (200-121
174n, 175n B . C ) 64n, 65n
3/5 573
Kung-sun Lung <5H£ffi (late 4th Machiavelli 21

century B.C.) 19, 20, 20n, 27, 33, Mao K ' u n 2&*|i (1512-1601) 74
43, 61, 62, 66n, 66, 90, 117, 144, Mencius £ i (?372-?289 B.C.) 7, 8,
161, 171, 172, 173, 174, 175, 176, 17, 18, 19n, 23,25,46, 50,51,52,
187, 199, 207, 208, 218, 227, 230, 59, 67, 69, 70, 181, 184, 197n,
235, 246, 313, 345, 355, 362, 363, 207, 272, 281, 451
364, 404, 410, 416, 421, 432, 434, Meng Sheng % № (died 381 B . C . )
440, 460, 461, 462, 463, 464, 478, 4,5
482 Meno 417, 418
Moore, G . E . 52
Lao-tzu (? c. 300 B.C.) 8, 179, Mo-tzu i S i (late 5th century B . C . )
261, 262, 454 3, 5, 5n, 6n, 7, 7n, 16, 17, 22n,
Lau, D . C. 19n, 198n 25, 33, 45, 50, 52, 59, 60, 65n,
Li Chuang-tzii commentator 434 66, 67, 69, 70, 71, 72, 230, 235,
L i Sstt (died 208 B . C . ) 430 236, 241, 243, 246, 248, 262, 329,
L i T'ao (1115-1184) 69 335, 374, 410, 412, 420, 423
Li-tsung mm (1225-1264) 86
L i Yu-shu 380 Needham, Joseph 6n, 54n, 387,
Liang Ch'i-ch'ao (1873- 389, 394, 412
1929)71,72, 73,93,95, 171,387,
388 P'eng Meng WSk (4th century B . C . )
Ling, King of Chou mmji (571- 193
545 B . C . ) 9 Pi Yuan * g n (1730-1797) 70, 73,
L i u Ch'ang SIM 293, 307, 308, 309 74, 84, 87, 101, 106, 129, 204,
L i u Chu 87 285, 299, 309, 430
L i u Hsiang SOft (79-8 B.C.) 20n, Plato 302, 307, 417
65n, 172n, 199, 479 Po Y u im 9
L i u Hsin 81 f* (died A . D . 23) 65n
L i u Hsu 9№ 87 Rubin, V . A . 8n
L i u Shih-p'ei M№№ (1884-1919) Ryle, Gilbert 44
219, 274, 326
L i u Ts'un-yan WtfE. 219, 326, Shang Yang i S « 61
395, 427 Shao-lien 'J>*M (c. 500 B . C . ) 412
L u Sheng fe№ (c. 300) 66, 67, 67n, Shehadi, Fadlou 25n
68n, 70, 175, 176 Shen Kua i t f ö (1031-1095) 69, 87,
L u Te-ming Wm\ (died 627) 416 350, 372, 377
L u Wen mm 74 Shen Pu-hai (died 337 B . C . )
Luan Tiao-fu 68, 73, 75, 198
82, 85,86, 88,89,91,92, 94, 352, Shih P'u Sffl 9
364, 387 Shun S 22, 346, 423
574 Indexes
Sivin, Nathan 372, 379, 380 Tuan Yii-ts'ai 182n

Socrates 418 Tung-fang Shuo JK#DS (?154-?93
Ssü-ma Piao (died 306) 66, B . C . ) 173
76, 82, 176, 344, 364, 374, 492 Tzu-chan 9
Sun Hsing-yen Mmi (1753-1818) Tzu-ch'an 8n, 9
70 Tzu-chou Chih-fu - T - f f l ^ 16
Sun Tz'ü-chou Jfefcft 4n, 52 Tzu-hsi -T-S 9
Sun Yi-jang Mmm (1848-1908) 70, Tzu-k'ung ? ? L 8
70n, 71, 72, 77, 81, 84, 95, 99,
101, 104, 119, 129, 144, 167n, Waley, Arthur 19n, 197n
204, 211, 219, 221, 229, 252, 263, Wang Chung ffit (1745-1794) 70
309, 329, 332, 365, 474, 477 Wang Ch'ung ^% (27-c. 100) 174,
Sung Hsing 7^fif (4th century B . C . ) 223
181 Wang Ming £ B J 65n
Sung Lien 5fcflK (1310-1381) 69 Wang P i (226-249) 66
Swain, Richard H . 464 Wang Yin-chih (1766-1834)
293, 437, 491
Ta-lien 412 Watanabe Takashi 4n, 5n,
Tai Wang MW. (1837-1873) 434 6n, 8n
Takata Atsushi rftffli? 6n Wittgenstein, Ludwig 44
T a n Chieh-fu 88, 91, 92, Wu, Han Emperor (140-87
101, 171,230, 358, 474 B . C ) 64n
T a n g Μ 346 Wu, Empress « B (684-704) 169
T a n g Chün-yi m^Wt (1909-1978) Wu, King « 3 E 346
191n, 485, 489 Wu Hou {c. 300 B . C ) 22
T a n g Yao-ch'en « § l i 5 74 Wu Ju-lun (1840-1903) 70,
T a o Tsung-yi (r. 1323- 429
1403) 69 Wu K'uan M S (1435-1504) 74
Teng Hsi fPt/f (6th century B.C.) 67, Wu L u M i l (late 5th century B.C.)
344 52
Teng-ling-tzü δ Ρ Ι ^ ί (c. 300 B . C . ) Wu Yu-chiang ^ t t y X 68,74,75,77,
22, 23 83, 86, 89, 104n, 129, 211, 277,
Ting T u T Ä (990-1053) 195n 297, 311, 427, 429, 435, 485, 489
Ts'ien Tsuen-hsuin ü f f IM 392
Tsou Po-ch'i fffSiS (1819-1869) Yang Chu (c. 350 B . C ) 15, 16,
71,72 17, 18, 19, 23, 46, 51, 245, 246,
Tsou Yen Wfä (late 4th century 254, 280, 413, 415
B.C.) 20, 199 Yang Liang t§{& (fl. 818) 188
T u Kuo-hsiang & ! № 66n Yao m 16, 22, 34, 115, 231, 346,
T u Yü ttS (222-284) 358 369, 421, 422, 423
3/5 575
Yen Shih-ku S I A (581-645) 459 Y i i Hsing-wu ¥ « 5 (1896- )

Y i 3 ? 11 396
Y i Chih (late 4th century B.C.) Y u Yiieh filM (1821-1907) 3n, 465,
52 467, 474, 491
Ying-tsung ^ (1064-1067) 87 Yueh T a i or Y i ^ S / ^ (6th cen
Y u m 49, 188, 250, 423, 488 tury?) 68
3/5 577
3/5/3 SUBJECT
Aggression. See 'War'. which overlap with 'being': Yu wu

A priori: 37f, 188f and passim. Hsien 5fe 'having and lacking': 26, 339,
'beforehand* not defined: 62. 'A 341, 418; Shih fei 'being or not
priori' is conventional equivalent, not being this': 26, 38f, 40f, 117f; Jan $t
to be exactly equated with hsien: 'being so': 26, 35,40f; Tang g 'being
Preface. Wei k'o chih T{C nj^fl 'not yetplumb with': 203; Ts'un wang 'ffiC.
knowable' distinguished from pu k'o 'presence or absence': 340f; Wei g
chih ^ \ oj^d 'unknowable': 125. Chain 'being deemed': See 'Deeming'; Wei
of definitions establishes circle as $\ 'calling': 25, 145-147; Wei ff£
known 'a priori': 57f, 342f. Another 'being specifically': 116-119, 362-
establishes moral as desired 'a priori' 364, 447f.
on behalf of the world: 47-49, 247f, See 'Essence', 'Copula'.
271. Hardness but not whiteness of Benefit and harm (li hat Îj^f). Defined:
stone referred to 'a priori': 37, 222, 48, 282f. Their weighing: 45-47,
404, 406. Idea of pillar known 'a 252-255, 321-323, 332. And Indivi
priori', of hammer not: 37f, 224f, dualist conception of 'life': 281-283.
428-431. And metaphor of the wall: Weighing of benefit disliked by
223f Mencius: 281
Adducing (yuan $g): 342,483f. Defined: See 'Weighing'.
483 Benevolence (jen £ : ) 270-272. Defined:
Ambiguity. Analyses of ambiguous 49, 270. Definitions of other schools:
words: 35, 323-336. Words changing 64 (n. 79) 179, 272. Slight importance
meaning with context: 354f. Ambi for later Mohists: 49, 50, 252.
guous idioms exposed by parallelism: Whether internal or external: 50, 451
41-43, 487-489 'Bigger pick*. See Ta-ch'u.
Analects (Lun yu mWr)> °f Confucius: Binomes, in vocabulary of dialectical
passim. On right use of names: 197f. chapters: 164f
Chapter titles: 243. On knowing
whether you know: 402 Body. Relation to concept of hsing JflJ
'shape': 194, 282. Shape and con
Analogy. Inference by: 351. Referring sciousness: 201, 282. Shape and
as presenting analogue: 285-287. characteristics: 194, 471 f, 475
Analogies (lei gj) in EC: 258f
Buddhism, and sense-perception: 404,
See 'Definition', 'Inference', 'Illus
417. And Neo-Confucianism: 65
tration', 'Kinds', 'Parallelising'.
(n. 79)
Arabic philosophy, 'Being' in: 25
Astronomy. And geometry of right- Camera obscura: 377
angled triangle: 369-371. Assumed as Canons (ching jjgg) and Explanations (chin
basic science: 53, 371. Greek and shuo'lgfifo), 261-457. Date of: 23.
Chinese explanations of length of Organised infivesequences: 30, 229-
shadow varying with latitude: 370f. 235, 361, 365,476. Order of Canons at
Possible loss of a Mohist document: three stages of text: 87-94. Dates of
23, 53, 57. Possible Mohist document three stages: 88f, 95. Numbering in
in Huai-nan-tzu: 369-371 present edition: 90, 388. Artificiality
See 'Gnomon'. of division into two parts: 94, 229.
Explanations perhaps originally oral:
Being. Relation of Western concept to 24. Canons and explanations in Han
Indo-European grammar of 'to be': Fei tzu : 24. In Kuan-tzu : 366f. Word
25f. Arabic contribution to Western ching originally used of the 10 Mohist
terminology: 25. Chinese words doctrines: 22, 243f
578 Subjects
See 'Mo-tzû', 'Dialectical chapters*, time: 365-367. On knowing: 268,417.

'Expounding the Canons'. On hard and white: 172f. On chang
Cause. Ku ifa 'reason/cause', 189f. ing times: 21 f, 367. Relativism in:
Defined: 38, 263-265. Shih 'to 39f, 148f, 340, 403,441, 446-448, 453
commission* or 'to cause', the senses 455-457. On pointing out: 459-461.
distinguished: 324. Causal relations Later Mohist criticisms of its rela
'necessary* (pi not 'appropriate' tivism: 398. Lists sophisms: 61.
(yi H): 54f, 263f, 301, 324, 41 If. Account of Mohist sects: 22
Explanations in Mohist sciences are Ch'iin-shu chih-yao i p t ^ f g ^ : 211
causal: 54, 371. Problems of causa Circle (yuan J | , [U) and square (fang
tion: 424-427 "Jj) 225f. Both words stative verbs:
t
Certainty (pi jft). Of knowledge: 39, 301 226. Defined: 58, 309. Not drawn
See 'Necessity'. exactly by instruments: 309. Sphere
Change. Hua 'transform': 299, 333f. and cube: 226. Fang as rectangle: 309.
Defined: 295. In Hsùn-tzû : 197. Circle known 'a priori': 56, 342f.
Relation to pien §|§ 'alter' and yi ^ Definition derived by chain of
'substitute for': 214. Awareness of definitions fromyo 'like': 57f, 306.
changing times: 21 f, 365-367, 422f. "A circle is nowhere straight" known
Logical necessity as invulnerable to by explanation: 56, 347f. So is
change: 22, 33f, 299-301. Change and "Something square will not rotate":
doubt: 33, 360f 327f. Several standards for circul
arity: 35f, 316f. Same in large and
Changes, Book of (Yi-ching ^ffi), 53, small circles: 476. Tallying of stone
55, 207, 473 and wooden squares: 437. Circular
Chan-kuo-ts'e 222 (n. 13) motion: 297
Ch'en Pjf[, revolt of corvée workers in : 9 Class, logical. And Men ffc 'total': 265f,
Cheng Hp, politics in 6th century B.C. : 9 363. And lei ® 'kind': 169, 266.
Cheng tzù t'ung l E ^ f f i 204 Common names as 'classifying' (lei) :
Chiyun 195 (n. 60), 204 325. Social class: 6-8
See 'Kinds', 'Unit and total'.
Chia-tzù hsin-shu H ^ f W 380 (n. 348)
CKien-fu-lun №£:m' 483 (n. 624) Collective and distributive: 37, 337f,
362-364, 466-468
Ch'in §j§, Mohists in: 5, 5 (n. 7).
Suppression of philosophers : 3, 64 Colours, the five: 342f. Black and white:
Chou Jj§, decline of dynasty: 8. Revolt 340, 344-346, 443
#
of craftsmen: 9. Confucius and the Complement (ti ]S|), one name as
Chou tradition: 10-12 complement of another: 38, 203-206,
291f, 300, 339-341, 356f. Relation
Chou # JifS: 393 (n. 373), 435 (n. 513),
necessary: 240f, 300. One-way or
483 (n. 626) two-way: 38, 205, 300, 356f. Corrup
Chou pet suan ching j i f f ^ t ® : 308 tion of the graph: 84, 204
(n. 161), 369, 371 See 'Implication'.
Ch*u ^g, state: 4, 6, 427 Conditions, necessary and both neces
CKu tz'û : 214 sary and sufficient: 38, 54, 263-265,
Chuang-tzû JS-ïS passim. Chapter titles : 324. External conditions, and moral
243. Parallels with dialectical chapters : behaviour: 50
76, 268,488. A chain of definitions in: Conformity with superiors (shang t'ung
63, 262, 268, 281. Ox, horse and dog nIRI) 4, 13, 78, 289-292
in disputation: 183, 218, 438. Pillar:
225. 'Staying*: 178. Loan-naming: Confucianism, and fatalism: 14, 490f.
358, 447f. Ch'ing ff 'essentials': Scepticism: 13f. Music: 11. Righ
179-181. Definitions of space and teousness: 50,272. Gentleman: 7, 10.
3/5 579
The nameyw f^: 6. Deference: 403. Deeming (wei f§): 38. Being deemed X
Appeal to ancient authority: 1 If, 22, and being X : 117, 126, 209f, 462f.
423 Being deemed 'non-horse': 126.
Consciousness. See 'Intelligence*. Supposing: 39, 210, 268. And know
ing: 39, 210, 268. Ambiguity of wei:
Contradiction. And fan № 'converse': 333f
39,169, 319. And hat ^ 'be inconsis
tent with': 188, 449. And hsiang ch'ii Defence. See 'War'.
j f g £ : 'exclude each other': 38, 342. Deference: 403f
And pet 'illogical, self-contradic Definition. Perhaps called yi^: 32, 326.
tory': 199f. Hsun-tzu's "use what one And standard: 36, 317. And idea: 36,
accepts to show that what one rejects 317. And essential characteristics: 47,
is pet": 235. Sophists "adorning 179, 181, 248. Chain of definitions
phrases in order to make each other providing 'a priori' knowledge: 47-49,
pet": 20f. Examples of refutation as 57f, 262. A chain in Chuang-tzu: 63.
pei: 402, 445f, 448f, 451, 452, 453 Lost Mohist definitions: 45, 59, 235f.
Converse (fan <R) 38f, 184f, 330f, 354, Canons of Al-75 are all definitions:
397,447f. Defined: 318f. And contra 230, 261, 274 (n. 76), 313, 318 (n.
diction: 39, 169, 319 187). Contrastive definitions: 261,
267, 271, 275, 297. Definitions never
Copula, Indo-European: 25f. Copula repeated: 164, 236. Of concrete
tive relation and yeh tfe: 26,114, 153, words: 217, 350, 430f, 445. Influence
155-158. And wei fl£: 116-119, 362f, of Mohist definitions on other schools:
446-448. And wei jg: 33, 116-118, 51, 62f. Definition in Hsiin-tzu: 63,
209f, 462f, 467f 261 f, 269, 273-281. In Han Fei tzu:
Counting. T'i 'unit' and Men 261f, 272. In dictionaries: 78, 217,
'total': 29, 265f. Levels of counting: 307, 350, 430f. Hui Shin's definition
37, 266, 330f, 363, 43If, 461. Decimal by analogue and differentia: 261,
place system: 432. Counting rods: 444f. How were ox and horse defined ?
432. 'Two' as typical number more 438
than one: 265, 329, 450
Demonstrative pronouns. Grammar of:
Craftsmen, in politics of states: 9f. 120-123, 160f. 'That' and 'this' in
Mohists as: 6-8, 10-15 thought of Chuang-tzu: 39f, 441. In
Crane, stock example of loan-naming: Mohist logic: 40, 440f
219, 340f, 357-359, 446-448. The See 'This'.
graph § : 2 1 9 f
Description, one of four later Mohist
Criterion (yin 0) 145, 214-216, 354f, disciplines: 35-44, 262-269, 336-364,
368, 471f, 491f. Defined: 36, 316. 469-494. Distinguished from logic:
And standard: 36, 214f, 316f, 337, 31, 43, 336f, 488f
345-347
Desire and dislike (yii wu $£§§). Senses
Crossbows, machine for shooting simul analysed: 47, 332. Direct or after
taneously: 304 (n. 153) weighing: 47, 254, 332. Basis of later
Curtain, a supposed device for raising it Mohist ethical system: 47, 283, 332.
mechanically: 392 Taoist preference for desirelessness:
413. Desire and nurture of life: 52,
282f, 413-415. "The essential desires
Death. Survival: 14, 281. Both alive and are few": 181f. The desired and
dead at moment of death: 58, 341 disliked 'a priori': 47-51, 247f
Decimal place system: 432 Destiny. See 'Fatalism'.
Deduction. True disputation as deduc Dialectical chapters (Mo-tzu ch. 40-45).
tion from definitions of names: 31, Origin and composition: 23f. Textual
37, 43, 398f problems: 73-110. Grammar: 111-
580 Subjects
165. Largely unintelligible from Han: Elements, Five (wu hsing JjLff): 53-55,
65. Studied in 3rd and 4th centuries 372f. Ascendancies rejected: 55,411f.
A . D . : 66f. Lu Sheng's edition of And 'Six Storehouses': 362, 411
Canons: 66f. Inaccessible from T'ang Elevating worth (shang hsien jSjJf), 13
to Ming: 68f, 76. Revival of studies Emotions. And duty: 48. Calm: 280.
from 17th century: 70-72 Love as emotion: 48
See 'Canons', 'Expounding the canons',Ends, ethical. Wei @ 'to be/or' defined:
'Names and objects', 'Mo-tzu'. 48, 321-323. Wei basis of ethical
Differentia 221, 261, 350, 445 reasoning: 45, 321-323. Being for
Dimensionless (wu hou 4H£j|£). A S theme persons: 45. And means: 45f, 254.
of disputation: 19, 172-175, 306. Hui Kantian 'end in itself: 51
Shih on: 58, 302, 306. And geome Erh ya i f J£: 350
trical point: 58, 306. And starting- Essence, Aristotelian, and Chinese
point: 58, 310, 315f. Dimensioned ch'ing = 27, 179, 181f, 248,
defined: 305 454. And existence: 26
Disciplines, the four organised in the Ethics: 44-52, 243-259, 269-292. The
Canons: 30f, 229-235 most important of the four dis
Disputation (pien J§f), 316-323, 398- ciplines: 47. Ethical theory and moral
457. Defined: 39, 318. Origins: 19. preaching: 24. Later Mohist rational
An early programme of: 20f. Topics isation of: 44-52. Conceived in terms
of: 19, 172-175. Wider and narrower of benefit to individuals: 51
senses of: 31, 319, 476. Strict dis See 'Ends', 'Weighing', 'Desire',
putation only between X and non-X : 'Benefit', 'Love', 'Benevolence',
37-39, 120, 319, 402f. Rejected by 'Righteousness', 'Intent', 'Utility',
Chuang-tzü: 21, 39f, 403, 441, 457. 'Egalitarianism'.
Attitudes of Confucians and Legalists : Excluded middle: 39, 319, 403
21
Existence. See 'Essence'. 'Being*.
Distributive. Grammar of distributive Explanations (Ching shuo Iffitft). See
particles: 127-136 'Canons'.
See 'Collective'.
Expounding the canons (Yii ching flf $g):
Documents, Book of (Shu-ching ftfäH): 243-259. Meaning of title: 106, 243.
15, 352 (n. 274), 407 (n. 421) Problem of recovering it from Ta-
Dog, stock example of object with two ch'u: 101-108. Place in summa: 23f,
names: 218f, 358f, 447f. "A white dog 30. Earliest document in summa: 23,
is black": 61, 344, 493. "Whelp- 243. Grammatical differences from
killing is not dog-killing": 423f, 488. rest: 24, 113, 147, 152. A possible
"A dog could be deemed a sheep": further fragment: 274
219f, 359, 447. "A whelp is not a
dog": 61, 219, 245 (n. 3), 423 Family, and equality of love: 12, 49, 69.
Dreaming. Defined: 280 Filial piety: 49, 275. Duties to, as
Duration. See 'Time'. one's portion: 49
Fang yen # g : 204, 329
Economics, 397f. Fatalism. Confucian: 14. Mohist anti-
See 'Price'. fatalism: 14. Anti-fatalist arguments:
Economy, in expenditure (chieh yung 33f, 420,490f. And determinism: 490f
Hjjffl): 11. In funerals (chieh tsang Fire, whether its heat is objective: 416f
Force, mechanical: 387
wm-- ii
Fragmentation, of Expounding the canons
Egalitarianism: Mohist equality of love
and Names and obiects: 101-110. Of
but inequality of 'portions': 48f, 52 Explanations A 22-39: HOf, 280-291.
Egoism. See 'Individualism'.
3/5 581
Elsewhere: 362 (n. 314), 411 (n. 433), (Chieh Lao fg&): 21, 149, 179, 192,
388 261 f, 272, 429
Han-shih wai-chuan ^|^^jv^: 20, 459
Gentleman (chiin-tzu ;U-JO, a Confucian Hard and white (chien pai 8 £ £ 0 : 355f,
but not a Mohist ideal: 7 365f, 368,404-408. Chien-pai defined:
Geometry. Definitions: 301-316. They 171, 313. Is a technical term for
assume manual operations: 302. And mutual pervasion misunderstood by
astronomy: 369-371. And the 'a Kung-sun Lung tzii forger: 170-176.
priori': 53, 57f, 342f. Proofs: 56, 328, The forged Essay on hard and white
n
435. Ideal and real figures: 54, 302, (Chien pai lun MStw) i Kung-sun
309. References in logical and Lung tzU: 170f, 176, 355 (n. 284).
scientific contexts: 53 There was no pre-Han sophism of
See 'Circle', 'Line', 'Point', 'Space', hard and white: 172-176. Chien-pai
'Science'. as a theme in disputation: 19, 172-
175. Kung-sun Lung's 'White horse'
Glosses: 98f
is disputation on this theme: 173-175,
Gnomon: 53, 306 (n. 158), 369-371, 404
405 (n. 415) Head characters, of Explanations 95-98.
Government. Good government ulti Discovered by Liang Ch'i-ch'ao: 95.
mate purpose of disputation: 47, 476. Sometimes one place too late: 95f. In
Canons on ruler and subject: 289-292. present edition bracketed in either
Chih 'govern, set in order' defined: position: 96. Never more than two:
283. Its small place in later Mohist 97. Intact throughout: 96. Originally
summa: 8 written in margin: 96, 98f. Introduced
Grammar. Chinese grammar and after fragmentation of A 22-39: 101
Chinese thought: 25-30. Grammar of Heaven (Vien ^.): 252. Definition lost:
dialectical chapters: 111-165. Canons, 59, 236. For Mohists but not Con
Explanations and Names and objects fucians a personal deity: 14. Will of
assume a deliberate system of gram (t'ien chih 55-g): 4, 14, 23, 60, 243-
matical coventions: 24, 111-113. 246. Allegiance to Heaven rather than
System not observed in Expounding to ruler: 13. Reduced significance in
the canons: 24, 113 later Mohism: 23, 59f, 244. 'Heaven's
constants' (t'ien ch'ang ^ ^ ) : 60,
Graphs. Degree of graphic corruption
409f. Heaven and human nature:
has been exaggerated: 81. Systema
16-18, 23, 59, 244-246
tically corrupt graphs: 81-84. Fre
quency of variant radicals: 76-80. Hedonism: 17
Rarity of sound loans: 80f. Technical Horyaku f[fg edition of Mo-tziX: 74f,
terms liable to corruption: 168. 80, 85, 426 (n. 485), 429 (n. 495), 448
Tabooed graphs: 85-87, 273. Em (n. 571), 454 (n. 588)
press Wu's graphs: 169. Li-shu forms: Ho-kuan-tzA J H x J ^ : 68
82f. Bronze forms: 83. Mohist graphic
Horse. Stock example of name of a kind
conventions to differentiate words:
of thing: 32, 217f, 325. Kung-sun
77-80, 168. Later Mohist fondness
Lung's sophism of the * White horse':
for the 'man' radical: 78f, 168
19f, 43, 61, 173-176, 218, 235, 245f,
Greece: 8, 21, 56, 57, 307, 309, 370, 481 404, 440. 'Ch'in horse': 471-473.
'Blind horse': 354, 492f. Riding a
Han Fei tzti f^f;-^: passim. Canons horse: 485f, 491. One horse or two:
with explanations in: 24. Definitions 493f. Ox and horse: 217f, 298f, 363f,
in: 261 f, 272. On changing times: 22. 403, 438-440, 492-494. Distinguish
On the Mohist sects: 23. On li M ing characteristics of ox and horse
'pattern': 192, 429. Attitude to unmentioned: 217, 319, 431. Perhaps
disputation: 21. Interpreting Lao-tzU preserved in Huai-nan-tzU: 438
582 Subjects
Hsiao-ch'u /J\Jj£ (Mo-tzù ch. 45), Illustration (p'i g£). Defined: 483f
'Smaller pick*. Meaning of title: 101 Illustrations in Explanations introduced
(n. 43). Compiled from Names and by jo 'like': 98f. Are glosses, some
objects: lOlf, 108-110. Compilation times parenthetic or displaced: 98.
not later than last century B.c. : 65 Similar to the analogies in E C : 99
Hsùn-tzû ^-f*: passim. Attitude to Image (hsiangfl^= ^ ) . See 'Idea'.
disputation: 21, 173. On human
nature: 18. On the orders from Image in mirror. See 'Shadow'.
mineral to man: 282, 350 (n. 269). Implication, conceived as between a
On weighing of considerations: 233f, name and its complement: 38, 205,
320 (n. 194), 322. On knowing: 268f, 300f, 330f. 'Following from each
301,402. Onfivesenses: 415 (n. 445). other or excluding each other': 38,
On hsin ifr 'heart/mind' : 268, 415 342. 'Rejecting one but not the other':
(n. 445). On being two objects or a 38, 354f, 358. Necessity of: 38, 300f,
transformed object: 197. The three 330f. Relation to ku 'reason':
types of fallacy; 43f, 234f. Defini 264f. Grammar of: 141-145
tions : 63,261 f, 269,273,281. On hsing Individualists, 'Nurture of life' school
f f proceeding': 178, 228. Nomina of Yang Chu: 15-17. Accused of wei
lism: 32. Naming conventional: 39. wo M3k 'egoism*: 245 (n. 10), 246.
Linking names: 191, 355. Sentence Introduced problem of human nature:
or proposition: 63f. Using names 16f. Relation to Mohist ethic of
to point out: 458. Dependence of
benefiting individuals: 51. Mohist
Right use of names (ch. 22, Cheng Ming
reactions to challenge: 16, 244-246,
jE^g) on the Mohist disputation:
269, 280-284, 413-415
63f. Its organisation based on the four
See 'Life*, 'Nature', 'Benefit'.
disciplines of the Canons: 233-235
'Weighing'.
Huai-nan-tzù ^ ê ^ i : passim. On astro Induction: 351
nomy: 306 (n. 158), 369-371. On
space and time: 365. On the One : 203. Inferring (t'ui $ 0 : 483f. Defined: 483.
On Yang Chu's doctrines: 17, 246. In Hsiin-tzu: 350 (n. 269). Tui lei
On milu deer and louse as stock ff^^H 'extending from kind to kind':
illustrations: 220, 221, 351. Parallels 220, 350f
with dialectical chapters: 76, 492 Infinite (wu ch'iung Mohist
(n. 648). On characteristics of ox and conception of: 58, 436, 448f. Ch'iung
horse: 438. On weighing of considera |g 'limit' defined: 294. Hui Shih's
tions: 254 sophisms: 58, 266, 449. The infinite
not a t'i fj§ 'unit': 58, 266. Infinite
Hypostatisation, discouraged by Chinese
divisibility: 306, 433
language: 29, 461
See 'Space'.
Idea (yi ^c), as mental picture: 40, 213, Innovation, defended against Con
429,471. And image: 213f, 405f, 428f. fucians: 11, 278, 442
Of pillar and hammer: 37f, 224f, Intelligence, the (chih £fl), faculty by
428-231. And definition: 36, 317, As which one knows: 32, 60, 247f, 267-
standard: 316f. Its verbal formula 269. Defined: 267. And mind: 59f,
tion: 207, 276, 483. Its communica 268. Quiescent in sleep: 281
tion: 213, 276, 406, 409, 480. Having
idea distinguished from knowing: 40, Intent (chih <§). Definition lost: 45.
471,473 Ethics of intent and act: 50, 249, 271,
275f
Identity (ch'ung J J ) : 197, 218, 334f,
406, 408f, 423, 475 'Jack and Jill' (Tsang and Huo 3fl),
Illness, stock example of event with typical proper names: 226f, 251, 257,
multiple causes: 56, 226, 324, 359, 361 325. Humblest persons: 227, 249,
3/5 583
484-487. "Jack has three ears": 208, 170f, 176, 364. Other misunderstand
227, 432, 478 (n. 615) ings of Canons: 61 (n. 78), 67, 355
Jen wu chih A^ftS- 2 0 n
( - 52) (n. 284), 439 (n. 529). Possible use of
Lu Sheng's edition of Canons: 67. On
Killing, of criminals: 42f, 251. Of chicken sophism: 432
innocent: 251. "Whelp-killing is not K'ung-ts'ung-tzil JIM^: 66 (n. 82), 20
dog-killing": 423f, 488. "Killing (n. 52), 227 (n. 17)
robbers is not killing people": 42f, Kuo-yii B i g : 277 (n. 84), 347 (n. 259)
234, 487-489
Kinds (lei ^ ) . Called by common Ladder, wheeled: 393-395
names: 36, 325. Sameness and being Language. Relation of logic to Chinese
of a kind: 36, 334, 336,475, 438. And and Indo-European language struc
likeness: 32. 'Proceeding' (hsing fr) to ture: 25-30. Later Mohist pursuit of
kind: 37, 40f, 177f, 348f, 480-483. linguistic clarity: 111-113, 161-164.
'Extending' (t'ui $£) from kind to Grammatical regularity preferred to
kind: 220, 350f. Wider and narrower: idiom: 161-164
350f. Lei not translatable as 'class':
See 'Grammar', 'Ambiguity', 'Say
169. And differentia: 350f. And cKii
ing', 'Name', 'Sentence', 'Descrip
{g 'separating off': 183, 478
tion'.
Knowing (chih Definitions: 266-
Lao-tzU pg-?: 8, 66, 342 (n. 247), 413,
269. Graphic differentiation of chih :
418
77. Distinguished from perceiving:
32, 267, 415f. And from having an Law, in Cheng: 9
idea: 40, 471-473. Implies temporal Laws of nature: 54—56
duration: 32-34, 267, 415f. Three Learning, defended against Taoists: 452
sources of: 32, 30, 327-329, 359.
Legalists (fa chia $c||£): 21 f, 59. Ignored
Four objects of: 32, 30, 327f, 359,361.
in dialectical chapters: 61. And com
Four objects basis of organisation of
parison of title and performance: 198,
Canons: 231-234. Its certainty: 39,
211
268. And doubt: 33, 360f. And
supposing: 39, 268-300. Knowing See Han Fei tzii.
'a priori': 37f, 57f, 342f, 428-431, Lever: 54. Ch'iian ^ 'positional advan
448f. "Knowing whether one knows": tage' as leverage: 386-392
402. "Knowing what one does not Li-chi jBMB: 339 (n. 240), 412 (n. 435)
know": 417f. In Hsun-tzii: 269. In
Lieh-tzU 16, 17, 66, 67, 76, 86,
Chuang-tzu: 268, 417. Foreknow
174, 246, 389, 420f, 434, 460
ledge in Lun-heng: 223
Life (sheng ££). Defined: 280. Its value
Ku H collation of Mo-tzU: 75 not independent of desire: 52, 413-
Kuan-tza : 59,101 (n. 43), 207-209, 415. Individualist 'nurture of life':
328, 338 (n. 237), 340, 366f, 451 15-17, 280-284. Vegetable life: 282.
Kui-ku-tzU fefi-î 68 And human nature: 16, 281
Kung-sun Lung tzH -&J£f| Light. Reflection: 377, 378f. And
Essay on
the white horse: 19,117,161,187,404. shadow: 372, 374. No conception of
Essay on pointing things out: 19, 161,rays: 372
457-468. Whether these written by Likeness (jo Jj), of objects called by the
Kung-sun Lung: 463. Rest of essays same name: 32, 325, 443f. Unde
forged between A.D. 300 and 600: fined: 57. Equality as likeness in
19, 66 (n. 82), 67, 170f. Parallels with quantity: 57, 306. Likeness as
Canons: 76, 81,176, 353 (n. 275), 355 starting-point for definition of circle:
(n. 284), 437 (n. 528). Parallels con 57. Being like and being 'so' (jan $&):
fined to forged essays: 61. Forger's 57, 35, 316f. Likeness to standard:
misunderstanding of 'hard and white': 35f, 316f. Relation to sameness: 336
584 Subjects
See 'Analogy', 'Sameness'. essential desires: 181 f. Knowing: 268·

Line, geometrical. And ch'ih 'foot/ On bad disputation: 185, 208, 275·
measured length': 302, 314. Outline Approximation to a syllogism in: 275.
or circumference (ch'ii hsiieh |M/C): Ox and horse: 138, 218. Li №
297, 303f, 31 If. Straightness defined: 'pattern': 192
307. Curve: 87, 297, 376-378
Mao 3^ edition of Mo-tzu : 74f, 85, 211,
Liu t'ao 266 (n. 58)
240, 343 (n. 253), 426 (n. 485), 448
Logic. And Chinese language: 25-30. (n. 571), 454 (n. 588)
Relation tofirstand fourth of Mohist Mathematics. See 'Counting', 'Geome
disciplines: 31, 43, 231 f, 489. Logic try', 'Measurement', 'Science'.
ally significant grammatical particles:
Meaning, and chih 'pointing out':
141-145. Whether there are universal
33, 457-468
forms of thought: 169. Logical ques
tions and factual questions: 39, 234f. Measurement. Ch'ih 'footrule/
Later neglect of Mohist logic: 64-69 measured foot': 302-304. Naming
See 'A priori', 'Complement', 'Con compared to measuring: 27, 444.
tradiction', 'Converse', 'Deduction', Measurement in astronomy: 54, 371.
'Induction', 'Inferring', 'Necessity',In technology: 54, 303, 304 (n. 153),
'Quantification', 'Reason', 'Self- 379 (n. 341). Absence of measure
ments in scientific sections: 54.
reference', 'Syllogism', 'Excluded
Measuring to compare: 315f, 339
middle', 'Implication'.
Mechanics: 269, 279, 385-398, 421.
Louse, stock example of shared name:
Concepts of: 385-387
221 f, 357. As parasite moving with
its host: 222, 298, 339, 341 See 'Force', 'Lever', 'Pillar', 'Pulley',
'Steelyard', 'Weight and pull'.
Love {at §£). Definition lost: 45, 48,
235, 241, 'Love' not perfect equiva Medicine, as a proto-science: 56
lent for at: 12. Not treated as an See 'Illness'.
emotion: 48. Self-love: 12, 48, 51, Mencius (Meng-tzu 35.-7*): passim. Good
251, 256, 271, 414. Universal love ness of human nature: 17f, 181, 491.
•(Men at ffe$£): 4, 10, 11, 48-51, Righteousness not external: 50, 451.
246-259, 270f. Answers to objections The ch'ing \y\ 'essentials' of humanity:
to: 448-450, 487-489. Love and 181. Liang chih J^5>fl 'native under
benevolence: 49, 252, 270f. Love is standing': 338 (n. 235). Finds
for individuals: 49, 51, 271. Love of inconsistency in Mohist view of love:
Jack and Jill: 227, 247, 249, 485f. 52. Quality of argumentation in: 19
Love of men simply as men is for all (n. 46) Technical vocabulary shared
men: 345f, 486, 491. Is equal for all with Later Mohist ethics: 46, 255.
men: 48, 51, 249-251, 257. Additional A possibly Mohist definition of
love for the more beneficent man: 48f, 'good': 51
250-252. Loving someone not incom Metaphysics. Later Mohist indifference
patible with killing him: 251, 487- to: 53, 59f. 'Metaphysical crisis' in
489. Love disinterested: 48, 257
4th century B . C . : 15-22
Lu state: 10 Mien-miao-ko f$M>|83 edition of Mo-
Lu g| edition of Mo-tzu: 74, 75, 85, tzu: 74, 426 (n. 485), 448 (n. 571),
426 (n. 485) 454 (n. 588)
Lu-shih ch'un-ch'iu passim. Military chapters (Mo-tzu ch. 52-71):
Individualist chapters: 15-17, 282f. 4, 7, 70, 79, 84, 145, 164, 169, 269.
Weighing of benefit: 282f, 322. Proto- Date of: 4 (n. 4). Cited: 303, 304
scientific thinking: 373, 411. Five (n. 153), 308 (n. 160), 379 (n. 341)
Elements: 411. Causation: 324. Im
Milu deer, as stock example: 220f, 350f,
perfections of sages: 346f. The
357, 363, 414f
3/5 585
Mind, and chih $Q 'intelligence': 60. naming (chia flg): 220, 358f, 471f.
Hsin ifr 'heart' not used in Canons: Linking (/i'Sf) ot names: 37,150,191,
59f, 109, 268. Used in Names and 354f, 424. And sentence: 25, 35, 40,
objects: 59, 109, 480, 488, 490 209, 354. And pronunciation of
See ' Hsun-tzti* 'Intelligence'.
t
sounds: 32, 200, 325f, 422f. Com
Mirrors. Plane: 381. Concave: 376-378, pound names: 187f, 362-364, 480,
382-384. Convex: 384f. Multiple: 482. Ambiguous names: 35, 323-336.
381. Burning mirror: 55 Ming includes all words: 35, 197.
Naming compared with measuring:
Moment (wu chin 'durationless'). 27, 444. Is by convention: 39f, 233,
Conceived as commencement of 235. Names which praise or blame:
period: 58, 295. Shared by two 44, 272f, 284, 327. Different things
periods: 58. Paradoxes: 58, 289, 299, sharing name: 35, 221 f. Different
341, 436. All space present at every names for same thing: 218f. 'School
moment: 34, 365, 368 of Names' (ming chia 19, 67.
Mohist school (Mo chia >3St|0. Organis 'Right use of names' (cheng ming
ed community: 3f, 10, 209. Origin of ]Ε%χ)'. 197f, 440. Knowing of names
the name Mo: 6. Ten doctrines: 3-5, and of objects: 327f, 408f, 417f, 422f
10-15, 22, 241. Three sects: 3, 22f. See 'Nominalism', 'Performance and
Based on craftsmen: 6-8, 65. Dis title'.
appearance after 221 B . C . : 3, 64. And
Names and objects (Ming shih ^^):
rationality: 11, 14f, 21. And military
469-494. Problem of recovering it
engineering: 4, 6, 7. And innovation: y
from Ta-ch ii and Hsiao-ch'ii: 101,
11, 278, 442. And traditional religion:
108-110. Later than the summa: 24.
14f, 59
Known to Hsun-tzu: 63. Discovery
Money: 8 of the proposition: 24, 25, 35, 40.
Mo-tzii 35·^. In Han bibliography: 65, Thought of: 40-44. Theme is the art
74, 95. Extant editions: 74f. Harvard- of description, not logic: 30, 40, 43,
Yenching concordance based on 470, 488f. Linguistic peculiarities:
Sun's emended text: 81. Wu Yii- 113, 120, 143-145
chiang's apparatus criticus: 74f. CoreNature (hsing ^k). Confucian problem
chapters expounding the ten doctrines of its goodness or badness: 17f.
(ch. 8-37): 3. Three versions of each, Individualist origins of the problem:
from three sects: 3. Only ch. 1-13 in 16f, 244. Dismissed by Mohists: 23,
circulation from T'ang to early Ming: 59, 244-246, 280f. In Chuang-tzu : 63.
68f. Yu manuscript of these: 68. In Kuan-tza: 451
Yueh Tai's lost commentary on
them: 68. Recovery of complete text Necessity (pi : passim. Defined: <34,
in Ming: 69. All editions derive from 299f. Applies to one kind of relation:
Sung exemplar: 85f. Ch'ing commen 330f. Logical and causal not dis
tators: 70-72 tinguished: 263f, 300. A later Mohist
discovery: 62. Eternity of the logically
See 'Dialectical chapters', 'Military necessary: 22, 33f, 293, 299f. Un
chapters'. recognised by Hsun-tzu: 64. Charac
Music. Rejected by Mohists: 11, 46, teristic of logic and science, not of
252 description and ethics: 34, 54f.
Mu t'ien-tza chuan U^^f'flf: 89 Belongs to the realm of names: 189,
300, 420. Contrasted with ku [5J
Names (ming fe). Three types in 'being inherent in the situation': 189,
Canons: 32f, 325f. Further types in 301, 324, 420. Requiring (tat f#): 202.
Names and objects: 40, 472. A "name The necessary and the merely
which passes on" (kuo ming lUtfe): 'appropriate' (yi U ) : 330f, 412, 420
400f. Naming (ming <^r): 199. Loan- See 'Certainty'.
586 Subjecti
Nominalism, Mohist: 29f, 32f, 287, 325, tion: 404, 417. Perception of heat as
444. In Hsün-tzü: 32. No realism in objective: 416f. Five senses: 415
ancient China: 33 Performance and title (hsing ming Jfcfe)
See 'Names'. 198
North and South: 227, 400f Pillar, as stock example: 224f, 235, 291 f,
Number. Grammatical: 28-30, 486, 493 428-431, 471, 473, 'Pillaring', in
See 'Counting'. mechanics: 386, 396
Point, geometrical. See 'Dimensionless',
Objects (shih Jf): 196-199, 200-202. 'Starting-point'.
Conceived as concrete and particular: Pointing (chin fa) Meanings of chin:
32, 35, 196f, 325. Science as the 457-460. Pointing out and referring
explaining of objects: 30, 53 by name: 286, 405-407, 422, 458.
See 'Names', 'Things'. Knowing without being able to point
Odes or Songs, Book of (Shih-ching out: 408. Impossible to point out a
property in one object: 405-407.
'USD: 15> 3 2 9 » 3 7 3 (n- 33 °)» 3 9 0 "What we point out we do not arrive
(n. 362), 428 (n. 494)
at": 286 (n. 108), 460. Kung-sun
Optics: 372-385. Mohist optics prima Lung's Essay on pointing things out:
rily the study of shadows: 372. 457-468
Concerned with borders of light, not
with rays: 372. Lii-shih ch'un-ch'iu on 'Portion' in life, field of action (fen ft,
optics: 373. Shen Kua on: 69, 372, £ ) : 46, 49, 52, 255, 270f, 275
377. Conceived geometrically: 54, 372 Praising and blaming: 44, 284, 327
See 'Camera obscura', 'Mirror', Predicate: 28. As 'root' (ken $g) of
'Shadow'. sentence: 28, 40f, 292, 475, 477, 481
Ox, stock example of 'X or non-X': Price. Depends on supply and demand:
217f, 318 8, 397f. Its Tightness not relative:
See 'Horse'. 340f.
'Proceeding' (hsing ff), from what is so
Paradoxes. See 'Sophisms'. of an instance to what is so of the
Parallelising (mou fä): 483-494. De kind: 37, 40f, 43, 177f, 348f, 447f,
fined: 483f 480-483. And 'road' metaphor: 227.
Parents. See 'Family'. In Hsun-tzu: 178, 234, 228. And
'staying' (chih it): 177f
Part and whole. Embraced under t'i fHl
'Programme of disputation' ascribed to
'unit' and Men ffc 'total': 265, 335,
Tsou Yen: 20f, 199, 207, 213, 459,
363. Sameness of whole but difference
479 (n. 618)
of parts: 470f, 475. Paradox of point
ing out part from whole: 460-468 Proof. And shuo |ft 'explanation': 216,
See 'Unit and total'. 317. Geometrical proof: 56, 328, 435
Properties. Distinction from their
Particulars. See 'Objects', 'Universals
and particulars'. possessor grammatically unclear in
Classical Chinese: 27. The 'having'
Pattern (li J$), organisation, or structure: (yu ^*) of shape and colour: 26, 282,
191f. As ordering of names in
341. What ox and horse 'have': 438
sentence: 41, 192, 475f, 480-482. As
See 'Quality'.
principle organising thought: 192,
321, 452. In Han Fei tzü: 192, 429. Proposition. See 'Sentence'.
In Lii-shih ch'un-ch'iu: 192. In Psychological terminology of Canons: 60
Neo-Confucianism: 191 Pulley: 389, 391 f
Perception. And knowing: 32, 267f,
415f. Seeing: 329. Hearing: 329. Quality. And mao & 'features': 194-196,
Buddhist criticism of sense-percep 223. And jaw jfe being so': 194-196.
3/5 587
Indo-European adjective and Chinese Sentence (tz'a f$): 207-209, 480-483.

intransitive verb: 25-27. Shuo-yiian Difficulty in recognising as different
and Hsiin-tza use chuang )[£ 'charac from complex name in Chinese
teristics': 195, 197, 445 language: 25,64. Unrecognised before
See 'Substance and qualities', 'This Names and objects: 35, 40, 354, 480-
and so'. 483. Not distinguished from proposi
tion: 469, 473. In Hsiin-tzti : 63f. And
Quantification. Of subject and of object: idea: 207, 275, 473, 483. And saying:
27f, 41, 349. 'Some' and 'all' defined: 287. Has pattern (li Eg): 41,192, 475f,
57, 127, 236, 294, 471 f. Grammatical 480-482. Pause and phrase-position:
distributives: 127-135 41, 53, 480, 482. And 'linking' {li |ff):
191, 354f. Predicate as its 'root': 28,
Rationality. Of Mohists: 4,14f, 21f, 24f, 40f, 292, 475, 477, 481. Parallelism of
59, 66, 70. Progress in schools of 4th sentences: 40-43, 480, 483-494.
century B . C . : 19-22. And Confucian Grammar of sentence in dialectical
ism: 14, 21. And Chuang-tzu: 18, 21, chapters: 113-116.
61. And Hsun-tzu: 21, 64
Shadow: 372-385. Mohist optics is the
Rays, optical: 372 study of shadows: 372. Silhouette,
Reason (ku Defined: 263-265. Its not image: 372, 381. Whether shadow
relation to ku [5}: 189f. Necessary and moves: 373f. As interference pheno
both necessary and sufficient condi menon: 374. Inversion in concave
tions: 263f. Reason and cause: 264. mirror: 376-378
Reason for action not ku 4j( but wei
Shih-chi &|5: 124, 274 (n. 75), 291,
45, 322
332, 380 (n. 348), 412 (n. 435),. 459
See 'Rationality', 'Logic', 'Implica
Shih-ming g £ : 200 (n. 75), 217, 222,
tion', 'Pattern'. 325, 430f
Referring (chit Ig), picking out an
Shih-tzn p^: 256, 327, 365
object by name, defined: 286f. And
pointing out: 286, 405-407, 422, 458. Shuo-wen te^: 77f, 211, 217, 276
And praising or blaming: 284. (n. 79), 297 (n. 133), 307, 325, 360
'Referring arbitrarily' (k'uang chii (n. 302, 304), 384 (n. 355), 430, 435
: 437f, 451. One kind of calling: (n. 512)
326f. "One uses names to refer to Shuo-yuan |9;^Q: 285 (n. 103), 308
objects": 483 (n. 161), 444f
Similarity. See 'Sameness', 'Likeness'.
Saying (yen g ) . Defined: 286-288
Sleep: 280f
Science. Mediaeval and modern: 53,
'Smaller pick'. See Hsiao-ch'u.
373. Mohist science based on causal
Snake: 259. Graph = № 79, 258
explanation: 54, 371. Ignores Yin and
(n. 55). 407 (n. 420). Hibernation:
Yang and Five Elements: 55, 372f,
407f, 419
411f. Not mathematised: 54, 389.
Does not propose laws of nature: 55f. 'So' (jan fk). See 'Likeness', 'This'.
Founded in geometry: 54, 56-58, 371, Sophisms. "Heaven and earth are one
372. No special experimental equip unit": 266. "A white horse is not a
ment: 377. And craftsmen: 8 horse": 19f, 43, 61, 173-176, 218,
See 'Optics', 'Mechanics', 'Econo 235, 245f, 363, 440. "Jack has three
mies', 'Astronomy', 'Cause'. ears": 208, 227, 432, 478 (n. 615).
"The eye does not see": 61, 416.
'Second List' (of sophisms in Chuang-tzU "A white dog is black": 61, 344, 493.
ch. 33): 61 "A shadow does not shift": 61, 373f.
Self-reference and self-contradiction: "The South has no limit yet does
199f, 403f, 445f, 452, 453 have a limit": 61, 449. "The carpen
ter's square is not square . . .": 309.
Senses, five: 415 (n. 455)
588 Subjects
"The meaning is not the meaning": (Hsiao che $ ^ ) : 471, 473. May be
457-468. "Fire is not hot": 61, 416. several for one thing: 316. Fitting
"A hair will pull 1,000 chun": 421. exactly or as more appropriate: 36,
"A stick one foot long if you take 337, 344-348
away a half every day . . .": 61, 433. Starting-point (tuan 263-266, 310-
"A whelp is not a dog": 61, 219, 245 315, 432f. Defined: 310. Only point
(n. 3), 423. "A chicken has three
recognised by Mohists: 58, 302. Tuan
feet": 432. "An orphan colt has never
not synonymous with 'point': 302,
had a mother": 434. "I go to Yiieh
today but arrived yesterday": 436. 310. Coinciding starting-points: 58.
"A yellow horse and a black ox are Tuan as sentence-pause: 41, 479
three": 364, 440. "A dog may be (n. 617), 480f. As starting-point of
deemed a sheep": 219f, 359, 447. argument in Hsun-tzn: 20 (n. 53). In
"What is pointed out we do not arrive optics: 376
at": 286 (n. 108), 460. "The dimen- See 'Dimensionless'.
sionless cannot be accumulated . . .": State. Emergence from Chou feudalism:
302, 306. "The sun at noon is 4. Mohist theory of: 13
simultaneously declining . . .": 58, See 'Government', 'Conformity with
341. "The pillar has the ox": 225, superiors', 'Elevating worth'.
235. "Mountains come out of holes":
'Staying' (chih \\-) for a time, of name
312. "A tortoise is longer than a t
in object: 361, 177f, 294f, 342f, 456f.

snake": 357. No pre-Han sophism of
Defined: 34, 298f. 'Fixing' for a time
'hard and white': 171-176. "Ox and
of name on object: 36, 177, 337, 345f,
horse are not ox" is not a sophism:
460. And logical necessity: 3()f, 34,
440. Sophisms of Hui Shih: 19, 61.
300, 336, 400. Characteristic of
Of 'Second List': 61
description and ethics: 30f, 34, 232,
Sophists (pien che 19, 61f, 66f 400. 'Fixing' the criterion and the
See 'Sophisms'. kind: 346, 348, 350
Space (yii Mohist definition: 34, Steelyard, Chinese and Tibetan: 389.
293f. Other definitions: 365. Abstrac Its counterpoise: 385, 389f
tion of space from the spatially Stone, stock example of thing within
extended: 34, 294, 366. Geometrical which properties are differentiated:
space and Mohist world-picture: 53. 222f
Relation to time: 365-369 Strength (It jj). Defined: 279. And
See 'Geometry', 'Infinite', 'Time', physical force: 387
'Dimensionless', 'Sophisms'. Strips, writing: 88f, 94. Evidence of
Species. See 'Kinds'. transposition: 89, 101-110. Of num
ber of characters to a strip: 89f, 94,
Sphere: 57, 226, 435
104, 106-109.
Spirits (kuei $2,). Revered by Mohists:
Subject, grammatical. And topic: 27.
14, 257, 59, 492. Man's spirit survives
death: 281 Its frequent absence in Chinese
sentence: 27
Square (fang See 'Circle'. See 'Quantification'.
Ssu-k'u ch'uan-shu edition of Substance and quality. Western con
Mo-tzu: 75
cepts of: 202
Ssu-ma fa ЦЩ,]^: 254, 276 See 'Properties', 'Quality'.
Stages, three, in text of Canons: 88 Summa, later Mohist (Expounding the
Standard (fa fe): 35f, 344-346, 437. canons, Canons and Explanations) :
Defined: 35, 316f. And criterion: 23f, 30
36, 214f, 316f, 337, 345-347. And 'Summings-up' of Canons in Part Β
definition: 36, 317. And exemplifier (shuo tsai f&Ê X)· Present from
3/5 589
Stage 1: 94. Perhaps eliminated from stematic corruption: 81-84. Graphic

A 88-98: 95 corruption: 76-87. Transpositions:
Sun-tzuffi^: 292, 328, 412 88f, 93-95, 100, 101-110. Glosses:
98f. Fragmentation: lOOf, 101-110
Sung 7^, state: story of its defense by
Mo-tzu: 6 Things (wu !$j) distinguished from shih
t
ff 'objects': 196f, 210, 468. Abstrac

Syllogism: 44, 55, 275
tions as wu: 29, 210, 247f
Taboos: graphs tabooed under Sung: 'This' (shih ft) and 'so' (jan £*). Shih
85-87, 273 (n. 73), 333 (n. 226), 435 as demonstrative 'this/the one in
(n. 512) question': 120. Shih 'the one in
Ta-ch'ii (Mo-tzu ch. 44), 'Bigger question' and tz'a jfcfc 'the one here':
pick'. Meaning of title: 101 (n. 43). 120. Shih and the kind of thing
Collection of Expounding the Canons ('what it is'): 120. Jan 'so' and jo ^
and Names and objects fragments: 'like': 35, 57, 342. Shih andjanbelong
101-110. Not later than 1st century with nominal and verbal sentence;
B.C.: 65. 17th century commentary by 26, 117f. Shih fei 'is this or is not':
Fu Shan: 70 117. As the alternatives in disputation:
37f, 120, 319, 403, 475f. When
Tai-p'ing ching ·χψ$_: 65 (η. 80)
translatable as 'right and wrong':
T'ang Jg edition of Mo-tzu: 74
120f. Treating what is not this as
T'ang-ts'e-hsien ']&M$L edition of 'this': 454-457. Nai shih and pu
Mo-tzu: 75, 258 (n. 55) shih ^f\;§k: 118. Whether what is 'so'
Tao (tao J3). Word not used in meta of the instance here is so of 'this':
physical sense in dialectical chapters: 41f, 348, 469, 471, 485-491. Shih and
59, 227 jan in Chuang-tzu : 39, 455. Chih shih
See 'Road'. ftUgr 'knowing what it is': 120, 342
Taoism. 18, 62, 367, 438. And naming: (n. 249), 405 (n. 411). Relation to
197. And taking the lower position: Western concept of 'being': 26.
454. And learning: 452. And initiating Sameness in being this and sameness
action: 442. Chuang-tzu's rejection of in being so: 40, 475, 477
disputation: 21, 193, 403, 441, 445- See 'Demonstrative pronouns',
447, 453, 455-457. And desireless- 'Being'.
ness: 413. Neo-Taoism of 3rd Truth. Hypostatisation of, in West: 29.
century A.D.: 66. And Nothing: 418. And tang ^ 'be plumb with the fact':
Han revolutionary Taoism: 65 (n. 80) 39, 202f. And k'o oj 'admissible': 39.
See Lao-tzu, Chuang-tzu, Lieh-tzu. And hsin ||f 'trustworthy': 275f. And
Taoist Patrology (Tao-tsang 65 shih Jfe 'this', jan j& 'so': 120f.
(n. 80), 69, 74-76, 80, 81, 85, 239, Self-contradiction of 'All propositions
343(n.253),426(n.485), 440 (n. 537), are true/false': 446
451 (n. 577), 464. Text of dialectical Ts'ang chieh p'ien ^Jjflg: 430
chapters reproduced: 499-526 Tso-chuan 8-10, 333, 358, 403,
Ten theses of Mo-tzu, the doctrines 407 (n. 421)
expounded in the ten triads of core
chapters, Mo-tzu ch. 8-37: If, 11-15, Unit (t'i f g ) and total (chien ^). Ti
241. Definitions of all words in them defined: 265f. Definition of chien lost:
are lost: 23 5f, 241. Known as 'canons': 235, 241. Applied to love: 45, 48f,
22, 243-246 254f, 270f, 448f. To pointing out:
Teng Hsi tzu SPtfrf: 20, 227 406, 461 f, 467f. To sacrifice of part
Textual criticism: of dialectical chapters: for whole: 45f, 253-255
73-110. Dangers of piecemeal emen See 'Counting', 'Part and whole',
dation: 73, 81, 167, 190, 217. Sy 'Class'.
590 Subjects
Universals and particulars. Problem of, Wei revolt of craftsmen in: 9

and grammatical number: 29f. No 'Weighing' (ch'uan $|), of benefit and
universals in pre-Han philosophy: harm: 45-47, 184, 252, 321-323.
33. 'Objects' (shin JJ) are particular: Defined: 253f. The directly desired
32, 35, 196f, 325 (cheng jE) and the desired after
See 'Nominalism'. weighing: 46f, 184, 254, 332. Sacrifice
Utility. Utilitarian attitude of Mohism: of part for sake of whole: 45, 253-255.
10, 12, 281. Of Individualism: 281. In Individualist school: 46, 254, 322.
Rejected by Mencius: 281 In Hsun-tzu: 233f, 322f. In Mencius :
See 'Benefit'. 46, 255
See 'Unit and total'.
Virtue (te $8). In Chuang-tzu: 63. Not Weight and pull, in mechanics: 386
a later Mohist concept: 59 Wen-hsien t'ung-k'ao ^MM^^ °9
Vocabulary, of dialectical chapters. (n. 98), 392

Simple except for technical terms: 77, Wu ^ manuscript of Mo-tzu: 74, 84,
161,167. Simplicity disguised by rare 195, 211, 426 (n. 485), 440 (n. 537),
graphs: 77, 161. Technical terms: 451 (n. 577)
167-216. Glossary: 547-570
Yen Vie lun ftftift: 173, 338 (n. 238)
Wall. Far side of wall as metaphor for Yi-li Han manuscript of: 89
the inaccessible to observation: 37, Yin ^ and Yang in Chinese proto-
223f, 248, 321f, 342f sciences: 53f, 55 (n. 75). Ignored by
War. Mohists as military engineers: 3,4, later Mohists: 55, 372f
6, 7. Against aggressive but not Yin Wen tzu -PX^: 66 (n. 82), 188
defensive war: 4, 12, 273 Yu ^tf manuscript of Mo-tzu ch. 1-13:
Way. See 'Tao'. 68, 82, 86, 274
A. C. G R A H A M is Professor of Classical
Chinese at the School of Oriental and
African Studies, London University. He
is the author of Two Chinese philosophers:
Ch'eng Ming-tao and Ch'eng Yi-ch'uan
(Lund Humphries, 1958), The book of
Lieh-tzu (John Murray, i960), The pro
blem of value (Hutchinson's University
Library, i960), Poems of the late T'ang
(Penguin Classics, 1965), and studies in
the Sinological journals on the thought,
the textual criticism and the grammar of
Chinese philosophical literature.
ISBNs
C U P 962-201-142-X
SOAS 0 7286 0025 0

Graham A.C. Later Mohist Logic, Ethics and Science, 1978

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Graham A.C. Later Mohist Logic, Ethics and Science, 1978

Uploaded by

Copyright:

Available Formats

Later Mohist Logic

Ethics and Science

All rights reserved

Published in the Far East by The Chinese University Press,

Set in Monotype Imprint

1/1 The Mohist philosophy 3

1/2/2/3 The Canons at Stage 2 91

1/3 Grammar Ill

1/4/4 Chih l h and hsing ft .. 177

1/4/38 Yi% 214

1/6 The organisation of the Canons and Explanations 229

PART II. TEXT AND TRANSLATION

2/3 The fragments of Expounding the canons 243

2/4 The Canons and Explanations 261

2/4/1/2/4 Saying (A 29-33) 284

PART III. APPENDICES

3/1 Abbreviations 497

3/2 The Taoist Patrology text of Mo-tzu ch. 40-45 499

3/3 The arrangement of the Canons at Stage 2 527

3/4 Bibliography 531

is extraordinarily clear and strict, confining each particle to a single gram­

The present edition passes over many older suggestions without

ledge my especial gratitude to Nathan Sivin, who collaborated with me on

THE MOHIST PHILOSOPHY

1/1/1 T H E HISTORICAL BACKGROUND

1/1/1/1 Early Mohism

('The will of Heaven', 'Elucidating the spirits'), philosophy ('Rejecting

• Lii-shih ch'un-ch'iu ch. 1/5 (HsU 95/1-96/6).

another name of obscure origin; unless it was a surname (and no other

city gates. The chapters about military engineering in Mo-tzu, whether

Throughout Mo-tzu one notices that chiin-tzu St"?* gentleman' and

1/1/1/2 The metaphysical crisis in the 4th century

momentum. The individualism expounded in the Lii-shih ch'un-ch'iu is a

"That which first engenders it is Heaven, that which by nurturing fulfils

current, that it is a mixture of good and bad, or neutral, or good in some

With the posing of deeper questions and the multiplication of compet­

actually describe and operate the apparatus of disputation, the later

SL » » M m m > m^m · · *mnmk ° ·

man"; K'ung Ch'uan, who as the descendant of Confucius is presented as the

1/1/1/3 Later Mohism

"The followers of Hsiang-li Ch'in's disciple W u Hou, and the Mohists of

hand, develops a new theory of description.

1/1/2 L A T E R M O H I S T L O G I C , ETHICS A N D SCIENCE

1/1/2/1 Logic and the Chinese language

1/1/2/1/2. The distinctions between different senses of the Indo-

even then continued to see it as a distinction between existence and essence

1/1/2/1/3. In Classical Chinese there is a radical difference between

1/1/2/1/4. In Classical Chinese there is some inconvenience in

1/1/2/1/5. In Chinese there is no presumption that the subject

when omitted has to be understood. Moreover we have a verbal copula:

1/1/2/1/6. Indo-European languages indicate time and number by

has trouble rendering into English without committing oneself to tense or

1/1/2/1/7. In Chinese one can place a verb in a nominal position and

1/1/2/2 The four objects of knowledge

1/1/2/3 Knowing, names and objects, change and necessity

1/1/2/4/1. The art of description in the Canons

'non-ox'), alternatives of which to affirm one of certain things is to affirm

name things as we please, I have only to transfer the name 'horse' to

1/1/2/4/2. The art of description in Names and objects

(6) Being without robbers is not being without people.

Since the Mohists were moralists before everything else, it is not

music is commonly supposed to be beneficial, but according to Mohism

A 26 " 'Benefit/ is what one is pleased to get."

is extraordinarily clear and strict, confining each particle to a single gram

With the posing of deeper questions and the multiplication of compet

The Mohists do not formulate their solutions in physics in mathe