Professional Documents
Culture Documents
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
ISBNs
The Chinese University Press 962-201-142-X
School of Oriental and African Studies 0 7286 0025 0
PREFACE xi
PART I. INTRODUCTION
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
INTRODUCTION
1/1
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.
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
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).
8 The Mohist Philosophy
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).
10 The Mohist Philosophy
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).
12 The Mohist Philosophy
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).
14 The Mohist Philosophy
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.
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
16 The Mohist Philosophy
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
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)
44
Lieh-tzu ch. 7.
46
Huai-nan-tzu ch. 13 (Liu 13, 9B/10).
18 The Mohist Philosophy
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.
20 The Mohist Philosophy
"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<
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
22 The Mohist Philosophy
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.
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.
24 The Mohist Philosophy
(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
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 ) .
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).
B 74 ftPp^A . . . «83£0TFP1 o
"He asks about all men. I love all whom he asks about. (Therefore I love
all men.)"
N O 15 ^AA-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."
6 5
Cf. G(5) 39-48.
1/1 29
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.
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).
32 The Mohist Philosophy
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.
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.
34 The Mohist Philosophy
"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
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
38 The Mohist Philosophy
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«S^iii).
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
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
42 The Mohist Philosophy
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
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
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:
48 The Mohist Philosophy
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).
50 The Mohist Philosophy
68
Mencius 6A/4.
i/i 51
69
Mencius 7B/25.
52 The Mohist Philosophy
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 jf^E. 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 ^IjK 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
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
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).
56 The Mohist Philosophy
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
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
7 8
For evidence that parallel passages in Kung-sun Lung tzu are borrowed and
misunderstood, cf. G(2) 156-164.
62 The Mohist Philosophy
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
# ± 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
64 The Mohist Philosophy
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 gC^I]
i/i 65
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
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
68 The Mohist Philosophy
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 №$5·^ possessed only
the shorter version and supposed the rest of the chapters to be lost. Huang
Chen K M (graduated 1256) and Sung Lien %9k ( A . D . 1310-1381) describe
only the text in three chiian, and the excerpts from Mo-tzu in the Shuo-fu
ftff of T'ao Tsung-yi (ft. A . D . 1360-1368) are all from the first 13
99
chapters.
Fortunately the complete Mo-tzu had survived in the Taoist Patrology,
which was printed under the Sung and again in 1445. Eventually someone
noticed it among the 1,476 titles in the later edition. The complete text
returned to general circulation with the L u S edition of 1552 and the
T'ang St edition of 1553, the latter based on the Ming Taoist Patrology
100
and the former descended from the same Sung exemplar. The dialect-
96
Meng-ch'ipi-Van ^gg^Egfe (Hu Tao-ching ^ ^ 1 ^ (1956) item 44), translated
Graham and Sivin 144-146.
97
Ho-nan Ch'eng-shih yi-shu fflffilfitgflf BSS (ch. 18), 254/9-14.
98
Chun-chai tu-shu chih (as n. 91) (ch. 10), 707; (ch. 11), 722. L i T'ao, ap.
Wen-hsien t'ung-k'ao ^iDcffi^ ch. 212 (WYWK, p. 1740 C/12-16).
99
Tzu liieh SPPY 3/3A. Chih-chai shu-lu chieh-t'i ]g]|g£3$#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.
70 The Mohist Philosophy
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 3^IL#}
( 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
1 0 4
Ch'en Li 12, 13A-14A.
72 The Mohist Philosophy
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.
T E X T U A L 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
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
1/2/1/2/2. 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.
1/2/1/2/4. 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 (№).
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.
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.
1/2/1/3/3. £ , ± (§ 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.
1/2/1/3/9. 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).
6
Luan (1957) 152-154.
7
Cf. p. 68 above.
1/2 83
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 3
Pi Yuan 10, 7A/8.
1/2 85
1 4
Luan (1957) 100.
16
Wu Yu-chiang 13, 5A/3.
86 Textual Problems
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.
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 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 $kWt strips, which were bound before writing, has
21
disproved this argument. There is a very strong counter-argument; there
22
is one copy of the Yi li with as many as 100-110 characters to a strip,
but the columns of paper manuscripts are hardly longer than those of
printed editions. At Stage 2 the 23 characters of B 9 were followed by the
9 characters of B 49, which implies columns allowing room for at least 32
characters. Luan, who had made this calculation, admitted that he could
not find columns of as many as 30 characters in any of the Tun-huang
23
manuscripts known to h i m . It is therefore likely that Stage 2 belongs to
the 3rd century B . C . and is not much later than Stage 1. Stage 3 on the
other hand is the work of a much later scribe or printer who was not trying
to understand the text, merely fitting it to the shorter columns of the
manuscript or block.
21
Wu-wei Han chien 58. The early Han silk manuscripts found at Ma-wang-tui
JH3y$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.
2 4
Cf. G(6) 134-136.
2 5
Luan (1957) 53.
2 6
Cf. p. 107 below.
1/2 91
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
3 1
Luan (1957) 53.
1/2 93
3 2
Cf. § 1/2/2/2/3.
94 Textual Problems
3 3
Luan (1957) 9.
1/2 95
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.
1/2/2/5/4. (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
(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.
4 0
Cf. p. 109 below.
1/2 101
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
102 Textual Problems
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
(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.
106 Textual Problems
(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.
108 Textual Problems
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).
110 Textual Problems
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
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 ttS^ifc-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
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
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
"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'
6
Cf. §1/3/10.
1/3 117
B2 SI
'All count as milu deer'
B3
B 12
B72
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
"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
1/3/4/7. 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."
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
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
1/3/5/5. (7) mo ^ 'none'. This has been treated with the distri
11
butives.
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".
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 IB^I '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
1 4
Cf. § 1/3/12/1/1 and p. 128 below.
1/3 127
1/3/6 DISTRIBUTION
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^-tfa^I 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."
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."
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
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
'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
1/3/7 QUESTIONS
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
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
1/3/10 QUOTATION
(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 ' .
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.
" '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 . . .").
"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.
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±
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
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
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$]^ 'the purpose of X is to . . . \
1/3/12/4/3. We also find the combinations yu yi (A 68, 86,
N O 12), wu yi (A 73, B 34), but not in their ordinary senses 'have/lack
the means'. The yi seems always to refer to what one uses, judges by, in
comparing things or deciding what they are. The yi is dropped when the
phrase is negated:
A 68
"To some extent coinciding, to some extent not."
A 86 ... o
"Having respects in which they are similar. . . . Not having the similarity."
B 47 x±mm °
"It is not that one thinks of the heat of the fire as belonging to oneself."
B 71 o
"To suppose that it fits the fact is necessarily ill-considered."
It is possible that in both cases wei has dropped out of the text, but
this usage is attested elsewhere in Mo-tzii (ch. 39, Sun 189/9 ftKIA-fe
" A l l think him an excellent man").
1/3/12/4/5. In other cases the instrument with yi is all that is explicit
of a clause with a dropped verb, and its significance is only plain when
one recognises the type of verb:
(1) Preposition yi as in A 3 X 9& J ' UffiWiHb ° ("As for the
'intelligence', it is the means by which one knows"):
B 46 %w^m№ o
"When deemings are linked, one cannot be sure either that they will be of
a kind or that they will conflict. . . . For example, someone deemed a fu,
if you use yung (to link with it: yung fu 'bold fellow') is not being deemed
a fu (husband); something deemed chu, if you use mat yi (to link with it:
mat yi chii 'buy coat and shoes') is being deemed chii (shoes)."
1/3 151
B 41 M± "answer him"
№&>!%%i "answer that you do not know it".
In an obscure phrase in B 44 (f£$^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 ^(^ib)^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 )
"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
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
" 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
1/3/14 STYLE
37
Cf. p. 118 above.
3 8
Cf. pp. 137, 138 above.
1/3 163
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
Postscript
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
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.
170 Technical Terminology
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.
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
A66^a,;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.
172 Technical Terminology
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' (^fJ^IHSHir). 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).
174 Technical Terminology
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 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
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
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
26
Cf. § 1/5/14.
178 Technical Terminology
1/4/5 C H I H Ifc
EC 3 №mmm&m% °
"No external condition (what is under external control) can make me more
,,
beneficial.
1/4/6 CH'lNGift
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.
180 Technical Terminology
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
(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'.)
» 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
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
4 0
M . 33 def. 9.
4 1
M . 2691/4, 5.
184 Technical Terminology
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«
1/4/9 F A N {&
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
"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."
186 Technical Terminology
1/4/10 F A N G ft
"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.
188 Technical Terminology
'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)
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 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.
" 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)
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
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
1/4/17 Li®
1/4/18 L I A N G M, P ' I E N M, C H O U Ml A N D Y I %
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
" . . . 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).
194 Technical Terminology
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."
1/4/20 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).
196 Technical Terminology
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
"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
198 Technical Terminology
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
1/4/22 MING#
7 3
Quoted p. 20 above.
200 Technical Terminology
1/4/24 S H E N G 8&
1/4/25 SHENG S
1/4/26 SHIHS
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>iȣ 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 » . . . ° J^UST · °
1/4/27 TAI?#
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)
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.
204 Technical Terminology
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-
206 Technical Terminology
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).
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)
82
Changes, Great appendix A 12, B 5.
83
Mencius 2A/2, cf. 3B/9.
208 Technical Terminology
The Lii-shih ch'un-ch'iu has a chapter entitled Yin tz'u which begins:
Ch. 18/5 (Hsu 841/2-4) IHMKfilffifti' * £ ( = « £ ) * « № ·
"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
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").
210 Technical Terminology
1/4/33 W u ^
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.
212 Technical Terminology
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
" 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."
X $i Y IRI (ift) 'Doing X is required by doing Y ' (B 58, 67, 68, 82)
The formula 'Xyii Y t'ung* with X and Y representing actions occurs
in three Canons (B 67, 68, 82) and one Explanation (B 58), 'Xyii Y t'ung
shuo' with X and Y as propositions in one Explanation (B 82). In all cases
the meaning is 'Doing X is required by doing Y ' , 'You cannot do Y without
doing X * .
B 6 7 ^*m±#*m^±m ·
" Y o u cannot allow that an ox and a horse are not an ox without also
disallowing it."
B82*r*ihJHr«^*itJHIft*«
"Therefore you cannot show that 'this* is not confined without showing
that it is."
In the second case the fuller phrasing of Y is explained by the fact
that it is always Y that is primary and X that is involved in it.
1/4 213
1/4/37 Y l M A N D H S I A N G ffi
"With whatever is the same in kind or the same in ch'ing (what a thing is
in itself), the image of the thing which the senses present is the same
(more literally, 'the senses* image-ing of the thing is the same). Therefore
when put side by side and compared, being confusable and alike they
interchange. This is why they are given a conventional name in common
and so determine each other.**
8 8
There is also a more specific word for 'image*, hsiang ffi ( = S ) ,
also used both verbally (B 38, 'imagine*) and nominally (B 57 'image'). In
both cases hsiang is used with yi, in the latter in the combination yi hsiang
'idea or image*, which appears in Han Fei tzu used verbally:
Ch. 20 (Ch*en 368) A * J i £ * - t b » f f n l # ^ * ^ # ' & * B U « X £ m ° tt
"Men seldom see a live elephant, but when they find a dead elephant's
bones they resort to its picture to imagine it alive. Therefore everything
which men use to form an idea or image is called a hsiang (elephant/image)."
For the preoccupation of Mohists and others with the danger of ideas
being confused by their verbal formulations (tz'u 8$), 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.
214 Technical Terminology
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
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
1/5/1 INTRODUCTION
1/5/2 Ox A N D HORSE
1
Cf. Sun 212/6, 8, -1 : 220/-2 : 221/8, 9.
218 The Stock Examples
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
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.
220 The Stock Examples
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 » MHffn^E3a °
"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
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
222 The Stock Examples
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
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
"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.
224 The Stock Examples
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
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
226 The Stock Examples
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."
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
1/5/13 N O R T H A N D SOUTH
1/5/14 T H E ROAD
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'
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.
232 The Organisation of the 'Canons' and 'Explanations'
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
4
Cf. pp. 63, 64 above.
234 The Organisation of the 'Canons' and 'Explanations'
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 ["^A^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 PM^InM 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
" '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.
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.
236 The Organisation of the 'Canons' and 'Explanations'
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
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
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
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*.
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.
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
250 The Fragments of ' Expounding the Canons'
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 ° »
"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*^igA;£gA-iii»
252 The Fragments of 'Expounding the Canons'
33
Ä 5 W / h A * » 3 » / h A ± W ^ A f t · ülKÄÄ^-tbffiig±[*] gÄ
("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."
EC 8
T C IA/9-IB/4 tkm»±^mmmn±m r* j ° m#*&fc»##i§
2/3 253
«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).
"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."
254 The Fragments of ' Expounding the Canons'
"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
"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
256 The Fragments of 'Expounding the Canons'
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*ssa^A-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".
(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".
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
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
258 The Fragments of 'Expounding the Canons'
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).
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'
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).
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)
266 The 'Canons' and 'Explanations'
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 ' °
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.
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,
268 The 'Canons' and 'Explanations'
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^
(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).
(it) o ^ X T S ^ 6 8
mmm^mz ° °
A 9 I f ' SH& °
7
(it) o »ia#«»M«^a«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. 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."
A 10 ft » °
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 ) .
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 *
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,
274 The 'Canons' and 'Explanations'
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) Iftl^ii^-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.
A 13 # ' WBfo °
78
Chih restored from A 8, which is perfectly parallel.
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 ° > ° (... °)
276 The 'Canons' and 'Explanations'
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
^iMlfk (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).
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).
278 The 'Canons' and 'Explanations'
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".
A 19 ' ± » a i f f i £ 0 f JS-fe °
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 24 ^ » E^ifnJ^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.
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" ^EAWftl, 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
282 The 'Canons' and 'Explanations'
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 ^ElTuMftl), 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."
Mohist thought, but they are the nourishing or thwarting of the vital
impulses:
9
Ch. 1/2 (Hsu 1, 8A/3) » Sf&ttM
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.
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*
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
286 The 'Canons' and 'Explanations'
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
288 The 'Canons' and 'Explanations'
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.
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.
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. ..."
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.
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).
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.
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
292 The 'Canons' and 'Explanations'
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
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
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).
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."
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 °
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,
296 The 'Canons' and 'Explanations'
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.
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.
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.
298 The 'Canons' and 'Explanations'
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.
A 51 & > ^ B i f e o
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).
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
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
304 The 'Canons' and 'Explanations'
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
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
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."
306 The 'Canons' and 'Explanations'
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. . .
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)
(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".
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
310 The 'Canons' and 'Explanations'
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).
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 63 H 1 6 7 9
^PR^-tfe 0
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.
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* :
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).
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).
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).
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.
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.
A 72 S i i f r & M l o
9
A 73 *i& 187
^ "J^ ^ oRfe o
9 1 8 8
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).
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
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
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
322 The 'Canons' and 'Explanations'
"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
A 76 B ° JS » t °
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
324 The 'Canons' and 'Explanations'
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
2 0 3
Delete (Kao, L i u , Liang).
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 · *
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
326 The 'Canons' and 'Explanations'
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
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.
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
"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.
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:
328 The 'Canons' and 'Explanations'
2 1 1
A 81 N ° ( « ) * # » Mo
(W) o > m& o #fui > m& o
2/4 329
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.
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
330 The 'Canons' and 'Explanations'
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).
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 )
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
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.
334 The 'Canons' and 'Explanations'
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 ,
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
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 InJI^I '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 J^Isirf "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".
336 The 'Canons' and 'Explanations'
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
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—'^iffijSlI
"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 -^E&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
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
A 93 R * ™ —*Jffl247 o
w j o i , ; ^ ® ' HiJf^im »
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."
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 ^ , №
344 The 'Canons' and 'Explanations'
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
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 97 j h H a S O J i o
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 ^imum ·{%m®U)™ » m%&№&
260
262
r£'£@f££ °
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
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."
B 2an 2 6 5
» m m m ± m ° mfe{2L)*±™ r J <B 3r 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).
350 The 'Canons' and 'Explanations'
' 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."
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.
352 The 'Canons' and 'Explanations'
B3 2 7 1
[ « B 2 ]m £ » r — » B J » r ^ j » r * » J B J »reis(8a)
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
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
Corrected from the Canon.
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."
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
H. ( 5 ) 284
SSI» —r^yefi » ] R * f f t S £ o
2 8 5
C. "Even when one cannot be dropped without the other they are two.
Explained by: seeing and appearing, one and two, length and breadth.
356 The 'Canons' and 'Explanations'
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.)"
294
Ku 'the thing as it inherently is', the noun corresponding to the adver
bial ku 'inherently' of B 3 (§ 1/4/14).
358 The 'Canons' and 'Explanations'
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.
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
B 9 HxZSx» »iswmftA*n±»o °
(»)o*«^»r«fJ-lb«a±' J298^O(^)*^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."
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):
$C^ife, # 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'.
B 11 £J&£— · ^ 3
<38> S o t f t f t ( f i ) « » o
0 9 3 1 0
# u
8 1 2
< B 4 3 ^T/JcdtJ » T A A J i № r ( & ) * * J · > <B43£±
814
· r j m r * J · ( 5 g § § « j s & ± « o · mm · >
362 The 'Canons' and 'Explanations'
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 *
/
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.)"
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'.
"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.)
366 The 'Canons' and 'Explanations'
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
(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
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'.
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.
B is *t&#s °
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'."
" 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 № *
370 The 'Canons' and 'Explanations'
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
"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.
372 The 'Canons' and 'Explanations'
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]".
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» »
»tr^a > 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:
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.
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).
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.
376 The 'Canons' and 'Explanations'
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
Emended on grounds of parallelism.
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."
"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
378 The 'Canons' and 'Explanations'
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."
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.
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).
(№,) mm™ss^a№iE^^
° I E « « I * * » ° xftfr >ftst*°
' «[4X](W)*InJJb o ^ # Z ( ^ ) * ^
347
^»fr^» 3 4 8
344
Tao, 'turn over', as in B 19.
380 The 'Canons' and 'Explanations'
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
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
382 The 'Canons' and 'Explanations'
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."
(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
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.
(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
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
3 5 6
Corrected from the head character.
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
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
390 The 'Canons' and 'Explanations'
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.
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
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*.
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
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
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
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).
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 The 'Canons' and 'Explanations'
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.)"
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.
400 The 'Canons' and 'Explanations'
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."
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' " .
402 The 'Canons' and 'Explanations'
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.
B 35 mmm&>jK*vt»№&m °
4 o 8
ssii^»(^)*K ^a±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."
B 36 rn^mw^ ° tsfffe °
'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."
4 1 0
B 38 »itfc—«*ffS£nJS£ * ! S £ D U 1 ( ^ ) * # *
(*H) o r ^ « ? s 4 1 1
>w@j^i3fafc« »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.
" 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) :
*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:
$K = W: (both proposed by Kao Heng) and possibly # = H ,
although this would not materially affect the sense (here Kao unneces
sarily emends to tung 'winter'). It is clear that the graph -& cannot
in this context represent the particle yeh ; and its only appearance in
a similar context is in E C App. 13 KSffi^Eifa "Its analogy is: dead
snake", where there is actually a variant № . Moreover the phrases
both occur in B 50, which has other indications of being
about snakes. It is remarkable that the text as we have it should use
the graph ifa both for she$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).
408 The 'Canons' and 'Explanations'
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."
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".
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
410 The 'Canons' and 'Explanations'
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
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.
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
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:
414 The 'Canons' and 'Explanations'
(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/g^o <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.
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."
416 The 'Canons' and 'Explanations'
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
Corrected from the head character.
4 4 8
Sun emends the graph (of which this is the only example in the dialec
tical chapters) to tu $1 'observe'; others take it as t'un 'assemble*
(T'an) or ch'un #6 'pure* (Kao).
449
Yi 'deem* ( = yi wet % as in B 71 (§ 1/3/12/4/4).
4 6 0
The pattern is * . . . yeh, fei. . .* (without final yeh, as in B 3), 'It is
that. . ., it is not that ...'(§ 1/3/5/7). For woyu 'my own possession',
cf. Lu-shih ch'un-ch'iu ch. 1/3 (Hsu l , H A / 5 ) ^ S ^ * J W f i f D * I S f c ,
#FA^c "Now my life likewise as my possession benefits me greatly'*,
Chuang-tzu ch. 22 (Kuo 739/3) ^ # ^ 5 W i f e , " I f my own
person is not my possession who possesses it?**.
"Fire is not hot*' (X^$&) 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
451
B 48 20jfc№ [ I X ] · o
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.
418 The 'Canons' and 'Explanations'
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."
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
^ 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).
4 6 0
as in A 83.
46i, 462 Restored on grounds of parallelism.
420 The 'Canons' and 'Explanations'
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".
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" (-^^E
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
ȣ 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).
422 The 'Canons' and 'Explanations'
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".
in the present, the object taken as example resides in the past. (Like
'Perhaps within the city gates?' and in the runaway Jack.)"
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"
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 '^ow-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
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.
"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
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
o^agJJUc498 » ^ ^ - t f e # ^ 4 9 4 o
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:
^$P@f!S* "Yang-yang-ly anxious, not knowing anythingfixed".T h e
Mohist would probably write the word as yang-jan since he seems to
avoid repeating syllables (p. 165 above).
4 9 5 4 9 6 4 9 7
B58te ±;«*^ »tftffir^IfflJ » r S i f f J *
f§ (*)(it) *m &mi&m »r *im jft°fcmmmm· m&№
498 499 500
1
mmmm™ ' rm(#)*frJ 5 0 2
ft°
4 9 5
Sun and most other editors start the Canon two places later, at the cost
of inflating the summing-up of B 16 to the unintelligible phrase tk&
But once it is recognised that the presence of a head character
at first or second place in the Explanation is a firm rule (§ 1/2/2/5/4),
there can be no doubt that W u Ju-lun and W u Yu-chiang are right in
starting the Canon at ch'ui 'hammer'. The graph is written in its
standard form ($£) 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.
430 The 'Canons' and 'Explanations'
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 » < Br> " 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
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."
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&—» »^ifn-*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
B 6i Ri«nb»w^rfn^^i* 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
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»*
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.
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
Emended following Sun.
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
Emended following Sun.
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".
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&»^as-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.
438 The 'Canons' and 'Explanations'
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
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^^I), 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
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."
"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
E · Piffn^W » A ^ * i b o « ^ j f n * * # £ o SljTJPfCPg » * * * * · m
9 9
$ A « A # nU№&& * ftA^AS <# ·
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
442 The 'Canons' and 'Explanations'
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.
Restored on grounds of parallelism.
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."
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.
444 The 'Canons' and 'Explanations'
'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).
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.
Restored on grounds of parallelism.
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".
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 °
450 The 'Canons' and 'Explanations'
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
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.
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.
583
B 78 tt ±^IS^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
B79
#B£SMil» ^m\- · ' *«I2Nl» ft^»ife ·
5 8 6
The head character has been wrongly provided with the radical.
5 8 7
B80^(«)*ffi ^S'tSft^rSJ ·
( % o B S S M ' *ft»ft»*5gsB&* o ft±d6te*Ste#»H*ifeft ·
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).
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.
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: ,
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*.
"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
460 The 'Canons' and 'Explanations'
"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).
462 The 'Canons' and 'Explanations'
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
9 692
XT«st^ 9
m^imvtfmm · mm^ ?
r » j -tfe# » X T ± W « 4 o r ^ j t t t ' xTzmmfo ·
59s
KiXTzm
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 ?
466 The 'Canons' and 'Explanations'
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".
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?
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 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&* ^:'^iclll 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 £ & °
N0 3
sn*i%» T C 3B/2-4... ^at-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 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."
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
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" . . .
474 Names and Objects
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
Emended on grounds of parallelism.
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.
not as one, therefore they are different things. The wood of willow-wood
and the word of peach-wood are the same thing . . .
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
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»*
(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,
480 Names and Objects
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
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
482 Names and Objects
we took the ti to be the name 'pillar' applied to the object. The sentence
grows from this root to its full length as
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
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.
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 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
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.
488 Names and Objects
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
N O 16
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.
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°
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'."
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
Restored on grounds of parallelism.
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
Emended on grounds of parallelism.
6 5 0
Delete as an accidental repetition.
6 5 1
Emended on grounds of parallelism.
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'."
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
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
The Taoist Patrology Text
4a
3/2 505
The Taoist Patrology Text
3/2 507
*** i
The Taoist Patrology Text
3/2 509
The Taoist Patrology Text
3/2 511
The Taoist Patrology Text
+
+
3/2 513
WW H№ tu&steéК
The Taoist Patrology Text
u* tf№t ^ ^ ^ ^
3/2 515
The Taoist Patrology Text
4
3/2
The Taoist Patrology Text
4¥ ^ α V ^ 4
Vψ ^»•WP» VfcF^
The Tacist Patrology Text
3/2 521
\
The Taoist Patrology Text
3/2
The Taoist Patrology Text
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^^^i« « 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
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
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
INDEXES
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
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
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
*. *
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.
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
# 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
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
# 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
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
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.
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
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
*
(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
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
3/5/3 SUBJECT
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-^i 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
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
"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
ISBNs
C U P 962-201-142-X
SOAS 0 7286 0025 0