You are on page 1of 290

Firt published by New Left Books, 1975

Revised edition published by Verso 1988


Third editon published by Verso 1993
Paul Feyerabend 1975, 1988, 1993
All rights resered
Vero
UK: 6 Meard Steet, London WI V 3HR
USA: 29 West 35th Street, New York, NY 1 01-2291
Verso is the imprint of New Left Boks
Britsh Librar Cataloging in Publicaton Data available
ISBN 0-86091 -481 -X
ISBN 0-86091 -646-4
US Librar of Congress Cataloging in Publicaton Data available
Typeset by Keyboard Serices, Luton
Printed and bound in Great Britain by
Biddies Ltd, Guildford and King's Lynn
Contents
Preface vii
Preface to the Third Edition I
Intoducton to the Chinese Editon 1
Analycal Index 5
Introducton 9
Parts1-20 14
Postscript on Relativism 268
Index 273
Preface
In 1 970 Imre Lakatos, one of the best friends I ever had, corered me
at a party. 'Paul,' he said, 'you have such strange ideas. Why don't you
write them down? I shall write a reply, we publish the whole thing and
I promise you - we shall have lots of fun.' I liked the suggeston and
stared working. The manuscript of my par of the bok was fnished
in
1
972 and I sent it to London. There it disappeared under rather
mysterious circumstances. lmre Lakatos, who loved dramatc
gestures, notfed Interol and, indeed, Interol found my manu
script and retured it to me. I reread it and made some fnal changes.
In Februar
1
974, only a few weeks after I had fnished my revision, I
was informed of Imre's death. I published my par of our common
enterrise without his response. A year later I published a second
volume, Science in a Free Societ, containing additonal material and
replies to critcism.
This histor explains the form of the book. It is not a systematc
teatse; it is a letter to a friend and addresses his idiosyncrasies. For
example, Imre Lakatos was a ratonalist, hence ratonalism plays a
large role in the book. He also admired Popper and therefore Popper
occurs much more frequently than his 'objectve importance' would
warrant. Imre Lakatos, somewhat jokingly, called me an anarchist
and I had no objecton to puttng on the anarchist's mask. Finally,
lmre Lakatos loved to embarrass serious opponents with jokes and
irony and so I, too, ocasionally wrote in a rather ironical vein. An
example is the end of Chapter 1: 'anything goes' is not a 'principle' I
hold - I do not think that 'principles' can be used and fruitfully
discussed outside the concrete research situaton they are supposed
to affect - but the terrifed exclamaton of a ratonalist who takes a
closer look at histor. Reading the many thorough, serious,
longinded and thoroughly misguided critcisms I received after
publicaton of the frst English editon I often recalled my exchanges
with Imre; how we would both have laughed had we been able to read
these effusions together.
The new editon merges parts ofAgainstMethodwith excerts from
Scie
ce in a Free Societ. I have omitted material no longer of interest,
vm
PREFACE
added a chapter on the tal of Galileo and a chapter on the noton of
reality that seems to be required by the fact that knowledge is part of a
complex historical process, eliminated mistakes, shortened the
argument wherever possible and freed it from some of its earlier
idiosyncrasies. Again I want to make two points: frst, that science can
stand on its own feet and does not need any help from ratonalists,
secular humanists, Marxists and similar religious movements; and,
secondly, that non-scientfc cultures, procedures and assumptons
can also stand on their own feet and should be allowed to do so, if this
is the wish of their representatves. Science must be protected from
ideologies; and soietes, especially democratc societes, must be
protected from science. This does not mean that scientsts cannot
proft from a philosophical educaton and that humanity has not and
never will profit from the sciences. However, the profts should not
be imposed; they should be examined and freely accepted by the
partes of the exchange. In a democracy scientfc insttutons,
research programes, and suggestons must therefore be subjected
to public contol, there must be a separaton of state and science just
as there is a separaton between state and religious insttutons, and
science should be taught as one view among many and not as the one
and only road to truth and reality. There is nothing in the nature of
science that excludes such insttutonal arrangements or shows that
they are liable to lead to disaster.
None of the ideas that underlie my argument is new. My
interpretaton of scientfc knowledge, for example, was a tviality for
physicists like Mach, Boltmann, Einstein and Bohr. But the ideas of
these great thinkers were distorted beyond recogniton by the rodents
of neopositvism and the competng rodents of the church of'critcal'
ratonalism. Lakatos was, afer Kuhn, one of the few thinkers who
notced the discrepancy and ted to eliminate it by means of a
complex and very interestng theory of ratonality. I don't thk he has
succeeded in this. But the attempt was worth the efort; it has led to
interestng results in the history of science and to new insights into
the limits of reason. I therefore dedicate also this second, already
much more lonely version of our common work to his memor.
Earlier material relatng to the problems in this book is now
collected in my Philosohicl Paper.
1
Farewell to Reaon
2
contains
historical material, especially from the early history of ratonalism in
the West and applicatons to the problems of today.
Berkeley, September 1 987
1. 2 vos, Cambrdge, 1 981 .
2. London, 1 987.
Preface to the Third Editon
Many things have happened since I frst published Against Method
(AM for short). There have been dramatc politcal, soial and
ecological changes. Freedom has increased - but it has brought
hunger, insecurity, nationalistc tensions, wars and staightforard
murder. World leaders have met to deal with the deterioraton of our
resources; as is their habit, they have made speeches and signed
agreements. The agreements are far from satisfactory; some of them
are a sham. However, at least verbally, the environment has become a
world-wide concer. Physicians, developmental agents, priests
working with the poor and disadvantaged have realized that these
people know more about their conditon than a belie f in the universal
excellence of science or organized religion had assumed and they
have changed their actons and their ideas accordingly (liberaton
theology; primary environmental care, etc.). Many intellectuals have
adapted what they have leared at universites and special schools to
make their knowledge more efcient and more humane.
On a more academic level historians (of science, of culture) have
started approaching the past in its own terms. Already in 1933, in his
inaugural lecture at the College de France, Lucien Febvre had
ridiculed writers who, 'sittng at their desks, behind mountains of
paper, having closed and covered their windows', made profound
judgements about the life of landholders, peasants and farmhands. In
a narrow feld historians of science tried to reconstuct the distant
and the more immediate past without distortng it by modem beliefs
about truth (fact) and ratonality. Philosophers then concluded that
the various forms of ratonalism that had ofered their serices had
not only produced chimaeras but would have damaged the sciences
had they been adopted as guides. Here Kuhn's masterpiece played a
dec
isiv
e role.
1
It led to new ideas. Unfortunately it also encouraged
lots of trash. Kuhn's main terms ('paradigm', 'revoluton', 'normal
I . The Structure ofScientic Reolutions, Chicago, 1 962.
X PREFACE TO THE THI RD EDI TI ON
science', 'prescience', 'anomaly', 'puzzle-solving', etc.) tured up in
various forms of pseudoscience while his general approach confused
many writers: fnding that science had been freed from the fetters of a
dogmatc logic and epistemolog they tied to te it down again, this
tme with soiologcal ropes. That tend lasted well into the early
seventes. By contast there are now historians and sociologsts who
concentate on partculars and allow generalites only to the extent
that they are supported by soiohistorical connectons. 'Nature', says
Bruno Latour, referring to 'science in the making' is 'the
consequence of [a] settlement' of'contoversies'.
2
Or, as I wrote in
the frst editon of A: 'Creaton of a thing, and creaton plus full
understanding of a corea ida of the thing, are ve ofe pars ofone
and the same indiisible proces and cannot be separated without
bringing the proess to a stop.'
3
Examples of the new approach are Andrew Pickering, Constraing
Qark, Peter Galison, How Exermets End, Martn Rudwick, The
Great Deonian Contrery, Athur Fine, The Shak Game and
others. 4 There are studies of the various taditons (religious,
stylistc, Ratonage, etc.) that influenced scientsts and shaped their
research;
5
they show the need for a far more complex account of
scientfc knowledge than that which had emerged from positvism
and similar philosophies. On a more general level we have the older
work of Michal Polanyi and then Putam, van Fraassen, Cartwright,
Marcello Pera
6
and, yes, lmre Lakatos, who was sufciently
optmistc to believe that histor herself-a lady he took ver seriously
-ofered simple rules of theor evaluaton.
In sociology the attenton to detail has led to a situaton where the
problem is no longer why and how 'science' changes but how it keeps
together. Philosophers, philosophers of biology especially, suspected
for some tme that there is not one entty 'science' with clearly
defned principles but that science contains a geat variety of (high
level theoretcal, phenomenologcal, experimental) approaches and
that even a partcular science such as physics is but a scattered
collecton of subjects (elastcity, hydrodynamics, rheology, thermo
dynamics, etc., etc.) each one containing contar tendencies
(example: Prandtl vs Helmholtz, Kelvin, Lamb, Rayleigh; Truesdell
2. 5cimceiaZuiea,Milton Keynes, 1 987, pp. 4 and 98f.
3. London, 1 975, p. 26, repeated on p. 1 7 of the present editon - original
emphasis.
4. All Chicago University Press.
5. An example is Mario Biagioli, CmCeanio,forhcoming.
6. 5cimcean Rhetenc,forhcoming.
PREFACE TO THE THI RD EDI TI ON XI
vs Prandtl; Birkhof vs 'physical commonsense'; Kinsman illustatng
all
trend
s - in hydrodYamics). For some authors this is not only a
fact; it is also desirable. 7 Here again I contibuted, in a small way, in
Chapters 3, 4 and
1 1
of A,8 in secton
6
of my contributon to
Lakatos and Musgrave's Crticism and the Growth of Knowledge
(critcism of the uniformity of paradigms in Kuhn)
9
and already in
1 9
62, in my contibuton to the Delaware Studies fr the Philosohy of
Scie
ce.
10
Unity further disappears when we pay attenton not only to breaks
on the theoretcal level, but to experiment and, especially, to modem
laborator science. As Ian Hacking has shown in his pathbreaking
essay Rereseting and lnteein
y
1
1
and as emerges from Pickering's
Sciece a Praice and Culture,
2
terms such as 'experiment' and
'obseraton' cover complex proesses containing many stands.
'Facts' come from negotatons between diferent partes and the
final product - the published report - is influenced by physical
events, dataprocessors, compromises, exhauston, lack of money,
natonal pride and so on. Some microstudies of laborator science
resemble the 'New Jouralism' of Jimmy Breslin, Guy Talese, Tom
Wolfe and others; researchers no longer sit back and read the papers
in a certain field; they are not content with silent visits to laboratories
either -they walk right in, engage scientsts in conversaton and make
things happen (Kuhn and his collaborators started the proedure in
their interviews for the histor of quantum mechanics). At any rate
we are a long way from the old (Platonic) idea of science as a system of
statements growing with experiment and obseraton and kept in
order by lastng ratonal standards.
A is stll partly propositon oriented; however, I also had my
sane moments. My discussion of incommensurability, for example,
does not 'reduce the diference to one of theor' as Pickering
writes. 1
3
It includes art forms, perceptons (a large part of Chapter
1 6
is about the tansiton from Greek geometric art and poet to the
classical period), stages of child development and asserts 'that the
views of scientsts and especially their views on basic matters are
often as diferent from each other as are the ideologies of diferent
7. J. Dupre, 'The Disunit ofScience, Miad,92, I 983.
8.
Present editon. Taken over unamended from frst editon.
9.
I.
Lakaros and A. Musgave (eds), Cnticum aad thc Cmmth e/Kaemlcdgc,
Cam
bridge, I965.
I O.
'How
to be a God Empiricist', Dcmman5tadic,Vol. 2, I963.
I I.
Cambridge, I 983.
I 2. A.
Pickering (ed.), 5cicca FrmiccaodCaltarc,Chicago, I 992.
1 3.
ibid., p. I 0.
Xtt PREFACE TO THE THI RD EDI TI ON
cultures'.14 In this connecton I examined the practcal aspects of
logic, the way, that is, in which ideas are related to each other in
ongoing research rather than in the fnished products (if there ever
are such products). My discussion of the many events that consttute
what is being obsered15 and especially my discussion of Galileo's
telescopic discoveries1
6
agree with the requirements of the new
laboratory soiology except that Galileo's 'laboratory' was rather
small by comparison. This case shows, incidentally, that like the
older philosophies of science the new microsociolog is not a
universal account but a descripton of prominent aspects of a special
period. It does not matter. A universal descripton of science at any
rate can at most ofer a list of events.17 It was different in antquit.
It is clear that the new situaton requires a new philosophy and,
above all, new terms. Yet some of the foremost researchers in the
area are stll asking themselves whether a partcular piece of research
produces a 'discovery', or an inventon', or to what extent a
(temporary) result is 'objectve'. The problem arose in quantum
mechanics; it is also a problem for classical science. Shall we
contnue using outoded terms to describe novel insight or would it
not be better to start using a new language? And wouldn't poets and
jouralists be of great help in fnding such a language?
Secondly, the new situaton again raises the queston of 'science'
vs democracy. For me this was the most important queston. 'My
main reason for writng the book', I say in the Intoducton to the
Chinese Editon, 18 'was humanitarian, not intellectual. I wanted to
support people, not to "advance knowledge."' Now if science is no
longer a unit, if different parts of it proceed in radically diferent ways
and if connectons between these ways are ted to partcular research
episodes, then scientfc projects have to be taken individually. This
is what goverment agencies started doing some tme ago. In the late
sixtes 'the idea of a comprehensive science policy was gradually
abandoned. It was realized that science was not one but many
enterrises and that there could be no single policy for the support of
all of them. 19 Goverent agencies no longer fnance 'science', they
fnance partcular projects. But then the word 'scientfc' can no
longer exclude 'unscientfc' projects - we have to look at matters in
1 4. A, frt editon, p. 274.
I 5. ibid., pp. I 49f. Reprinted in the present edition.
I 6. Chapter 8 to I 0 of the present editon.
I 7. Cf. my contbuton to the I 92 Erasmus Sympsium, ' Has the Scientfc
View of the World a Spcial Status Compared With Other Views?', forhcoming.
I 8. Contained in the present editon.
1 9. J. Ben-David, 5ntcCmmth,Berkeley, 1 99I , p. 525.
PREFACE TO THE THI RD EDI TI ON Xttt
detail. Are the new philosophers and sociologists prepared to
consider this consequence of their research?
There have been many other changes. Medical researchers and
techn
ologists have not only invented useful instruments (such as
those employng the principles of fbre optcs which in many contexts
replace the more dangerous methods of X-ray diagnostic) but have
become more open towards new (or older) ideas. Only twenty years
ago the idea that the mind afects physical well-being, though
supported by experience, was rather unpopular - today it is
mainstream. Malpractce suits have made physicians more careful,
sometmes too careful for the good of their patent, but they have also
forced them to consult alteratve opinions. (In Switzerland a
belligerent plurality of views is almost part of culture-and I used it
when arranging public confrontatons between hardheaded scientsts
and 'alteratve' thinkers.2< However, here as elsewhere, simple
philosophies, whether of a dogmatc or a more liberal kind, have their
limits. There are no general solutions. An increased liberalism in the
defniton of 'fact' can have grave repercussions/' while the idea
that tuth is concealed and even pererted b
i
the processes that are
meant to establish it makes excellent sense. 2 I therefore again war
the reader that I don't have the intenton of replacing 'old and
dogmatc' principles by 'new and more libertarian ones'. For
example, I am neither a populist for whom an appeal to 'the people' is
the basis of all knowledge, nor a relatvist for whom there are no
'tuths as such' but only truths for this or that group andor
individual. All I say is that non-experts often know more than experts
and should therere be consulted and that prophets of tuth (including
those who use arguments) more often than not are carried along by a
vision that clashes with the ver events the vision is supposed to be
exploring. There exsts ample evidence for both parts of this
asserton.
A case I already mentoned is development: professionals dealing
with
the ecological, social and medical parts of developmental aid
have by now realized that the impositon of 'ratonal' or 'scientfc'
procedures, though occasionally benefcial (removal of some
parasites and infectous diseases), can lead to serious material and
20. Cf. the series edited by Christian Thomas and myself and published by the
Verla
g der F achvereine, Zurich, 1 983-87.
21 .
Cf.
Peter W. Huber, Galileo' Reege, New York, 1991 .
22.
For a
fctional account, cf. Tom Wolfe's The Bonfre of the Vaniti, New York,
1987.
xiv PREFACE TO THE THI RD EDI TI ON
spiritual problems. They did not abandon what they had leared in
their universites, however; they combined this knowledge with local
beliefs and customs and thereby established a much needed link with
the problems of life that surround us everhere, in the First,
Second, and Third Worlds.
The present editon contains major changes (Chapter 1 9 and part of
Chapter
1 6
have been rewritten, the old Chapter 20 has been
omitted), additons (a paragraph here, a paragraph there), stylistc
changes ( I hope they are improvements) and correctons as well as
additons in the references. As far as I am concered the main ideas
of the essay (i.e. the ideas expressed in italics in the Intoducton to
the Chinese Editon) are rather tivial and appear tivial when
expressed in suitable terms. I prefer more paradoxcal forulatons,
however, for nothing dulls the mind as thoroughly as hearing familiar
words and slogans. It is one of the merits of deconstucton to have
undermined philosophical commonplaces and thus to have made
some people think. Unfortunately it afected only a small circle of
insiders and it afected them in ways that are not always clear, not
even to them. That's why I prefer Nestoy, who was a great, popular
and funny deconstructeur, while Derrida, for all his good intentons,
can't even tell a stor.
Rome,July 1 992
Introduction
to the Chinese Edition
This book proposes a thesis and draws consequences from it. The
thesis is: the eents, procedure and reults that constitute the sciences hae
no common straure; there are no elements that occur in every
scientfc investgaton but are missing elsewhere. Concrete develop
ments (such as the overthrow of steady state cosmologies and the
discovery of the stucture of DNA) have distnct features and we can
often explain why and how these features led to success. But not
ever discover can be accounted for in the same manner, and
proedures that paid off in the past may create havoc when imposed
on the future. Successful research does not obey general standards; it
relies now on one tick, now on another; the moves that advance it
and the standards that defne what counts as an advance are not
always known to the movers. Far-reaching changes of outlook, such
as the so-called 'Coperican Revoluton' or the 'Darinian
Revoluton', afect diferent areas of research in diferent ways and
receive diferent impulses from them. A theor of science that
devises standards and structural elements for a/ scientfc actvites
and authorizes them by reference to 'Reason' or 'Ratonality' may
impress outsiders - but it is much too crude an instrument for the
people on the spot, that is, for scientsts facing some concrete
research problem.
In this book I tr to support the thesis by historical examples. Such
supp
ort does not etablish it; it makes it plausible and the way in which
it is reached indicates how future statements about 'the nature of
science' may be undermined: given any rule, or any general
statement about the sciences, there always exst developments which
are praised by those who support the rule but which show that the
rule does more damage than good.
One consequence of the thesis is that scientic successes cannot be
elained in a simple way. We cannot say: 'the structure of the atomic
nucleus was found because people did A, B, C ... ' wherl A, B and C

re procedures which can be understood independently of their use


lD nuclear physics. All we can do is to give a historical account of the
1
2
AGAI NST METHOD
details, including soial circumstances, accidents and personal
idiosyncrasies.
Another consequence is that the success of'sciece' cannot be used a
an armet fr treating a yet unsoled probles in a standrdied way.
That could be done only if there are proedures that can be detached
from partcular research situatons and whose presence garantees
success. The thesis says that there are no such procedures. Referring
to the success of 'science' in order to justf, say, quantfing human
behaviour is therefore an argment without substance. Quantfica
ton works in some cases, fails in others; for example, it ran into
difficultes in one of the apparently most quanttatve of all sciences,
celestal mechanics (special region: stability of the planetar system)
and was replaced by qualitatve (topological) consideratons.
It also follows that 'non-scietic' procedure cannot be pushed aid b
arumet. To say: 'the procedure you used is non-scientfic, therefore
we cannot tust your results and cannot give you money for research'
assumes that 'science' is successful and that it is successfl because it
uses uniform procedures. The first part of the asserton ('science is
always successful') is not tue, if by 'science' we mean things done by
scientsts - there are lots of failures also. The second part - that
successes are due to uniform procedures- is not tue because there
are no such proedures. Scientsts are like architects who build
buildings of diferent sizes and different shapes and who can be
judged only afer the event, i.e. only after they have finished their
stucture. It may stand up, it may fall down - nobody knows.
But if scientfc achievements can be judged only after the event
and if there is no abstact way of ensuring success beforehand, then
there exsts no special way of weighing scientfc promises either -
scientsts are no better of than anybody else in these matters, they
only know more details. This means that the public cn paricpate in the
discussion without disturbing eisting roa to succes (there are no such
roads). In cases where the scientsts' work afects the public it even
should partcipate: first, because it is a concered party (many
scientfic decisions afect public life); secondly, because such
partcipaton is the best scientfic educaton the public can get-a full
democratzaton of science (which includes the protecton of
minorites such as scientsts) is not in conflict with science. It is in
conflict with a philosophy, often called 'Ratonalism', that uses a
frozen image of science to terrorize people unfamiliar with its
practce.
A consequence to which I allude in Chapter 1 9 and which is
closely connected with its basic thesis is that there can be many dieet
kind ofsciece. People startng from diferent soial backgrounds will
I NTRODUCTI ON TO THE CHI NESE EDI TI ON 3
approach the world in diferent ways and lear diferent things about
it. People surived millennia before Wester science arose; to do this
they had to know their surroundings up to and including elements of
astronomy. 'Several thousand Cuahuila Indians never exhausted the
natural resources of a desert region in South Califoria, in which
today only a handful of white families manage to subsist. They lived
in a land of plenty, for in this apparently completely barren territor,
they were familiar with no less than sixty kinds of edible plants and
twenty-eight others of narcotc, stmulant or medical properes'. 1
The knowledge that preseres the lifestyles of nomads was acquired
and is presered in a non-scientfic way ('science' now being modem
natural science). Chinese technolog for a long tme lacked any
Wester-scientfic underpinning and yet it was far ahead of
contemporar Wester technolog. It is true that Wester science
now reigns supreme all over the globe; however, the reason was not
insight in its 'inherent ratonality' but power play (the colonizing
natons imposed their ways of living) and the need for weapons:
Wester science so far has created the most efcient instuments of
death. The remark that without Wester science many 'Third World
natons' would be starving is correct but one should add that the
toubles were created, not alleviated by earlier forms of 'develop
ment'. It is also tue that Wester medicine helped eradicate
parasites and some infectous diseases but this does not show that
Wester science is the only taditon that has good things to ofer and
that other forms of inquir are without any merit whatsoever. Firt
world science is one science among many; by claiming to be more it ceases
to be an instrument of research and turs into a (politcal) pressure
group. More on these matters can be found in my book Farewell to
Reason.
2
My main motve in writng the book was humanitarian, not
intellectual. I wanted to support people, not to 'advance knowledge'.
People all over the world have developed ways of surviving in partly
dangerous, partly agreeable surroundings. The stories they told and
the actvites they engaged in enriched their lives, protected them and
gave them meaning. The 'progress of knowledge and civilizaton'-as
the process of pushing Wester ways and values into all comers of the
globe
is being called-destoyed these wonderful products of human
ingenuity and compassion without a single glance in their directon.
'Pro
gress of knowledge' in many places meant killing of minds.
Tod
ay old traditons are being revived and people try again to adapt
I. C. Lev-Stauss, 1hc5magcMiad,London, 1 966, pp. 4f.
2.
London, 1 987.
4 AGAI NST METHOD
their lives t o the ideas of their ancestors. I have tried t o show, by an
analysis of the apparently hardest parts of science, the natural
sciences, that science, properly understood, has no argument against
such a procedure. There are many scientsts who act accordingly.
Physicians, anthropologists and environmentalists are startng to
adapt their procedures to the values of the people they are supposed
to advise. I am not against a science so understood. Such a science is
one of the most wonderful inventons of the human mind. But I am
against ideologies that use the name of science for cultural murder.
Analycal Index
Being a Skech ofthe Main Armet
Intoducton
9
Sciec is an esetill anarchic eterre: theoreticl anarchim is more
humanitaran and more likel to ecourage proges than its lw-and-ordr
alteatie.
1 14
Ti i shown both b an eamination ofhitoncl eisod and b an
abstra analysi ofthe relation bewee ida and aion. Te onl prnciple
that des not inhibit prgess is: ayg goes.
2 20
For eample, we may use hypothee that contradia well-confred theore
andor wel-established eermetal reults. We may adance sciec b
prceeding counternduaiel.
3 24
Te conistec condition which dmand that new hyothee agee with
aceted teores is unreasonable becuse it preee the oldr theor, and
not the better theor. Hypothee cntradiaing well-confred theore ge
us
eidnce that cannot be obtained in at other way. Prolieation of
theore is beecial fr sciece, while unirit impair its crticl power.
Unirit alo ednger the free delomet ofthe indiidual.
4
33
T
here
is no ida, howee anciet and absurd, that is not capable of
impring our knowledge. Te whole histor ofthought is absored into
scece and i used fr impruing ee single theor. Nor i political
intererece reeaed. It may be needd to uercome the chafinism ofscece
that resists alteatie to the status quo.
s
39
No
theo
r eer agee with all the fas in its dmain, yet it i not alwas the
the
or
that
is to blme. Fas are cnstituted b olr idloge, and a clash
5
6 AGAI NST METHOD
betwee fas and theore may be proof ofprgres. It is alo a frt ste in our
attept to fnd the principle implicit in familiar obserational notions.
6 54
As an eample ofsuch an attept I eamine the tower argument which the
Aristotelians used to reute the motion ofthe earh. Te armet irole
natural interpretatons - idas so cosel conneaed with obserations that
it need a special efr to realie their eistece and to dterine their
contet. Galileo idntie the natural interretations which are inconsistet
with Coeics and relace the b other.
7 65
Te ne natural interretations constitute a new and highl abstra
obseration langage. The are introduced and concealed so that one fils
to notice the change that has take p
i
ace (method ofanamneis). Te contain
the ida ofthe relatvit of all moton and the law of circular inerta.
8 77
In ad ition to natural interretations, Galleo alo change sensatons that
see to ednger Coeics. He admits that there are such sesations, he
praise Coeics fr haing disregardd the, he caims to hae reoed
the with the help ofthe telescope. Howee, he off e no theoretcal
reasons why the telecoe shoul be eeaed to gve a tre piaure ofthe sA.
9 86
Nor de the initial experience with the telecoe prid such reaons. Te
frt telecoic obserations of the sA are indistina, indterinate,
cntradiaor and in conjia with what eeone cn see with his unaidd
ee. And, the onl theor that could hae helped to searate telecoic
ilusions frm vedical pheomea wa reuted b simple tets.
IO 103
On the other hand, thee are some telecoic pheomea which are plinl
Coeican. Galileo intrduce thee pheomea a indedt eidnce
fr Coeics while the situation is rathe that one reuted view -
Coeicnism - ha a cerain similrt with pheomea eerng fm
anothe reuted vie - the ida that telecpic pheomea are fithfl
image ofthe sA.
II 106
Such 'irational' method ofsuppor are needd becuse ofthe 'unee
delomet' (Mar, Lein) ofdi eet pars ofsciece. Coeicn ism and
othe esetial ingrediets ofmod sciece suried onl because reaon
wa frequetl oerled in their pat.
ANALYTI CAL I NDEX 7
12
123
Galieo s method works in other feld as well. For eample, it cn be used to
eliminate the eisting armets against materalism, and to put an ed to
the philosophical mindbod proble (the creponding scientfc
prbles reain untouched, howee). It de not fllow that it should be
unie
rall aplied.
13
125
The Church at the time ofGalileo not onl ket closer to reason as dned
the and, in par, ee now: it also considred the ethical and socal
consequece ofGalileo's views. Its indicmet ofGalileo wa rational and
onl op orunism and a lack ofperpecie can dmand a reision.
14 135
Galileo 's inquire fred onl a small par ofthe so-called Coeicn
Reolution. Ading the reaining eleet makes it still more dif clt to
reconcle the delomet with familiar prncple oftheor ealuation.
15 147
Te reults obtained so far sugest abolihing the ditincion betwee a
contet of discoer and a contet ofjustication, nors and facs,
obserational tes and theoreticl ters. None ofthee distincions plays a
role in scetic pracice. Attepts to efrce the would hae disatrous
consequece. Pop e's crticl rationalim fail fr the same reasons.
Appendi 1 159
16 164
Finall, the kind ofcomparson that undrlie most methodloge is possible
onl in some rather simple ces. It breaks dwn whe we tr to cmpare
non-scietic views with sciece and whe we considr the most aancd,
most geeal and theere most mythologcal pars ofsciece itsel
Appendi2
209
17
214
Neithe sciece nor rationalit are unieral measures ofecellece. Te are
P
aricla
r traditions, unaware oftheir historcal gunding.
18
230
Ye
t it is
possible to ealuate standrd ofrationalit and to improe the.
T
he prnciple of impreet are neither aboe tradition nor beond
cha
nge
and it is impossible to nail the dw.
8 AGAI NST METHOD
1 9 238
Sciece is neither a single tradition, nor the best tradition there is, ecet fr
peole who have become accstomed to its preece, its beets and its
disadantage. In a dmocac it should be searated from the state just a
churche are now searated fom the state.
20 252
The point ofview undrling this book is not the reult ofa wel-planned
train ofthought but ofarmets prompted b acidntal ecounte. Ange
at the wanton dtrcion ofcltural ahieeets from which we al could
have leared, at the conceited asurance with which some intelecuals
interee with the lives ofpeole, and contept for the treal phrae the
use to ebelih their misded wa and stil is the motive frce behind my
work.
Intoducton
Scece i an esetiall anarchic eterre: theoreticl anarchim i more
humanitaran and more lkel to ecurage proges than it lw-and-ordr
alteatie.
Ordnung ist heututage meistens dor,
wo nichts ist.
Es ist eine Mangelercheinung.
BRCHT
The following essay is written in the convicton that anarchism, while
perhaps not the most attactve political philosophy, is certainly
excellent medicine for eitemolog, and for the philosohy ofscece.
The reason is not difcult to find.
'Histor generally, and the histor of revoluton in partcular, is
always richer in content, more vared, more many-sided, more lively
and subtle than even' the best historian and the best methodologist
can imagine. 1 Histor is full of 'accidents and conjunctures and
curious juxtapositons of events'2 and it demonstates to us the
'complexty of human change and the unpredictable character of the
ultmate consequences of any given act or decision of men'.
3
Are we
really to believe that the naive and simple-minded rules which
methodologists take a their guide are capable of accountng for such
a 'maze of interactons'?4 And is it not clear that successful
l . 'History a a whole, and the history of revolutons in parcular, is always richer
in content, more varied, more multfor, more lively and ingenious than is imagined
by even the best pares, the most conscious vanguards of the most advanced clases'
(.I. Lenin, 'Lef-Wing Communism- An Infantle Disorder', 5clcucdKe,Vol. 3,
Lon
don, 1 96 7, p. 4 I). Lenin is addressing pares and revolutonar vanguards rater
than sie
ntsts and methoologists; the lesson, however, is the sme. Cf. fotote 5.
2.
Herbert Butterfield, 1hcmigIatcqrcmtieae/Huteq,New York, 1 965, p. 66.
3.
ibid., p. 21 .
4. ibid., p. 25, cf. Hegel, Fhiles@hicdCcchichtc, Kmc,Vol. 9, ed. Edward Gans,
B
erlin
, 1 837, p. 9: 'But what experience and histor teach us is this, that natons and
9
10 AGAI NST METHOD
paricpation i n a process of this kind i s possible only for a ruthless
opportunist who is not ted to any partcular philosophy and who
adopts whatever procedure seems to fit the occasion?
This is indeed the conclusion that has been drawn by intelligent
and thoughtful obserers. 'Two very important practcal conclusions
follow from this [character of the historical process],' writes Lenin, 5
contnuing the passage from which I have just quoted. 'First, that in
order to fulfil its task, the revolutonary class [i.e. the class of those
who want to change either a part of society such as science, or society
as a whole] must be able to master al forms or aspects of social
actvity without excepton [it must be able to understand, and to
apply, not only one partcular methodology, but any methodology,
and any variaton thereof it can imagine] . . . ; second [it] must be
ready to pass from one to another in the quickest and most
unexpected manner.' 'The exteral conditons', writes Einstein,
6
'which are set for [the scientst] by the facts of experience do not
permit him to let himself be too much restricted, in the constucton
of his conceptual world, by the adherence to an epistemological
system. He, therefore, must appear to the systematc epistemologist
as a type of unscrupulous opportunist. . e .' A complex medium
containing surrising and unforeseen developments demands
complex procedures and defies analysis on the basis of rules which
goverments have never leared anything from histor, or acted according to rles that
might have derived from it. Every period has such peculiar circumstances, is in such an
individual state, that decisions will have to be made, and decisions caaonly be made, in
it and out of it.' - 'Very clever'; 'shrewd and very clever'; 'NB' writes Lenin in his
marginal notes to this passage. (CellcctcdKerH,Vol. 38, London, 1961 , p. 307.)
5. ibid. We see here very clearly how a few substtutions can tur a politcal lesson
into a lesson for mcthedle.This is not at all surrising. Methodolog and politcs are
both means for moving from one historical stage to another. We also see how an
individual, such as Lenin, who is not intimidated by traditonal boundaries and whose
thought is not tied to the ideolog of a particular profession, can give useful advice to
everyone, philosophers of science included. In the 1 9th centur the idea of an elastic
and historically infored methodolog was a matter of course. Thus Erst Mach
wrote in his book Erkatais aadImam, Neudrck, Wissenschaftliche Buchgesell
schaft, Darstadt, 1 980, p. 20: 'It is often said that research cannot be taught. That is
quite corect, in a certain sense. The schemata of fora/logic and of iadauaclogic are
oflittle use for the intellectual situations are never exactly the same. But the examples
of great scientists are very suggestve.' They are not suggestive because we can abstact
rules from them and subject future research to their jurisdiction; they are suggestive
because they make the mind nimble and capable of inventing entirely new research
traditions. For a more detailed account of Mach's philosophy see my essay Farmcll te
Rcmea, London, 1 987, Chapter 7, as well as Vol. 2, Chapters 5 and 6 of my
Fhiles@hicalFapc,Cambridge, 1 981 .
6. Alber Einstein, Zlbcn Eiastcia: Fhiles@h5citist, ed. P.A. Schilpp, New
York, 1 951 , pp. 683f.
I NTRODU CTI ON 11
have been set up in advance and without regard to the ever-changing
conditons of history.
Now it is, of course, possible to simplif the medium in which a sientst
works by simplifing its main actors. The history of science, after all,
does not just consist of facts and conclusions drawn from facts. It also
contains ideas, interpretations offacts, problems created by conflictng
interpretatons, mistakes, and so on. On closer analysis we even fi nd
that science knows no 'bare facts' at all but that the 'facts' that enter our
knowledge are already viewed in a certain way and are, therefore,
essentally ideatonal. This being the case, the history of science will be
as complex, chaotc, full of mistakes, and entertaining as the ideas it
contains, and these ideas in tum will be as complex, chaotc, full of
mistakes, and entertaining as are the minds of those who invented
them. Conversely, a little brainwashing will go a long way in makng the
histor of science duller, simpler, more uniform, more 'objectve' and
more easily accessible to treatent by strict and unchangeable rules.
Scientfic educaton as we know it today has precisely this aim. It
simplifes 'science' by simplifing its partcipants: frst, a domain of
research is defned. The domain is separated from the rest of history
(physics, for example, is separated from metaphysics and from
theology) and given a 'logic' of its own. A thorough taining in such a
'logic' then conditons those working in the domain; it makes their
aions more uniform and it freezes large parts of the historical proces
as well. Stable 'facts' arise and persevere despite the vicissitudes of
histor. A essental part of the training that makes such facts appear
consists in the attempt to inhibit intuitons that might lead to a
blurring of boundaries. A person's religion, for example, or his
metaphysics, or his sense of humour (his natural sense of humour and
not the inbred and always rather nasty kind of jocularity one fnds in
specialized professions) must not have the slightest connecton with
his scientfc actvity. His imaginaton is restained, and even his
language ceases to be his own. This is again reflected in the nature of
scientfc 'facts' which are experienced as being independent of
opinion, belief, and cultural background.
It is thus possible to create a traditon that is held together by strict
rules, and that is also successful to some extent. But is it dsirable to
sup
port
such a traditon to the exclusion of everything else? Should
we tran
sfer to it the sole rights for dealing in knowledge, so that any
resu
lt
that has been obtained by other methods is at once ruled out of
court? And did scientsts ever remain within the boundaries of the

raditons they defned in this narrow way? These are the questons I
Intend to ask in the present essay. And to these questions my answer
will be a frm and resounding NO.
12
AGAI NST METHOD
There are two reasons why such an answer seems t o be
appropriate. The frst reason is that the world which we want to
explore is a largely unknown entty. We must, therefore, keep our
optons open and we must not restict ourselves in advance.
Epistemological prescriptons may look splendid when compared
with other epistemological prescriptons, or with general principles
but who can guarantee that they are the best way to discover, not just
a few isolated 'facts', but also some deep-lying secrets of nature? The
second reason is that a scientfc educaton as described above (and as
practsed in our schools) cannot be reconciled with a humanitarian
atttude. It is in conflict 'with the cultvaton of individuality which
alone produces, or can produce, well-developed human beings';
7
it
'mais by compression, like a Chinese lady's foot, ever part of
human nature which stands out prominently, and tends to make a
person markedly diferent in outline'
8
from the ideals of ratonality
that happen to be fashionable in science, or in the philosophy of
science. The attempt to increase liberty, to lead a full and rewarding
life, and the corresponding attempt to discover the secrets of nature
and of man, entails, therefore, the rejecton of aluniversal standards
and of all rigid taditons. (Naturally, it also entails the rejecton of a
large part of contemporar science.)
It is surrising to see how rarely the stultfing efect of 'the Laws
of Reason' or of scientfc practce is examined by professional
anarchists. Professional anarchists oppose any kind of resticton and
they demand that the individual be permitted to develop freely,
unhampered by laws, dutes or obligatons. Ad yet they swallow
without protest all the severe standards which scientsts and logicians
impose upon research and upon any kind of knowledge-creatng and
knowledge-changing actvity. Occasionally, the laws of scientfc
method, or what are thought to be the laws of scientfc method
by a partcular writer, are even integrated into anarchism itself.
'Aarchism is a world concept based upon a mechanical explanaton
of all phenomena,' writes Kropotkin. 9 'Its method of investgaton is
7. John Stuan Mill, 'On Libeny', in 1hcFhiles@eQeha5tmuMi/l,ed. Marhall
Cohen, New York, 1 961 , p. 258.
8. ibid., p. 265.
9. Peter Alexeivich Kroptkin, 'Modem Science and Anarchism', Kpetkias
RmlatiqFamphlcc,ed. R.W. Badwin, New York, 1 970, pp. 1 502. 'It is one of
Ibsen's great distnctons that nothing was vaid for him but science.' B. Shaw, kte
Mcthmcmh,New York, 1 921 , p. xcvii. Commentng on these and similar phenomena
Stndberg writes (atibamam ): 'A generaton that had the courage to get rid of Go,
to crsh the stte and church, and to ovenhrow soiety and morality, stll bwed before
Science. And in Science, where freedom ought to reign, the order of the day was
"believe in the authorites or of with your head".'
I NTRODUCTI ON 13
that of the exact natural sciences ... the method of inducton and
dedu
cton.' 'It is not so clear,' writes a modem 'radical' professor at
Colu
mbia
,
10
'that scientfc research demands an absolute freedom
of speech and debate. Rather the evidence suggests that certain kinds
of unfreedom place no obstacle in the way of science . .. .'
There are certainly some people to whom this is 'not so clear'. Let
us, therefore, start with our outline of a anarchistc methodology
and a corresponding anarchistc science. There is no need to fear
that the diminished concer for law and order in science and society
that characterizes an anarchism of this kind will lead to chaos. The
human nerous system is too well organized for that. 11 There may,
of course, come a tme when it will be necessary to give reason a
temporary advantage and when it will be wise to defend its rules to the
exclusion of everything else. I do not think that we are living in such a
tme today.1
2
10. R.P. Wolf, 1hcFme/Liboalum, Boston, 1 968, p. 15. For a critcism of
Wolf see fotote 52 of my essay 'Against Method', in Miaaceta5twic ia thc
Fhilm@he/5occ,Vol. 4, Minneapolis, 1 970.
I I . Even in undetermined and ambigous situatons, uniformit of acton is son
achieved and adhered to tenaciously. See Muzafer Sherif, DcFqcheh e/5ecml
Nem,New York, 1 964.
12. This was my opinion in 1 970 when I wote the frst version of this essay. Times
have changed. Considering some tendencies in US educaton ('politcally corect',
academic menus, etc.), in philosophy (postoderism) and in the world at large I think
that
reason should now be given greater weight not because it is and always was
fundamental but because it seems to be needed, in circumstances that ocur rather
frequently today (but may disappear tomorow), to create a more humane approach.
1
This i shoa both b an eamination ofhistorcl eisod and b an
abstraa analsis ofthe relation betwee ida and aion. The onl prnciple
that de not inhibit prgres is: anything goes.
The idea of a method that contains fnn, unchanging, and
absolutely binding principles for conductng the business of science
meets considerable difcult when confronted with the results of
historical research. We fnd, then, that there is not a single rle,
however plausible, and however firmly grounded in epistemolog,
that is not violated at some tme or other. It becomes evident that
such violatons are not accidental events, they are not results of
insufcient knowledge or of inattenton which mght have been
avoided. On the contary, we see that they are necessary for pro
gress. Indeed, one of the most striking features of recent
discussions in the history and philosophy of science is the
realizaton that events and developments, such as the inventon of
atomism in antquit, the Coperican Revoluton, the rise of
modem atomsm (kinetic theory; dispersion theory; stereochemst;
quantum theory), the gradual emergence of the wave theory of
light, occurred only because some thinkers either dcdd not to be
bound by certain 'obvious' methodological rles, or because they
unwittingl broke them.
This liberal practce, I repeat, is not just a faa of the history of
science. It is both reasonable and absolutel necesar for the growth of
knowledge. More specifcally, one can show the following: given any
rle, however 'fundamental' or 'ratonal', there are always cir
cumstances when it is advisable not only to ignore the rle, but to
adopt its opposite. For example, there are circumstances when it is
advisable to intoduce, elaborate, and defend a hoc hypotheses, or
hypotheses which contadict well-established and generally accepted
experimental results, or hypotheses whose content is smaller than the
14
ONE IS
c
onten
t of the exstng and empirically adequate alteratve, or self
inc
onsistent hypotheses, and so on. 1
There are even circumstances -and they ocur rather frequently
when arumet loses its forard-looking aspect and becomes a
hindrance to progress. Nobody would claim that the teaching of small
childre is exclusively a matter of argument (though argument may
enter into it, and should enter into it to a larger extent than is
customary}, and almost everyone now agrees that what looks like a
result of reason - the mastery of a language, the exstence of a richly
artculated perceptual world, logical ability - is due partly to
indoctrinaton and partly to a process of growth that proeeds with the
force of natural law. And where arguments do seem to have an efect,
this is more often due to their physical reetition than to their 5eantic
contet.
Having admitted this much, we must also concede the possibility
of non-argumentatve growth in the adult as well as in (the theoretcal
parts of institutions such as science, religion, prosttuton, and so on.
1 . One of the few thinker to undertand this feature of the development of
knowledge was Niels Bohr: ' . . . he would never t to outline any fnished picture, but
would patently go through all the phases of the development of a problem, startng
from some apparent paradox, and gradually leading to its elucidaton. In fact, he never
regarded achieved results in any other ligt than a startng pint for furer
exploraton. In speculatng about the prospects of some line of investgaton, he would
dismiss the usual consideraton of simplicity, elegance or even consistenc wth the
remark that such qualites can only be proprly judged a/o[my italics] the event . . . .'
L. Rosenfeld in Nich ehr. HuLqaad Kerk 11 sc q huFnd aadCelhap,
S. Rosental (ed.), New York, 1 967, p. 1 1 7. Now science is never a completed proess,
therefore it is always 'before' the event. Hence simplicity, elegance or consistenc are
anonecessar conditons of (scientfc) practce.
Consideratons such a these are usually critcied by the chidish remark that a
contadicton 'entails' everything. But contadictons do not 'entail' anything unless
people use them in cerain ways. And people wll use them as entailing everything only
if they accept some rather simple-minded rles of derivaton. Scientst propsing
theories wth logical faults and obtaining interestng result wth their help (for
example: the results of early fors of the calculus; of a geomet where lines consist of
points, planes of lines and volumes of planes; the predictons of the older quantum
theor and of early fors of the quantum theor of radiaton - and so on) evidently
proeed according to diferent rules. The critcism therefore falls back on its authors
unless it can be show that a logically decontaminated sience has better result. Such
a demonstaton is impossible. Logically perfect versions (if such versions exst)
usually arrive only long after the imperfect versions have enriched science by their
contibutons. For example, wave mechanics was not a 'logical reconstucton' of
pre
ceding
theories; it was an attempt to presere their achievements and to solve the
phy
sical
problems that had arisen from their use. Both the achievements and the
pro
blem
s were produced in a way ver diferent from the ways of those who want to
sub
ject
evert
hing to the tyranny of'logic'.
16 AGAI NST METHOD
We certainly cannot take i t for granted that what i s possible for a
small child - to acquire new modes of behaviour on the slightest
provoaton, to slide into them without any notceable effort - is
beyond the reach of his elders. One should rather expect that
catastophic changes in the physical environment, wars, the
breakdown of encompassing systems of morality, politcal revolu
tons, will tansform adult reacton patters as well, including
imporant patters of argumentaton. Such a tansformaton may
again be an entrely natural process and the only function of a ratonal
argument may lie in the fact that it increases the mental tension that
preceded and caused the behavioural outburst.
Now, if there are events, not necessarily arguments, which cuse us
to adopt new standards, including new and more complex forms of
argumentaton, is it then not up to the defenders of the status quo to
provide, not just counter-arguments, but also contary cuse?
('Virtue without terror is ineffectve, ' says Robespierre.) And if the
old forms of argumentaton tur out to be to weak a cause, must not
these defenders either give up or resort to stonger and more
'irratonal' means? (It is ver difcult, and perhaps entrely
impossible, to combat the effects of brainwashing by argment.)
Even the most puritanical ratonalist will then be forced to stop
reasoning and to use proagand and coecion, not because some of his
reasons have ceased to be valid, but because the psychologcal conditions
which make them effectve, and capable of influencing others, have
disappeared. And what is the use of an argument that leaves people
unmoved?
Of course, the problem never arises quite in this form. The
teaching of standards and their defence never consists merely in
puttng them before the mind of the student and making them as cear
as possible. The standards are supposed to have maxal causal
ef cac as well. This makes it very diffcult indeed to distnguish
between the logcal frce and the mateal ea of an argment. Just as a
well-trained pet will obey his master no matter how great the
confusion in which he fnds himself, and no matter how urgent the
need to adopt new patters of behaviour, so in the very same way a
well-trained ratonalist will obey the mental image of his master, he
will conform to the standards of argumentaton he has leared, he
will adhere to these standards no matter how great the confusion in
which he fnds himself, and he will be quite incapable of realizing that
what he regards as the 'voice of reason' is but a cusal afe-ef a of the
taining he had received. He will be quite unable to discover that the
appeal to reason to which he succumbs so readily is nothing but a
political manoeur.
ONE 17
That interests, forces, propaganda and brainwashing techniques
play a much greater role than is commonly believed in the growth of
our knowledge and in the growth of science, can also be seen from an
analysis of the relation between ida and acion. It is often taken for
granted that a clear and distnct understanding of new ideas
precedes, and should precede, their formulaton and their insttu
tonal expression. Firt, we have an idea, or a problem, the we act, i.e.
either speak, or build, or destroy. Yet this is certainly not the way in
which small children develop. They use words, they combine them,
they play with them, untl they grasp a meaning that has so far been
beyond their reach. And the inital playful actvity is an essental
prerequisite of the fnal act of understanding. There is no reason why
this mechanism should cease to functon in the adult. We must
expect, for example, that the ida ofliberty could be made clear only
by means of the very same actons, which were supposed to ceate
liberty. Creaton of a thing, and creaton plus full understanding of a
corec ida of the thing, are ve ofe pars ofone and the same indiisible
prces and cannot be separated without bringing the proess to a
stop. The process itself is not guided by a well-defned programme,
and cannot be guided by such a programme, for it contains the
conditons for the realizaton of all possible programmes. It is guided
rather by a vague urge, by a 'passion' (Kierkegaard). The passion
gives rise to specifc behaviour which in tum creates the cir
cumstances and the ideas necessary for analysing and explaining the
process, for making it 'ratonal'.
The development of the Coperican point of view from Galileo to
the 20th century is a perfect example of the situaton I want to
describe. We start with a stong belief that runs counter to
contemporary reason and contemporary experience. The belief
spreads and fnds support in other beliefs which are equally
unreasonable, if not more so (law of inerta; the telescope). Research
now gets defected in new directons, new knds of instuments are
built, 'evidence' is related to theories in new ways untl there arises an
ideolog that is rich enough to provide independent arguments for
any partcular part of it and mobile enough to fnd such arguments
whe
never they seem to be required. We can say today that Galileo
w
as
on the right track, for his persistent pursuit of what once seemed
to
be a silly cosmolog has by now created the material needed to
de
fend
it against all those who will accept a view only if it is told in a
ce
rtain way and who will tust it only if it contains certain magical
p

rase
s, called 'observatonal reports'. And this is not an excepton

t I
s
the
normal case: theories become clear and 'reasonable' only afer
In
coh
erent parts of them have been used for a long tme. Such
18 AGAI NST METHOD
unreasonable, nonsensical, unmethodical foreplay thus turs out
to be an unavoidable preconditon of clarity and of empirical
success.
Now, when we attempt to describe and to understand develop
ments of this kind in a general way, we are, of course, obliged to
appeal to the exstng forms of speech which do not take them into
account and which must be distorted, misused, beaten into new
patters in order to fit unforeseen situatons (without a constant
misuse of language there cannot be any discover, any progress).
'Moreover, since the taditonal categories are the gospel of everday
thinking (including ordinar scientfic thinking) and of everday
practce, [such an attempt at understanding] in efect presents rules
and forms of false thinking and acton - false, that is, from the
standpoint of (scientfic) common sense.'
2
This is how dialeaicl
thinking arises as a form of thought that 'dissolves into nothing the
detailed determinatons of the understanding',
3
formal logic
included.
(Incidentally, it should be pointed out that my frequent use
of such words as 'progress', 'advance', 'improvement', etc., does
not mean that I claim to possess special knowledge about what
is good and what is bad in the sciences and that I want to impose
this knowledge upon my readers. Everone cn read the ters in
his own way and in accordance with the taditon to which he
belongs. Thus for an empiricist, 'progress' will mean transiton
to a theor that provides direct empirical tests for most of its basic
assumptons. Some people believe the quantum theor to be a
theor of this kind. For others, 'progress' may mean unific"ton
and harmony, perhaps even at the expense of empirical adequac.
This is how Einstein viewed the general theor of relatvity. And
n theis is that anarchism helps to achiee proges in any one of
the senses one care to choose. Even a law-and-order science will
succeed only if anarchistc moves are ocasionally allowed to take
place.)
It is clear, then, that the idea of a fixed method, or of a fed theor
of ratonality, rests on too naive a view of man and his soial
surroundings. To those who look at the rich material provided by
histor, and who are not intent on impoverishing it in order to please
their lower instncts, their craving for intellectual security in the form
of clarity, precision, 'objectvity', 'truth', it will become clear that
there is only one principle that can be defended under all
2. Herben Marcuse, RcmeaaadRnelatiea,London, 1 941 , p. 1 30.
3. Hegel, Kuscha/dLegk,Vol. 1, Hamburg, 1 965, p. 6.
ONE
19
circumstances and in all stages of human development. It i s the
principle: anything goes.
This abstract principle must now be examined and explained in
concrete detail.
2
For eample, we may use hypotheses that contradia well-confred theore
andor well-established eerimetal results. We may aance sciece b
proceeding counterinduaie/.
Examining the principle in concrete detail means tacing the
consequences of 'counterrules' which oppose familiar rules of the
scientfc enterrise. To see how this works, let us consider the rule
that it is 'experience', or the 'facts', or 'experimental results' which
measure the success of our theories, that agreement between a theor
and the 'data' favours the theor (or leaves the situaton unchanged)
while disagreement endangers it, and perhaps even forces us to
eliminate it. This rule is an important part of all theories of
confrmaton and corroboraton. It is the essence of empiricism. The
'counterrule' corresponding to it advises us to introduce and
elaborate hyotheses which are inconsistent with well-established
theories andor well-established facts. It advises us to proceed
counterinduaie/.
The counterinductve procedure gives rise to the following
questons: Is counterinducton more reasonable than inducton? Are
there circumstances favouring its use? What are the arguments for it?
What are the arguments against it? Is perhaps inducton always
preferable to counterinducton? And so on.
These questons will be answered in two steps. I shall frst examine
the counterrule that urges us to develop hyotheses inconsistent with
accepted and highly confirmed theores. Later on I shall examine the
counterrule that urges us to develop hyotheses inconsistent with
well-establishedfaas. The results may be summarized as follows.
In the first case it emerges that the evidence that might refute a
theor can often be unearthed only with the help of an incompatble
alteratve: the advice (which goes back to Newton and which is stll
ver popular today) to use alteratves only when refutatons have
already discredited the orthodox theor puts the cart before the
20
TWO 21
horse. Also, some of the most important formal propertes of a theor
are found by contrast, and not by analysis. A scientst who wishes to
maximize the empircal content of the views he holds and who wants
to
under
stand them as clearly as he possibly can must therefore
introduce other views; that is, he must adopt a pluralistic methodlog.
He must compare ideas with other ideas rather than with
'experence' and he must tr to improve rather than discard the views
that have failed in the competton. Proceeding in this way he will
retain the theories of man and cosmos that are found in Genesis, or i
the Pimander, he will elaborate them and use them to measure the
success of evoluton and other 'moder' views. He may then discover
that the theor of evoluton is not as good as is generally assumed and
that it must be supplemented, or entrely replaced, by an improved
version of Genesis. Knowledge so conceived is not a series of self
consistent theories that converges towards an ideal view; it is not a
gradual approach to the tuth. It is rather an ever increasing ocean of
mutuall incompatible alteatie, each single theor, each fair-tale,
each myth that is part of the collecton forcing the others into greater
artculaton and all of them contbutng, via this process of
competton, to the development of our consciousness. Nothing is
ever settled, no view can ever be omitted from a comprehensive
account. Plutarch or Diogenes Laertus, and not Dirac or von
Neumann, are the models for presentng a knowledge of this kind in
which the histor of a science becomes an inseparable part of the
science itself- it is essental for its further delomet as well as for
giving contet to the theores it contains at any partcular moment.
Experts and laymen, professionals and dilettanti, truth-freaks and
liars - they all are invited to partcipate in the contest and to make
their contributon to the enrichment of our culture. The task of the
scientist, however, is no longer 'to search for the truth', or 'to praise
god', or 'to systematize obseratons', or 'to improve predictons'.
These are but side effects of an actvity to which his attenton
is now mainly directed and which is 'to make the weaker case the
stron
ger' as the sophists said, and thereb to sustain the motion ofthe
whole.
The
second 'counterrule' which favours hypotheses inconsistent
with obserations, facts and eermetal results, needs no special
defence, for there is not a single interestng theor that agrees with all
the
know
n facts in its domain. The queston is, therefore, not
wh
eth
er
counterinductve theories should be admitted into science;
th
e
ques
ton is, rather, whether the existing discrepancies between
the
or
and
fact should be increased, or diminished, or what else
sh
oul
d be
done with them.
22 AGAI NST METHOD
To answer this queston it suffces t o remember that obseratonal
reports, experimental results, 'factual' statements, either contain
theoretcal assumptons or aser them by the manner in which they
are used. (For this point cf. the discussion of natural interpretatons
in Chapters
6
f.) Thus our habit of saying 'the table is brown' when
we view it under normal circumstances, with our senses in good
order, but 'the table seems to be brown' when either the lightng
conditons are poor or when we feel unsure in our capacity of
obseraton expresses the belief that there are familiar cir
cumstances when our senses are capable of seeing the world 'as it
really is' and other, equally familiar circumstances, when they are
deceived. It expresses the belief that some of our sensor
impressions are veridical while others are not. We also take it for
granted that the material medium between the object and us exerts
no distortng influence, and that the physical entty that establishes
the contact - light - carries a tue picture. All these are abstact,
and highly doubtful, assumptons which shape our view of the world
without being accessible to a direct critcism. Usually, we are not
even aware of them and we recognize their efects only when we
encounter a entrely diferent cosmology: prejudices are found by
contast, not by analysis. The material which the scietist has at his
disposal, his most sublime theories and his most sophistcated
techniques included, is stuctured in exactly the same way. It again
contains principles which are not known and which, if known,
would be extemely hard to test. (As a result, a theor may clash
with the evidence not because it is not correct, but because the
evidence is contaminated.)
Now-how can we possibly examine sometg we are using all the
te? How can we analyse the terms in which we habitually express
our most simple and staightforard obseratons, and reveal their
presuppositons? How can we discover the kind of world we
presuppose when proceeding as we do?
The answer is clear: we cannot discover it from the insid. We need
an eteal standard of critcism, we need a set of alteratve
assumptons or, as these assumptons will be quite general,
consttutng, as it were, an entre alteratve world, we need a dream
world in ordr to discue the fature ofthe real world we think we inhabit
(and which may actually be just another dream-world). The frst step
in our critcism of familiar concepts and procedures, the frst step in
our critcism of 'facts', must therefore be an attempt to break the
circle. We must invent a new conceptual system that suspends, or
clashes with, the most carefully established obseratonal results,
confounds the most plausible theoretcal principles, and introduces
TWO 23
per
ceptons that cannot form par of the existng perceptual world. 1
This step is again counterinductve. Counterinducton is, therefore,
always reasonable and it has always a chance of success.
In the following seven chapters, this conclusion will be developed
in greater detail and it will be elucidated with the help of historical
examples. One might therefore get the impression that I recommend
a new methodology which replaces inducton by counterinducton
and uses a multplicit of theories, metaphysical views, fair-tales
instead of the customar pair theor/ obseraton.
2
This impression
would cerainly be mistaken. My intenton is not to replace one set of
general rules by another such set: my intenton is, rather, to convince
the reader that al methodloge, ee the most obious one, hae their
limits. The best way to show this is to demonstate the limits and even
the irratonalit of some rules which she, or he, is likely to regard as
basic. In the case of inducton (including inducton by falsifcaton)
this means demonstatng how well the counterinductve procedure
can be suppored by argument. Always remember that the
demonstratons and the rhetorics used do not express any 'deep
convictons' of mine. They merely show how easy it is to lead people
by the nose in a ratonal way. A anarchist is like an undercover agent
who plays the game of Reason in order to undercut the authorit of
Reason (Truth, Honest,Justce, and so on).
3
I . 'Clashes' or 'suspends' is meant to be more general than 'contadicts'. I shall say
that a set ofideas or actons 'clashes' with a conceptual system ifit is either inconsistent
with it, or makes the system appear absurd. For details cf. Chapter 16 below.
2. This is how Professor Eran McMullin interpreted some earlier papers of
mine.
See 'A Taxonomy of the Relatons beteen Histor and Philosophy of Science',
Mioo
ceta5tadic,Vol. 5, Minneapolis, 1 971 .
3.
'Dada', says Hans Richter in D=: Zuaad Zati-Zu, 'not only had no
pro
gra
mme, it was against all programmes.' This dos not exclude the skilful defence
of pro
gra
mmes to show the chimerical character of any defence, however 'ratonal'. (In
e
sam
e way an actor or a playwright could prouce al the outer manifestatons of
dee
p lo
ve' in order to debunk the idea of'deep love' itelf. Example: Pirandello.)
3
Te consistec condition which dmand that ne hypotheses agee with
aceted theories is unreasonable because it preee the oldr theor, and
not the better theor. Hypothee contradicing well-confred theores ge
us eidnce that cannot be obtained in any other way. Prolieration of
theore is beecial fr sciece, while unirit impair its crtical power.
Uniorit also ednger the fee delomet ofthe indiidual.
In this chapter I shall present more detailed arguments for the
'counterrule' that urges us to introduce hypotheses which are
inconsitet with well-established theores. The arguments will be
indirect. They will start with a critcism of the demand that new
hypotheses must be consistent with such theories. This demand will
be called the consistec cndition. 1
Prma face, the case of consistency conditon can be dealt with in a
few words. It is well known (and has also been shown in detail by
Duhem) that Newton's mechanics is inconsistent with Galileo's law
offree fall and with Kepler's laws; that statstcal thermodynamics is
inconsistent with the second law of the phenomenological theor
that wave optcs is inconsistent with geometical optcs; and so on.
1
Note that what is being asserted here is logical inconsistency; it may
well be that the differences of predicton are too small to be detected
by experiment. Note also that what is being asserted is not the
inconsistec of, say, Newton's theor and Galileo's law, but rather the
inconsistency of some consequeces of Newton's theor in the domain
of validity of Galileo's law, and Galileo's law. In the last case, the
situaton is especially clear. Galileo's law asserts that the acceleraton
of free fall is a constant, whereas applicaton of Newton's theor to
the surface of the earth gives an acceleraton that is not constant but
I. The consistency condition goes back to Aristotle at least. It plays an imprtant
pan in Newton's philosophy (though Newton himself constantly violated it). It is taken
for granted by many 20th-centur scientists and philosophers of science.
2. Piere Duhem, 1hcimaod5tmuarc e/Fhsica/ 1hceq, New York, 1 962,
pp. 180f.
24
THREE 25
dcre
ases (although imperceptibly) with the distance from the centre
of the
earth.
To speak more abstactly: consider a theory T' that successfully
describes the situation inside domain D' . T' agrees with a fnite
numb
er of obseratons (let their class be F) and it agrees with these
obseratons inside a margin M of error. Any alteratve that
contradicts T' outside F and inside M is supported by exactly the
same obseratons and is therefore acceptable if T' was acceptable (I
shall assume that Fare the only obseratons made). The consistency
conditon is much less tolerant. It eliminates a theory or a hypothesis
not because it disagrees with the facts; it eliminates it because it
disagrees with another theory, with a theory, moreover, whose
confrming instances it shares. It thereby makes the as yet untested
part of that theory a measure of validity. The only diference between
such a measure and a more recent theory is age and familiarity. Had
the younger theory been there frst, then the consistency conditon
would have worked in its favour. 'The frt adequate theory has the
right of priority over equally adequate aftercomers.'
3
In this respect
the effect of the consistency conditon is rather similar to the efect of
the more traditonal methods of transcendental deducton, analysis
of essences, phenomenological analysis, linguistc analysis. It
contibutes to the preseraton of the old and familiar not because of
any inherent advantage in it but because it is old and familiar. This is
not the only instance where on closer inspecton a rather surrising
similarity emerges between moder empiricism and some of the
school philosophies it attacks.
Now it seems to me that these brief consideratons, although
leading to an interestng taaical critcism of the consistency
conditon, and to some frst shreds of support for counterinducton,
do not yet go to the heart of the matter. They show that an alteratve
to the accepted point of view which shares its confrming instances
cannot be eliminated by factual reasoning. They do not show that such
an alteratve is accetable; and even less do they show that it should be
used. It is bad enough, a defender of the consistency conditon might
point
out, that the accepted view does not possess full empirical
support. Adding new theories of an equall unsatisfaaor charaaewill
not improve the situaton; nor is there much sense in trying to relace
the accepted theories by some of their possible alteratves. Such
replacement will be no easy matter. A new formalism may have to be
lea
red
and familiar
p
roblems may have to be calculated in a new
3. C. Tresdell, 'A Program Toward Rediscovering the Rational Mechanics of
the
Age
of Reason', Zrchac/rthcHutee/Exm5cicc,Vol. 1 , p. 14.
26 AGAI NST METHOD
way. Textbooks must be rewritten, university curricula readjusted,
experimental results reinterpreted. And what will be the result of all
the efort? Another theory which from an empirical standpoint has no
advantage whatsoever over and above the theory it replaces. The only
real improvement, so the defender of the consistency conditon will
contnue, derives from the adition ofnew facs. Such new facts will
either support the current theories, or they will force us to modif
them by indicatng precisely where they go wrong. In both cases they
will precipitate real progress and not merely arbitary change. The
proper proedure must therefore consist in the confrontaton of the
accepted point of view with as many relevant facts as possible. The
exclusion of alteratves is then simply a measure of expediency: their
inventon not only does not help, it even hinders progress by
absorbing tme and manpower that could be devoted to better things.
The consistency conditon eliminates such fruitless discussion and it
forces the scientst to concentrate on the facts which, after all, are the
only acceptable judges of a theory. This is how the practsing scientst
will defend his concentraton on a single theory to the exclusion of
empirically possible alteratves.
It is worthwhile repeatng the reasonable core of this argument.
Theories should not be changed unless there are pressing reasons for
doing so. The only pressing reason for changing a theory is
disagreement with facts. Discussion of incompatble facts will
therefore lead to progress. Discussion of incompatble hypotheses
will not. Hence, it is sound procedure to increase the number of
relevant facts. It is not sound procedure to increase the number of
factually adequate, but incompatble, alteratves. One might wish to
add that formal improvements such as increased elegance, simplicity,
generality, and coherence should not be excluded. But once these
improvements have been carried out, the collecton of facts for the
purpose of tests seems indeed to be the only thing left to the scientst.
And so it is - provided facts eist, and are tailable indedtl of
whether or not one consid alteatie to the theor to be teted. This
assumpton, on which the validity of the foregoing argument depends
in a most decisive manner, I shall call the assumpton of the relatve
autonomy of facts, or the autonomy prnciple. It is not assered by this
principle that the discovery and descripton of facts is independent of
al theorizing. But it is asserted that the facts which belong to the
empirical content of some theory are available whether or not one
considers alteratves to thi theory. I am not aware that this very
important assumpton has ever been explicitly formulated as a
separate postulate of the empirical method. However, it is clearly
implied in almost all investgatons which deal with questons of
THREE 27
confrmation and test. All these investgatons use a model in which a
sing
le
theor is compared with a class of facts (or obseraton
statements) which are assumed to be 'given ' somehow. I submit that
this is much too simple a picture of the actual situaton. Facts and
theories are much more intmately connected than is admitted by the
autonomy principle. Not only is the descripton of ever single fact
dependent on some theor (which may, of course, be ver diferent
from the theor to be tested), but there also exst facts which cannot
be unearthed except with the help of alteratves to the theor to be
tested, and which become unavailable as soon as such alteratves are
excluded. This suggests that the methodological unit to which we
must refer when discussing questons of test and empirical content is
consttuted by a whole set ofparl uerlapping, fatuall aquate, but
mutuall inconsistent theorie. In the present chapter only the barest
outlines will be given of such a test model. However, before doing
this, I want to discuss an example which shows ver clearly the
functon of alteratves in the discover of critcal facts.
It is now known that the Brownian partcle is a perpetual moton
machine of the second knd and that its exstence refutes the
phenomenological second law. Brownian moton therefore belongs
to the domain of relevant facts for the law. Now could this relaton
between Brownian moton and the law have been discovered in a
direa manner i.e. could it have been discovered by an examinaton of
the obseratonal consequences of the phenomenological theor that
did not make use of an alteratve theor of heat? This queston is
readily divided into two: ( I ) Could the releance of the Brownian
partcle have been discovered in this manner? (2) Could it have been
demonstated that it actually reute the second law?
The answer to the fi rst queston is that we do not know. It is
impossible to say what would have happened if the knetc theor had
not been introduced into the debate. It is my guess, however, that in
that case the Brownian partcle would have been regarded as an
oddity - in much the same way as some of the late Professor
Ehrenhaft's astounding efects were regarded as an oddity, and that it
would not have been given the decisive positon it assumed in
contemporar theor. The answer to the second queston is simply
No.
Consider what the discover of an inconsistency between the
phenomenon of Brownian moton and the second law would have
req
uired. It would have required: (a) measurement of the exact
motion of the partcle in order to ascertain the change in its knetc
energy plus the energy spent on overcoming the resistance of the
uid; and (b) precise measurements of temperature and heat transfer
1D
the
surrounding medium in order to establish that any loss
28 AGAI NST METHOD
occurring there was indeed compensated by the increase in the
energy of the moving partcle and the work done against the fluid.
Such measurements are beyond experimental possibilites; 4 neither
the heat transfer nor the path of the partcle can be measured with the
desired precision. Hence a 'direct' refutaton of the second law that
considers only the phenomenological theor and the 'facts' of the
Brownian moton is impossible. It is impossible because of the
structure of the world in which we live and because of the laws that
are valid in this world. And as is well known, the actual refutaton
was brought about in a ver diferent manner. It was brought about
via the kinetc theor and Einstein's utlizaton of it in his calculaton
of the statstcal propertes of Brownian moton. In the course of
this proedure, the phenomenological theor (T') was incororated
into the wider context of statstcal physics (T) in such a manner
that the consistec condition wa violated, and it was only the that
crucial experiments were staged (investgatons of Svedberg and
Perrin).5
4. For details cf. R. Fiinh, ZsFhsik,Vol. 81, 1933, pp. 1 43f.
5. For these investgations (whose philosophical background derives from
Boltmann) cf. A. Einstein, hctigatieaseathc 1hceqe/thcmmaiaaMetiea,ed. R.
Fiinh, New York, 1 956, which contains all the relevant paper by Einstein and an ex
haustive bibliography by R. Fiinh. For the experimental work of J. Perrin, see Dic
Z temc,Leipzig, 1920. For the relation between the phenomenological theor and the
kinetic theor of von Smoluchowski, see 'Experimentell nachweisbare, der iiblichen
Thenodynamik widersprechende Molekularphanomene', FhsikalischcZs. , Vol. 8,
1 91 2, p. 1069, as well as the brief note by K.R. Popper, 'Ireversibility, or, Entopy
since 1 905', ntsheamal/rthcFhiles@he/5cicc, Vol. 8, 1 957, p. 1 51 , which
summarizes the essental arguments. Despite Einstein's eph-making discoveries
and von Smoluchowski's splendid presentation of their consequences (Oom d
Manc5melachemski, Cracow, 1 927, Vol. 2, pp. 226f, 31 6f, 462f and 530f, the
present situaton in thenodynamics is extremely unclear, especially in view of the
contnued presence of some ver doubtful ideas of reducton. To be more specifc, the
attempt is frequently made to detenine the entropy balance of a complex statistiml
proess by reference to the (refuted) phemelegicalJaw after which fuctuations are
insered in an a hecfashion. For this cf. my note 'On the Possibility of a Perpetuum
Mobile of the Second Kind, Miadg MattcraodMcthed,Minneaplis, 1 966, p. 49, and
my paper 'In Defence of Classical Physics', 5tadic io thcHisteq aadFhiles@h e/
5cicc, I , No. 2, 1970.
It ought to be mentioned, incidentally, that in 1 903, when Einstein stared his work
in thenodynamics, there existed empirical evidence suggesting that Brownian moton
could not be a molecular phenomenon. See F.M. Exner, 'Notiz zu Browns
Molekularbewegung',Ann. Fhs. , No. 2, 190, p. 843. Exner claimed that the moton
was of orders of magnitude beneath the value to be expected on the equipartion
principle. Einstein (hctigatieasiothc1hceqe/thcmmiaaMmt,pp. 63f, esp.
p. 6 7) gave the following theoretcal explanation of the discrepancy: 'since an obserer
operating with defnite means of obseration in a defnite manner can never perceive
the actual path transversed in an arbitarily small time, a cerain mean veloity will
THREE 29
It seems to me that this example is typical of the relaton between
fairly general theories, or points of view, and the 'facts'. Both the
relevance and the refutng character of decisive facts can be
established only with the help of other theories which, though
factually adequate,6 are not in agreement with the view to be tested.
This being the case, the inventon and artculation of alteratves may
have to precede the producton of refutng facts. Empiricism, at least
in some of its more sophistcated versions, demands that the
empirical content of whatever knowledge we possess be increased as
much as possible. Hence the invention ofalteaties to the vie at the
cntre ofdiscssion constitute an essential par ofthe epircal method.
Conversely the fact that the consistency conditon eliminates
alteratves now shows it to be in disagreement not only with
scientfic practce but with empiricism as well. By excluding valuable
tests it decreases the empirical content of the theories that are
permitted to remain (and these, as I have indicated above, will usually
be the theories which were there frst); and it especially decreases the
number of those facts that could show their limitatons. This is how
empiricists (such as Newton, or some proponents of what has been
called the orthodox interretaton of quantum mechanics) who
defend the consistency conditon, being unaware of the complex
nature of scientfc knowledge (and, for that matter, of any form of
knowledge) are voiding their favourite theories of empirical content
and thus turing them into what they most despise, v. metaphysical
doctrines.7
John Stuar Mill has given a fascinatng account of the gradual
transformaton of revolutonar ideas into obstacles to thought.
When a new view is proposed it faces a hostle audience and excellent
reasons are needed to gain for it an even moderately fair hearing. The
reasons are produced, but they are often disregarded or laughed out
of cour, and unhappiness is the fate of the bold inventors. But new
generatons, being interested in new things, become curious; they
consider the reasons, pursue them further and groups of researchers
initate detailed studies. The studies may lead to surrising successes
(they also raise lots of diffcultes). Now nothing succeeds like
always appear to him as an instantaneous veloity. But it is clear that the veloity

scenained thus coresponds to no objectve propery of the moton under


Inv
estiga
tion.' Cf. also Mary Jo Nye, MelcmlarRcali,London, 1972, pp. 98f.
6.
The
condition offactual adequacy will be removed in Chapter 5.
7.
The most dramatic confiration of the onhodox view which made its empirical
nature obvious came by way of Bell's theorem. But Bell was on the side of Einstein, not
of Bohr, whom he regarded as an 'obscurantist'. Cf. Jeremy Berstein, Qaaatam
Fre//c,Princeton, 1991 , pp. 3f (for Bell's background) and p. 84 (for 'obscurantst').
30 AGAI NST METHOD
success, even if it is success surrounded by difcultes. The theor
becomes acceptable as a topic for discussion; it is presented at
meetngs and large conferences. The diehards of the status quo feel
an obligaton to study one paper or another, to make a few grumbling
comments, and perhaps to join in its exploraton. There comes then a
moment when the theor is no longer an esoteric discussion topic for
advanced seminars and conferences, but enters the public domain.
There are intoductor texts, popularizations; examinaton questons
start dealing with problems to be solved in its terms. Scientsts from
distant felds and philosophers, trying to show of, drop a hint here
and there, and this often quite uninformed desire to be on the right
side is taken as a further sign of the importance of the theor.
Unfortunately, this increase in importance is not accompanied by
better understanding; the ver opposite is the case. Problematc
aspects which were orignally introduced with the help of carefully
constucted arguments now become basic principles; doubtful points
tum into slogans; debates with opponents become standardized and
also quite unrealistc, for the opponents, having to express
themselves in terms which presuppose what they contest, seem to
raise quibbles, or to misuse words. Alteratves are stll employed but
they no longer contain realistc counter-proposals; they only sere as
a background for the splendour of the new theor. Thus we do have
success - but it is the success of a manoeuvre carried out in a void,
overcoming diffcultes that were set up in advance for easy soluton.
A empirical theor such as quantum mechanics or a pseudo
empirical practce such as modem scientfic medicine with its
materialistc background can of course point to numerous achieve
ments but any view and any practce that has been around for some
tme has achievements. The question is whose achievements are
better or more important and this queston cannot be answered for
there are no realistc alteratves to provide a point of comparison. A
wonderful inventon has tured into a fossil.
There exist numerous historical examples of the process I have
just described and various authors have commented on it. The most
important recent author is Professor Thomas Kuhn. In his Structure
ofScietic Reolutions, 8 he distinguishes between science and pre
science and, within science, between revolutons and normal science.
Pre-science, according to him, is pluralistc throughout and
therefore in danger of concentatng on opinions rather than on
things (Bacon made a similar point). The two components of mature
8. Chicago, 1 962.
THREE 31
science perfectly agree with the two stages mentoned above except
that Kuhn doubts that science or, for that matter, any actvity that
claims to produce factual kowledge can do without a nonnal
component. Fossils, he seems to say, are needed to give substance to
the debates that occur in the revolutonar component - but he adds
that the latter cannot advance without alteratves. Two earlier
authors are Mill and Niels Bohr. Mill gives a clear and compelling
descripton of the tansiton from the early stage of a new view to its
orthodoxy. Debates and reasoning, he writes, are features
belonging to periods of transiton, when old notons and feelings have
been unsettled and no new dotrines have yet succeeded to their
ascendanc. At such tmes people of any mental actvity, having given
up their old beliefs, and not feeling quite sure that those they stll
retain can stand unmodifed listen eagerly to new opinions. But this
state of things is necessarily transitor: some partcular body of
dotrine in tme raUies the majority round it, organizes soial
insttutons and modes of acton conforably to itself, educaton
impresses this new creed upon the new generaton without the metal
prcese that hae led to it and by degrees it acquires the ver same
power of compression, so long exercised by the creeds of which it had
taken place.
9
A account of the alteratves replaced, of the proess of
replacement, of the arguments used in its course, of the stength of
the old views and the weakesses of the new, not a 'systematc
account' but a historcl acunt ofeah stage ofknowledge, can alleviate
these drawbacks and increase the ratonality of one's theoretcal
commitents. Bohr's presentaton of new discoveries has precisely
this patter; it contains preliminar summaries sureying the past,
moves on to the 'present state of kowledge' and ends up by making
general suggestons for the future.
10
Mill's views and Bohr's procedure are not only an expression of
their liberal atttude; they also refect their convicton that a pluralism
of ideas and fonns of life is an essental par of any ratonal inquir
concering the nature of things. Or, to speak more generally:
Unanimit ofoinion may be ftting fr a rgid church, fr the frghteed or
geedy viaims ofsome (anciet, or mod) myth, or fr the weak and
9. 'Autobiography', quoted from Esstial KerH e/eha 5taanMill, ed. M.
Ler
er, New York, 1 965, p. 1 1 9; my emphasis.
1 0.
For
a more detailed account cf. my Fhiles@himlFapc, Vol. I , Chapter 1 6,
se
ction
6.
32 AGAI NST METHOD
willing fllowe ofsome trant. Varet ofoinion i necessar fr objecie
knowledge. And a method that ecourages varet is also the onl method
that is compatible with a humanitaran outlook. (To the extent to which
the consistency conditon delimits variety, it contains a theologcal
element which lies, of course, in the worship of 'facts' so
characteristc of nearly all empiricism. 1 1 )
I I . . It is interesting to see that the platitudes that directed the Protestants to the
Bible are often almost identical with the platitudes which direct empiricists and other
fundamentalists to thcir foundation, viz. experience. Thus in his Nmam Oqaoam
Bacon demands that all preconceived notions (aphorism 36), opinions (aphorisms
42ft), even merd (aphorisms 59, 121 ), 'be adjured and renounced with fr and
solemn resolution, and the understanding must be completely freed and cleared of
them, so that the access to the kingdom of man, which is founded on the sciences, may
resemble that to the kingdom of heaven, where no admission is conceded, except to
children' (aphorism 68). In both cases 'disputation' (which is the consideraton of
alteratives) is critcized, in both cases we are invited to dispense with it, and in both
cases we are promised an 'immediate perception', here, of God, and there of Nature.
For the theoretcal background of this similarity cf. my essay 'Classical Empiricism', in
R.E. Butts (ed.), 1hcMcthedlegimlHcritagce/Nmteo,Oxford and Toronto, 1970. For
the strong connections between Puritanism and modem science see R.T. Jones,
ociuaodMeds,Califoria, 1 965, Chapters 5-7. A thorough examinaton of the
factors that influenced the rise of modem empiricism in England is found in R.K.
Menon, 5cicc, 1cchoeleaod5ecicio5ntcthCtaqEoglaod,New York, 1 970
(book version of the 1938 article).
4
There
is no ida, howeer ancient and absurd, that is not capable of
impruing our knowledge. The whole histor ofthought is absorbed into
science and is used fr impruing eer single theor. Nor is political
intererence reeaed. It may be needd to uercome the chauinism ofscience
that resists alteaties to the status quo.
This fnishes the discussion of part one of counterinducton dealing
with the inventon and elaboraton of hypotheses inconsistent
with a point of view that is highly confrmed and generally accepted.
The result was that a thorough examinaton of such a point of
view may involve incompatble alteratives so that the (ewtonian)
advice to postpone alteratves untl after the frst difculty has arisen
means puttng the cart before the horse. A scientst who is inter
ested in maximal empirical content, and who wants to understand
as many aspects of his theory as possible, will adopt a plural
istc methodology, he will compare theories with other theories
rather than with 'experience', 'data', or 'facts', and he will t to
improve rather than discard the views that appear to lose in the
competton. 1 For the alteratves, which he needs to keep the
contest going, may be taken from the past as well. As a matter of fact,
they may be taken from wherever one is able to fnd them -
from ancient myths and modem prejudices; from the lucubratons
of experts and from the fantasies of cranks. The whole history
of a subject is utlized in the attempt to improve its most recent
and most 'advanced' stage. The separaton between the history
of a
science, its philosophy and the science itself dissolves into thin
I . It is, therefore, imponant that the alteratives be set against each other and not
be isolated or emasculated by some for of 'demythologization'. Unlike Tillich,
BItmann and their followers, we should regard the world-views of the Bible, the
G!lgamesh epic, the Iliad, the Edda, as fully fledged altmatncsmelegcwhich can be
used to modif, and even to replace, the 'scientifc' cosmologies of a given period.
33
34 AGAI NST METHOD
air and so does the separaton between science and non-science.
2
This positon, which is a natural consequence of the arguments
presented above, is frequently attacked - not by counter-arguments,
which would be easy to answer, but by rhetorical questons. 'If any
metaphysics goes,' writes Dr Hesse in her revew of an earlier essay
of mine,
3
'then the queston arises why we do not go bak and exploit
2. An account and a tuly humanitarian defence of this psiton can be found inJ.S.
Mill's OLi&.Popper's philosophy, which some people would like to lay on us as the
one and only humanitarian rationalism in existence toay, is but a pale reflecton of Mill.
It is specialized, foralistc and elitist, and devoid of the concer for individual
happiness that is such a characteristic feature of Mill. We can undertand its
peculiarites when we consider (a) the background oflogical positivism, which plays an
imporant role in the Legice/5citcDucq,(b) the unrelentng puritanism of its
author (and of most of his follower), and when we remember the influence of Harriet
Taylor on Mill's life and on his philosophy. There is no Harriet Taylor in Popper's life.
The foregoing arguments should also have made it clear that I regard proliferaton not
just as an 'exteral catalyst' of progress, as Lakatos suggests in his essays ('History of
Science and Its Ratonal Reconstuctons', esteo5tadic, Vol. 8, p. 98; 'Popper on
Demarcaton and Inducton', MS, 1970, p. 21 ), but as an essental par ofit. Ever since
' Explanaton, Reducton, and Empiricism' (Miooceta 5tadic, Vol. 3, Minneapolis,
1 962), and especially in' How to Be a Good Empiricist' (Dcmmarc5tadic,Vol. 2, 1 963), I
have argued that alteratves increase the empirical content of the views that happen to
stand in the centre of atenton and are, therefore, 'occcsaqpars' of the falsifing
proess (Lakatos, ' History', fn. 27, describing his own positon). In 'Reply to Critcism'
(esteo5tadm,Vol. 2, 1 965) I pointed out that 'the principle of proliferaton not only
recommends inventon of omalteratves, it also prevents the eliminaton of eld
theories which have been refuted. The reason is that such theories contibute to the
content of their victorious rivals' (p. 224). This agrees with Lakatos' obseratonofl971
that 'alteratves are not merely catalysts, which can later be removed in the ratonal
reconstucton' ('History', fn. 27), ocqtthat Lakatos attbutes the psychologistc view
to me and my mmviews to himself. Considering the argment in the text, it is clear that
the increasing separaton of the history, the philosophy of science and of science itselfis a
disadvantage and should be tenninated in the interest of all these three disiplines.
Otherise we shall get tons of minute, precise, but uterly baren results.
3. Mary Hesse, Ratie, No. 9, 1 967, p. 93; cf. B.F. Skinner, qeodFrccd aod
Digi,New York, 1 971 , p. 5: 'No moem physicist would tum to Aristode for help.'
This is neither te, nor would it be an advantage if it were te. Aristotelian ideas
influenced research long after they had allegedly been removed by early modem
astonomy and physics - any history of 1 7th- or 1 8th-century science wshow that
(example: John Heilbronn's marellous Elcunci io thc I7th aod Ith Ctanc,
Berkeley and Los Angeles, 1 979). They resurfaced in biolog, in the theroynamics
of open systems and even in mathematcs. Aristode's theory ofloomoton (which has
the consequence that a moving object has no precise length and that an object having a
precise loaton must be at rest) was more advanced than the Galilean view and
showed that ideas which in our tme emerged from empirical research can be obtained
by a careful analysis of the problems of the contnuum (details on this point in Chapter
8 of my FarmchteRcmeo,London, 1 987). Here as elsewhere the propagandist of a
naive scientsm give themselves the air of presenting argument when all they do is
spread unexamined and ill-conceived rmour.
FOUR 35
the objectve critcism of moder science available in Aristotelianism,
or indeed in Voodo?' -and she insinuates that a critcism of this kind
wo
uld
be altogether laughable. Her insinuaton, unfortunately,
assumes a great deal of ignorance in her readers. Progress was often
achieved by a 'crtcism from the past', of precisely the kind that is
now dismissed by her. After Aristotle and Ptolemy, the idea that the
earth
moves - that stange, ancient, and 'entrely ridiculous',
'
Pythagorean view -was thrown on the rubbish heap of histor, only
to be
revived by Copercus and to be forged by him into a weapon for
the defeat of its defeaters. The Hermetc writngs played an
important Qart in this revival, which is stll not sufciently
understood, 5 and they were studied with care by the great Newton
himself.
6
Such developments are not surprising. No idea is ever
examined in all its ramifcatons and no view is ever given all the
chances it deseres. Theories are abandoned and superseded by
more fashionable accounts long before they have had an opporunity
to show their virues. Besides, ancient doctines and 'primitve'
myths appear stange and nonsensical only because the informaton
they contain is either not known, or is distorted by philologists or
anthropologists unfamiliar with the simplest physical, medical or
astonomical knowledge. 7 Voodoo, Dr Hesse's piece d reistance, is a
4. Ptolemy, 5atm, quoted afer the translaton of Manitus, Dc Cmadim
FtelmsHaadachdrZstmemic,Vol. I , Leipzig, 1 963, p. 1 8.
5. For a positve evaluaton of the role of the heretc writngs during the
Renaissance cf. F. Yates, CierdaemaeaadthcHmctic1rmitiea,London, 1963,
and the literature given there. For a critcism of her psiton cf. the artcles by Mar
Hesse and Edward Rosen in Vol. 5 of the Miaaceta5tadic/rthcFhiles@e/5o,
ed. Roger Stuewer, Minnesota, 1970; R.S. Westan andj.E. McGuire, Hmctimm
aadthc5otcRnlatiea, Los Angeles, Clark Memorial Librar, 1 977, a well as
Bran Vickers, eame/MMHuteq,51 , 1 979.
6. Cf. J.M. Keynes, 'Newon the Man', in Essays and Sketches in Biogaphy, New
York, 1956, and, in much greater detail, McGuire and Rattansi, 'Newon and the
"P
i
pes of Pan" ', NetcaadRcrde/thcR@al5eoc,Vol. 21 , No. 2, 1966, pp. 108t.
For more detailed accounts cf. Frank Manuel, 1hcRclipea e/IsmNmtea,Oxford,
197
4, R.S. Westfall's monumental biogaphy, NnoatRct,Cambridge, 1 980, with
liter
ature,
as well as Chapters x and xi of R. Popkin, 1hc 1hird Fer ia
5ntcth-Ctaq1heaght,Leiden and New York, 1 992.
7. For the scientfc content of some myths cf. C. de Santillana, 1hc Onpa e/
5o
tq
c
!heaght,New York, 1961 , especially the Prologe. 'We can see then', writes
de
San
tllana, 'how so many myths, fantatc and arbitar in semblance, of which the
Gre
ek
tale
of the Argonauts is a late ofspring, may provide a terinolog of image
r
ot
ifs,
a kind
of code which is beginning to be broken. It wa meant to allow those who
new
(a) to
deterine unequivoally the positon of given planet in respect to the
r, t

the
frament, and to one another; () to present what knowledge there wa of
e
fabn
c of
the world in the for of tales about 'how the world bgan'. There are two
re
aso
ns
why
this code wa not discovered earlier.
One is the fr conviction of
36 AGAI NST METHOD
case in point. Nobody knows it, everbody uses it as a paradigm of
backwardness and confusion. And yet Voodoo has a frm though stll
not sufficiently understood material basis, and a study of its
manifestatons can be used to enrich, and perhaps even to revise, our
knowledge of physiolog. 8
A even more interestng example is the revival of taditonal
medicine in Communist China. We start with a familiar develop
ment:
9
a great countr with great traditons is subjected to Wester
dominaton and is exploited in the customar way. A new generaton
recognizes or think it recognizes the material and intellectual
superiority of the West and taces it back to science. Science is
imported, taught, and pushes aside all taditonal elements. Scientfc
chauvinism tiumphs: 'What is compatble with science should live,
what is not compatble with science, should die' .
10
'Science' in this
context means not just a specific method, but all the results the
method has so far produced. Things incompatble with the results
must be eliminated. Old style doctors, for example, must either be
historians of science that science did not start before Greece and that scientifc results
can only be obtained with the scientfc method as it is practised today (and as it was
foreshadowed by Greek scientsts). The other reason is the astronomical, geological,
etc., ignorance of most Assyriologists, Aegptologists, Old Testament scholars, and so
on: the apparent primitvism of many myths is just the refection of the primitve
astonomical, biological, etc., etc., knowledge of their collector and translator. Since
the discoveries of Hawkins, Marshack, Seidenberg, van der Waerden (Cmmctqaad
Zlgcbra iaZaotCiriluatieas, New York, 1 983) and others we have to admit the
existence of an interatonal palaeolithic astronomy that gave rise to schols,
obseratories, scientfc taditons and most interesting theories. Thes theories,
which were expressed in siological, not in mathematical, ters, have left their traces
in sagas, myths, legends and may be reconstcted in a twofold way, by goingfTard
into the present from the material remains of Stone Age astonomy such as marked
stones, stone obseratories, etc., and by going bk into the past from the literar
remains which we fnd in sagas, legends, myths. An example of the frt method is A.
Marhack, 1hcRoue/Ciriluatiea,New York, 1 972. An example of the second is de
Santllana-von Dechend, Hamlct :Mid,Boston, 1969.
8. Cf. Chapter 9 of Levi-Strauss, 5tmuaralZathpele, New York, 1 96 7. For
the physiological basis ofVooo cf. C.R. Richter, 'The Phenomenon ofUnexplained
Sudden Death', in Gantt (ed.), 1hcFhsielepmlmue/Fqchiatq, as well as W.H.
Cannon, edi Chaagc ia Faia, Haago, Fcar aadRagc, New York, 1915, and
' "Voodo" Death', |nZmcncaaZathr@elegist,n.s., xliv, 1 942. The detailed biological
and meteorological obseratons made by so-called 'primitives' are repored in
Levi-Strauss, 1hc5magcMiad,London, 1 966.
9. R.C. Croizier, 1raditieaalMcdiciaciaMed Chiaa,Cambridge, Mass., 1 968.
The author gives a ver interesting and fair account of developments with numerous
quotatons from newspaper, book, pamphlets, but is ofen inhibited by his respect for
20th-centur science.
10. Chou Shao, 1 933, as quoted in Croizier, op. cit., p. 109. Cf. also D.W.Y.
Kwok, 5citismiaChiacc1heaght,New Haven, 1 965.
FOUR 37
removed from medical practce, or they must be re-educated. Herbal
medicine, acupuncture, moxibuston and the underlying philosophy
are a thing of the past, no longer to be taken seriously. This was the
atttude up to about 1954 when the condemnaton of bourgeois
elements in the Minist of Health started a campaign for the revival
of taditonal medicine. No doubt the campaign was politcally
inspired. It contained at least two elements, v. (1) the identfcaton
ofWestem science with bourgeois science and (2) the refusal of the
party to exempt science from politcal superision 1 1 and to grant
experts special privileges. But it provided the counterforce that was
needed to overcome the scientfc chauvinism of the tme and to make
a plurality (actually a duality) of views possible. (This is an important
point. It often happens that pars of science become hardened and
intolerant so that proliferaton must be enforced from the outside,
and by politcal means. Of course, success cannot be garanteed -
see the Lysenko afair. But this does not remove the need for non
scientfc contols on science.)
Now this politcally enforced dualism has led to most interestng
and puzzling discoveries both in China and in the West and to the
realizaton that there are efects and means of diagnosis which
modem medicine cannot repeat and for which it has no explanaton.
It revealed sizeable lacunae in Wester medicine. Nor can one expect
that the customary scientfc approach will eventually fnd an answer.
In the case of herbal medicine the approach consists of two steps. 1
2
First, the herbal concoction is analysed into its chemical consttuents.
Then the specifc efects of each constituent are determined and the
total efect on a partcular organ explained on their basis. This
neglects the possibility that the herb, taken in its entrety, changes the
state of the whole organism and that it is this new state of the whole
organism rather than a specifc part of the herbal concocton, a 'magic
bullet', as it were, that cures the diseased organ. Here as elsewhere
knowledge is obtained from a multplicity of views rather than from
the determined applicaton of a preferred ideolog. And we realize
that proliferaton may have to be enforced by non-scientfc agencies
whose power is sufcient to overcome the most powerful scientfc
insttutions. Examples are the Church, the State, a politcal party,
public discontent, or money: the best single entty to get a modem
scientst away from what his 'scientfc conscience' tells him to pursue
is stll the dlar (or, more recently, the Swiss franc).
I I . For the tensions between 'red' and 'exer' cf. F. Schurmann, Idele aad
0aaintieoioCemmaoistChiaa,Berkeley, 1966.
1 2. Cf. M.B. Kreg, CrcMcdiciac,New York, 1964.
38 AGAI NST METHOD
Pluralism of theories and metaphysical views is not only important
for methodolog, it is also an essental part of a humanitarian outlook.
Progressive educators have always tried to develop the individuality
of their pupils and to bring to fruiton the partcular, and sometmes
quite unique, talents and beliefs of a child. Such an educaton,
however, has very often seemed to be a futile exercise in day
dreaming. For is it not necessary to prepare the young for life a it
actuall is? Does this not mean that they must lear one pariclar set of
views to the exclusion of everything else? And, if a trace of their
imaginaton is stll to remain, will it not fnd its proper applicaton in
the arts or in a thin domain of dreams that has but little to do with the
world we live in? Will this procedure not fnally lead to a split between
a hated reality and welcome fantasies, science and the arts, careful
descripton and unrestained self-expression? The argument for
proliferaton shows that this need not happen. It is possible to retain
what one might call the freedom of artstc creaton and to use it to the
full, not just as a road of escape but as a necessary means for
discovering and perhaps even changing the features of the world we
live in. This coincidence of the part (individual man) with the whole
(the world we live in), of the purely subjectve and arbitary with the
objectve and lawful, is one of the most important arguments in
favour of a pluralistc methodology. For detils the reader is advised
to consult Mill's magnifcent essay On Libert.
1
3
1 3. Cf. my account of this essay in Vol. 1 , Chapter 8 and Vol. 2, Chapter 4 of my
Fhiles@hicalFapm.Cf. also Appendix I of the present essay.
5
No theor ee agee with al the facts in its dmain, yet it is not always the
theor that is to blame. Facs are constituted b oldr idologies, and a cash
between facs and theores may be proof ofprgess. It is also a .frt ste in our
attempt to fnd the prnciples implicit in fmiliar obseational notions.
Considering now the inventon, elaboraton and the use of theories
which are inconsistent, not just with other theories, but even with
eeriments, facs, obserations, we may start by pointng out that no
single theor ee agee with all the known facs in its dmain. And the
trouble is not created by rumours, or by the result of sloppy
procedure. It is created by experiments and measurements of the
highest precision and reliability.
It will be convenient, at this place, to distinguish two diferent
kinds of disagreement between theor and fact: numerical disagree
ment, and qualitatve failures.
The first case is quite familiar: a theor makes a certain numerical
predicton and the value that is actually obtained differs from the
predicton made by more than the margin of error. Precision
instruments are usually involved here. Numerical disagreements
abound in science. They give rise to an 'ocean of anomalies' that
surrounds ever single theor. 1
Thus the Coperican view at the tme of Galileo was inconsistent
with
facts so plain and obvious that Galileo had to call it 'surely
false'.
2
'There is no limit to my astonishment,' he writes in a
I . For the 'ocean' and varous ways of dealing with it, cf. my 'Reply to Critcism',
8estenStadic,Vol. 2, 1 965, pp. 224f.
2. Galileo Galilei, 1hcssao, quoted in S. Drake and C.D. O'Malley (eds), 1hc
Contr
m
onthcCemctse/II,London, 1 960, p. 1 85. The 'surely false' refers to the
con
dem
nation by Church authorites. But as will be explained in the course of the book
an
d
espe
cially in Chapter 1 3, the condemnation was based in par on the 'philosophical

bsurdity' of the idea of a moving earth, i.e. on its empirical failures and its theoretical
m

dequacy. See also the next quotation and fotnote. 'As to the system of Ptolemy',
Wtes Galileo on this point (184), 'neither Tycho, nor other astronomers, nor even
39
40 AGAI NST METHOD
later work,3 'when I reflect that Aristarchus and Copericus were
able to make reason so conquer sense that, in defance of the latter,
the former became mistress of their belief.' Newton's theory of
gravitaton was beset, from the very beginning, by difcultes serious
enough to provide material for refutation. 4 Even quite recently and in
the non-relatvistc domain it could be said that there 'exst numerous
discrepancies between obseraton and theory'.
5
Bohr's atomic
model was intoduced, and retained, in the face of precise and
unshakeable contary evidence.
6
The special theory of relatvity was
retained despite Kaufmann's unambiguous results of 1 906, and
despite D.C. Mler's experiment.
7
The general theory of relatvity,
Copericus could clearly refute it, inasmuch as a most imporant argument taken from
the movement of Mar and Venus always sto in their way.' The 'most imprant
argment' and Galileo's resoluton are discussed in Chapter 9 and 10.
3. Galileo Galilei, Dmgac Ceacmiag thc 1m ChiqerU 5stm, Berkeley,
1 953, p. 328.
4. According to Newon the 'mutual actons of comets and planet upn one another'
give rise to 'some inconsiderable ireglarites . . . which will be apt to increase, tll the
system want a reformaton', Qtim,New York, 1 952, p. 402. What Newon means i
that gavitaton disturbs the planet in a way that is likely to blow the planetar system
apan. Babylonian data as used by Ptolemy show that the planetar system has remained
stable for a long tme. Newon concluded that it wa being periodically 'refored' by
divine interentons: Go act as a stabilizing force in the planetar system (and in the
world a a whole, which is constantly losing moton through proesses such as inelastc
collisions). One of the 'iregularites' considered by Newon, the great inequalit of
Jupiter and Satur (Fnaopm,tans!. Motte, ed. Cajori, Berkeley, 1 934, p. 397) was
show by Laplace to be a perioic disturbance with a large priod. Then Poincare found
that the series developments customar in the calculatons ofen diverged afer they had
show some convergence while Brhns discovered that no quanttatve methods oter
than series expansions could resolve then-body problem. This was te end of the purely
quanttatve perio in celestal mechanics (details in J. Moser, Ann e/Mathatim
5tadic,Vol. 77, 1 973, Princeton). See also M. Ryabov ,ZaEltaq5anqe/Cchtml
Mcchaaic,New York, 1 961 , for a surey and quanttatve results of various methods of
calculaton. The qualitatve approach is briefly described on pp. 1 26f. Thus ittok more
than two hundred yea before one of the many difcultes of this rather successfl
theor wa fnally resolved.
5. Brower-Clemence, Mcthede/CclctalMcchaaic,New York, 1 961 . Also R.H.
Dicke, 'Remarks on the Obseratonal Basis of General Relatvit', in H.Y. Chiu and
W.F. Holman (eds), CrmitatieaaadRcmtni,New York, 1 964, pp. 1-1 6. For a more
detailed discussion of some of the difcultes of classical celestal mechanics, cf.
J. Chay, La1hmncdmrcmtnitctmMchaaiqaclctc,Vol. 1 , Chapter 4 and 5, Paris,
1 928.
6. Cf. Max Jammer, 17cCeacqtaalDncl@mte/QmatamMcchaaio,New York,
1 966, secton 22. For an analysis cf. secton 3c/2 of Lakatos, 'Falsifcaton and the
Methodolog of Scientfc Research Programmes', in Lakatos-Musgrave (eds),
CnticismaadthcCmmthe/Kaemlcdgc, Cambridge, 1 970.
7. W. Kaufmann, '
U
ber die Konstitton des Elektons',Ann. Fs.,No. 1 9, 196,
p. 487. Kaufmann stated his conclusion quite unambigously, and in italics: '1hc
FI VE 41
though surrisingly successful in a series of occasionally rather
dramatc tests, 8 had a rough time in areas of celestal mechanics
diferent from the advance of the perihelion of Mercur.
9
In the
sixes the arguments and observatons of Dicke and others seemed
to en
danger even this predicton. The problem is stll unresolved.
1
0
roalts e/thcmcmartsarcaetmpatiblc= iththc /aadmtalmsamptieae/Leru
aad Eiastcia.' Lorent's reacton: ' . . . it seems ver likely that we shall have to
relinquish this idea altogether' (1hmqe/Elcctmas,second edition, p. 2 13). Ehrenfest:
'Kaufmann demonstates that Lorent's deforable electon is rled out by the
measurements' ('Zur Stabilitatsfrage bei den Bucherer-Langevin Elektonen', Fs.
Zs.,Vol. 7, 1 96, p. 302). Poincare's reluctance to accept the 'new mechanics' of
Lorent can be explained, at least in par, by the outcome of Kaufmann's expriment.
Cf 5cicc aadMcthm, New York 1 960, Bok III, Chapter 2, section v, where
Kaufmann's expriment is disussed, the conclusion being that the 'principle of
relatvity . . . cannot have the fundamental imprance one was inclined to ascribe to
it'. Cf. also St Goldberg, 'Poincare's Silence and Einstein's Relativity', ntuheaml
/rthcHisteqe/5ncc,Vol. 5, 1 970, pp. 73f, and the literature given there. Einstein
alone regarded the results as 'improbable because their basic assumptons, from which
the mass of the moving electron is deduced, are not suggested by theoretical systems
which encompass wider complexes of phenomena' (Jahmm dRmiktniO taad
Elcktmitt,Vol. 4, 1 907, p. 349). Miller's work was studied by Lorent for many year,
but he could not fnd the trouble. lt was only in 1 955, twenty-fveyearafer Miller had
fnished his expriments, that a satsfactor account of Miller's results was found. Cf.
R.S. Shankland, 'Converations with Einstein', Zm.eam. Fs. , Vol. 31 , 1 963, pp.
47-57, espcially p. 51 , as well as footnotes 1 9 and 34; cf. also the inconclusive
discussion at the 'Conference on the MichelsonMorley Expriment', Zstpsiml
eamal,Vol. 68, 1 928, pp. 341 ff.
Kaufmann's experiment was analysed by Max Planck and found to be not decisive:
what had stopped Ehren fest, Poincare and Lorent did not stop Planck. Why? My con
jecture is that Planck's fr belief in an objective reality and his assumpton that
Einstein's theor was about such a reality made him a litle more critical. Details in
Chapter 6 ofEiie Zahar, Eiastcia'sRnelatiea,La Salle, Ill., 1 989.
8. Such as the test of the efects of gravity upon light that was carried out in 1 91 9 by
Eddingon and Crommelin and evaluated by Eddingon. For a colourful descripton of
the event and its impact, cf. C.M. Will, asEiastciaRight,New York, 1 986, pp. 75f.
9. Chazy, op. cit., p. 230.
10. Repeating considerations by Newcomb (repred, for example, in Chazy, op.
cit., pp. 20411), Dicke pointed out that an oblateness of the sun would add classical
terms to Mercur's motion and reduce the excess (compared with Newon's theor)
advan
ce of its perihelion. Measurements by Dicke and Goldenberg then found a
diference of 52 km between the equatorial and polar diameter of the sun and a
cor
esponding reduction of three seconds of arc for Mercur - a sizeable deviation
from
the relativistc value. This led to a considerable controversy concering the
accu
racy
of the Dicke-Goldenberg experiment and to an increase in the number of
non-Einsteinian theories of gravitation. Technical details in C.M. Will, 1hmqaad
=cnmt ia Crmitatieaal Fsio, Cambridge, 1 981 , pp. 1 76f, a ppular surey
In
clud
ing
later developments in mEiastciaRight,Chapter 5. Note how a new theor
(Ei
nste
in's theor of gravitation) which is theoretically plausible and well confred
can
be
endangered by exploiting its 'refuted' predecessor and carring out appropriate
e
xp
erim
ents. Cf. also R.H. Dicke, op. cit.
42 AGAI NST METHOD
On the other hand there exist numerous new tests, both inside the
planetar system and outside of it
1 1
that provide confrmatons of a
precision unheard of only twenty years ago and unimagined by
Einstein. In most of these cases we are dealing with quanttatve
problems which can be resolved by discovering a better set of numbe
but which do not force us to make qualitatve adjustents.
1 2
I I . Tests outside the planetar system (cosmology, black holes, pulsar) are
needed to examine alteratves that agree with Einsteinian relatvity inside the solar
system. There now exists a considerable number of such alteratives and special steps
have been taken to classif them and to elucidate their similarites and diferences. Cf.
the introducton to C.M. Will, op. cit.
I 2. The situaton just described shows how silly it would be to approach science
from a naive-falsifcatonist perspectve. Yet this is precisely what some philosopher
have been tring to do. Thus Herbert Feigl (Miaaceta5tadic,Vol. 5, I 971, p. 7) and
Karl Popper (Oqcuirc Kaemlcdgc, p. 78) have tied to tm Einstein into a naive
falsifcatonist. Feigl writes: 'If Einstein relied on "beauty", "harony", "symmety",
"elegance" in constctng . . . his general theor ofrelativity, it must neverheless be
remembered that he also said (in a lectre in Prague in I 920- I was present then as a
ver young stdent): "If the obseratons of the red shift in the specta of massive star
don't come out quanttatively in accordance with the principles of general relatvity,
then my theor will be dust and ashes".' Popper writes: 'Einstein . . . said that if the red
shift efect . . . was not obsered in the case of white dwarfs, his theor of general
relatvity would b refuted.'
Popper gives no source for his stor, and he most likely has it from Feigl. But Feigl's
stor and Popper's repetton conflict with the numerous ocasions where Einstein
emphasizes the 'reason of the mater' ('die Vemunf der Sache') over and abve
'verifcaton oflittle efects' and this not only in casual remarks during a lectre, but in
writng. Cf. the quotaton in fotote 7 above, which deals with difcultes of the
special theor of relatvity and precedes the talk at which F eigl was present. Cf. also the
leter to M. Besso and K. Seelig as quoted in G. Holton, 'Infuences on Einstein's
Early Work', Oqaa, No. 3, 1966, p. 242, and K. Seelig, ZlbcnEiastoa, Zurich,
1 96, p. 271 . In 1 952 Bor wrote to Einstein (emEiastciaLcttm,New York, 1 971 ,
p. 1 9, dealing with Freundlich's analysis of the bending of light near the sun and the
red shif): 'It really loks as if your forula is not quite corect. It loks even wore in
the case of the red shif [the crucial case refered to by Feigl and Popper); this is much
smaller than the theoretcal value towards the cente of the sun's disk, and much larger
at the edges . . . . Could this be a hint of non-linearity?' Einstein Oeter of1 2 May 1 952,
op. cit., p. 192) replied. 'Freundlich . . . does not move me in the slightest. Even if the
deflecton oflight, the perihelia) movement or line shift were unknown, the gravitaton
equatons would stll be convincing because they avoid the inertal system (the
phantom which afects everhing but is not itself afected). It u rcal straagc that
hamaaboagarcae~alda/tethcstmagctaqmts mhilcthqarcalmsiacliacdte
moctimatc mcmariagmracic' (my italics). How is this conflict (between Feigl's
testimony and Einstein's writngs) to be explained? It cannot be explained by a chaagcin
Einstein's atttude. His disrespectful attitde towards obseraton and experiment was
there from the ver beginning, as we have seen. It might be explained either by a
mistake on Feigl's part, or else as another instance of Einstein's 'opporunism' - cf
text to footnote 6 of the Iatmdauiea.
On the last page (p. 91 ) of his
[
bodic5picllc aadallgciacRclatirituthceric,
FI VE 43
The second case, the case of qualitatve failures, is less familiar,
but of much greater interest. In this case a theory is inconsistent
not with a recondite fact, that can be unearthed with the help of
complex equipment and is known to experts only, but with
circ
umstances which are easily notced and which are familiar to
every
one.
The first and, to my mind, the most iportant example of an
inconsistency of this kind is Parmenides' theory of the unchanging
and homogeneous One. This theory illustates a desire that has
propelled the Wester sciences from their incepton up to the present
tme -the desire to find a unity behind the many events that surround
us. Today the unity sought is a theor rich enough to produce all the
accepted facts and laws; at the tme of Parmenides the unity sought
was a substance. Thales had proposed water,
1
3 Heraclitus fire,
Anaxmander a substance which he called the apeiron and which
could produce all four elements without being identcal with a single
one of them. Parmenides gave what seems to be an obvious and
rather tivial answer: the substance that underlies everyhing that is is
Being. But this tivial answer had surprising consequences. For
example, we can assert that (frst principle) Being is and that (second
principle) Not Being is not. Now consider change and assume it to be
fundamental. Then change can only go from Being to Not Being. But
according to the second principle Not Being is not, which means that
there is no fndamental change. Next consider diference and
assume it to be fundamental. Then the diference can only be
between Being and Not Being. But (second principle) Not Being is
not and therefore there exst no diferences in Being - it i a single,
unchanging, contuous blok. Parmenides knew of course that
people, himself included, perceive and accept change and diference;
but as his argument had shown that the perceived proesses could
not be fundamental he had to regard them as merely apparent, or
deceptve. This is indeed what he said - thus antcipatg al those
scientsts who contasted the 'real' world of science with the everyday
world of qualites and emotons, declared the latter to be 'mere
appearance' and tied to base their arguments on 'objectve'
experiments and mathematcs exclusively. He also antcipated a
Brunswick, 1922, Einstein writes: 'If the red shift of the spctal lines caused by the
gravitatonal potental did not exst, then the general theor of relatvit would be
unt
enab
le.' Does this conflict with Einstein's cavalier atttude towards obseraton a
desc
ribed
above? It does not. The passage speaks of thcrcdshnot of eHmatieme/it.
13.
The
following account i higy speculatve. Details in Vols I and 2 ofW.K.C.
Gu
thrie,
Z Hitteq e/CrcckFhiles@h, Cambridge, 1 962 and 1 965, as well a in
Cha
pters
1 , 2 and 3 of my FarmchteRcmea.
44 AGAI NST METHOD
popular interretation of the theor of relatvity which sees all events
and transitons as already prearranged in a four-dimensional
contnuum, the only change beinf the (deceptve) jourey of
consciousness along its world line.
1
Be that as it may, he was the
frst to propose a conseraton law (Being is), to draw a boundar line
between reality and appearance (and thus to create what later
thinkers called a 'theor of knowledge') and to give a more
satsfactor foundaton for contnuity than did 19th- and 20th
centur mathematcians who had to invoke 'intuiton'. Using
Parenides' arguments Aristotle constucted a theor of space and
moton that antcipated some ver deep-lying propertes of quantum
mechanics and evaded the difcultes of the more customar (and
less sophistcated) interretaton of a contnuum as consistng of
indivisible elements.
15
Parmenides' theor clashes with most
modem methodological principles -but this is no reason to disregard
it.
A more specifc example of a theor with qualitatve defects is
Newton's theor of colours. According to this theor, light consists of
rays of diferent refrangibility which can be separated, reunited,
refracted, but which are never changed in their interal consttuton,
and which have a ver small lateral extension in space. Considering
that the surface of mirrors is much rougher than the lateral extension
of the rays, the ray theor is found to be inconsistent with the
exstence of mirror images (a is admitted by Newton himselO: if light
consists of rays, then a mirror should behave like a rough surface, i.e.
it should look to us like a wall. Newton retained his theor,
eliminatng the difculty with the help of an a hoc hypothesis: 'the
reflecton of a ray is efected, not by a single point of the reflectng
body, but by some power of the body which is evenly diffused all over
its surface'.
1
6
I 4. A vivid descripton of the Parmenidean flavour of the theor of relatvity is
given by H. Weyl, Fhiles@he/MathatioaadNataral5cicc, Princeton, I 949, p.
I 1 6. Einstein himself wrote: 'For us who are convinced physicists the distncton
between past, present and future has no other meaning than that of an illusion, though
a tenacious one.' CencpeadaccmccMichclcc e,Paris, I 979, p. 3 I 2. Cf. alsop. 292.
In a word: the events of a human life are 'illusions, though tenacious ones'.
I 5. For Aristode cf. the essay quoted in Chapter 4, footote 3. Modem atempt to
get contnuity out of collectons of indivisible elements are repned in A. Grenbaum,
'A Consistent Concepton of the Extended Linear Contnuum as an Aggregate of
Unextended Elements', Fhilm@ e/5cicc, No. I 9, I 952, pp. 283f. Cf. also W.
Salmon (ed.), Ze:Farxc,New York, I 970.
1 6. Sir Isaac Newon, Qtio,Bok 2, pan 3, propsition 8, New York, 1 952, p.
266. For a discussion of this aspect of Newon's method cf. my essay, 'Classical
Empiricism', Fhiles@himlF@m,Vol. 2, Chapter 2.
FI VE 45
In Newton's case the qualitatve discrepancy between theory and
fact was removed by an a hoc hypothesis. In other cases not even this
very fimsy manoeuvre is used: one retains the theory and tres to fret
its shortcomings. An example is the atttude towards Kepler's rule
according to which an object viewed through a lens is perceived at the
point at which the rays travelling from the lens towards the eye
intersect. 1 7

w

se n obje --
..-"et
The rule implies that an object situated at the focus will be seen
infnitely far away.
'But on the contary,' writes Barrow, Newton's teacher and
predecessor at Cambridge, commentng on this predicton, 18 'we
are assured by experience that [a point situated close to the focus]
appears variously distant, according to the diferent situatons of the
eye . . . . And it does almost never seem further of than it would be if
it were beheld with the naked eye; but, on the contary, it does
sometmes appear much nearer . . . . All which does seem repugnant
to our principles.' 'But for me,' Barrow contnues, 'neither this nor
any other difculty shall have so great an infuence on me, as to make
me renounce that which I know to be manifestly agreeable to reason. '
sen object
in infinity
real
object
at fous
1 7. Johannes Kepler, d Iitcdiea Faralipema, ehaaac Kqlo, Ccammcltc
hcrkc,Vol. 2, Munich, 1 939, p. 72. For a detailed discussion of Kepler's rule and its
influence see Vasco Ronchi, Qti:s:1hc5cicce/Iuiea, New York, 1957, Chapter
43f. Cf. also Chapter 9-1 1 below.
1 8.
Lcctieac XV/ Caatabnm ia 5chelie pablicu habitm ia qaibas Qticemm
Fh

emeagaiamRatieac imctigaatara peatar,London, 1669, p. 1 25. The


p
assag
e is used by Berkeley in his attack on the traditional, 'objectivistic' optics (aEssa
!
em
ards
aNcm1hcee/Iisieo,Works, Vol. I , ed. Frazer, London, 1 901 , pp. 1 37f.
46 AGAI NST METHOD
Barrow metions the qualitatve difcultes, adding that he will not
abandon the theor. This is not the usual procedure. The usual
procedure is to forget the difficultes, never to talk about them, and to
proceed as if the theor were without fault. This atttude is ver
common today.
Thus the classical electodynamics of Maxwell and Lorentz
implies that the moton of a free partcle is self-accelerated.
Considering the self-energy of the electon one obtains divergent
expressions for point-charges while charges of finite extension can be
made to agree with relatvity onl by adding untestable stesses and
pressures inside the electon.
1
The problem reappears in the
quantum theor, though it is here partally covered up by
'renormalizaton'. This procedure consists in crossing out the results
of certain calculatons and replacing them by a descripton of what is
actually obsered. Thus one admits, implicitly, that the theor is in
touble while formulatng it in a manner suggestng that a new
principle has been discovered.
2
0
Small wonder when philosophic
ally unsophistcated authors get the impression that 'all evidence
points with merciless definiteness in the . . . directon . . . [that] all
the processes involving . . . unknown interactons conform to the
fundamental quantum law.'
2
1
A stiking example of qualitatve failure is the status of classical
mechanics and electodynamics after Boltzmann's equipartton
theorem. According to this theorem energ is equally distibuted
over all degrees of freedom of a (mechanical or electodynamic)
system. Both atoms (which had to be elastc to rebound from the walls
of a container and from each other) and the electromagnetc field had
infinitely many degrees of freedom which meant that solids and the
I 9. Cf. W. Heitler, 1hcQaaotam1hceqe/Radiatieo,Oxford, I 95+,p. 3 I .
Z0. Renormalization has in the meantime become the basis of quantum feld
theory and has led to predictions of surprising accuracy (repor with literature in A.
Pais, Iomardeaod, Oxford, I 986).This shows that a point of view which, looked at
from afar, appears to be hopelessly wrong may contain excellent ingredients and that
its excellence may remain unrevealed to those guided by strict methodologcal rules.
Always remember that my examples do not criticize science; they criticize those who
want to subject it to their simpleminded rules by showing the disasters such rules
would create. Each of the examples of footnotes 3~I 7can be used as a basis for case
studies of the kind to be carried out in Chapters 6I Z (Galileo and the Coperican
Revolution). This shows that the case ofGalileo is not 'an exception characterizing the
beginning of the so-called scientfc revoluton' (G. Radnitky, 'Theorienpluralismus
Theorienmonismus', in Diemer Meisenheim (ed.), Do Mcthed- aod 1hcen-
plaralismasiodusscha/, I 97I , p. I 6+)but is picalof scientifc change at all
times.
ZI . Rosenfeld in ObsmatieoaodIotcqrctatieo,London, I 957,p. 4.
FI VE 47
electomagnetc fields should have acted as insatable sinks of energy.
Yet ' [a]s so often in the histor of science, the confict between simple
and generally known facts and current theoretcal ideas was
recoged only slowly'.
22
Another example of modem physics is quite instructve, for it
might have led to an entrely diferent development of our knowledge
concering the microcosm. Ehrenfest proved a theorem according to
which the classical electon theor of Lorentz taken in conjuncton
with the equipartton theorem excludes induced magnetsm.
2
3 The
reasoning is exceedingly simple; according to the equipartton
theorem, the probability of a given moton is proportonal to exp
(
-U/Rn, where Uis the energy of the moton. Now the rate of work
of an electon moving in a constant magnetc field B is, according to
Lorentz, W= Q(E + VxB). V, where Q is the charge of the moving
partcle, V its velocity and E the electic field. This magnitude
reduces to QEV which means that the energ and, therefore, the
probability remains unafected by a magnetc field. (Given the proper
context, this result stongly supports the ideas and experimental
findings of the late Felix Ehrenhaft.)
Occasionally it is impossible to surey all the interestng
consequences, and thus to discover the absurd results of a theor.
This may be due to a deficiency in the exstng mathematcal
methods; it may also be due to the ignorance of those who defend the
theor. Under such circumstances, the most common procedure is to
use an older theor up to a certain point (which is ofen quite
arbitar) and to add the new theor for calculatng refnements.
Seen from a methodological point of view the procedure is a veritable
nightare. Let us explain it using the relatvistc calculaton of the
path of Mercur as an example.
The perihelion of Mercur moves along at a rate of about 560''
per centur. Of this value, 5026" are geometic, having to do wth the
movement of the reference system, while 531" are dynamical, due to
perturbatons in the solar system. Of these perturbatons al but the
22. K. Gottried, V.F. Weisskopf, Ceacqc e/FartitIc Fsic, Vol. I , Oord and
New York, 1 984, p. 6.
23. The difculty was realized by Bohr in his dotoral thesis, cf. Niels Bhr,
Ce/lc
ucd erH, Vol. I, Amsterdam, 1 972, pp. 1 58, 381 . He pinted out that the
veloity changes due to the changes in the exteral feld would equalize afer the field
was established, so that no magnetc efects could arise. Cf. also Heilbron and T.S.
Kuhn, 'The Genesis of the Bohr Atom', Huteriml5tadiciatbcFhsiml5occ,No. I ,
1 9
69,
p. 221 .
The argment i n the text is taken from 1bcFqamaaLcuarc, Vol. 2,
C
alif
oria
and London, 1 965, Chapter 34.6. For a somewhat clearer account cf. R.
Be
cke
r,
1hce
ricdElcktmitt,Leipzig, 1949, p. 132.
48 AGAI NST METHOD
famous 43" are accounted for by classical mechanics. This is how the
situation is usually explained.
The explanaton shows that the premise from which we derive the
43" is not the general theor of relatvity plus suitable inital
conditons. The premise contains classical physics in adition to
whatever relatvistc assumptons are being made. Furthermore, the
relatvistc calculaton, the so-called 'Schwarschild soluton', does
not deal with the planetar system as it exsts in the real world (i.e. our
own asymmetric galaxy); it deals with the entrely fctonal case of a
central symmetcal universe containing a singularity in the middle
and nothing else. What are the reasons for employing such an odd
conjuncton of premises?
The reason, according to the customar reply, is that we are
dealing with approxmatons. The formulae of classical physics do
not appear because relatvity is incomplete. Nor is the centrally
symmetric case used because relatvity does not ofer anything better.
Both schemata flow from the general theor under the special
circumstances realized in our planetar system pruidd we omit
magnitudes to small to be considered. Hence, we are using the
theor of relatvity throughout, and we are using it in an adequate
manner.
Now in the present case, making the required approxmatons
would mean calculatng the full n-body problem relatvistcally
(including long-term resonances between diferent planetar orbits),
omittng magnitudes smaller than the precision of obseraton
reached, and showing that the theor thus curtailed coincides with
classical celestal mechanics as corrected by Schwarzschild. This
procedure has not been used by anyone simply because the
relativistc n-body problem has as yet withstood soluton. When the
argument started, there were not even approxmate solutons for
important problems such as, for example, the problem of stability
(one of the frst great stumbling blocks for Newton's theor). The
classical part of the explanans, therefore, did not occur just for
convenience, it wa absolutel necesar. And the approxmatons made
were not a result of relatvistc calculatons, they were introduced in
order to make relatvity fit the case. One may properly call them a hoc
approximations.
2
4
24. Today the so-called parametrized post-Newtonian foralism satisfes most of
the desiderata outlined in the text (details in C.M. Will, 1hce).My pint is that this
was a later achievement whose absence did not prevent scientists from arguing, aod
a
q
iog mcll, about the new ideas. Theories are not only used a premises for
derivations; they are even more frequently used as a general background for novel
FI VE 49
Ad hoc approximatons abound in modem mathematcal physics.
They play a ver important part in the quantum theor of fields and
they
are an essental ingredient of the correspondence principle. At
the
moment we are not concered with the reasons for this fact, we
are only concered with its consequences: ad hoc approxmatons
conceal, and even eliminate, qualitatve difcultes. They create a
false impression of the excellence of our science. It follows that a
philosopher who wants to study the adequacy of science as a picture
of the world, or who wants to build up a realistc scientfic
methodology, must look at modem science with special care. In most
cases modem science is more opaque, and more deceptve, than its
1 6th- and 1 7th-centur ancestors have ever been.
As a final example of qualitatve difcultes I menton again the
heliocentric theor at the tme of Galileo. I shall soon have occasion
to show that this theor was inadequate both qualitatvely and
quanttatvely, and that it was also philosophically absurd.
To sum up this brief and ver incomplete list: wherever we lok,
whenever we have a little patence and select our evidence in an
unprejudiced manner, we fnd that theories fail adequately to
reproduce certain quantitatie reults, and that they are qualitatiel
incompetet to a surprising degree. Science gives us theories of great
beauty and sophistcaton. Modem science has developed mathe
matcal stuctures which exceed anything that has exsted so far in
coherence generality and empirical success. But in order to achieve
this miracle all the exstng troubles had to be pushed into the relation
between theor and fact, 25 and had to be concealed, by a hoc
hypotheses, a hoc approxmatons and other procedures.
guesses whose foral relaton to the basic assumptons is difcult to aserain. 'I must
. . . confess', writes Descanes in his DisceanceaMcthed(Librar of Liberal Ar, 1 965,
p. 52), 'that the pwer of nature is so ample and so vat, and these principles [the
theoretical principles he had developed for his mechanical univere) so simple and so
general, that I almost never notce any partcular efect such that I do not see right away
that it can [be made to confor to these principles) in many diferent ways; and my
greatest difculty is usually to discover in which of these ways the efect is derived.'
Modem theoretcal physicists fnd themselves in exacdy the same situaton.
25. Von Neumann's work in quantum mechanics is an especially instctve
example of this proedure. In order to arrive at a satsfactor prof of the expansion
theorem in Hilben Space, von Neumann replaced the quasi-intuitve notons ofDirac
(and Bohr) by more complex notons of his own. The theoretical relations between the
new notions are accessible to a more rigorous teatent than the theoretcal relations
between the notions that preceded them ('more rigorous' from the point of view of von
Neumann and his followers). It is diferent with their relaton to experimental
proedures. No measuring instments can be specifed for the great majority of
obserables (Wigner, Zmcnmaeamale/Fhsic, Vol. 31 , 1 963, p. 1 4), and where
specifcation is possible it becomes necessary to modif well-known and unrefuted
so AGAI NST METHOD
This being the case, what shall we make of the methodological
demand that a theory must be judged by experience and must be
rejected if it contadicts accepted basic statements? What atttude
shall we adopt towards the various theories of confrmaton and
corroboraton, which all rest on the assumpton that theories can be
made to agree with the known facts, and which use the amount of
agreement reached as a principle of evaluaton? This demand, these
theories, are now all seen to be quite useless. They are as useless as a
medicine that heals a patent only if he is bacteria-free. In practce
they are never obeyed by anyone. Methodologists may point to the
importance of falsifcatons - but they blithely use falsifed theories,
they may sermonize how important it is to consider all the relevant
evidence, and never menton those big and drastc facts which show
that the theories they admire and accept may be as badly off as the
older theories which they reject. In praaice they slavishly repeat the
most recent pronouncements of the top dogs in physics, though
in doing so they must violate some very basic rules of their trade.
Is it Jossible to proceed in a more reasonable manner? Let us
see!
2
According to Hume, theories cannot be ded from facts. The
demand to admit only those theories which follow from facts leaves
us without any theory. Hence, science a we know it can exst only if
we drop the demand and revise our methodology.
According to our present results, hardly any theory is consistet with
the facts. The demand to admit only those theories which are
consistent with the available and accepted facts again leaves us
without any theory. (I repeat: without any theor, for there is not a
single theory that is not in some trouble or other.) Hence, a science as
we know it can exst only if we drop this demand also and again revise
our methodology, now admitting counternduction in adition to
admitting unsupored hypothee. The right method must not contain
laws in an arbitrary way or else to admit that some quite ordinary problems of quantum
mechanics, such as the scatering problem, do not have a soluton ( .M. Cok, eamal
of MathaticalFhsio,Vol. 36, 1957). Thus the theory becomes a veritable monster of
rigour and precision while its relation to experience is more obscure than ever. It is
interesting to see that similar developments occur in 'primitive thought'. 'The most
striking features of Nupe sand divining', writes S.F. Nader in NapcRcligea, 1954,
p. 63, 'is the contast between its pretentous theoretical framework and its primitive
and slipshod application in practice.' It does not need a science to produce
Neumannian nightmares.
26. The existence of qualitatve difcultes, or 'pockets of resistance' (St
Augustine, Ceatraaliaaam, V, xiv, 51 Migc, Vol. 4), was used by the Church
fathers to defuse objectons which the science of their time had raised against par of
the Christan faith, such as the dotrine of the cororeal resurection.
FI VE 51
any
rules that make us choose between theories o the bais of
f
licat
ion. Rather, its rules must enable us to choose between
theo
ries which we have already tested and which are falied.
To proeed further. Not only are facts and theories in constant
disharmony, they are never as nearly separated as everone makes
them out to be. Methodological rules speak of 'theories', 'obsera
tons' and 'exerimental results' as if these were well-defned objects
whose propertes are easy to evaluate and which are understood in
the same way by all scientsts.
However, the material which a scientst auall has at his disposal,
his laws, his exerimental results, his mathematcal techniques, his
epistemological prejudices, his atttude towards the absurd con
sequences of the theories which he accepts, is indetennate in many
ways, ambiguous, and neer full searated from the historcl bak
gound. It is contaminated by principles which he does not know
and which, if known, would be extemely hard to test. Questonable
views on cogniton, such as the view that our senses, used in normal
circumstances, give reliable informaton about the world, may invade
the obseraton language itself, consttutng the obseratonal ters
as well as the distncton between veridical and illusor appearance.
As a result, obseraton languages may become ted to older layers of
speculaton which afect, in this roundabout fashion, even the most
progressive methodolog. (Example: the absolute space-tme frame
of classical physics which was codifed and consecrated by Kant.)
The sensor impression, however simple, contains a component that
expresses the physiological reacton of the perceiving organism and
has no objectve correlate. This 'subjectve' component often merges
with the rest, and forms an unstuctured whole which must be
subdivided from the outside with the help of counterinductve
procedures. (A example is the appearance of a fed star to the
naked eye, which contains the efects of irradiaton difracton,
diffusion, restricted by the lateral inhibiton of adjacent elements of
the retna and is further modifed in the brain.) Finally, there are the
auxliar premises which are needed for the derivaton of testable
conclusions, and which occasionally form entre auiliar sciece.
Consider the case of the Coperican hypothesis, whose inventon,
defence, and partal vndicaton runs counter to almost ever
meth
odological rule one might care to think of today. The auxliar

ciences here contained laws describing the propertes and the


mfuence of the terrestial atosphere (meteorolog); optcal laws
dealing with the stucture of the eye and of telescopes, and with the
behavour of light; and dynamcal laws describing moton in moving
systems. Most importantly, however, the auxliar sciences contained
52 AGAI NST METHOD
a theor of cogniton that postulated a certain simple relaton
between perceptons and physical objects. Not all auxliar dis
ciplines were available in explicit form. Many of them merged with
the obseraton language, and led to the situaton described at the
beginning of the preceding paragraph.
Consideraton of all these circumstances, of obseraton terms,
sensor core, auxliar sciences, background speculaton, suggest
that a theor may be inconsistent with the evidence, not because it is
incorrect, but becuse the eidnce is contaminated. The theor is
threatened because the evidence either contains unanalysed sensa
tons which only partly correspond to exteral processes, or because
it is presented in terms of antquated views, or because it is evaluated
with the help of backward auxliar subjects. The Coperican theor
was in touble for all these reasons.
It is this historc-physiological charaer ofthe eidnce, the fact that it
does not merely describe some objectve state of afairs but alo
erese subjeaie, mythical, and long-frotte views concering this
state of afairs, that forces us to take a fresh look at methodolog. It
shows that it would be extemely imprudent to let the evidence judge
our theories directly and without any further ado. A staightorard
and unqualified judgement of theories by 'facts' is bound to eliminate
ideas simpl becuse the t not ft into the framework ofsome olr
cosmolog. Taking experimental results and obseratons for granted
and puttng the burden of proof on the theor means taking the
obseratonal ideology for granted without having ever examined it.
(Note that the experimental results are supposed to have been
obtained with the greatest possible care. Hence 'taking obseratons,
etc., for granted' means 'taking them for granted after the most
careful examinaton of their reliability': for even the most careful
examinaton of an obseraton statement does not interfere with the
concepts in which it is expressed, or with the stucture of the sensor
image.)
Now-how can we possibly examine something we use all the tme
and presuppose in ever statement? How can we critcize the terms in
which we habitually express our obseratons? Let us see!
The first step in our critcism of commonly-used concepts is to
create a measure of critcism, something with which these concepts
can be compared. Of course, we shall later want to know a little more
about the measuring-stck itself; for example, we shall want to know
whether it is better than, or perhaps not as good as, the material
examined. But in order for thi examinaton to start there must be a
measuring-stck in the first place. Therefore, the first step in our
critcism of customar concepts and customar reactons is to step
FI VE 53
outside the circle and either to invent a new conceptual system, for
example a new theor, that clashes with the most carefully
esta
blished obseratonal results and confounds the most plausible
theoretcal principles, or to import such a system from outside
science, from religion, from mythology, from the ideas of incom
petents, 27 or the ramblings of madmen. This step is, again,
counterinductve. Counterinducton is thus both a fa- science
could not exist without it -and a legitmate and much needed mue in
the game of science.
27. It is interestng to see that Philolaos, who disregarded the evidence of the
s
enses
and set the earh in moton, was 'an unmathematical confusionist. It wa the
co
nfusio
nist who found the courage lacking in many great obserer and
m
athem
atically well-infored sientists to disregard the immediate evidence of the
se

ses in order to remain in agreement with principles he frly believed.' K. von


Fnt, Cmodpmbl~cdCcchichtcdaotik usschq,Berlin-New York, 1971 , p.
I 65. 'It is therefore not surprising that the next step on this path was due to a man
wose writings, as far as we know them, show him as a talented stylist and ppularer
W
_
th ocasionally interesting idea of his own rather than as a profound thinker or exact

ent
ist,
ibid., p. 184. Confsionist and superfcial intellectuals mmcahcmwhile the
d

eep
'
thinke
r dcdinto the darker regions of the status quo or, to express it in a
lff
ere
nt way, they remain stuck in the mud.
6
As an eample ofsuch an attept I eamine the tower argment which the
Arstotelians used to reute the motion ofthe earh. Te armet inole
natural interpretatons idas so cosel conneced with obserations that
it need a specal efr to realie their eistece and to dterine their
contet. Galileo idntie the natural interretations which are inconsistet
with Coeics and relce the b other.
It seems to me that [Galileo) suffers greatly from contnual
digressions, and that he does not stop to explain all that is relevant at
each point; which shows that he has not examined them in order, and
that he has merely sought reasons for parcular efects, without
having considered . . . frt causes . . . ; and thus that he has built
without a foundaton.
DESCARTES
I am (indeed) unwilling to compress philosophical dotines into the
most narrow kind of space and to adopt that stf, concise and
graceless manner, that manner bare of any adorment which pure
geomencians call their own, not uttering a single word that has not
been given to them by strict necessity . . . I do not regard it as a fault to
talk about many diverse things, even in those teatses which have only
a single topic . . . for I believe that what gives grandeur, nobility, and
excellence to our deeds and inventons does not lie in what is
necessar-though the absence of it would be a great mistake -but i
what is not . . . .
GAL! LEO
But where common sense believes that ratonalizing sophists have the
intenton of shaking the ver fundament of the commonweal, then it
would seem to be not only reasonable, but perissible, and even
54
sr x
praiseworthy to aid the god cause with sham reasons rather than
leaving the advantage to the . . . opponent.
KANT1
55
As
a
concrete illustraton and as a basis for further discussion, I shall
now briefly describe the manner in which Galileo defused an
important argument against the idea of the moton of the earth.
I say 'defused', and not 'refuted', because we are dealing with a
changing conceptual system as well as with cerain attempts at
conc
ealment.
According to the argument which convinced Tycho, and which is
used against the moton of the earth in Galileo's own Trattato dll
sfera, obseraton shows that 'heavy bodies . . . falling down from on
high, go by a staight and vertcal line to the surface of the earth. This
is considered an irrefutable argument for the earth being motonless.
For, ifit made the diural rotaton, a tower from whose top a rok was
let fall, being carried by the whirling of the earth, would tavel many
hundreds of yards to the east in the te the rock would consume in
its fall, and the rok ought to stike the earth that distance away from
the base of the tower. '
2
I . The three quotatons are: Descares, letter to Mersenne of 1 1 October 1 638,
Oorc, I I , p. 380. Galileo, letter to Leopld of Toscana of 1 640, usually quoted
under the ttle 5alCaadrLaaarc,Qoc,Favoro, VIU, p. 491 . For a detailed discussion
of Galileo's style and its connection with his natral philosophy cf. L. Olschki, Cm
aadsciacZcit.Ccchichtcdamsprachlichmissscha/lichLitoatar,Vol. H, Halle,
1927, reprinted Vaduz, 1965. The letter to Leopold is quoted and discussed on
pp. 455f.
Descares' letter is discussed by Salmon as an example of the issue beteen
ratonalism and empiricism in 'The Foundatons of Scientfc Inference', Miadaad
Cesmes,
ed. Colodny, Pittsburgh, 1 966, p. 1 36. It should rather be regarded as an
example of the issue beteen dogmatc methodologies and opprunistc methodolo
gies, bearing in mind that empiricism can be as strict and unyielding as the most
rigorous tpes of ratonalism.
The Kant quotaton is from the Cntiqace/FarcRcmea,8777, 8f (the quotaton was
brought to my atention by Professor Stanley Rosen's work on Plato's 5mpesiam).
Kant continues: 'However, I would think that there is nothing that goes less well
together with the intenton of assertng a god cause than subterfuge, conceit, and
dec
eption. I
one could take only this much for granted, then the banle of speculatve
r
easo
n . . .
would have been concluded long ago, or would son come to an end. Thus
the
pu
rity
of a cause often stands in the inverse proporton to its tth . . . .' One should
als
?
note
that
Kant explains the rise of cnilmtieaon the basis of disingenuous moves

htc
h
'have the functon to raise mankind above its crde past', ibid., 776, 1 4f. Similar
tde
as
ocur in his account of world histor.
2.
Dialegac,op. cit., p. 1 26.
56 AGAI NST METHOD
In considering the argument, Galileo at once admits the
correctess of the sensory content of the observaton made, v. that
'heavy bodies . . . falling from a height, go perpendicularly to the
surface of the earth'.
3
Considering an author (Chiaramont) who
sets out to convert Copericus by repeatedly mentoning this fact, he
says: 'I wish that this author would not put himself to such trouble
tying to have us understand from our senses that this moton of
falling bodies is simple staight moton and no other kind, nor get
angry and complain because such a clear, obvious, and manifest thing
should be called into queston. For in this way he hints at believing
that to those who say such moton is not staight at all, but rather
circular, it seems they see the stone move visibly in an arc, since he
calls upon their senses rather than their reason to clarif the efect.
This is not the case, Simplicio; for just as I . . . have never seen nor
ever expect to see, the rock fall any way but perpendicularly, just so
do I believe that it appears to the eyes of everyone else. It is, therefore,
better to put aside the appearance, on which we all agree, and to use
the power of reason either to confrm its reality or to reveal its
fallacy. '4 The correctess of the observaton is not in queston. Wat
is in queston is its 'reality' or 'fallacy'. Wat is meant by t
expression?
The queston is answered by an example that occurs in Galileo's
next paragraph, ' from which . . . one may lear how easily anyone
may be deceived by simple appearance, or let us say by the
impressions of one's senses. This event is the appearance to those
who tavel along a steet by night of being followed by the moon, with
steps equal to theirs, when they see it go gliding along the eaves of the
roofs. There it looks to them just as would a cat really running along
the tles and puttng them behind it; an appearance which, if reason
did not intervene, would only too obviously deceive the senses.'
In this example, we are asked to start with a sensory impression and
to consider a statement that is forcefully suggested by it. (The
suggeston is so strong that it has led to entre systems ofbelief and to
rituals, as becomes clear from a closer study of the lunar aspects of
witchcraft and of other cosmological hypotheses.) Now 'reason
intervenes'; the statement suggested by the impression is examned,
and one considers other statements in its place. The nature of the
impression is not changed a bit by this actvity. (This is only
approxmately tue; but we can omit for our present purpose the
complicatons arising from an interacton of impression and
3. ibid., p. 1 25.
4. ibid., p. 256.
SI X 57
propositon.) But it enters new obseraton statements and plays new,
better or worse, parts in our knowledge. What are the reasons and the
meth
ods which regulate such an exchange?
To start with, we must become clear about the nature of the total
phenomenon: appearance plus statement. There are not two acts -
one, notcing a phenomenon; the other, expressing it with the help of
the appropriate statement but onl one, viz. saying in a certain
observatonal situaton, ' the moon is following me', or, ' the stone is
falling straight down'. We may, of course, abstractly subdivide this
process into parts, and we may also ty to create a situaton where
statement and phenomenon seem to be psychologically apart and
waitng to be related. (This is rather difcult to achieve and is
perhaps entrely impossible.) But under normal circumstances such a
division does not occur; describing a familiar situaton is, for the
speaker, an event in which statement and phenomenon are frmly
glued together.
This unity is the result of a proess ofleaming that starts in one's
childhood. From our ver early days we lear to react to situatons
with the appropriate responses, linguistc or otherise. The teaching
procedures both shape the 'appearance', or 'phenomenon', and
establish a frm conneaion with words, so that fnally the phenomena
seem to speak for themselves without outside help or extaneous
knowledge. They are what the associated statements assert them to
be. The language they 'speak' is, of course, infuenced by the beliefs
of earlier generatons which have been held for so long that they no
longer appear as separate principles, but enter the ter of everday
discourse, and, after the prescribed training, seem to emerge from
the things themselves.
At this point we may want to compare, in our imaginaton and quite
abstractly, the results of the teaching of diferent languages
incororatng diferent ideologies. We may even want consciously to
change some of these ideologies and adapt them to more 'modem'
points of view. It is ver difcult to say how this will alter our situaton,
unles we make the further assumpton that the quality and stucture
of sensatons (perceptons), or at least the quality and stucture of
those sensatons which enter the body of science, is independent of
their linguistc expression. I a ver doubtful about even the
approxmate validity of this assumpton, which can be refuted by
si
mple
examples, and I am sure that we are depriving ourselves of
n
ew
and
surrising discoveries as long as we remain within the limits
de
fn
ed
by it. Yet, I shall for the moment, remain within these limits .
.
aking
the additonal simplifing assumpton, we can now
d1sn
ngu
ish
between sensatons and those 'mental operatons which
58 AGAI NST METHOD
follow so closely upon the senses',
5
and which are so frmly
connected with their reactons that a separaton is difcult to achieve.
Considering the origin and the efect of such operatons, I shall call
them natural interretations.
In the history of thought, natural interretatons have been
regarded either as a prori preuppositions of science, or else as
preudice which must be removed before any serious examinaton can
begin. The frst view is that of Kant, and, in a very diferent manner
and on the basis of ver diferent talents, that of some contemporar
linguistc philosophers. The second view is due to Bacon (who had
predecessors, however, such as the Greek sceptcs).
Galileo is one of those rare thinkers who wants neither forever to
retain natural interretatons nor altogether to eliminate them.
Wholesale judgements of this kind are quite alien to his way of
thinking. He insists upon a ctical discussion to decide which natural
interretatons can be kept and which must be replaced. This is not
always clear from his writngs. Quite the contar. The methods of
reminiscence, to which he appeals so freely, are designed to create
the impression that nothing has changed and that we contnue
expressing our obseratons in old and familiar ways. Yet his atttude
is relatvely easy to ascertain: natural interretatons are necesar.
The senses alone, without the help of reason, cannot give us a tue
account of nature. What is needed for arriving at such a tue account
are 'the . . . senses, accompanied b reasonini.
6
Moreover, in the
arguments dealing with the moton of the earth, it is this reasoning, it
is the connotaton of the obseraton terms and not the message of the
senses or the appearance that causes trouble. 'It is, therefore, better
to put aside the appearance, on which we all agree, and to use the
power of reason either to confrm its reality or to reveal its fallacy.'
7
Confrming the reality or revealing the fallacy of appearances means,
however, examining the validity of those natural interretatons
which are so intmately connected with the appearances that we no
longer regard them as separate assumptons. I now tum to the frst
natural interretaton implicit in the argment from falling stones.
According to the Coperican view as presupposed in the tower
argment the moton of a falling stone should be 'mixed staight
and-circular'. 8 By the 'moton of the stone' is meant not its moton
relatve to some visible mark in the visual feld of the obserer, or its
5. Francis Bacon, NmamOqaaam,Introduction.
6. Dialegc,op. cit., p. 255, my italics.
7. ibid., p. 256.
8. ibid., p. 248.
SI X 59
observed moton, but rather its moton in the solar system or in
(abs
olute) space, i.e. its real motion. The familiar facts appealed to in
the argument present a different kind of moton, a simple vercal
moton. This refutes the Coperican hypothesis only if the concept
of moton that occurs in the observaton statement is the same as the
concept of moton that occurs in the Copercan predicton. The
observaton statement 'the stone is falling staight down' must,
therefore, refer to a movement in (absolute) space. It must refer to a
real
moton.
Now, the force of an 'argument from observaton' derives from the
fact that the observaton statements involved are firmly connected
with appearances. There is no use appealing to observaton if one
does not know how to describe what one sees, or if one can ofer one's
descripton with hesitaton only, as if one had just leared the
language in which it is formulated. Producing an observaton
statement, then, consists of two very diferent psychological events:
(1) a clear and unambiguous sesation and (2) a clear and
unambiguous connecion between this sensaton and parts of a
language. This is the way in which the sensaton is made to speak. Do
the sensatons in the above argument speak the language of real
moton?
They speak the language of real moton in the context of 1 7th
centur everday thought. At least, this is what Galileo tells us. He
tells us that the everday thinking of the tme assumes the 'operatve'
character of all moton, or, to use well-known philosophical terms, it
assumes a naie realism with repect to motion: except for occasional
and unavoidable illusions, apparent moton is identcal with real
(absolute) moton. Of course, this distncton is not exlicitly drawn.
One does not first distnguish the apparent moton from the real
moton and then connect the two by a correspondence rule. One
rather describes, perceives, acts towards moton as if it were already
the real thing. Nor does one proceed in this manner under all
circumstances. It is admitted that objects may move which are not
seen to move; and it is also admitted that certain motons are illusory
(cf. the example of the moon mentoned earlier in this chapter).
Apparent moton and real moton are not always identfied. However,
ere are paradigmatic cae in which it is psychologically very difcult,
tf not plainly impossible, to admit decepton. It is from these
para
dig
atc cases, and not from the exceptons, that naive realism
with respect to moton derives its stength. These are also the
situatons in which we first lear our kinematc vocabular. From our
very childhood we lear to react to them with concepts which have
naive realism built right into them, and which inexticably connect
60 AGAI NST METHOD
movement and the appearance of movement. The motion of the
stone in the tower argument, or the alleged moton of the earth, is
such a paradigmatic case. How could one possibly be unaware of the
swif moton of a large bulk of matter such as the earth is supposed to
be! How could one possibly be unaware of the fact that the falling
stone traces a vastly extended trajector through space! From the
point of view of 1 7th-centur thought and language, the argument is,
therefore, impeccable and quite forceful. However, notce how
theores ('operatve character' of all motion; essental correctess of
sense reports) which are not formulated explicitly, enter the debate in
the guise of observable events. We realize again that such events are
Trojan horses which must be watched most carefully. How is one
supposed to proceed in such a stcky situaton?
The argument from falling stones seems to refute the Coperican
view. This may be due to an inherent disadvantage ofCopemicanism;
but it may also be due to the presence of natural interretatons which
are in need of improvement. The frst task, then, is to discuer and to
isolate these unexamined obstacles to progress.
It was Bacon's belief that natural interretatons could be
discovered by a method of analysis that peels them of, one afer
another, untl the sensor core of ever observaton is laid bare. This
method has serious drawbacks. First, natural interretatons of the
kind considered by Bacon are not just add to a previously exstng
feld of sensations. They are instrumental in constituting the feld, as
Bacon says himself. Eliminate all natural interretatons, and you
also eliminate the ability to think and to perceive. Second,
disregarding this fundamental functon of natural interretatons, it
should be clear that a person who faces a perceptual feld without a
single natural interretaton at his disposal would be compleel
disoreted, he could not even star the business of science. The fact
that we d start, even after some Baconian analysis, therefore shows
that the analysis has stopped prematurely. It has stopped at precisely
those natural interretatons of which we are not aware and without
which we cannot proceed. It follows that the intenton to start from
scratch, after a complete removal of all natural interretatons, is self
defeatng.
Furthermore, it is not possible even parl to unravel the cluster of
natural interretatons. At frst sight the task would seem to be simple
enough. One takes observaton statements, one after the other, and
analyses their content. However, concepts that are hidden in
observaton statements are not likely to reveal themselves in the more
abstract parts of language. If they do, it will stll be difcult to nail
them down; concepts, just like percepts, are ambiguous and
SI X 61
dependent on background. Moreover, the content of a concept is
determined also by the way in which it is related to percepton. Yet,
ho
w can this way be discovered without circularity? Perceptons must
be identfied, and the identfing mechanism will contain some of the
very same elements which gover the use of the concept to be
investgated. We never penetate this concept completely, for we
always use part of it in the attempt to fnd its consttuents. There is
only one way to get out of this circle, and it consists in using an
eteal measure ofcomparison, including new ways of relatng concepts
and percepts. Removed from the domain of natural discourse and
from all those principles, habits, and atttudes which consttute its
form oflife, such an exteral measure will look stange indeed. This,
however, is not an argument against its use. On the contrary, such an
impression of stangeness reveals that natural interpretatons are at
work, and is a frst step towards their discovery. Let us explain this
situaton with the help of the tower example.
The example is intended to show that the Coperican view is not in
accordance with 'the facts'. Seen from the point of view of these
'facts', the idea of the moton of the earth is outlandish, absurd, and
obviously false, to menton only some of the expressions which were
frequently used at the tme, and which are stll heard whenever
professional squares confront a new and counter-factual theory.
This makes us suspect that the Coperican view is an exteral
measuring rod of precisely the kind described above.
Let us therefore tum the argument around and use it as a dtecing
dice that helps us to discover the natural interpretatons which
exclude the moton of the earth. Turing the argument around, we
frt asser the moton of the earth and the inquire what changes will
remove the contradicton. Such an inquiry may take considerable
tme, and there is a good sense in which it is not finished even today.
The contadicton may stay with us for decades or even centuries.
Stll, it must be upheld untl we have finished our examinaton or else
the examinaton, the attempt to discover the antediluvian com
ponents of our knowledge, cannot even start. This, we have seen, is
one of the reasons one can give for retaining, and, perhaps, even for

neting, theories which are inconsistent with the facts. Ideological


mgredients of our knowledge and, more especially, of our
obs
erva
tons are discovered with the help of theories which are
refuted by them. The are disc(ered counternductie/.
Let me repeat what has been asserted so far. Theories are tested,
and possibly refuted, by facts. Facts contain ideological components,
older views which have vanished from sight or were perhaps never
fonnulated in an explicit manner. Such components are highly
62 AGAI NST METHOD
suspicious. First, because of their age and obscure origin: we do not
know why and how they were introduced; secondly, because their
ver nature protects them, and always has protected them, from
critcal examination. In the event of a contadicton between a new
and interestng theor and a collecton of firmly established facts, the
best procedure, therefore, is not to abandon the theor but to use it to
discover the hidden principles responsible for the contadicton.
Counterinducton is an essental part of such a process of discover.
(Excellent historical example: the arguments against moton and
atomicity of Parmenides and Zeno. Diogenes of Sinope, the Cynic,
took the simple course that would be taken by many contemporar
scientsts and all contemporar philosophers: he refuted the
arguments by rising and walking up and down. The opposite course,
recommended here, has led to much more interestng results, as is
witessed by the histor of the case. One should not be too hard on
Diogenes, however, for it is also reported that he beat up a pupil who
was content with his refutaton, exclaiming that he had given reasons
which the pupil should not accept without additonal reasons of his
own.'
Having discered a partcular natural interretaton, how can we
eamine it and tet it? Obviously, we cannot proceed in the usual way,
i.e. derive predictons and compare them with 'results of obsera
ton'. These results are no longer available. The idea that the senses,
employed under normal circumstances, produce correct reports of
real events, for example reports of the real moton of physical bodies,
has been removed from all obseratonal statements. (Remember
that this noton was found to be an essental part of the ant
Coperican argument.) But without it our sensor reactons cease to
be relevant for tests. This conclusion was generalized by some older
ratonalists, who decided to build their science on reason only and
ascribed to obseraton a quite insignificant auxliar functon.
Galileo does not adopt this proedure.
If one natural interretaton causes trouble for an attactive view,
and if its elimination removes the view from the domain of
obseraton, then the only acceptable procedure is to use other
interretatons and to see what happens. The interretaton which
Galileo uses restores the senses to their positon as instrments of
exploraton, but onl with repea to the realit ofrelatie motion. Moton
'among things which share it in common' is 'non-operatve', that is,
'it remains insensible, imperceptble, and without any efect
9. Hegel, Ierlcaag Udic Ccchichtc d Fhiles@hic, l, ed. C. L. Michele!,
Berlin, 1 84, p. 289.
SI X 63
whatever'
. 1 0
Galileo's frst step, in his joint examinaton of the
Cope
rican doctine and of a familiar but hidden natural interpreta
tion, consists therefore in relacing the latter b a di eret interretation.
In other words, he introduce a new obseration langage.
This is, of course, an entrely legitmate move. In general, the
observaton language which enters an argument has been in use for a
long tme and is quite familiar. Considering the stucture of common
idioms on the one hand, and of the Aristotelian philosophy on the
other, neither this use nor this familiarity can be regarded as a test of
the underlying principles. These principles, these natural interpreta
tons, occur in every descripton. Extaordinary cases which might
create difcultes are defused with the help of 'adjustor words',
1 1
such as 'like' or 'analogous', which divert them so that the basic
ontolog remains unchallenged. A test is, however, urgently needed.
It is especially needed in those cases where the principles seem to
threaten a new theory. It is then quite reasonable to intoduce
alteratve obseraton languages and to compare them both with the
original idiom and with the theory under examinaton. Proceeding in
1 0. Diahgc,op. cit., p. 1 71 . Galileo's kinematc relatvism is not consistent. In the
passage quoted, he propses the view ( 1) that shared moton ha aeg umhatseno.
'Moton,' he says, 'in so far as it is and acts as moton, to that extent exst relatvely to
things that lack it; and among things which all share equally in any moton, it dos not
act and is as ifit did not exist' (. 1 1 6); 'Whatever moton comes to be attibuted to the
earth must necessarily remain imperceptble . . . so long a we lok only at terestial
objects' (. 1 14); ' . . . moton that is common to many moving things is idle and
inconsequental to the relaton of those movables among themselves . . .' (. 1 1 6). On
the other hand, (2) he also suggests that 'nothing . . . mmciaastraightliacqaatarc.
The moton of all celestal objects is in a circle; ships, coaches, hores, birds, all move
in a circle around the earth; the motions of the pars of animals are all circular; in sum
we are forced to assume that only grmiadenamand lniasanammove apparendy in a
straight line; but even that is not cerain as long as it ha not been proven that the earth
is at rest' (. 19). Now, if (2) is adopted, then the lose part of systems moving in a
straight line will tend to describe circular paths, thus contadictng (1). It is this
inconsistency which has prompted me to split Galileo's argument into two steps, one
dealing with the relatvity of moton (only relatve moton uaeticc,the other dealing
with ineral laws (and only inertial moton lcmcthcrclatieabctmcthc panse/aqst
aaq ctcd~ assuming, of course, that neighbouring ineral motons are approximately
parallel). For the two steps of the argument, see the next chapter. One must also realize
that
acce
ptng relatvity of moton for ineral paths means giving up the ipctasthmq,

hie provides an (inner) cause for motons and therefore assumes an absolute space
m which this cause becomes manifest. This Galileo seems to have done by now, for his
argu
ment
for the existence of'boundless' or 'perpetual' motons which he oudines on
pp.
1 47f of the Dialegacappeals to motons which are neutal, i.e. neither natural nor
for
ced,
and
which may therefore (?) be assumed to go on for ever.
1 . J.L.
Austn, 5scaad5sibilia,New York, 1 964, p. 74. Adjustor words play
an Im
por
ant role in the Aristotelian
philosophy.
64 AGAI NST METHOD
this way, we must make sure that the comparison i s fair. That is, we
must not critcize an idiom that is supposed to functon as an
observaton langage because it is not yet well known and is,
therefore, less strongly connected with our sensor reactons and less
plausible than is another, more 'common' idiom. Superfcial
critcisms of this kind, which have been elevated into an entre
'philosophy', abound in discussions of the mind-body problem.
Philosophers who want to introduce and to test new views thus fnd
themselves faced not with arments, which they could most likely
answer, but with an impenetrable stone wall of well-entrenched
reaions. This is not at all different from the atttude of people
ignorant of foreign langages, who feel that a certain colour is
much better described by 'red' than by 'rosso'. As opposed to such
attempts at conversion by appeal to familiarity ('I know what pains
are, and I also know, from intospecton, that they have nothing
whatever to do with material processes!'), we must emphasize that a
comparatve judgement of observaton langages, e.g. materialistc
observaton langages, phenomenalistc observaton langages,
objective-idealistc observaton langages, theological observaton
langages, etc., can start only whe all ofthem are spoke equall
fuetl.
Let us now contnue with our analysis ofGalileo's reasoning.
7
The new natural interretations constitute a ne and highl abstract
obseration language. The are intrduced and concealed so that one fils to
notice the change that has take place (method ofanamneis). Te contain
the ida ofthe relatvity of all moton and the law of circular inerta.
Galileo replaces one natural interretaton by a very diferent and as
yet (1 630) at least partly unnatural interpretaton. How does he
proceed? How does he manage to introduce absurd and counter
inductve assertons such as the asserton that the earth moves, and yet
get them a just and attentve hearing? One antcipates that arguments
will not sufce - an interestng and highly important limitaton of
ratonalism - and Galileo's utterances are indeed arguments in
appearance only. For Galileo uses proagand. He uses pschological
trck in additon to whatever intellectual reasons he has to ofer.
These ticks are very successful: they lead him to victory. But they
obscure the new atttude towards experience that is in the making,
and postpone for centuries the possibility of a reasonable philosophy.
They obscure the fact that the experience on which Galileo wants to
base the Coperican view is nothing but the result of his own fertle
imaginaton, that it has been inveted. They obscure this fact by
insinuatng that the new results which emerge are known and
conceded by all, and need only be called to our attenton to appear as
the most obvious expression of the truth.
Galileo 'reminds' us that there are situatons in which the non
operative character of shared motion is just as evident and as firmly
believed as the idea of the operatve character of all moton is in other

ircumstances. (This latter idea is, therefore, not the only natural
Interpretation of moton.) The situatons are: events in a boat, in a
smoothly moving carriage, and in other systems that contain an
obserer and permit him to carry out some simple operatons.
Sag
ed: There has just ocurred to me a certain fantasy which
pas
sed
through my imaginaton one day while I was sailing to Aleppo,
65
66 AGAI NST METHOD
where I was going as consul for our count . . . . I f the point of a pen
had been on the ship during my whole voyage from Venice to
Alexandretta and had had the property of leaving visible marks
of its whole tip, what tace - what mark - what line would it have
left?
Simplico: It would have left a line extending from Venice to there;
not perfectly staight - or rather, not lying in the perfect arc of a circle
- but more or less fluctuatng according as the vessel would now and
again have rocked. But this bending in some places a yard or two to the
right or left, up or down, i length of many hundreds of miles, would
have made little alteraton in the whole extent of the line. These would
scarcely be sensible, and, without an error of any moment, it could be
called part of a perfect arc.
Saged: So that if the fluctuaton of the waves were taken away and
the moton of the vessel were calm and tranquil, the tue and precise
moton of that pen would have been an arc of a perfect circle. Now if I
had had that same pen contnually in my hand, and had moved it only a
little sometmes this way or that, what alteratons should I have
brought into the main extent of this line?
Simplicio: Less than that which would be given to a staight line a
thousand yards long which deviated from absolute staightess here
and there by a flea's eye.
Saged: Then if an artst had begn drawing with that pen on a
sheet of paper when he left the port and had contnued doing so all the
way to Alexandrena, he would have been able to derive from the pen's
moton a whole narratve of many fgres, completely traced and
sketched in thousands of directons, with landscapes, buildings,
animals, and other things. Yet the actual real essental movement
marked by the pen point would have been only a line; long, indeed,
but ver simple. But as to the artst's own actons, these would
have been conducted exactly the same as if the ship had been
standing stll. The reason that of the pen's long moton no trace would
remain except the marks drawn upon the paper is that the gross
moton from Venice to Alexandrena was common to the paper, the
pen, and everything else in the ship. But the small motons back and
forth, to right and left, communicated by the artst's fingers to the pen
but not to the paper, and belonging to the forer alone, could thereby
leave a tace on the paper which remained statonar to those
motons.
1
I . Dmegc,op. cit., pp. I 7 I f.
Or
S EVEN
Saliati: + . . Imagine yourelf i n a boat with your eyes fied on a
point of the sail yard. Do you think that because the boat is moving
along briskly, you will have to move your eyes i order to keep your
vision always on that point of the sail and follow its moton?
Simplico: I a sure that I should not need to make any change at all;
not just as to my vision, but ifl had aimed a musket I should never have
to move it a hairsbreadth to keep it aimed, no matter how the boat
moved.
Saliati: And this comes about because the moton which the ship
confer upon the sail yard, it confers also upon you and upon your
eyes, so that you need not move them a bit in order to gaze at the top of
the sail yard, which consequently appears motonless to you. (And the
rays of vision go from the eye to the sail yard just as if a cord were ted
between the two ends of the boat. No a hundred cords are ted at
different fied points, each of which keeps its place whether the ship
moves or remains stlli
67
It is clear that these situatons lead to a non-operatve concept of
moton even within common sense.
On the other hand, common sense, and I mean 1 7th-centur
Italian-arsan common sense, also contains the idea of the oeatie
character of all moton. This latter idea arises when a limited object
that does not contain too many parts moves in vast and stable
surroundings; for example, when a camel tots through the desert, or
when a stone descends from a tower.
Now Galileo urges us to 'remember' the conditons in which we
assert the non-operatve character of shared moton in this case also,
and to subsume the second case under the frst.
Thus, the first of the two paradigms of non-operatve moton
mentoned above is followed by the asseron that -'It is likewise tue
that the earth being moved, the moton of the stone in descending is
actually a long stetch of many hundred yards, or even many
thousand; and had it been able to mark its course in motonless air or
upon some other surface, it would have left a ver long slantng line.
But that part of all this moton which is common to the rock, the
2. ibid., pp. 249f. That phenomena of scmoton depend on rcwncmoton has
been
assered by Euclid in his Otic, Theon red. par. 49f. A old scholion of par. 50
uses
the
example of a boat leaving the harbour: Heiberg, vii, 283. The example is
repeated by Copericus in Bok 1 , Chapter viii, of DcRnl.It was a commonplace in
mediaeval optics. Cf. Witelo, Fmpcuna,iv par 1 38 (Basel, 1 572, p. 1 80).
68 AGAI NST METHOD
tower, and ourselves remains insensible and as if it did not exst.
There remains obserable only that part in which neither the tower
nor we are partcipants; in a word, that with which the stone, in
falling, measures the tower.'
3
And the second paradigm precedes the exhoraton to 'tansfer this
argument to the whirling of the earth and to the rok placed on top of
the tower, whose moton you cannot discer because, in common
with the rock, you possess from the earth that moton which is
required for following the tower; you do not need to move your eyes.
Next, if you add to the rock a downward moton which is peculiar to it
and not shared by you, and which is mixed with this circular moton,
the circular porton of the moton which is common to the stone and
the eye contnues to be imperceptble. The staight moton alone is
sensible, for to follow that you must move your eyes downwards.'4
This is stong persuasion indeed.
Yielding to this persuasion, we now quite automaticall start
confounding the conditons of the two cases and become relatvists.
This is the essence of Galileo's tcker! As a result, the clash
between Copercus and 'the conditons afectng ourselves and
those in the air above us'
5
dissolves into thin air, and we fnally
realize 'that all terrestial events from which it is ordinarily held that
the earth stands stll and the sun and the fied stars are moving would
necessarily appear just the same to us if the earth moved and the
other stood stll'.
6
3. ibid., pp. 1 72f.
4. ibid., p. 250.
5. Ptolemy, 5atm,i, 1, p. 7.
6. Dialegc,op. cit., p. 41 6: cf. the DialegacCeacmiag1m Nm 5ci,tansl.
Henr Crew and Alfonso de Salvio, New York, 1958, p. 1 64: 'The same experiment
which at frst glance seemed to show one thing, when more carefully examined, asures
us of the contar.' Professor McMullin, in a critque of this way of seeing things,
wants more 'logical and biogaphical justfcaton' for my asseron that Galileo not
only argued, but also cheated ['A Taxonomy of the Relaton between Histor and
Philosophy of Science', Miaaceta5tadic,Vol. 5, Minneapolis, 1 971 , p. 39], and he
objects to the way in which I let Galileo intoduce dynamical relatvism. According to
him 'what Galileo argues is that since his opponent alrcaq interrets obseratons
made in such a context [movements on boat] in a "relatvistc" way, how can he
consistendy do otherwise in the case of obseratons made on the earth's surface?'
(ibid., p. 4). This is indeed how Galileo argues. But he argues so against an oppnent
who, according to him, 'feels a great repugnance towards recognizing this non
operatve quality of moton among the things which share it in common' (Dialegac,op.
cit., p. 1 71 ), who is convinced that a bot, apan from having relatve motons, h
aHelatc pesitieasaadmetieasmcd(cf. Aristode, Physics, 208b8f), and who at any rate
ha developd the an of using diferent notons on different ocaions without rnning
into a contadicton. Now if thuis the psiton to be attacked, then showing that an
S EVEN 69
Let us now look at the situation from a more abstract point of view.
We start with two conceptual sub-systems of'ordinary' thought (see
the
following table). One of them regards moton as an absolute
process which always has effects, effects on our senses included. The
description of this conceptual system given here may be somewhat
idealized; but the arguments of Copericus' opponents, which are
quoted b Galileo himself and, according to him, are 'very
plausible', show that there was a widespread tendency to think in its
terms, and that this tendency was a serious obstacle to the discussion
of alteratve ideas. Occasionally, one finds even more primitve ways
of thinking, where concepts such as 'up' and 'down' are used
absolutely. Examples are: the asserton 'that the earth is too hea, to
climb up over the sun and then fall headlong back down again', or
the asserton that 'after a short tme the mountains, sinking
downward with the rotaton of the terrestal globe, would get into
such a positon that whereas a little earlier one would have had to
climb steeply to their peaks, a few hours later one would have to stoop
and descend in order to get there'.
9
Galileo, in his marginal notes,
calls these 'utterly childish reasons which] sufced to keep imbeciles
believing in the fty of the earth'
1
and he tk it unnecessary 'to
bother about such men as those, whose name is leion, or to take notce
of their fooleries'. 1 1 Yet it is clear that the absolute idea of moton
was 'well-entenched', and that the attempt to replace it was bound to
encounter stong resistance.
1 2
opponent has a relative idea of motion, or frequently uses the relatve idea in his
everyday affairs, is not at all 'prof of inconsistency in his own "paradigm" '
(McMullin, op. cit., p. 4). I t just reveals one par of that paradigm without touching
the other. The argument turs into the desired prof only if the absolute noton is
either suppressed or spirited away, or else identifed with the relatvistc noton - and
this is what Galileo actually dos, though sureptitiously, as I have ted to show.
7. Dialegac,op. cit., p. 328.
8.
ibid., p. 327.
9.
ibid., p. 330.
10.
ibid., p. 327.
I I .
ibid., p. 327, italics added.
1 2.
The idea that there is an absolute directon in the universe has a very
int
erestin
g history. It rest on the strcture of the gravitatonal feld on the surface of
the
earth,
or of that par of the earth which the obserer knows, and generalizes the
experiences made there. The generalizton is only rarely regarded as a separate
hyp
othes
is, it rather enter the 'grammar' of common sense and gives the ters 'up'
a
nd
'down' an absolute meaning. (This is a 'natural interretation', in precisely the
sense that was explained in the text above.) Lactantus, a Church father of the fourth
century, appeals to this meaning when he asks (DnimIastitatieac, I l l , De Fas
Sapientia): 'Is one really going to b so confused as to asume the exstence of humans
whose feet are above their heads? Where trees and fuit grow not upwards, but
70 AGAI NST METHOD
The second conceptual system is built around the relatvity of
moton, and is also well-entenched in its own domain of applicaton.
Galileo aims at replacing the frst system by the second in all cases,
terrestial as well as celestal. Naive realism wth respect to moton is
to be cmpletel eliminated.
Now, we have seen that this naive realism is on ocasions a
essental part of our obseratonal voabular. On these occasions
(aradigm 1), the obseraton langage contains the idea of the
efcacy of all moton. Or, to express it in the material mode of
speech, our experience in these situatons is an experience of objects
Paraig/: Moton of compact
objects in stable suroundings of
great spatal extension -
deer obsered by the hunter.
Natural interretatio:
Al moton is operatve.
Falling stone
pre
!
Earth at rest
Moton of earth
predics
!
Oblique moton
of stone
Paraig I: Moton of objects in
boats, coaches and other
moving systems.
Natural interretatio:
Only relatve moton is operatve.
Falling stone
pre
!
No reltie
moton between
starng point
and earth
Moton of earth
predics
!
No relatve
moton between
starng point
and stone
dowwards?' The same use of language is presuppsed by that 'mass of untutored
men' who raise the queston why the antpodeans are not falling of the earh (Pliny,
NataralHuteq,II, pp. I 61-, cf. also Ptolemy, 5atm,I, 7). The attempts ofThales,
Anaxmenes and Xenophanes to fnd suppon for the earh which prevent it from
falling 'down' (Aristode, DcCecle,294al21) shows that almost all early philosopher,
with the sole excepton of Anaximander, shared in this way of thinking. (For the
Atomists, who assume that the atoms originally fall 'down,' cf. Jammer, Ccqu e/
5pc, Cambridge, Mass., I 953, p. l I .) Even Galileo, who thoroughly ridicules the
idea of the falling antpodes (Dialegc, op. cit., p. 33 I), ocaionally speaks of the
'upper half of the moon', meaning that pan of the mon 'which is invisible to us'. And
let us not forget that some linguistc philosophers of today 'who are to stupid to
recognize their ow limitatons' (Galileo, op. cit., p. 327) want to revive the absolute
meaning of 'up-down' at leat lecal. Thus the power over the minds of his
contemporaries of a primitve conceptual frame, assuming an anisotopic world, which
Galileo had also to fght, must not be underestmated. For an examinaton of some
apects of Britsh common sense at the tme of Galileo, including atonomical
common sense, see E.M.W. Tillyard, 1hcElccthaa erldFictarc, London, 1 963.
The agreement between popular opinion and the centally symmetc univere is
frequently assened by Aristode, e.g. in DcCecle,p. 308a23f.
S EVEN 71
which move absolutely. Taking this into consideraton, it is apparent
that Galileo's proposal amounts to a partal revision of our
obseraton langage or of our experience. A experence which
partly contraics the idea of the moton of the earth is tured into an
experience that cnfnns it, at least as far as 'terrestial things' are
concered.
1 3
This is what auall hapes. But Galileo wants to
persuade us that no change has taken place, that the second
conceptual system is already universally knorn, even though it is not
universally used. Salviat, his representatve in the Dialoge, his
opponent Simplicio and Sagredo the intelligent layman all connect
Galileo's method of argmentaton with Plato's theor of anamneis
a clever tactcal move, typically Galilean one is inclined to say. Yet we
must not allow ourselves to be deceived about the revolutonar
development that is actually taking place.
The resistance against the assumpton that shared moton is non
operatve was equated with the resistance which forgotten ideas
exhibit towards the attempt to make them known. Let us accept this
interretation of the resistance! But let us not forget its eistece. We
must then admit that it resticts the use of the relatvistc ideas,
confning them to par of our everday experence. Outsid this part,
i.e. in interlanetar space, they are 'forgotten' and therefore not
actve. But outside this part there is not complete chaos. Other
concepts are used, among them whose ver same absolutstc
concepts which derve from the frst paradigm. We not only use them,
we must also admit that they are entrely adequate. No difcultes
arise as long a one remains within the limits of the frst paradigm.
'Experience', i.e. the totality of all facts from all domains, cannot
force us to carr out the change which Galileo wants to intoduce.
The motve for a change must come from a different source.
It comes, frst, from the desire to see 'the whole [correspond] to its
parts with wonderful simplicity',
1 4 as Copericus had already
1 3. Dialegac,op. cit., pp. 1 32 and 416.
1 4. ibid., p. 341 . Galileo quotes here from Copericus' address to Pope Paul III in
De
Rn
elatieaibas,cf. also Chapter I 0 and the NanatieFrima(quoted from E. Rosen,
Tree C@mima1rcatisc,New York, 1959, p. 1 65): 'For all these phenomena appear
t be linked most nobly together, as by a golden chain; and each of the planets, by its
psition, and order, and ever inequality of its moton, bear witess that the earh
moves and that we who dwell upn the globe of the earh, instead of acceptng its
cha
ng
es
of position, believe that the planets wander in all sors of motons of their
ow
n.'
Note that empirical reasons are absent from the argument and have to be, for
Co
per
icus himself admit (Cemmtarielas,op. cit., p. 57) that the Ptolemaic theor is
'consistent with the numerical data'.
72 AGAI NST METHOD
expressed himself. It comes from the 'typically metaphysical urge'
for unity of understanding and conceptual presentaton. And the
motve for a change is connected, secondly, with the intenton to
make room from the moton of the earth, which Galileo accepts and is
not prepared to gve up. The idea of the moton of the earth is closer
to the frst paradigm than to the second, or at least it was at the tme of
Galileo. This gave stength to the Aristotelian arguments, and made
them plausible. To eliminate the plausibility, it was necessar to
subsume the frst paradigm under the second, and to extend the
relatve notons to all phenomena. The idea of anamneis functons
here as a psychologcal crutch, as a lever which smooths the process
of subsumpton by concealing its exstence. As a result we are now
ready to apply the relatve notons not only to boats, coaches, birds,
but to the 'solid and well-established earth' as a whole. And we have
the impression that this readiness was in us all the tme, although it
took some efort to make it conscious. This impression is most
certainly erroneous: it is the result of Galileo's propagandistc
machinatons. We would do better to describe the situaton in a
diferent way, as a change of our conceptual system. Or, because we
are dealing with concepts which belong to natural interpretatons,
and which are therefore connected with sensatons in a very direct
way, we should describe it as a change ofeerece that allows us to
accommodate the Copercan doctine. It is this change which
underlies the tansiton from the Aristotelian point of view to the
epistemolog of modem science.
For experience now ceases to be the unchangeable fundament
which it is both in common sense and in the Aristotelian philosophy.
The attempt to support Copericus makes experience 'fluid' in the
ver same manner in which it makes the heavens fluid, 'so that each
star roves around in it by itself. 15 An empiricist who starts from
experience, and builds on it without ever looking back, now loses the
ver ground on which he stands. Neither the earth, 'the solid, well
established earth', nor the facts on which he usually relies can be
trusted any longer. It is clear that a philosophy that uses such a fluid
and changng experience needs new methodologcal principles
which do not insist on an asymmetic judgement of theories by
experience. Classical physic intuitvely adopts such principles; at least
its great and independent thinkers, such as Newton, Faraday,
Boltzmann proceed in this way. But its ofcial dcrne stll clings to
the idea of a stable and unchangng basis. The clash between this
I 5. Dialegc,op. cit., p. 1 20.
SEVEN 73
doctrine and the actual procedure is concealed by a tendentous
presentaton of the results of research that hides their revolutonary
origin and suggests that they arose from a stable and unchanging
source. These methods of concealment start with Galileo's attempt
to introduce new ideas under the cover of anamnesis, and they
culminate in Newton.
1
6
They must be exposed if we want to arrive at
a better account of the progressive elements in science.
My discussion of the ant-Coperican argument is not yet
complete. So far, I have tried to discover what assumpton will make a
stone that mues alongid a muing to1er appear to fall 'staight down',
instead of being seen to move in an arc. The assumpton, which I
shall call the relatiit principle, that our senses notce only relatve
moton and are insensitve to a moton which objects have in common
was seen to do the tick. What remains to be explained is 1hy the stone
stays 1ith the to1er and is not left behind. In order to save the
Coperican view, one must explain not only why a moton that
preseres the relaton among visible objects reains unnoticed, but
also, why a common moton of various objects does not affect their
relaton. That is, one must explain why such a moton is not a causal
agent. Turg the queston around in the manner explained in the
text to footote 10, page 63 of the last chapter, it is now apparent
that the ant-Coperican argument described there rests on t1o
natural interpretatons: v, the eiteolgcal assumption that
absolute moton is always noticd, and the dynamicl principle that
objects (such as the falling stone) which are not interfered with
assume their natural moton. For Aristotelians the natural moton of
an object not interfered with is ret, i.e. constancy of qualites and
of positon.
1
7 This corresponds to our own experience where things
have to be pushed around to move. The discovery of seeds, bacteria,
viruses would have been impossible without a frm belief in the
qualitatve part of the law - and it confrmed it in a most impressive
way. Using this law scientsts inferred that a stone dropped from a
tow
er situated on a moving earth would be left behind. Thus the
relatvity principle must be combined with a new law of inerta in such
a fashion that the moton of the earth can stll be asserted. One sees at
once
that the following law, the principle ofcircular ineria as I shall call
it, provides the required soluton: an object that moves with a given
angular velocity on a frictonless sphere around the cente of the
earth will contnue moving with the same angular veloity for ever.
16.
'Classical Empiricism', op. cit .
.
1 7.
This is the geeal account of moton. In the smelepml account we have
Circ
ular moton above and up-and-down motons on earh.
74 AGAI NST METHOD
Combining the appearance of the falling stone with the relatvity
principle, the principle of circular inerta and some simple
assumptons concering the compositon of veloites,
1
8 we obtin an
argument which no longer endangers Copericus' view, but can be
used to give it partal support.
The relatvity principle was defended in two ways. The frst was by
showing how it helps Copericus: this defence is a hoc but not
objectonable, because necessar for revealing natural interret
tons. The second was by pointng to its functon in common sense,
and by surrepttously generalizing that functon (see above). No
independent argument was given for its validity. Galileo's support for
the principle of circular inerta is of exactly the same kind. He
intoduces the principle, again not by reference to experiment or to
independent obseraton, but by reference to what everone is
already supposed to know.
Simplica: So you have not made a hundred tests, or even one? And
yet you so freely declare it to be cerain? . . .
Saliat: Without experiment, I am sure that the efect will happen
as I tell you, because it must happen that way; and I might add that you
yourelf also know that it cannot happen otherise, no matter how you
may pretend not to know it . . . . But I am so handy at picking people's
brains that I shall make you confess this in spite of yourself.
1
9
Step by step, Simplicia is forced to admit that a body that moves,
without fricton, on a sphere concentic with the cente of the earth
1 8. These assumptons were not at all a matter of course, but conficted with some
ver basic ideas of Arstotelian physics. The principle of circular inera is related to the
impetus theor, but not identcal with it. The impetus theor retains the idea that it
needs a force to bring about change, but it puts the force inside the changng object.
Once pushed, an object contnues moving in the same way in which a heated object
stays war-both contain the cause of their new state. Galileo moifes this idea in two
ways. First, the circular moton is supposed to g on forever while an object kept
moving by impetus will gradually slow down, just a a heated object, its analogue,
gradually becomes colder. The argument for this modifcaton is given in the text
below; it is purely rhetorical. Secondly, the eteral circular motons must proeed
without a cause: if relatve motons are not operatve, then intoducing a moton with
the same cente and the same angular veloity as a circular motion upheld by impetus
cannot eliminate forces: we are on the way from impetus to momentum (cf. A. Maier,
Dic Ierla/o Calilos im I1. ahrhaad, Rome, 1 949). All these changes are
overlooked by those who assume that the tansiton was the simple result of a new and
better dynamics and that the dynamics was already available, but had not yet been
applied in a detennined way.
1 9. Dmlegc,op. cit., p. 145.
SEVEN 75
will carr out a 'boundless', a 'peretual' moton. We know, of
course, especially after the analysis we have just completed of the
non-operatve character of shared moton, that what Simplicio
accepts is based neither on experiment nor on corroborated theory. It
is a daring new suggeston involving a temendous leap of the
imaginaton.
20
A little more analysis then shows that this suggeston
is connected with experiments, such as the 'experiments' of the
Dicori
2
1
by a hoc hypotheses. (The amount of fricton to be
eliminated follows not from independent investgatons - such
investgatons commence only much later, in the 1 8th century - but
from the result to be acheved, v. the circular law of inerta.)
Viewing natural phenomena in this way leads to a re-evaluaton of all
experience, as we have seen. We can now add that it leads to the
inventon of a nw kn ofeerece that is not only more sophstcated
but alo far more speclatie than the experience of Aristotle or of
20. For a Coperican the only leap involved was the identficaton of the earh a a
celestal object. According to Aristotle celestal objects move in circles and 'a boy that
moves in a circle ha neither heaviness nor lighmess for it cannot change its distance
from the cente, neither in a natural nor in a forced way'. DcCecle,269b34f.
21 . Incidentally, many of the 'expriences' or 'experiment' used in the argments
about the moton of the earth are entrely ficttous. Thus Galileo, in his 1ratlatedlm
5/oa(Qoc,Vol. II, pp. 21 I f), which 'follows the opinion of Aristotle and of Ptolemy'
(p. 223), uses this argment against a rotation of the earth: ' . . . object which one lets
fall from high places to the gound such a a stone from the top of a tower would not fall
towards the fot of that tower; for during the tme which the stone coming rectlinearly
towards the gound, spends in the air, the earth, escaping it, and moving towards the
east would receive it in a par far removed from the foot of the tower iaoanthcsamc
maaaoia hichasteacthatudp cd/mmthcmmte/arapid mmiagshipI/ aet/all
temard its/ot, bat merc teard thc stm' (p. 224). The italicized reference to the
behaviour of stones on ships is again used in the Dialegc(p. 1 26), when the Ptolemaic
arguments are discussed, but it is no longer accepted a correct. 'It seems to be an
appropriate tme,' says Salviat (ibid., p. 1 80), 'to take notce of a cerain generosity on
the par of the Copericans towards their adveraries when, wth perhaps to much
liberality, they concede as te and corect a number of experiments which their
opponents have never made. Such for example is that of the boy falling from the mast
of a ship while it is in moton . . . . ' Earlier, p. 1 54, it is implied rather than obsered, that
the stone will fall to the fot of the mast, even if the ship should be in moton while a
possible experiment is discussed on p. 186. Brno (LaCadlcCm,QocIlaliaac,
I, ed. Giovanni Gentle, Bari, 1907, p. 83) takes it for granted that the stone warrive
at the
foot of the mast. It should be noted that the problem did not readily lend itelf to
an
experimental soluton. Experiments were made, but their results were far from
conclusive. Cf. A. Armitage, 'The Deviaton of Falling Boies, Zaaa e/5ci, 5,
1941 -7, pp. 342f, and A. Koyt, MctqhsicaadMcmart,Cambridge, 1968, pp.
89f.
The tower argment can be found in Aristotle, DcCecle, 296b22, and Ptolemy,
5otis, i, 8. Copericus discusses it in the same chapter of DcRnel,but tes to
defuse it in the next chapter. Its role in the Middle Ages is described in M. Clagett, 1hc
5ci
cce/McchaaiciathcMidlcZgc,Madison, 1959, Chapter 10.
76 AGAI NST METHOD
common sense. Speaking paradoxcally, but not incorrectly, one may
say that Galeo itets an eerence that has metaphysical ingediets. It
is by means of such an exerience that the tansiton from a geostatc
cosmolo to the point of view of Copericus and Kepler is
achieved. 2
22. Alan Chalmers, in an interestng and well-argued paper ('The Galileo That
Feyerabend Missed: An Improved Case Against Method' inJ.A. Schuster and R.R.
Yeo (eds), 1hcFelitio aadRhctence/5citcMcthed, Dordrecht, 1 986, pp. I1),
distnguishes 'between Galileo's contbutons to a new science, on the one hand, and
the queston of the soial conditons in which that science is developed and practised,
on the other', admits that 'propaganda' (though much less than I suggest) may have
been part of his attempt to change the latter, but emphasizes that it does not affect the
forer. 'The main source for Galileo's contibuton to science itself, says Chalmer,
'is his 1mNm5cicc'.This is the work Ishould have studied to explore Galileo's
proedure. But the 1meNm5ciccdo not deal with the topic I was disussing, v. the
tansiton to Copericus. HocGalileo used proedures rather different from those of
his later work. Lynn Thordike, who shares Chalmer' evaluaton of the Dialeguc,
wished that Galileo had written a systematc textbok on that subject (Huteq e/
MapcaadEmmtal5occ,Vol. 6, New York, 1 941 , pp. 7 and 62: 'Galileo might
have done better to write a systematc textbook than his provoatve dialogues'). Now
for such a textbook to have substance it would have to b as general as its Aristotelian
rival and it would have to show how and why Aristotelian concepts needed to be
replaced at the most elementar level. Aristotelian concepts, though abstact, were
closely related to common sense. Hence it was necessar to replace some common
notons by other (I am now speaking about what Chalmer calls 'perceptual relatvit'
-p. 7). Two questons arise: how big were the changes? and was propaganda (rhetoric,
were 'iratonal moves') needed to carr them out? My answer to the latter queston is
that discoure atemptng to bring about major conceptual changes is a noral part of
science, common sense, and cultural exchange (for the latter cf. Chapter 1 6 and
Chapter 1 7, item vi, 'open exchange'), and that it differ from the discoure carried out
mithiaa more or less stable framework. Peronally, I am quite prepared to make it part
of ratonalit. But there exst philoophical shols that oppse it or call it incoherent
(cf. Chapter 1 0 of FarmcllteRcmwhich discusses some of Hilar Putam's views).
Usiagthctmiaele e/thccscheeh I speak ofGalileo's 'tcker', etc. And I add that
sience contains ingedients that ocasionally need such 'tcker' to become
acceptable. The difference between the 5cicc and the Dialeguc, therefore, is not
between sience and soiolog but between technical changes in a narow feld and
basic changes, realistically interpreted. My answer to the frt queston is that
perceptual relativity, though acknowledged by many scholar (and by Aristotle
himself), was not a common possession (Galileo points out that even some ofhis fellow
scientsts stumbled at this pint) and thus had to be argued for. This is not at all
surprising, as my discussion of qualitatve difcultes in Chapter 5 shows. Besides, is it
really te that a taveller on a boat sees the harbour as receding as if it were removed
by some stange force? I conclude that Galileo's 'tcker' was necessr for a proper
understanding of the new cosmolog, that it is 'tcker' only for philosophies that set
narow conditons on conceptual change and that it should be extended to areas stll
restcted by such conditons (in Chapter 1 2 1 argue that the mind-boy problem is one
such area).
8
In
tition to natural interretations, Galileo also change sensatons that
see to ednger Coeicus. He amits that there are such sesations, he
praie Coeics fr haing disregardd the, he caims to hae reoed
the with the help ofthe telescope. Howee, he off er no theoretcal
reaons why the telecoe should be eeced to ge a tre picure ofthe sk.
I repeat and summarize. A argument is proposed that refutes
Copercus by obseraton. The argument is inverted in order to
discover the natural interpretatons which are responsible for the
contadicton. The ofensive interpretatons are replaced by others,
propaganda and appeal to distant, and highly theoretcal, parts of
common sense are used to defuse old habits and to enthrone new
ones. The new natural interpretatons, which are also formulated
explicitly, as auxliary hypotheses, are established partly by the
support they give to Copericus and partly by plausibilit considera
tons and a hoc hypotheses. A entrely new 'experience' arises in
this way. There is as yet no independent evidence, but this is no
drawback; it takes tme to assemble facts that favour a new
cosmolog. For what is needed is a new dynamics that explains both
celestal and terrestial motons, a theory of solid objects, aero
dynamics, and all these sciences are stll hidden in the future.
1
But
their task i now wel-dned, for Galileo's assumptons, his a hoc
hypotheses included, are sufciently clear and simple to prescribe
the directon of future research.
Let it be noted, incidentally, that Galileo's procedure drastcally
reduces the content of dynamics. Aristotelian dynamics was a general
I . Galileo's circular law is not the right dynamics. It fts neither the epicycles
which stll ocur in Copericus, nor Kepler's eUipses. In fact, it is refted by both. Stll,
Galileo regards it a an essental ingredient of the Coperican pint of view and ties to

emo
ve
boies, such as comets, whose moton quite obviously is not circular, fom
Int
erl
anetar space. In hisAsse'Galileo talked about comets [and interreted them
as iUusions, similar to rainbows) in order to protect the Coprican system from
possible falsifcatons.' P. Redondi, CalilceHoctic,Princeton, 1987, pp. 145, 3 1 .
77
78 AGAI NST METHOD
theor of change, comprising locomoton, qualitatve change,
generaton and corrupton, and it could also be applied to mental
processes. Galileo's dynamcs and its successors deal with locomotion
only, and here again just with the loomoton of matter. Other
kinds of moton are pushed aside with the promssor note (due to
Demoritos) that locomoton will eventually be capable of explaining
all moton. Thus a comprehensive empirical theor is replaced by a
narrow theor plus a metaphysics of moton/ just as an 'empirical'
experience is replaced by an experience that contains speculatve
elements. Counternduaion, however, is now seen to play an important
role both vis-a-vis theories and vi-a-vis facts. It clearly aids the
advancement of science. This concludes the consideratons begun in
Chapter 6. I now tum to another part of Galileo's propaganda
campaign, dealing not with natural interretatons but with the
sesor core of our obseratonal statements.
Replying to an interlocutor who expressed his astonishment at the
2. The so-called scientifc revoluton led to atounding disoveries and
considerably extended our knowledge of physics, physiolog, and astonomy. This was
achieved by pushing aside and regarding as irelevant, aade/maea-mtt,those
facts which had suppned the older philosphy. Thus the evidence for witchcraft,
demonic pssession, the existence of the devil, etc., wa disregarded tqethowith the
'superttions' it once confred. The result was that 'towards the close of the Middle
Ages science was forced away from human psycholog, so that even the great
endeavour of Erasmus and his friend Vives, as the best representatves of humanism,
did not sufce to bring about a reapproachment, and psychopatholog had to tail
centuries behind the developmental tend of general medicine and surger. As a
matter of fact . . . the divorcement of medical sience from psychopatholog wa so
defnite that the latter was always totally relegated to the domain of theolog and
ecclesiastc and civil law - two felds which naturally became funher and funher
removed from medicine . . . .' G. Zilboorg, MD, 1hcMcdiml Maa aad thc itch,
Baltmore, 1 935, pp. 3f and 70f. Astonomy advanced, but the knowledge of the
human mind slipped back into an earlier and more primitve stage. Another example is
astolog. 'In the early stages of the human mind,' writes A. Comte (Cean d
Fhiles@hicFesitnc, Vol. III, pp. 273-80, ed. Litte, Paris, 1 836), 'these connecting
links between astonomy and biolog were studied from a ver different point of view,
batat lcmtthey were studied and not lef out of sight, as is the common tendency in our
ow tme, under the restcting influence of a nascent and incomplete positvism.
Beneath the chimerical belief of the old philosophy in the physiological influence of
the stars, there lay a stong, though confused recogniton of the truth that the facts of
life were in some way dependent on the solar system. Like all primitive inspiratons of
man's intelligence this feeling needed rectfcaton by psitve sience, but not
destucton; thoug unhappily in sience, as in plitcs, it is often hard to reorganize
without some brief period of ovenhrow.' A third area is mathematcs. Aristotle had
developed a highly sophistcated theor of the contnuum that overcame the
difcultes raised by Zeno and antcipated quantum theoretcal ideas on moton (see
footote I5and text of Chapter 5). Most physicists retured to the idea of a contnuum
consistng of indivisible elements - if they considered such recondite matters, that is.
EI GHT 79
small number of Copemicans, Salviat, who 'act[s] the part of
Copericus', 3 gives the following explanaton: 'You wonder that
there are so few followers of the Pythagorean opinion [that the earth
moves] while I am astonished that there have been any up to this day
who have embraced and followed it. Nor can I ever suffciently
admire the outstanding acumen of those who have taken hold of this
opinion and accepted it as true: they have, through sheer force of
intellect, done such violence to their own senses as to prefer what
reason told them over that which sensible experience plainly showed
them to be the contary. For the arguments against the whirling [the
rotaton] of the earth we have already examined [the dynamical
arguments discussed above] are very plausible, as we have seen; and
the fact that the Ptolemaics and the Aristotelians and all their
disciples took them to be conclusive is indeed a strong argument of
their efectveness. But the experiences which overtly contadict
the annual movement [the movement of the earth around the sun] are
indeed so much greater in their apparent force that, I repeat,
there is no limit to my astonishment when I reflect that Aristarchus
and Copercus were able to make reason so conquer sense that
in defance of the latter, the former became mistess of their belief. '
;
A little later Galileo notes that 'they e Copericans] were
confdent of what their reason told them!' And he concludes his
brief account of the origins of Copericanism by saying that 'with
reason as his guide he [Copericus] resolutely contnued to afrm
what sensible experience seemed to contadict'. 'I cannot get over
my amazement', Galileo repeats, 'that he was constantly willing to
persist in saying that Venus might go around the sun and might
be more than six tmes as far from us at one tme as at another,
and stll look always equal, when it should have appeared forty tmes
larger.'
6
The 'experiences which overtly contradict the annual movement',
and which 'are much greater in their apparent force' than even the
dynamical arguments above, consist in the fact that 'Mars, when it is
close to us . . . would have to look sixty tmes as large as when it is most
distant. Yet no such difference is to be seen. Rather, when it is in
oppositon to the sun and close to us it shows itself only four or five
3. Dialegc,op. cit., pp. I 3I and Z56.
4. ibid., p. 3Z8. At other times Galileo speaks much more belligerently and
dog
matic
ally, and apparently without any awareness of the difculties mentioned here.
Cf.
his
preparatory notes for the letter to Grand Duchess Christina, Qoc, V, pp.
3
67d.
5.
ibid.,
p. 335.
6.
ibid. , p. 339.
80 AGAI NST METHOD
tmes as large as when, at conjuncton, it becomes hidden behind the
rays of the sun.'
7
'Another and greater difculty is made for us by Venus which, if it
circulates around the sun, as Copericus says, would now be beyond
it and now on this side of it, receding from and approaching towards
us by as much as the diameter of the circle it describes. Then, when it
is beneath the sun and very close to us, its disc ought to appear to us a
little less than forty tmes as large as when it is beyond the sun and
near conjuncton. Yet the diference is almost imperceptble.'
In an earlier essay, The Assayer, Galileo expressed himself stll
more bluntly. Replying to an adversary who had raised the issue of
Copercanism he remarks that 'neither Tycho, nor other astronome
nor ee Coeics culd cearl reute [Ptoley] inasmuch as a most
important argument taken from the movement of Mars and Venus
stood always i their way'. (This 'argument' is mentoned again in the
Dialoge, and has just been quoted.) He concludes that 'the two
systems' (the Coperican and the Ptolemaic) are 'surely false'.
8
We see that Galileo's view of the origin of Copericanism difers
markedly from the more familiar historical accounts. He neither
points to ne facs which ofer inductve supor to the idea of the
moving earth, nor does he menton any obseratons that would reute
the geocentic point of view but be accounted for by Copercanism.
On the contary, he emphasizes that not only Ptolemy, but
Copericus as well, is refuted by the facts,
9
and he praises
Aristarchus and Copericus for not having given up i the face of
7. ibid., p. 334.
8. 1hc Zsso,quoted from 1hcCeatmmeathcCemctse/I9I,op. cit., p. 1 85.
9. This refer to the perod before the end of the 1 6th centur; cf. Derek). de S. Prce,
'Conta-Copericus: A Critcal Re-Estmaton of the Mathematcal Planetar Theor
ofPtolemy, Copericus and Kepler', in M. Clagett (ed.), CntimlFmblsiathc Huteqe/
5occ,Madison, 1 959, pp. 1 97-21 8. Price deals only wth the kiaaticand the @timl
difcultes of the new views. (A consideraton of the dynamical difcultes would frther
stengthen his case.) He points out that 'under the best conditons a geostatc or
heliosttc system using eccentric circles (or their equivalent) wth cental epiccles can
account for all angular motons of the planets to an accurac better than 6' . . . exceptng
only the special theor needed to account for . . . Mercur and exceptng also the planet
Mars which shows deviatons up to 30' from a theor. [This is) certainly better than the
accurac ofl 0' which Copericus himself stted as a satsfactor goal for his ow theor'
which was difcult to test, especially in view of the fact that refracton (almost I o on the
horizon) was not taken into account at the tme of Copericus, and that the obseratonal
basis of the predictons was less than satsfactor.
Carl Schumacher (Uatmachaag bodicptelm uchc1hcencd aatoFmact,
Minster, 1 91 7) has found that the predictons concering Mercur and Venus made
by Ptolemy differ at most b an amount of 30' from those of Copericus. The
deviatons found beteen moem predictons and those of Ptolemy (and Copericus),
EI GHT 81
such tremendous difcultes. He praises them for having proceeded
counternduaie/.
This, however, is not yet the whole story.
For while it might be conceded that Copericus acted simply on
faith, it may also be said that Galileo found himself in an entrely
different positon. Galileo, after all, invented a new dynamics. And he
invented the telescope. The new dynamics, one might want to point
out, removes the inconsistency between the moton of the earth and
the 'conditons affectng ourselves and those in the air above us'. 1
0
And
the telescope removes the 'even more glaring' clash between the
changes in the apparent brightess of Mars and Venus as predicted on
the basis of the Coperican scheme and as seen with the naked eye.
This, incidentally, is also Galileo's own view. He admits that 'were it
not for the exstence of a superior and better sense than natural and
common sense to join forces with reason' he would have been 'much
more recalcitrant towards the Coperican system' Y The 'superior
and better sense' is, of course, the telecoe, and one is inclined to
remark that the apparently counterinductve procedure was a a
matter of fact inducton (or conjecture plus refutaton plus new
conjecture), but one based on a bette eerece, containing not only
better natural interpretatons but also a better sensory core than was
available to Galileo's Aristotelian predecessors.
1 2
This matter must
now be examined in some detail.
The telescope is a 'superior and better sense' that gives new and
more reliable evidence for judging astronomical matters. How is this
hypothesis examined, and what arguments are presented in its favour?
In the Sideus Nuncius, 13 the publicaton which contains his first
which in the case of Mercur may be as large as 7,are due mainly to wrong constants
and inital conditons, including an incorrect value of the constant of precession. For
the versatility of the Ptolemaic scheme cf. N.R. llanson, lsis, No. 5 ! , ! 960,pp. I 508.
!0. Ptolemy, 5atmis,i, 7.
I I . Dialegac,op. cit., p. 3Z8.
! Z. For this view cf. Ludovico Geymonat, Calilm Calilci, transl. Stllman Drake,
New York, I 965(frt Italian edition ! 957),p. I 8+.For the stor ofGalileo's inventon
and use of the telescope cf. R.S. Westfall, 'Science and Patronage', Isis,Vol. 76! 985,
pp. I I f. According to Westfall, Galileo 'saw the telescope more as an instument of
pato
nage than as an instument of astonomy' (p. Z6)and had to be pushed into some
ast
onomical applicatons by his pupil (and staunch Coperican) Castelli. Galileo's
tele
scopes were beter than other in circulation at the time and were much in demand.
But he frst satisfed the demands of potental patrons. Kepler, who complained about
the quality of telescopes (cf. next chapter, fotnote Z! and text) and who would have
lov
ed to possess a better instument, had to wait.
! 3.
!hc 5idcalMcsgo e/Calilce Calihl, transl. E. St Carlos, London, I 880,
reis
sued by Dawsons of Pall Mall, ! 960,p. I 0.
82 AGAI NST METHOD
telescopic obseratons, and which was also the first important
contibuton to his fame, Galileo writes that he 'succeeded (in
building the telescope) through a deep study of the theor of
refracton'. This suggests that he had theoreticl reaon for preferring
the results of telescopic obseratons to obseratons with the naked
eye. But the partcular reason he gives -his insight into the theor of
refracton - is not co"ec and is not sufcet either.
The reason is not correct, for there exst serious doubts as to
Galileo's knowledge of those pars of contemporar physical optcs
which were relevant for the understanding of telescopic phenomena.
In a letter to Giuliano de Medici of 1 October 161 0, H more than
half a year afer publicaton of the Sidreu Nuncus, he asks for a copy
of Kepler's Otic of 1 604,
1 5
pointng out that he had not yet been
able to obtain it in ltaly. Jean Tarde, who in 1 61 4 asked Galileo about
the constucton of telescopes of pre-assigned magnificaton, repors
in his diar that Galileo regarded the matter as a difcult one that he
had found Kepler's Otic of 161 1
16
so obscure 'that perhaps its own
author had not understod it' P In a letter to Licet, written two
years before his death, Galileo remarks that as far as he was
concered the nature of light was stll in darkness.
18
Even if we
consider such utterances with the care that is needed in the case of a
whimsical author like Galileo, we must yet admt that his knowledge
of optcs was inferior by far to that of Kepler.
1 9
This is also the
1 4. Galileo, Qoc,Vol. X, p. 441 .
I S. ZdIitcieaFaralipemaqaibasZstmaemimFanQtica1rmitar,Frankfun,
1 604, to be quoted from ehaaacKqlo, Ccammcltcokc,Vol. II, Munich, 1 939, ed.
Franz Hammer. This partcular work will be referred to as the 'optcs of 1 604'. 1t was
the only useful optcs that existed at the tme. The reason for Galileo's curiosit was most
likely the many references to this work in Kepler's reply to the 5idnNaaoas.For
the histor of this reply as well as a tanslaton cf. Kqlo'sCemmatimith Calilm's
5idcalMcsgo,transl. E. Rosen, New York, 1 965. The many references to earlier
work contained in the Cemmatieawere interreted by some ofGalileo's enemies as a
sign that 'his mask had been tom from his face' (G. Fugger to Kepler, 28 May 1 61 0,
Galileo, Qoc,Vol. X, p. 361) and that he (Kepler) 'had well plucked him', Maestlin to
Kepler, 7 August (Galileo, Qoc,Vol. X, p. 428). Galileo must have received Kepler's
Cemmatiea before 7 May (Qoc, X, p. 349) and he acknowledges receipt of the
printed Cemmatieain a letter to Kepler of 1 9 Augst (Qcrc,X, p. 421).
1 6. Di@tncc, Augsburg, 1 61 1 , okc, Vol. IV, Munich, 1 941 . This work was
written after Galileo's discoveries. Kepler's reference to them in the preface has been
translated by E. St Carlos, op. cit., pp. 37, 79f. The problem referred to by Tarde is
teated in Kepler's Di@tncc.
1 7. Geymonat, op. cit., p. 37.
1 8. Letter to Licet of 23 June 1 64. Qoc,VIII, p. 208.
1 9. Kepler, the most knowledgeable and most lovable of Galileo's contem
praries, gives a clear account of the reasons why, despite his superior knowledge of
EI GHT 83
conclusion of Professor E. Hoppe, who sums up the situaton as
fol
lows:
Galileo's asserton that having heard of the Dutch telescope he
reconstucted the apparatus by mathematcal calculaton must of
course be understod with a grain of salt; for i his wrtngs we do not
fnd any calculatons and the reprt, by letter, which he gives of his
frt efort says that no better lenses had been available; six days later
we fnd him on the way to Venice with a better piece to hand it a a gif
to the Doge Leonardi Donat. This dos not look like calculaton; it
rather loks like tal and eror. The calculaton may well have been of
a diferent kind, and here it succeeded, for on 25 August 1609 his
salar was increased by a factor of three.
20
Trial and error-this means that 'in the case of the telescope it was
eerece and not mathematcs that led Galileo to a serene faith in the
reliability of his device' .
21
This second hypothesis on the origin of
the telescope is alo supported by Galileo's testmony, in which he
writes that he had tested the telescope 'a hundred thousand tmes on
a hundred thousand stars and other objects'. 22 Such tests produced
great and surprising successes. The contemporar literature -
letters, books, gossip columns - testfies to the exaordinar
impression which the telescope made as a means of improving
ter etral vision.
optcal matter, he 'refrained from attemptng to constct the device'. 'You, however,'
he addresses Galileo, 'desere my praise. Puttng aside all misgivings you tured
direcdy to visual experimentaton' ( Cemmatiea,op. cit., p. I 8).It remains to add that
Galileo, due to his lack of knowledge in optics, had no 'misgivings' to overcome:
'Galileo . . . was totally ignorant of the science of optcs, and it is not to bold to assume
that this was a most happy accident both for him and for humanity at large', Ronchi,
5citqcChaage,ed. Crombie, London, I 963,p. 550.
Z0. DieCcchichtedQtik,Leipzig, I9Z6,p. 3Z.Hoppe's judgement concering
the inventon of the telescope is shared by Wolf, Zinner and others. Huyghens points
out that superhuman intelligence would have been needed to invent the telescope on
the basis of the available physics and geomet. Afer all, says he, we stll do not
understand the workings of the telesope. ('Dioptica', Hagii Qasmm Festama,
Ludg. Bat., I 903, I 63,paraphrased afer A.G. Kaster, CcchichtedMathatik,
Vol. IV, Gottngen, I 800,p. 60.)
Z! . Geymonat, op. cit., p. 39.
ZZ. Letter to Carioso, Z+May I 6I 6,Qere, X, p. 357.letter to P. Dini, I Z May
! 6
I I ,
Qcrc,IX, p. I06.'Nor can it be doubted that I, over a period of two years now,
h
ave
tested my instment (or rather dozens of my instments) on hundreds and
th
ous
ands of objects near and far, large and small, bright and dark; hence I do not see
ho
w it
can enter the mind of anyone that I have simple-mindedly remained deceived in
my
ob
seratons.' The hundreds and thousands of expriments remind one of
Hoke, and are most likely equally spurious. Cf. fotote 9of Chapter 9.
84 AGAI NST METHOD
Julius Caesar Lagalla, Professor of Philosophy in Rome, describes
a meetng of 1 6 April 1 61 1 , at which Galileo demonstated his
device: 'We were on top of the Janiculum, near the city gate named
after the Holy Ghost, where once is said to have stood the villa
of the poet Martal, now the property of the Most Reverend Malvasia.
By means of this instrument, we saw the palace of the most illus
trious Duke Altemps on the Tuscan Hills so distnctly that we
readily counted its each and every window, even the smallest;
and the distance is sixteen Italian miles. From the same place
we read the letters on the galler, which Sixtus erected in the Lateran
for the benedictons, so clearly, that we distnguished even the
periods cared between the letters, at a distance of at least two
miles.'
23
Other repors confrm this and similar events. Galileo himself
points to the 'number and importance of the benefts which the
instument may be expected to confer, when used by land or sea'.
24
23. Legalla, DephmemmuiaemelamamitclciimaaD.CalihCaliloaaac
itmm sascimtu phsica dupatatie (Venice, I 61 2), p. 8,quoted from E. Rosen, Oe
Namiage/the1elcc@e, New York, I 947, pl. 54. The regular reports (mist} of the
Duchy of Urbino on events and gossip in Rome contain the following notce of the
event: 'Galileo Galilei the mathematcian, arrived here from Florence before Easter.
Formerly a Professor at Padua, he is at present retained by the Grand Duke of
Tuscany at a salar of I ,0 scudi. He has obsered the moton of the star with the
e i, which he invented or rather improved. Against the opinion of all ancient
philosopher, he declares that there are four more star or planets, which are satellites
of Jupiter and which he calls the Medicean bodies, as well as two companions of
Satur. He has here discussed this opinion of his with Father Clavius, the Jesuit.
Thurday evening, at Monsignor Maavasia's estate outide the St Pancratus gate, a
high and open place, a banquet was given for him by Frederck Cesi, the marquis of
Montcelli and nephew of Cardinal Cesi, who was accompanied by his kinsman, Paul
Monaldeso. In the gatherng there were Galileo; a Fleming named Terentus;
Perio, of Cardina Cesi's retnue, [La) Gala, Professor at the Univerity here; the
Greek, who is Cardinal Gonzga's mathematician; Piffari, Professor at Siena, and as
many as eight others. Some of them went out expressly to perfor this obseraton,
and even though they stayed untl one o'clok in the moring, they stll did not reach an
agreement in their views' (quoted from Rosen, op. cit., p. 3 1).
24. 5ideal Mcsgo, op. cit., p. ii. According to Berellus (De Im 1ehii
Immterc, Hague, 1 655, p. 4), Prince Morit immediately realized the militar value of
the telescope and ordered that its inventon - which Berellus attibutes to Zacharas
Jansen - be kept a secret. Thus the telescope seems to have commenced as a secret
weapon and was tured to astonomical use only later. There are many antcipatons of
the telescope to be found in the literature, but they mosdy belong to the domain of
natural magic and are used accordingly. An example is Agrippa von Netesheim, who,
in his bok on ocult philosophy (writen 1 509, Book II, chapter 23), writes 'et ego novi
ex illis miranda confcere, et specula in qui bus quis videre poterit quaecunque voluerit
a longissima distanta'. 'So may the toy of one age come to be the precious teasure of
another', Henr Morley, OeLqe/CemelimZgnyavo Ncnchcim,Vol. II, p. 166.
EI GHT 85
The terestral success of the telescope was, therefore, assured.
Its applicaton to the star, however, was an entrely diferent
matter.
9
Nor de the initial experience with the telescoe provid such reasons. The
frt telecoic obserations of the sk are inditina, indterinate,
contradiaor and in confia with what everone cn see with his unaidd
ees. And the onl theor that could have helped to searate telescoic
illusions from verdical phenomena was reuted b simple tests.
To start with, there is the problem of telescopic vision. This problem
is diferent for celestal and terrestrial objects; and it was also thought
to be diferent in the two cases.
1
It was thought to be diferent because of the contemporar idea
that celestal objects and terrestial objects are formed from dif
ferent materials and obey different laws. This idea entails that
the result of an interacton of light (which connects both domains
and has special propertes) with terrestrial objects cannot, without
further discussion, be extended to the sky. To this physical idea
one added, entrely in accordance with the Aristotelian theor
of knowledge (and also with present views about the matter), the
idea that the senses are acquainted with the close appearance of
terrestial objects and are, therefore, able to perceive them distnctly,
even if the telescopic image should be vastly distorted, or disfgured
by coloured fringes. The stars are not known from close by. 2 Hence
we cannot in their case use our meor for separatng the
contibutons of the telescope and those which come from the object
I . This is hardly ever realized by those who argue (with Kastner, op. cit., p. I 33)
that 'one does not see how a telescope can be good and useful on the earth and yet
deceive in the sky'. Kastner's comment is directed against Horky. See below, text to
footnotes 9I 6of the present chapter.
Z. That the senses are acquainted with our everday suroundings, but are liable
to give misleading reports about objects outside this domain, is proved at once by
the aycaraacce/thcmeea.On the earth large but distant objects in familiar suround
ings, such as mountains, are seen as being large, and far away. The appearance
of the mon, however, gives us an entirely false idea of its distance and its
size.
86
NI NE 87
itself.
3
Moreover, all the familiar cues (such as background, overlap,
knowledge of nearby size, etc.), which consttute and aid our vision
on the surface of the earth, are absent when we are dealing with the
sky, so that new and surrising phenomena are bound to occur. 4 Only
a new theory of vision, containing both hypotheses concering the
behaviour of light within the telescope and hypotheses concering
the reacton of the eye under exceptonal circumstances, could have
bridged the gulf between the heavens and the earth that was, and stll
is, such an obvious fact of physics and of astonomical obseraton. 5
We shall soon have occasion to comment on the theories that were
available at the te and we shall see that they were unft for the task
and were refuted by plain and obvious facts. For the moment, I want
to stay with the obseratons themselves and I want to comment on
the contadictons and difcultes which arise when one ties to take
the celestal results of the telescope at their face value, as indicatng
stable, objectve propertes of the things seen.
Some of these difcultes already announce themselves in a report
of the contemporary Avi
l
which ends with the remark that 'even
though they (the partcipants in the gathering described) went out
expressly to perform this obseraton (of"four more stars or planets,
which are satellites of Jupiter . . . as well as of two companions of
Satur"7), and even though they stayed untl one in the moring,
they stll did not reach an agreement in their views. '
3. It i s not to difcult t o separate the letter of a familiar alphabet from a
background of unfamiliar lines, even if they should happen to have been written with
an almost illegible hand. No such separation is possible with letter which belong to an
unfamiliar alphabet. The pars of such letter do not hang together to for distnct
patters which stand out from the background of general (optca) noise (in the manner
described by K. Koflka, Fqchel. a., 19, 1922, pp. 55l iT, partly reprinted in M.D.
Veron (ed.), Eqmmts ia lisaalFocqtiea, London, 1966; cf. also the arcle by
Gottschaldt in the same volume).
4. For the imporance of cues such as diaphragms, crossed wires, background,
etc., in the localizaton and shape of the telescope image and the stange situatons
arising when no cues are present cf. Chapter I of Ron chi, Qtic,op. cit., espcialy
pp. l S I , 1 74, 1 89, 1 91 , etc. Cf. also R.L. Gregor, Ecaadraia,New York, 1966,
]assimand p. 99 (on the autokinetic phenomenon). F.P. Kilpatick (ed.), Eqleratieasia
!r
aasaie
aalFqchele,New York, 1961, contains ample material on what happens in
the absence of familiar cues.
5. It is for this reason that the 'deep study of the theor of refracton' which Galileo
pre
ten
ded to have carried out (text to footote 13 of Chapter 8) would have been quite
tmj
otfor establishing the usefulness of the telescope; cf. also fotote 16 of thl
present chapter.
6.
Details in Chapter 8, fotote 23.
7. This is how the ring of Satur was seen at the tme. Cf. also R.L. Gregor, Oc
Ia
tc
ig
tEc,p. 1 1 9.
88 AGAI NST METHOD
Another meetng that became notorious all over Europe makes the
situaton even clearer. About a year earlier, on 24 and 25 April 1610,
Galileo had taken his telescope to the house of his opponent, Magini,
in Bologna to demonstrate it to twenty-four professors of all facultes.
Horky, Kepler's overly-excited pupil, wrote on this occasion;
8
'I never
slept on the 24th or 25th April, da or night, but I tested the instrument
of Galileo's in a thousand ways, both on things here below and on
those above. Below it work wondrull; in the heavens it deceives one,
as some fed stars [Spica Virginis, for example, is mentoned, as well
as a terrestial fame] are seen double. 10 I have as witesses most
excellent men and noble doctors . . . and all have admitted the
instument to deceive . . . . This silenced Galileo and on the 26th he
sadly lef quite early in the moring . . . not even thanking Magini for
his splendid meal. . . .' Magini wrote to Kepler on 26 May: 'He has
achieved nothing, for more than twenty leared men were present; yet
nobody has seen the new planets distnctly (nemo perfecte vidit); he
will hardly be able to keep them.' 1 1 A few months later (in a letter
signed by Rufni) he repeats: 'Only some with sharp vision were
convinced to some extent.'
1 2
After these and other negatve reports
had reached Kepler from all sides, like a paper avalanche, he asked
Galileo for witesses: 1
3
'I do not want to hide it from you that quite a
few Italians have sent letters to Prague assertng that they could not see
those stars [the moons ofjupiter] with your own telescope. I ask myself
how it can be that so many deny the phenomenon, including those who
use a telescope. Now, ifl consider what occasionally happens to me,
then I do not at all regard it as impossible that a single person may see
what thousands are unable to see . . . e 14 Yet I reget that the
8. Galileo, Qoc,Vol. X, p. 342 (my italics, referring to the diference commented
upon above, between celesta and terestal obserations).
9. The 'hundreds' and 'thousands' of obseratons, tals, etc., which we find here
again are hardly more than a rhetorica flourish (corespnding to our 'I have told you a
thousnd tmes'). They cannot be used to infer a life ofincessant obseraton.
I 0. Here again we have a cae where exeral clues are missing. Cf. Ronchi,
Qtic,op. cit., as regards the appearance of flames, small lights, etc.
I I . Letter of26 May, Qoc,Ill.
1 2. ibid., p. 1 96.
1 3. Letter of9 August 1 61 0, quoted from CasparDck,ehaaacKqloia5cia
nq,Vol. I , Munich, 1 930, p. 349.
1 4. Kepler, who sufered from Polyopia ('instead of a single small object at a great
distance, two or three are seen by those who sufer from this defect. Hence, instead of
a single mon ten or more present themselves to me', Cemmatiea,op. cit., footote
94; cf. aso the remainder of the footote for furher quotatons), and who was familia
r
with Platter's anatomical investgations (cf. S.L. Polyak, 1hcRctiaa,Chicago, 1942,
pp. 1 34f for details and literature), was well aware of the need for a phsiehpm
l
mtiosme/treaemimlebsmatieas.
NI NE 89
confrmaton by others should take so long in turing up . . . .
Therefore, I beseech you, Galileo, give me witesses as soon as
possible . . . . ' Galileo, in his reply of 1 9 Augst, refers to himself, to
the
Duke ofToscana, and Giuliano de Medici 'as weU as many others
in Pisa, Florence, Bologna, Venice and Padua, who, however, remain
silent and hesitate. Most of them are entrely unable to distnguish
Jupiter, or Mars, or even the Moon as a planet . . . . ' 1 5 - not a ver
reassuring state of afairs, to say the least.
Today we understand a little better why the direct appeal to
telescopic vision was bound to lead to disappointent, especially in
the inital stages. The main reason, one already foreseen by Aristotle,
was that the senses applied under abnormal conditons are liable to
give an abnormal response. Some of the older historians had an
iing of the situaton, but they speak negatiel, they ty to exlain
the absece of satsfactor obseratonal reports, the puer of what is
seen in the telescope. 16 They are unaware of the possibility that the
obserers might have been disturbed by strong positie il usions also.
The extent of such illusions was not realized untl quite recently,
mainly as the result of the work of Ronchi and his school.
1 7
Here
sizeable variatons are reported in the placement of the telescopic
image and, correspondingly, in the obsered magication. Some
obserers put the image right inside the telescope making it change
15. Caspar-Dck, op. cit., p. 352.
1 6. Thus Emil Wohlwill, CammaadsciaKamp/ /rdicK@mibauchcLchrc,Vol.
I , Hamburg, 1 99, p. 288, writes: 'No doubt the unpleasnt result were due to the
lack of taining in telescopic obseraton, and the resticted field of vsion of the
Galilean telesope as well as to the absence of any pssibility for changing the ditance
of the glases in order to make them ft the pculiartes of the eyes of the leaed
men . . . . ' A similar judgement, though more dramatcally expressed, is found in Arur
Koesder's5lcewalkm, p. 369.
17. Cf. Ronchi, Qtic, op. cit.: Huteirc d m Lamioc, Pars, 1956; 5tem dl
Caaaechmlc,Vatcan City, 1964; Cntmd Fea&mtidll' Zmtimcdl'Ohim,Rome,
1964; cf. also E. Cantore's summar in Zrmncd'huteirc&so~,Decembr 1966,
pp
.
333f. I would like to acknowledge at this place that Professor Ronchi's
investgatons have greatly influenced m thinking on sientfic meto. For a brief
historical account of Galileo's work cf. Ronchi's arcle in A.C. Crombie (ed.),
5otcChaagc,London, 1963, pp. 542-1 . How little this field is explored becomes
clear from S. Tolansky's bok QtmlIllmieas,London, 1964. Tolansk is a physicist
who in his microsopic research (on crstals and metals) was distacted by one optcal
illusion afer another. He wites: 'This tured our interest to the analysis of other
situatons, with the ultmate unexpected disover that optcal illusions can, and do,
p
lay a ver real par in afectng many daily sientfic obseratons. This wared me to
be on the lokout and as a result I met more illusions than I had bargained for.' The
'illusions of direct vision', whose role in sientfic research is slowly bing redisovered
were well known to mediaeval writer on optcs, who teated them in spcial chapter
of their textbo ks. Moreover, they teated lens-images as pqmehglphenomena, as
90 AGAI NST METHOD
its lateral positon with the lateral positon of the eye, exactly as would
be the case with an after image, or a refex inside the telescope - an
excellent proof that one must be dealing with a 'illusion' .
1 8
Others
place the image in a manner that leads to no magnifcaton at all,
although a linear magnifcaton of over thirty may have been
promised.
1 9
Even a doubling of images can be explained as the
result of a lack of proper focusing. 20 Adding the many imperfectons
of the contemporary telescopes to these psychological diffcultes,
21
one can well understand the scarcity of satsfactory reports and one is
rather astonished at the speed with which the reality of the new
phenomena was accepted, and, as was the custom, publicly
ackowledged. 22 This development becomes even more puzling
results of a misapprehension, for an image 'is merely the appearance of an object
outide its place' as we read in John Pecham (cf. David Lindberg, 'The "Perpectva
Communis" of John Pecham', ZrchncI atmatieaalcd'histeircdscicc, 1 965, p. 51 ,
as weU as the last paragraph of Propositon iVl 9 of Pecham's Fmpcuna Cemmaau,
which is to be found in ehaFccham aadthc5cicc e/Qtic, D. Lindberg (ed.),
Madison, 1 970, p. 1 71).
1 8. Ronchi, Qtic,op. cit., p. 1 89. This may explain the frequendy uttered desire
to lok iasidthe telescope. No such problems arise in the cae of tm ctmlobjects
whose images are reglarly placed 'in the plane of the object' (ibid. , p. 1 82).
1 9. For the magifcaton ofGalileo's telescope cf. 1hc5UcalMcsmgo,op. cit.,
p. I I , cf. also A. Sonnefeld, 'Die Optschen Daten der Himmelsfemrohre von Galileo
Galilie',Je Raadchaa,Vol. 7, 1 962, pp. 207f. The old rle 'that the size, psiton
and arngement according to which a thing is seen depends on the size of the angle
through which it is seen' (R. Grosseteste, DcInd, quoted from Crombie, Rebm
Cmssnctc,Oxord, 1 953, p. 120), which gos back to Euclid, is almestalmsm g.l
stU remember my disappointent when, having built a reflector with an aUeged linear
magifcaton of about I SO, l found that the mon wa only about fve tmes enlarged,
and situated quite close to the oular (1937).
20. The image remains shar and unchaged over a considerable interal - the ,
lack of fousing may show itself in a doubling, however.
21 . The frt usable telescope which Kepler received from Elector Erst of Koln
(who in tum had received it from Galileo), and on which he baed his Nanatmd
eHmatuascqmumr msatcllibas,Frankfr, 1 61 1 , showed the star as sqmmand
intensely harcd(Cc. mc,I, p. 461). Erst von Koln himself wa unable to see
anything with the telesop and he aked Clavius to send him a better instment
(rcbniedlmFeatm Uanmita Crqemaa,530, f l 82r). Francesco Fontana, who
from 1 643 onwards obsered the phaes of Venus, notes an unevenness of the
boundar (ad infer mountains), cf. R. Wolf, Ccchichtcd Zstmaemic,Munich, 1 877,
p. 398. For the idiosyncrasies of contemporar telescops and descriptve literature cf.
Erst Zinner, Dmcchc aad Nicdm aduchcZstmaemuchc Iastmmtcd II bu I.
ahrhaad, Munich, 1 956, pp. 21 6-21 . Refer also to the author catalogue in the
second pan of the book.
22. Father Clavi us (letter of l 7 December 1 61 0, Qoc,X, p. 485), the astonomer
of the pwerful Jesuit Collegium Romanum, praises Galileo a the frt to have
obsered the mons of jupiter and he recognizes their realit. Magini, Grienberger,
and other son followed suit. It is clear that, in doing so, they did not proeed
NI NE 91
when we consider that many reports of even the best obserers were
either plainly fhe, and capable of being shown as such at the tme, or
else se/contraiaor.
Thus Galileo reports unevenness, 'vast protuberances, deep
chasms, and sinuosites'
23
at the inner boundar of the lighted part
of the moon while the outer boundary 'appear[s] not uneven, rugged,
and irregular, but perfectly round and circular, as sharly defned as
if marked out with a pair of compasses, and without the indentatons
of any protuberances and cavites'.
24
The moon, then, seemed to be
acco
rding to the methods prescribed by their own philosophy, or else they were ver
lax in the investgaton of the mater. Professor McMullin (op. cit., fotote 32) makes
much of this quick acceptance of Galileo's telescopic obseratons: 'The regular
periods obsered for the satellites and for the phases of Venus strongly indicated that
they were not arefacts of physiolog or optcs. There was surely no need for "auxliar
sciences" . . . ' 'There was no need for auxliar sciences,' wites McMullin, while using
himself the unexamined auxliar hypthesis that astonomical events are distn
guished from physiological events by their regularity and their interubjectvity. But
this hypothesis is /he, as is shown by the moon illusion, the phenomenon of fata
morgana, the rainbow, haloes, by the many microscopic illusions which are so vividly
described by Tolansky, by the phenomena of wtchcraf which surive in our textboks
of psycholog and psychiat, though under a different name, and by numerous other
phenomena. The hypthesis was also Hematebe/aheby Pecham, Witelo, and other
mediaeval scholar who had studied the regular and interubjectve 'illusions' created
by lenses, mirror, and other optcal contivances. In antquity the falseho of
McMullin's hypthesis was cemmeqme.Galileo explicidy disusses and repudiates it
in his book on comets. Thus a new theor of vision was needed, not just to tthe
Galilean obseratons, but also to provide aqmlsfor their astonomical realit. Of
coure, Clavi us may not have been aware of this need. This is hardly surrising. Afer
all, some of his sophistcated 20th-centur successor, such as Professor McMullin,
are not aware ofit either. In additon we must point out that the 'reglar prios' of the
mons of Jupiter were not as well known as McMullin insinuates. For his whole life
Galileo tied to deterine these perios in order to find beter ways of deterining
longitude at sea. He did not succeed. Later on the same problem retured in a
different for when the atempt to deterine the veloit of light wth more than one
moon led to conflictng results. This was found by Cassini shortly afer Romer's
disover-cf. I. B. Cohen, 'Romer and the firt deterinaton of the veloit oflight
(1676)', Isis,Vol. 31 (194), pp. 347f. For the atttude ofCiavius and the scientsts of
the Collegium Romanum cf. the ver interestng book CiaChimby Pasquale M.
d'Elia, S.J . , Cambridge, Mass., 1960. The early obseratons of the astonomer of the
Colle
gium
are contained in their own 'Nuncius Sidereus', Qoe,III/I, pp. 291-8.
23.
!he5idealMcsgo,op. cit., p. 8.
24.
op.
cit., p. 24. - cf. the drawng on page 97 which is taken from Galileo's
p
ubli
caton. Kepler in his Qticof 1604 wites (on the basis of obseratons wth the
un
aide
d
eye): 'It seemed as though something was missing in the circularit of the
ou
nno
st
peripher' (oke, Vol. II, p. 219). He returs to this asseron in his
e
mm
atiea
(op. cit., pp. 2811), critcizing Galileo's telescopic results by what he
him
self
had
seen wth the unaided eye: 'You ask why the mon's outerost circle does
n
ot
.
also
appear
irregular. I do not know how carefully you have thought about this
su
bJe
ct
or whe
ther your quer, as is more likely, is based on popular impression. For in
92 AGAI NST METHOD
full of mountains at the inside but perfectly smooth at the periphery,
and this despite the fact that the peri:hery changed as the result of the
slight libraton of the lunar body. The moon and some of the
planets, such as for example Jupiter, were enlarged while the
apparent diameter of the fied stars decreased: the former were
brought nearer whereas the latter were pushed away. 'The stars,'
writes Galieo, 'fied as well as erratc, when seen with the telescope,
by no means appear to be increased in magnitude in the same
proporton as other objects, and the Moon itself, gin increase of size;
but in the case of the stars such increase appears much less, so that
you may consider that a telescope, which (for the sake of ilustaton)
is powerful enough to magnif other objects a hundred tmes, will
scarcely render the stars magnified four or five tmes.'
26
my bok [the Oti of 1 60) I state that there was surely some imprfecton in that
outenost circle during full mon. Study the matter, and once again tell us, how it
loks to you . . . . ' Here the results of naked eye obseraton are quoted against Galileo's
telesopic repr- and wth prfectly god reason, as we shall see below. Te reader
who remember Kepler's plyopia (cf. fotote 14 to this chapter) may wonder how he
could tust his senses to such an extent. The reply is contained in the followng
quotaton (cHc,II, pp. 194f): 'When eclipses of the mon begin, I, who sufer from
this defect, become aware of the eclipse before all the other obserer. Long before te
eclipse stars, I even detect the directon from which the shadow is approaching, while
the others, who have ver acute vision, are still in doubt . . . . Te afore-mentoned
wavines of the mon [cf. the previous quotaton) stops for me when the mon
approaches the shadow, and the stongest par of the sun's rays is cut of . . . .' Galileo
has two explanatons for the contadictor appearance of the mon. The one involves a
lunar atosphere (ctgo,op. cit., pp. 26f). The other explanaton (ibid., pp. 25f),
which involves the tangental appearance of series of mountains lying behind each
other, is not really ver plausible as the distrbuton of mountains near the visible side
of the lunar globe dos not show the arangement that would be needed (this is now
even better established by the publication of the Russian mon photograph of 7
October 1 959; cf. Zdenek Kopal, Za Iatmauiea te thc5taqe/thcMa, Norh
Holland, 1 966, p. 242).
25. The libratons were notced by Galileo. C. G. Righini, 'New Light on Galileo's
Lunar Obseratons', in M.L Righini-Bonelli and R. Shea (eds), Rcmea, Emm
amMttimm ia thc5ntcRnlatiea, New York, 1 975, p. 59f. Thus it was not
sloppiness of obseratons but the phenomena themselves that misgided Galileo.
In two letter to the joural 5ncc(2 May and 10 October 1980) T.H. Whitaker
accused me of giving a misleading account of Galileo's obseratonal skill-I called him
a po r obserer when his lunar obseratons were in fact rater impressive. Te
accuston is refuted by the text to fototes 29 and 30 and by fotote 4 of the
present chapter. Witaker obviously thought my quotatons from Wolf (text to
fotote 28) reflected my own opinion. He also pint out that the copperlates
of
Galileo's obseratons are much better, from a moem pint of view, than the
wodcut which accompanied the Naaoas. This is te but dos not invalidate my
descripton of the debate which wa baed on the published account.
26. Mcsgo,op. cit., p. 38; cf. also the more detailed account in Di, op. cit.,
pp. 336f. 'The telescop, a it were, removes the heavens from us,' wrtes A. Chwalin
a
NI NE 93
The strangest features of the early history of the telescope emerge,
however, when we take a closer look at Galileo's piaures of the moon.
It needs only a brieflook at Galileo's drawings, and a photograph
of sim
ilar phases, to convince the reader that 'none of the features
rec
orded . . . can be safely identfied with any known markings of the
lunar landscape'P Looking at such evidence it is very easy to think
that 'Galileo was not a great astonomical obserer; or else that the
excitement of so many telescopic discoveries made b

him at that
tme
had temporarily blurred his skill or critcal sense' .
8
Now this asserton may well be true (though I rather doubt it i
view of the quite extaordina

obseratonal skill which Galileo
exhibits on other occasions). But it is poor i content and,
in his editon of Klcemc&, Dic KrcubmqagdrCctimc (Leipzig, 1 927, p. 9),
commentng on the decrease of the apparent diameter of all stars with the sole
excepton of the sun and the mon. Later on, the diferent magnifcaton of planets (or
comets) and fed stars was used as a means of distnguishing them. 'From experience,
I know', writes Herschel in the paper reportng his frst obseraton of Uranus (Fhil
1raas.,7I , I 78I , pp. 493f- the planet is here identfed as a cemct),'that the diameter
of the fed star are not proportonally magnifed with higher power, as the planets
are; therefore, I now put on the power of 460 and 932, and found the diameter of the
comet increased in proporton to the power, as it ought to be . . . .' It is noteworhy that
the rle did not invariably apply to the telescopes in use at Galileo's tme. Thus,
commentng on a comet ofNovember 1 61 8, Horato Grassi ('On the Three Comets of
I 6I 8',in 1hcCeatmme/thcCemctse/II,op. cit., p. I 7)point out 'that when the
comet was obsered through a telescope it sufered scarcely any enlargement', and he
infers, perfectly in accordance with Herschel's 'experience', that 'it whave to be said
that it is more remote from us than the mon . . . .' In his Zstmaemimlamacc(ibid., p.
80)he repeats that, according to the common experience of 'illustous astonomer'
from 'many pars of Europe' the comet obsered with a ver extended telescope
received scarcely any increment. . . . Galileo (ibid., p. I 77) accepts this as a fact,
critcizing only the conclusions which Grassi wants to draw from it. All these
phenomena refute Galileo's asserton (Asayer, op. cit., p. 20) that the telescope
'works always in the same way'. They also underine his theor of iradiaton (cf.
fotote 56 to this chapter).
Z7. Kopal, op. cit., p. Z07.
28. R. Wolf (Ccchichtc drZstmaemic, p. 396) remark on the por qualit of
Galileo's drawings of the moon (' . . . seine Abbildung des Mondes kann man . . . kaum
+ eine
Kare nennen'), while Zinner (Ccchichtcdr5tmwa&,Berlin, 1 931 , p. +73)
calls Galileo's obseratons of the moon and Venus 'typical for the obserations of a
beginner'. His picture of the mon, according to Zinner, 'has no similarit with the

oon' (ibid., p. +7Z). Zinner also mentons the much better qualit of the almost
Simultaneous obseratons made by the Jesuits (ibid., p. 473), and he finally asks
whether Galileo's obseratons of the moon and Venus were not the result of a fertle
ain, rather than of a careful eye ('sollte dabei . . . der Wunsch der Vater der
eobachtung gewesen sein?') - a pertnent queston, especially in view of the
p
henomena briefly described in footote 34 to this chapter.
29. The discover and identifcaton of the moons of jupiter were no mean achieve
me
nts, e
specia
lly as a useful stable suppor for the telescope had not yet been developed.
94 AGAI NST METHOD
I submit, not ver interestng. No new suggestons emerge for
additonal research, and the possibility of a tet is rather remote.
30
There are, however, other hyotheses which do lead to new
suggestons and which show us how complex the situaton was at
the tme of Galileo. Let us consider the following two.
Hyothesis I. Galileo recorded faithfully what he saw and in
this way lef us evidence of the shorcomings of the first tele
scopes as well as of the peculiarites of contemporar telescopic
vision. Interreted in this way Galileo's drawings are reports
of exacdy the same kind as are the reports emerpg from
the experiments of Stratton, Ehrismann, and Kohler
3
- except
that the characteristcs of the physical apparatus and the un
familiarity of the objects seen must be taken into account too.
32
We must also remember the many confictng views which were
held about the surface of the mon, even at Galileo's te,
33
and which may have infuenced what obserers saw.
34
What would
be needed in order to shed more light on the matter is an em
pirical collecton of all the early telescopic results, preferably
in parallel columns, including whatever pictorial representatons
30. The reaon, among other things, is the geat variaton of telesopic vision from
one obserer to the next, cf. Ronchi, Qtic,op. cit., Chapter IV.
31 . For a surey and some intoductor literatre cf. Gregor, op. cit., Chapter I I .
For a more detailed disussion and literatre cf. K.W. Smith and W.M. Smith,
FoqtieaaadMetiea, Philadelphia, 1 962, reprinted in par in M.D. Veron, op. cit.
The reader should also consult Ames' artcle 'Aniseikonic Glases', Eqhratieas ia
1raasmiemlFqcheQ,which deals with the change of ae~lvision caused by only
slighdy abnoral optcal conditons. A comprehensive account is given by I. Rok, 1hc
Natarce/FoqtaZ&ptatiea,New York, 1966.
32. Many of the old instments, and excellent descriptons of them, are stll
available. Cf. Zinner, DmtschcaadNicdLadirchcmtmaemischcIastmmtc.
33. For interestng informaton the reader should consult the relevant passges of
Kepler's Cemmatieaa well a of his 5emaiam(the latter is now available in a new
translaton by E. Rosen, who has added a considerable amount of background
material: Kqlo:5emaiam, ed. Rosen, Madison, 1967). The standard work for the
beliefs of the tme is stll Plutarch's Fc ea thcMa (it will be quoted from H.
Chemiss' translaton of MeraliaXI, London, 1967).
34. 'One describes the mon afer object one thinks one can perceive on its
surface' (Kaster, op. cit., Vol. I, p. 1 67, commentng on Fontana's obseratonal
reprs of 1 64). 'Maesdin even saw rain on the moon' (Kepler, Cemmatiea,op. cit.,
pp. 29f, presenting Maesdin's own obseratonal repor); cf. also da Vinci, notebooks
,
quoted from J.P. Richter, 1hcNetcbeeHe/Lmmrdd Iiao,Vol. II, New York, 1 970,
p. 1 6 7: 'If you keep the details of the spts of the moon under obseraton you will ofen
fnd great variaton in them, and this I myself have proved by drawing them. And this is
caused by the clouds that rise from the waters in the mon . . . . ' For the instability of the
images of unknown objects and their dependence on belief (or 'knowledge') cf.
Ronchi, Qtic,op. cit., Chapter I.
NI NE 95
have survived.
35
Subtractng instrumental peculiarites, such a
collecton adds fascinatng material to a yet-to-be-written history of
percepton (and of science).
36
This is the content of Hypothesis I.
Hypothesis II is more specifc than Hypothesis I, and develops it in
a certain directon. I have been considering it, with varying degrees of
enthusiasm, for the last two or three years and my interest in it has
been revived by a letter from Professor Stephen Toulmin, to whom I
am
grateful for his clear and simple presentaton of the view. It seems
to me, however, that the hypothesis is confronted by many diffculties
and must, perhaps, be given up.
Hypothesis II, just like Hypothesis I, approaches telescopic
reports from the point of view of the theor of perception; but it adds
that the practce of telescopic obseraton and acquaintance with the
new telescopic reports changed not only what was seen through the
telescope, but also what was seen with the naked re. It is obviously of
importance for our evaluaton of the contemporar atttude towards
Galileo's reports.
That the appearance of the stars, and of the moon, may at some
tme have been much more indefnite than it is today was originally
suggested to me by the exstence of various theories about the moon
which are incompatble with what everyone can plainly see with his
own eyes. Anaximander's theory of partal stoppage (which aimed to
explain the phases of the moon), Xenophanes' belief i the exstence
of different suns and different moons for different zones of the earth,
Heraclitus' assumpton that eclipses and phases are caused by the
tumin: of the basins, which for him represented the sun and the
moon - all these views run counter to the exstence of a stable and
plainly visible surface, a 'face' such as we 'know' the moon to possess.
The same is tue of the theory of Berossos which occurs as late as
Lucretus
38
and, even later, in Alhazen.
35. Chapter 15 of Kopal, op. cit., contains an interestng collecton of exactly this
kind. Wider scope has W. Schulz, DicZaschaaaagvo MdaadsoaCctaltia
MthesaadKaastdIlko,Berlin, 1912.
36. One must, of course, also investgate the dependence of what i s seen on the
current methods of pictorial representaton. Outside astonomy this was done by E.
Gomhrich,Ar aadIllasiea,London, 1 960, and L. Choulant,A Huteaadiblie@
e/Zaotemicolh astratiea, New York, 1 945 (tanslated, with additons, by Singer and
othe
rs),
who deals with anatomy. Astronomy has the advantage that one side of the
puz
zle,
viz. the stars, is fairly simple in structure (much simpler than the uterus, for
e
xam
ple) and relatively well known; cf. also Chapter 16 below.
37.
For
these theories and further literature cf. J.L.D. Dreyer, A Hute e/
Zstr
oaep
/mm1holcteKqlo, New York, 1953.
?B.
For
Berossos, cf. Toulmin's article in Isis, No. 38, 1 967, p. 65. Lucretus
Wtes (
0athcNatarc e/1hiap,trans!. Leonard, New York, 1 957, p. 21 6): 'Again, she
96 AGAI NST METHOD
Now such disregard for phenomena which for us are quite obvious
may be due either to a certain indiference towards the exsting
evidence, which was, however, as clear and as detailed as it is today, or
ele to a di erence in the eidnce itsel It is not easy to chose between
these alteratves. Having been influenced by Wittgenstein, Hanson,
and others, I was for some tme inclined towards the second version,
but it now seems to me that it is ruled out both by physiolog
(psycholog)
39
and by historical informaton. We need only
remember how Copericus disregarded the difcultes arising from
the variatons in the brightess of Mars and Venus, which were well
known at the tme. 4 And as regards the face of the mon, we see
that Aristotle refers to it quite clearly when observng that 'the stars
do not rl. For rolling involves rotton: but the "face", as it is called,
of the moon is always seen.'
41
We may infer, then, that the
ocasional disregard for the stbility of the face was due not to a lack
of clear impressions, but to some widely held views about the
unreliability of the senses. The inference is supported by Plutarch's
discussion of the matter which plainly deals not with what is seen
(except as evidence for or aganst certain views) but with certain
elanations of phenomena otherse asumed to be well kow:
4
2
'To
begin with,' he says, 'it is absurd to call the fgure seen in the mon an
afecton of vision . . . a conditon which we call bedazlement (glare).
Anyone who assers this does not obsere that this phenomenon
should rather have ocurred in relaton to the sun, since the sun lights
upon us keen and violent, and moreover does not explain why dull
and weak eyes discer no distncton of shape in the mon but her orb
for them has an even and fll light whereas those of keen and robust
may revolve upn herself I like to a ball's sphere -if perchance to b-I one half of her
dyed o'er with glowing light I and by the revoluton of that sphere I she may bget for
us her varing shapes I untl she trs that fer par of her I fll to the sight and opn
eyes ofmen . . . .'
39. Cf. text to fototes 50f of my 'Reply to Critcism', op. cit., p. 24.
4. In antquit the diferences in the magitudes of Venus and Mars were
regarded as being 'obvious to our eyes', Simplicius, DcCmle, II, 1 2, Heibrg, p. 50.
Polemarchus here considers the difcultes of Eudoxos' theor of homoentc
spheres, viz. that Venus and Mars 'appar in the midst of the retograde moement
many tmes brighter, so that [Venus) on monless night causes bies to trow
shadows' (objecton of Autolycus) and he may well b appaling to the pssibilit of a
decepton of the senses (which wa fequendy disussed by ancient shols). Aistode,
who must have been familiar with all these fact, dos not menton them anywhere i
DcCmleor in the Mn@hsm,though he gives an account of Eudoxos' system and of
the improvement of Polemarchus and Kalippus. Cf. fotote 7 of Chapter 8.
41 . DcCmle,29a25f.
42. op. cit., p. 37, cf. also S. Sambursky, 1hcFhsml edd e/hc CrccH, Ne
York, 1 962, pp. 244f.
NI NE
FIGURE I .
-
The shape of a lunar mountain and a walled plain, from Galileo
Side Naaoas,Venice, 1 61 0 (cf. p. I l l ).
97
98 AGAI NST METHOD
vision make out more precisely and distctly the patter of facial
features and more clearly perceive the variatons.' 'The unevenness
also entrely refutes the hypothesis,' Plutarch contues,4
3
'for the
shadow that one sees is not contnuous and confused, but is not badly
depictured by the words of Agesianax: "She gleams with fire
encircled, but within I Bluer than lapis show a maiden's eye I And
dainty brow, a visage manifest." In tuth, the dark patches submerge
beneath the bright ones which they encompass . . . and they are
thoroughly entwined with each other so as to make the delineaton of
the fgure resemble a paintng.' Later on the stability of the face is
used as an argument against theories which regard the moon as being
made of fre, or air, for 'air is tenuous and without configuraton, and
so it naturally slips and does not stay in place'. 4 The apearanc of
the moon, then, seemed to be a well-known and distct
phenomenon. What was in queston was the releance of the
phenomenon for astonomical theory.
We can safely assume that the same was tue at the tme of
Galileo.45
But the we must admit that Galleo 's obserations could be checked with
the naked ee and could in this way be eosed a illusor.
Thus the circular monster below the centre of the disk of the
moon 4 is well above the threshold of naked eye obseraton (its
43. ibid., cf however, fotote 1 7 to this chapter, Pliny's remark (Hut.Nm.,II, 43,
46) that the mon is 'now sptted and then suddenly shining clear', as weD as da Vinci's
report, refered to in fotote 34 to this chapter.
4. ibid., p. 50.
45. A stong argument ia/mearof this contenton is Kepler's descripton of the
mon in his Qtic of 160: he comments on the broken character of the boundar
beteen light and shadow (okc, II, p. 21 8) and describes the dark part of the mon
during an eclipse as loking like tom flesh or broken wod (ibid., p. 21 9). He retur to
these passages in the Cmatiea (op. cit., p. 27), where he tells Galileo that 'these
ver acute obseratons of yours do not lack the support of even my ow testmony. For
[in my] Qticyou have the halfmon divided by a wavy line. From this fact I deduced
peaks and depressions in the body of the moon. [Later on) I describe the mon during
an eclipse as looking like tom flesh or broken wod, with bright steaks penetatng into
the region of the shadow.' Remember also that Kepler critcizes Galileo's telescopic
reprts on the bais of his ow naked eye obseratons; cf. fotote 24 of this chapter.
46. 'There is one other point which I must on no account forget, which I have
notced and rather wondered at it. It is this: The middle of the Moon, as it seems, is
ocupied by a cerain cavity larger than all the rest, and in shape perfectly round. I have
looked at this depression near both the frt and the third quarter, and I have
represented it as well a I can in the second illustaton already given. It produces the
same appearance as to effects oflight and shade as a tact like Bohemia would produce
on the Earth, if it were shut in on all sides by ver lofty mountains aranged on
the
circumference of a perfect cirle; for the tact in the mon is walled in with peaks of
such enorous height that the furthest side adjacent to the dark poron of the mon i
NI NE 99
diameter is larger than 3
1
/z minutes of arc), while a single glance
con
vinces us that the face of the moon is not anywhere disfgured by a
blemish of this kind. It would be interestn. to see what
contemporar obserers had to say on the matter
4
or, if they were
artsts, what they had to draw on the matter.
I summarize what has emerged so far.
Galileo was only slightly acquainted with contemporar optcal
theor.
48
His telescope gave surrising results on the earth, and
these results were duly praised. Trouble was to be expected in the
sky, as we know now. Trouble promptly arose: the telescope
produced spurious and contadictor phenomena and some of its
results could be refuted by a simple look with the unaided eye. Only a
new theor of telescopic vision could bring order into the chaos
(which may have been stll larger, due to the diferent phenomena
seen at the tme even with the naked eye) and could separate
appearance from reality. Such a theor was developed by Kepler, frst
in 1 604 and then again in 1 61 1 .
49
According to Kepler, the place of the image of a punctform object
is found by frst tacing the path of the rays emerging from the object
according to the laws of (reflecton and) refracton untl they reach
the eye, and by then using the principle (stll taught today) that 'the
image will be seen in the point determined by the backward
seen bathed in sunlight before the boundary between light and shade reaches half way
across the circular space . . . '(csgo, op. cit., pp. 21 f. Ti desripton, I think,
defnitely refutes Kopal's conjecture of obseiatonal laxity. It is interestng to note the
difference between the wodcuts in the Naaou (. 1 31 , Figure I and Galileo's
original drawing. The woodcut correspnds quite closely to the desripton while the
original drawing with its impressionistc features ('Kaum eine Kare,' sys Wolf) is
vague
enough to escape the accusaton of gross obseiatonal error.
4 7. 'I cannot help wondering about the meaning of that large circular cavity in
what
I usually call the left comer of the mouth,' writes Kepler ( Cemmatiea,op. cit., p.
28), and then proeeds to make conjectures a to its origin (conscious effors by
intelligent beings included).
48. Contemporary academic optcs went beyond simple geometical constuctons
(which Galileo may have known) and included an account of mhuscwhen loking at
a mirror, or through a lens, or a combinaton oflenses. Exceptng irradiaton Galileo
now
here considers the propertes of telescopic isiea. Aristotelians writng afer
Galileo's telescopic obseiatons did. Cf. Redondi, op. cit., pp. l 69ff.
49. I have here disregarded the work of della Pora (Dc Rc/mieac) and of
Maurolycus who both antcipated Kepler in cerain respects (and are duly mentoned
by him). Maurolycus makes the imprtant step [Fhetumi dLamiac, tansl. Henry
Crew, New York, 194, p. 45 (on mirror) and p. 74 (on lenses)) of considering only
the
cusp
of the caustc; but a connecton with what is seen on dircuvision is stll not
est
ab
lishe
d. For the difcultes which were removed by Kepler's simple and ingenious
hy
p
thesi
s cf. Ronchi, Huteircd/ Lamirc,Chapter III.
100 AGAI NST METHOD
intersecton of the rays of vision from both eyes'50 or, in the case of
monocular vision, from the two sides of the pupil. 5 1 This rule, which
proceeds from the assumpton that 'the image is the work of the act of
vision', is partly empirical and partly eometical. It bases the positon
of the image on a 'metical triangle' 2 or a 'telemetic triangle'5
3
as
Ronchi calls it, 54 that is constucted out of the rays which fnally
arrive at the eye and is used by the eye and the mind to place the image
at the proper distance. Whatever the optcal system, whatever the
total path of the rays from the object to the obserer, the rule says that
the mind of the obserer utlizes its ve lat par onl and bases its
visual judgement, the percepton, on it.
The rule considerably simplifed the science of optcs. However, it
needs only a second to show that it is false: take a magnifing glass,
determine its fous, and look at an object close to it. The telemetic
tiangle now reaches beyond the object to infnity. A slight change of
distance brings the Keplerian image from infnity to close by and
back to infnity. No such phenomenon is ever obsered. We see the
image, slightly enlarged, in a distance that is most of the tme
identcal with the actual distance between the object and the lens.
The visual distance of the image remains constant, however much we
may var the distance between lens and object and even when the
image becomes distorted and, fnally, difuse. 55
50. okc, II, p. 72. The Qticof 1 60 has been partly tanslated into Geran by
F. Plehn, J Kqlm Cmadmg dgcemctmch Qtik, Leipzig, 1922. The relevant
pasages ocur in secton 2 of Chapter 3, pp. 38--48.
5 1 . ibid., p. 67.
52. 'Cum imago sit visus opus', ibid., p. 6. 'In visione tenet sensus communis
oulorm suorm distantam ex asuefactone, angulos vero ad illam distantam notat
ex sensu contoronis oulorm', ibid., p. 66.
53. 'Trianglum distantae mensorium', ibid., p. 67.
54. Qtic,op. cit., p. 4. One should also consult the second chapter of this b k
for the histor of pre-Keplerian optcs.
55. ibid., pp. 182, 202. This phenomenon was known to everone who had used a
magnifing glas only once, Kepler included. Which shows that disregard of familiar
phenomena dos not entail that the phenomena were seen diferently (cf. text to
fomote 4 to this chapter). Isaac Barow's account of the difculty of Kepler's rle
wa mentoned above (text to fotote I8 to Chapter 5). According to Berkeley (op.
cit., p. 1 4 I) 'this phenomenon . . . entrely subvers the opinion of those who wlhave
us judge of distances by lines and angles . . . . ' Berkeley replaces this opinion by his own
theor according to which the mind judges distances from the clarity or confusion of
the primar impressions. Kepler's idea of the telemetc triangle was adopted at once
by almost all thinker in the feld. It was given a fundamental positon by Descanes
according to whom 'Distantam . . . discimus, per mutuam quandam conspiratonem
oulorm' (Di@tn,quoted from RatiDcmnc5pccimFhiles@him,Amsterdam,
I 657, p. 87). 'But,' says Barow, 'neither this nor any other difculty shall . . . make me
renounce that which I know to be manifestly agreeable to reason.' It is this atttude
NI NE 101
This, then, was the actual situaton in 1610 when Galileo
published his telescopic fndings. How did Galileo react to it? The
answer has already been given: he raised the telescope to the state
of a 'superior and better sense'. 56 What were his reasons for doing
which was responsible for the slow advance of a sientfc theor of qcgmscand of
visual optcs in general. 'The reaon for this pculiar phenomenon,' writes Morit von
Rohr (Dmnllgwa @tuchclmtmmt,Berlin, 1934, p. 1 ), 'is to b sught in te
close connecton between the eye glas and the eye and it is impssible to give an
acceptable theor of eye glasses without undertanding what happns in the proess of
vision itself . . . .' The telemetic tiangle omits precisely this proess, or rather gives a
simplistc and false account of it. Te state of optcs at the beginning of te 20
centur is well desrbd in A. Gullstand's 'Appndices to Par I' of Helmholt's
1rcatucm FhsielqmlQtic,tansl. Southall, New Yor, 1 962, pp. 26l f. We read
here how a retur to the psycho-physiological proess of vision enabled physicists to
arrive at a more reanable account even of te physics of optcal imager: 'The reaon
why the laws of actua optcal imager have ben, s to spak, summoned to life b the
requirement of physiological optcs is due partly to the fact that b means of
tgnometical calculaton, tedious to be sure, but easy to prfor, it ha been
pssible for the optcal engineer to gt closer to the realites of his problem. Thus,
thanks to the labour of such men a Abb and his shol, technical optcs ha attained
it presnt splendid development; whereas, wit te sientfc mean available, a
comprehensive grasp of the intcate rlatons in te cae of the imagr in the eye h
been actually impssible.'
56. '0 Nichola Copricus, what a pleaure it would have been for you to se this
par of your system confred b so clear an experment!' writes Galileo, implying tat
the new telesopic phenomena are additonal suppr for Copricus (Dmgc,op.
cit., p. 339). The difernce in the apparance of planet and fed star (cf. fomote 26
to this chapter) he explains by te hypthesis that 'the ver instument of seeing [the
eye] intouces a hindrance of it ow' (ibid., p. 335), and that the telesop removes
this hindrance, v. immtiea,prmittng the eye to see the star and the planets a
they really are. (Mario Giuducci, a follower of Galileo, asrbd irdiaton to
refracton b moisture on the surface of the eye, Duunc0theCne/II,op. cit.,
p. 47.) This explanaton, plausible a it may seem (espciaiiy in view of Galileo's
atempt to show how irradiaton can be removed
b
means other than the telesop) is
not as
staightforard a one might wish. Gullstand (op. cit., p. 426) sys that 'owing
to the propres of te wave surface of the bundle of rays refracted in the eye . . . it is a
mathematcal impssibilit for any cros secton to cut the caustc surface in a smoth
cure
in the for of a circle concentc with the pupil'. Oer author pint to
'inhomogeneites in the various humour, and abve all in te crstalline len' (Ronchi,
Qtic, op. cit., p. I 0). Kepler gives tis account (Cmim , op. cit., pp. 33f):
'Poi
nt
sources of light tansmit their cones to the crstalline lens. There refacton
takes place, and behind the lens the cones again contct to a pint. But tis pint dos
not reach a far a the retna. Therefore, the light is dispred once more, and spreads
over a small area of retna, wherea it should impinge on a pint. Hence the telesop,
b intoucing another refracton, makes this pint coincide with te retna . . . .'
P

Iyak, in his classical work DeRetim,attibutes irradiaton parly to 'defect of the


d
J
op
tcal media and to the imperfect accommodaton' but 'chiefy' to the 'pculiar
St
uct
ural
consttuton of the retna itelf (p. 1 76), adding that it may b a fncton of
!
e
b
rain
also (p. 429). None of these hypteses cover mlthe facts know abut
Ir
ad
iato
n. Gullstand, Ronchi, and Polyak (if we omit his reference to the brain
102 AGAI NST METHOD
so? This queston brings me back to the problems raised by the
evidence (against Copericus) that was reported and discussed in
Chapter 8.
which can be made to explain anything we want) cannot explain the disapparance of
irdiaton in the telescope. Kepler, Gullstand and Ronchi also fail to give an account
of the fact, emphasized by Ronchi, that large objects show no iradiaton at their edges
('Anyone underaking to account for the phenomenon of iradiaton must admit that
when he loks at an electic bulb from afar so that it seems like a pint, he sees it
surounded by an immense crown of rays whereas from nearby he sees nothing at all
around it,' Otic, op. cit., p. I 05). We know now that large object are made defnite by
the lateral inhibitor interacton of retinal elements (which is further increased by
brain functon), cf. Ratliff, Mhaad,p. 1 46, but the variaton of the phenomenon
with the diameter of the object and under the conditions of telescopic vision remains
unexplored. Galileo's hypothesis received suppr mainly from its agreement with the
Coperican point of view and wa, therefore, largely mhec.
1 0
On the other hand, there are some telecoic pheomea which are plainl
Coerican. Galileo introduce thee pheomea as indednt eidnce
fr
Coerics while the situation is rather that one reuted view -
Coeicanism - has a cerain similart to pheomea emerng frm
another reuted view - the ida that telecoic pheomea are fithful
images ofthe sk.
According to the Coperican theor, Mars and Venus approach and
recede from the earth by a factor of I :6 or I :8, respectvely. (These
are approxmate numbers.) Their change of brightess should be
I :40 and I :60, respectvely (these are Galileo's values). Yet Mars
changes ver little and the variaton in the brightess of Venus 'is
almost imperceptble' .
1
These experiences 'overtly contadict the
annual movement [of the earth]'.
2
The telescope, on the other hand,
produces new and strange pheomea, some of them exposable as
illusor by obseraton with the naked eye, some contradictor, some
having even the appearance of being illusor, while the only theor
that could have brought order into this chaos, Kepler's theor of
\ision, is refuted by evidence of the plainest kind possible. But - and
with this I come to what I think is a central feature of Galileo's
procedure - there are telescoic pheomea, namely the telescopic
variaton of the brightess of the planets, which agree more closel with
Copeicus than d the results ofnaked-ee obseration. Seen through the
telescope, Mars does indeed change as it should according to the
Coperican view. Compared with the total performance of the
teles
cope this change is stll quite puzzling. It is just as puzzling as is
the Coperican theor when compared with the pre-telescopic
evid
ence. But the change is in harmony with the predictons of
Copericus. It is this harony iather than any deep understanding of
I .
The actual variatons of Mars and Venus are four magnitudes and one
magn
itude respectively.
2. Dialegc,op. cit., p. 328.
103
104 AGAI NST METHOD
cosmology and of optcs which for Galileo pruves Coeicus and the
veracit ofthe telescoe in terrestial as well as celestal matters. And it is
this harmony on which he builds an entrely new view of the universe.
'Galileo,' writes Ludovico Geymonat,
3
referring to this aspect of the
situaton, 'was not the first to tum the telescope upon the heavens,
but . . . he was the first to grasp the enormous interest of the things
thus seen. And he understood at once that these things fitted in
perfectly with the Coperican theor whereas they contadicted the
old astonomy. Galileo had believed for years in the tuth of
Copemicanism, but he had never been able to demonstate it despite
his exceedingly optmistc statements to friends and colleagues [he
had not even been able to remove the refutng instances, as we have
seen, and as he says himself] . Should direct proof [should even mere
agreeet with the evidence] be at last sought here? The more this
convicton took root in his mnd, the clearer to him became the
importance of the new instument. In Galileo's own mind faith in the
reliabilit of the telescope and recogniton of its importance were not
two searate as, rather, they were two aspeas ofthe same process.' Can
the absence of independent evidence be expressed more clearly?
'The Nuncius', writes Franz Hammer in the most concise account I
have read of the matter,4 'contains two unknowns, the one being
solved with the help of the other.' This is entrely correct, except that
the 'unknowns'"ere not so much unknown as known to be false, as
Galileo on occasions says himself. It is this rather peculiar situaton,
this harmony between two interestng but refuted ideas, which
Galileo exploits in order to prevent the eliminaton of either.
Exactly the same procedure is used to presere his new dynamics.
We have seen that this science, too, was endangered by obserable
events. To eliminate the danger Galileo intoduces fricton and other
disturbances with the help of a hoc hypotheses, teatng them a
tendencies dned by the obvious discrepancy between fact and
theor rather than as physical events elained by a theor of fricton
for which new and independent evidence mght some day become
3. op. cit., pp. 38f (my italics).
4. ehaaacKqlo, Ccammcltccdc,op. cit., Vol. I, p. 47. Kepler (Cemmatiea,
op. cit., p. 14) speaks of 'mutually self-supportng evidence'. Remember, however,
that what is 'mutually self-supportng' are to refuted hypotheses and aet to
hypotheses which have iaqmtsayenin the domain of basic statement. In a letter
to Herarth of 26 March 1 598, Kepler speaks of the 'many reaons' he wants to
adduce for the moton of the earth, adding that 'each of these reaons, taken for itself,
would fnd only scant belief (Cat-Qck, ehaaacKqloiasoanq,Vol. I,
Munich, 1 930, p. 68).
TEN lOS
available (such a theor arose only much later, in the 1 8th centur).
Yet the agreement between the new dynamics and the idea of the
moton of the earth, which Galileo increases with the help of his
method of anamneis, makes both seem more reasonable.
The reader will realize that a more detailed study of historical
phenomena such as these creates considerable difcultes for the
view that the tansiton from the pre-Copercan cosmology to that of
the 17th centur consisted in the replacement of refuted theories by
more general conjectures which explaied the refutng instances,
made new predictons, and were corroborated by obseratons
carried out to test these new predictons. And he will perhaps see the
merits of a diferent view which asserts that, whie the pre
Copercan astonomy was in truble (was confronted by a series of
refutng instances and implausibilites), the Coperican theor was in
ee geate trouble (was confronted by even more drastc refutng
instances and implausibilites); but that being in harony with stil
furhe inaquate theore it gained stength, and was retained, the
refutatons being made inefectve by a hoc hypotheses and clever
techniques of persuasion. This would seem to be a much more
adequate descripton of the developments at the tme of Galileo than
is ofered by almost all altertve accounts.
I shall now interrupt the historical narratve to show that the
descripton is not only fauall aquate, but that it is also peral
reasonable, and that any attempt to enforce some of the more familiar
methodologies of the 20th centur would have had disastous
consequences.
1 1
Such 'irational' methods ofsuppor are needd because ofthe 'uneen
deloment' (Man, Lenin) ofdi erent pars ofscence. Coeican ism and
other esential ingredients ofmodm science suried onl because reason
was fequentl uerled in their past.
A prevalent tendency in philosophical discussions is to approach
problems of knowledge sub spece aeteitatis, as it were. Statements
are compared with each other without regard to their histor and
without considering that they might belong to diferent historical
strata. For example, one asks: given background knowledge, inital
conditons, basic principles, accepted obseratons - what conclu
sions can we draw about a newly suggested hypothesis? The answers
var considerably. Some say that it is possible to determine degrees of
confirmaton and that the hypothesis can be evaluated with their help.
Others reject any logic of confirmaton and judge hypotheses by their
content, and by the falsificatons that have actually occurred. But
almost everone takes it for granted that precise obseratons, clear
principles and well-confirmed theories are already dciie that they
can and must be used here and now to either eliminate the suggested
hypothesis, or to make it acceptable, or perhaps even to prove it.
Such a procedure makes sense only if we can assume that the
elements of our knowledge - the theories, the obseratons, the
principles of our arguments - are timeless entitie which share the
same degree of perfecton, are all equally accessible, and are related
to each other in a way that is independent of the events that produced
them. This is, of course, an extremely common assumpton. It is
taken for granted by most logicians; it underlies the familiar
distincton between a context of discover and a context of
justfcaton; and it is often expressed by saying that science deals
with propositons and not with statements or sentences. However,
the procedure overlooks that science is a complex and heterogeneous
historcal proces which contains vague and incoherent anticipatons of
future ideologies side by side with highly sophistcated theoretcal
106
ELEVEN 107
systems and ancient and petifed fors of thought. Some of its
elements are available in the form of neatly written statements while
others are submerged and become known only by contast, by
comparison with new and unusual views. (This is the way in which
the inverted tower argment helped Galileo to discover the natural
interpretatons hostle to Copericus. And this is also the way in
which Einstein discovered certain deep-lying assumptons of
classical mechanics, such as the assumpton of the exstence of
infnitely fast signals. For general consideratons, cf. the last
paragraph of Chapter 5.) Many of the conflicts and contradictons
whch occur in science are due to this heterogeneity of the material,
to this 'unevenness' of the historical development, as a Marxst would
say, and they have no immediate theoretcal signifcance.
1
They
have much in common with the problems which arise when a power
1 . According to Mar, 'secondar' pars of the soial proess, such as demand,
artstc producton or legal relatons, may get ahead of material producton and drag it
along: cf. 1hcFmcqe/Fhiles@h but especially the Iatmduuiea te thc Cntquc e/
Felitiml Eae, Chicago, 1 91 8, p. 309: 'The unequal relaton between the
development of material producton and an, for instance. In general, the concepton of
progress is not to be taken in the sense of the usual abstacton. In the case of an, etc., it
is not so imponant and difcult to understand this disproporton as in that of practcal
soial relatons, e.g. the relaton between educaton in the U.S. and Europe. The really
difcult point, however, that is to be disussed here is that of the unequal development
of relatons of prouction as legal relatons.' Trotsky describes the same situaton:
'The gist of the matter lies in this, that the diferent aspects of the historical progress
economics, plitcs, the state, the growth of the working class - do not develop
simultaneously along parallel lines' ('The Schol of Revolutonar Stateg', speech
delivered at the general pany membership meetng of the Moscow Organizaton of
July 192 I , published in DcFintFivc\cane/thcCemmuautIatmial,Vol. II, New
York, 1 953, p. 5). See also Lenin, Lq-iag Cmuaum- aaIa/atilcDuerdr(op.
cit., p. 59), concerng the fact that multple causes of an event may be out of phase and
have an effect only when they ocur together. In a diferent for, the thesis of'uneven
development' deals with the fact that capitalism has reached diferent stages in
diferent counties, and even in diferent par of the same count. This second tpe
of uneven development may lead to invere relatons between the accompanying
ideolog
ies, s that efciency in proucton and radical plitcal ideas develop in invere
proportons. 'In civilized Europe, with its highly developed machine indust, it rich,
multfor culture and its consttutons, a point of histor has been reached when the
commanding bourgeoisie, fearing the growth and increasing stength of the
proletariat, comes out in suppon of everthing backward, moribund, and medieval. . . .
But all young Asia grows a mighty demoratc movement, spreading and gaining in
st
ength'
(Lenin, 'Backward Europe and Advanced Asia', CeIlcucderH,Vol. 1 9, op.
cit., pp.
99fl). For this ver interestng situaton, which deseres to be exploited for the
p
hilo
sophy of science, cf. A. C. Meyer, Liaum,Cambridge, 1957, Chapter 1 2 and L.
Alth
usser, Fer Mam, London and New York, 1 970, Chapters 3 and 6. The
P
hilos
ophical background is splendidly explained in Mao Tse-tung's essay O
Cea
tr
miui(5clcucdRcmiap,Peking, I 970, p. 70, especially secton IV).
108 AGAI NST METHOD
staton is needed right next to a Gothic cathedral. Occasionally, such
features are taken into account; for example, when it is asserted that
physical laws (statements) and biological laws (statements) belong to
diferent conceptual domains and cannot be directly compared. But
in most cases, and especially in the case obseraton vs theor, our
methodologies project the various elements of science and the
diferent historical strata they occupy on to one and the same plane,
and proeed at once to render comparatve judgements. This is like
arranging a fght between an infant and a grown man, and announcing
tiumphantly, what is obvious anyway, that the man is going to win (the
histor of science is full of inane critcisms of this kind and so is the
histor of psychoanalysis and of Marxism). In our examinaton of new
hypotheses we must obviously take the historical situaton into
account. Let us see how this is going to afect our judgement!
The geoentic hypothesis and Aristotle's theor of knowledge
and percepton are well adapted to each other. Percepton suppors
the theor of loomoton that entails the unmoved earth and it is in
tum a special case of a comprehensive view of moton that includes
loomoton, increase and decrease, qualitatve alteraton, generaton
and corrpton. This comprehensive view defnes moton as the
transiton of a form from an agent to a patent which terminates when
the patent possesses exactly the same form that characterized the
agent at the beginning of the interacton. Percepton, accordingly, is a
proess in which the form of the object perceived enters the
percipient as precisely the same form that characterized the object so
that the percipient, in a sense, assumes the propertes of the object.
A theor of percepton of this kind (which one might regard a a
sophistcated version of naive realism) does not permit any major
discrepancy between obseratons and the things obsered. 'That
there should be things in the world which are inaccessible to man not
only now, and for the tme being, but in principle, and because ofhis
natural endowment, and which would therefore never be seen by him
this was quite inconceivable for later antquit as well a for the
Middle Ages. '
2
Nor does the theor encourage the use of
2. F. Blumenberg, CCalilci,5idmsNuaous,Nhnchtreaan5tm,Vol.
I, Frankfur, 1965, p. 13. Aristotle himself wa more open-minded: 'The evidence
(concering celestal phenomena) is furished but scantly by sensatons, wherea
respectng perishable plants and animals we have abundant information, living as we do
in their midst . o . ', DcFan.Zaim.,64b26f. In what follows, a highly idealized account is
given oflater Aristotelianism. Unless oterise stated, the word' Aristotle' refer to this
idealizton. For the difculties in forming a coherent picture of Aristotle himccf.
Dnng, Zmtetclc,Heidelberg, 1 966. For sme differences beteen Aristotle and his
mediaeval followers cf. Wolfgang Wieland, Dic ZmtetcluchcFhsik,Gottngen, 1970.
ELEVEN /09
instruments, for they interfere with the processes in the medium.
These processes carry a true picture only as long as they are left
undisturbed. Disturbances create forms which are no longer
identcal with the shape of the objects perceived - they create
illusions. Such illusions can be readily demonstrated by examining
the images produced by curved mirrors,
3
or by crude lenses (and
remember that the lenses used by Galileo were far from the level of
perfection achieved today): they are distorted, the lens-images have
coloured fringes, they may appear at a place different from the place
of the object and so on. Astonomy, physics, psychology, epistemol
og-all these disciplines collaborate with the Aristotelian philosophy
to create a system that is coherent, ratonal and in agreement with the
results of observaton as can be seen from an examinaton of
Aristotelian philosophy in the form in which it was developed by
some mediaeval philosophers. Such an analysis shows the inherent
power of the Aristotelian system.
The role of observaton in Aristotle is quite interestng. Aristotle is
an empiricist. His injunctons against an overly-theoretcal approach
are as militant as those of the 'scientfc' empiricists of the 1 7th and
18th centuries. But while the latter take both the tuth and the
content of empiricism for granted, Aristotle explains the nature of
experience and why it is important. Experience is what a normal
observer (an observer whose senses are in good order and who is not
drunk or sleepy, etc.) perceives under normal circumstances (broad
daylight; no interference with the medium) and describes in an idiom
that fts the facts and can be understood by all. Experience is
imporant fr knowledge because, given normal circumstances, the
perceptons of the observer contain identcally the same forms that
reside in the object. Nor are these explanatons a hoc. They are a
direct consequence of Aristotle's general theor of moton, taken in
conjuncton with the physiological idea that sensatons obey the same
physical laws as does the rest of the universe. And they are confrmed
by the evidence that confrms either of these two views (the exstence
of distorted lens-images being part of the evidence). We understand
today a little better why a theor of moton and percepton which is
now
regarded as false could be so successful (evolutonar
explanaton of the adaptaton of organisms; movement in media).
3. Already a plain mirror gives rise t an interesting illusion. To notce it, frst
look at yourself in a plain mirror. You will see your face at its 'noral' size.
Then let some steam condense on the surface of the mirror and draw the out
lin
e
of y
our face in the steam. The outline will look about half the size of your
face.
110 AGAI NST METHOD
The fact remains that no decisive empirical argument could be raised
against it (though it was not free from diffcultes).
This harmony between human percepton and the Aristotelian
cosmology is regarded as illusor by the supporters of the moton of
the earth. In the view of the Copemicans there exst large-scale
processes which involve vast cosmic masses and yet leae no trace in
our experience. The exstent observatons therefore count no longer
as tests of the new basic laws that are being proposed. They are not
directly attached to these laws, and they may be entrely discon
nected. Tody, afer the success of modem science led to the belief
that the relaton between man and the universe is not as simple as is
assumed by naive realism, we can say that this was a correct guess,
that the observer is indeed separated from the laws of the world by
the special physical conditons of his observaton platform, the
moving earth (gravitational efects; law of inerta; Coriolis forces;
influence of the atosphere upon optcal observations; aberraton;
stellar parallax; and so on . . . ), by the idiosyncrasies of his basic
instrument of observaton, the human eye (irradiaton; after-images;
mutual inhibiton of adjacent retnal elements; and so on . . . ) as well
as by older views which have invaded the obseraton language and
made it speak the language of naive realism (natural interpretatons).
Observatons may contain a contributon from the thing obsered,
but this contributon merges with other efects (some of which we
have just mentoned), and it may be completely obliterated by them.
Just consider the image of a fxed star as viewed through a telescope.
This image is displaced by the efects of refracton, aberraton and,
possibly, of gravitation. It contains the spectrum of the star not as it is
now, but as it was some tme ago (in the case of exta-galactc
superovae the diference may be millions of years), and distorted by
Doppler effect, intervening galactc matter, etc. Moreover, the
extension and the interal structure of the image is entrely
determined by the telescope and the eyes of the obserer: it is the
telescope that decides how large the diffracton disks are going to be,
and it is the human eye that decides how much of the stucture of
these disks is going to be seen. It needs considerable skill and much
theor to isolate the contributon of the original cause, the star, and to
use it for a test, but this means that non-Aristotelian cosmologies can
be tested only after we have searated obseratons and laws with the
help of auxiliar sciences describing the complex processes that
occur between the eye and the object, and the even more complex
processes between the corea and the brain. We must subdiid what
we perceive to fnd a core that mirrors the stmulus and nothing else.
In the case of Copericus we need a new meteorolog (in the good old
ELEVEN Ill
FIGURE 2. Moon, age seven days (frst quaner).
112 AGAI NST METHOD
sense of the word, as dealing with things below the moon), a new
science of physiological otic that deals with the subjectve (mind) and
the objectve (light, medium, lenses, stucture of the eye) aspects of
vision as well as a new dynamic statng the manner in which the
moton of the earth might infuence the physical proesses at its
surface. Obseratons become relevant only afer the proesses
described by these new subjects have been inserted between the
world and the eye. The language in which we express our
obseratons may have to be revised as well so that the new cosmolog
receives a fair chance and is not endangered by an unnotced
collaboraton of sensations and older ideas. In sum: what is needd fr
a tet ofCoeicus is an etirel new world-view containing a new view of
man and ofhis capacitie ofknowing. 4
4. Bacon realized that scientfc change involves a reforaton not only of a few
ideas, but of a entre world-view and, perhaps, of the ver nature of humans. 'For the
senses are weak and erring', he writes in NmamOqaaam, Aphorsm 50. 'For man's
sense is falsely assened to be the standard of things; on the contar, all the
prceptions, bth of the senses and of the mind bear reference to man and not to the
universe, and the human mind resembles those uneven mirors which impan their
ow properes to diferent objects from which rays are emitted and diston and
disfgure them' (Aphorsm 41). Bacon repeatedly comments on the 'dullness,
incompetency and erors of the senses' (50) and perits them only to 'judge . . . the
experment' while it is the experment that functons as a judge 'of nature and the thing
itself (50). Thus when Bacon speaks of the 'unprejudiced senses' he dos not mean
sense-data, or immediate impressions, but reactions of a sense organ that hm bc
rcbailt in order to miror nature in the rght way. Research demands that thc tirc
hamaa miagbc rcbailt. This idea of a physical and mental refor of humanity has
religious features. A 'demolishing branch' (1 1 5), an 'expiator proess', a 'purifcaton
of the mind' (69) must precede the accumulaton of knowledge. 'Our only hope of
salvation is to begin the whole labour of the mind again' (Preface) but only 'afer having
cleansed, polished, and levelled its surface' (I I 5). Preconceived notions (36), opinions
(42f, even the most common words (59, 1 21) 'must be abjured and renounced with
fr and solemn resoluton . . . so that the access to the kingdom of man, which is
founded on the sciences, may resemble that to the kingdom of heaven, where no
admission is conceded except to children' (68).
A refor of man is necessar for a corect science -but it is not sufcient. Science,
according to Bacon, not only orders events, it is als supposed to give physical reasons.
Thus Ptolemy and Copericus give us 'the number, situaton, moton, and peros of
the stars, as a beautful outside of the heavens, whilst the fesh and the entails are
wanting; that is, a well fabricated system, or the physical reasons and foundatons for a
just theor, that should not only solve phenomena, as almost any ingenious theor may
do, but show the substance, motions and infuences of the heavenly boies as they
really are.' Zdamte/Lcamiag,Chapter 4, quoted from Wiley Bok, New York,
1 94, p. 85. Cf. als the Nmam Oqaaam, op. cit., p. 371 : 'For let no one hope to
deterine the queston whether the eanh or heaven revolve in the diural moton,
unless he have frst comprehended the nature of spontaneous moton': the new man
needs a new physics in order to give substance to his astonomy. Galileo did not
succeed in providing such a physics.
ELEVEN 113
It is obvious that such a new world-view will take a long tme
appearing, and that we may never succeed to formulate it in its
entrety. It is extremely unlikely that the idea of the moton of the
e
arth
will at once be followed by the arrival, in full formal splendour,
of
all the sciences that are now said to consttute the body of'classical
physics' . Or, to be a little more realistc, such a sequence of events is
not
only extemely unlikely, it is impossible in princple, given the
nature of humans and the complextes of the world they inhabit.
Today Copercus, tomorrow Helmholtz - this is but a Utopian
dream. Yet it is only afer these sciences have arrived that a test can be
said to make sense.
This need to wait, and to igore large masses of critcal
obseratons and measurements, is hardly ever discussed in our
methodologies. Disregarding the possibility that a new physics or a
new astonomy might have to be judged by a new theor of knowledge
and might require entrely new tests, empirically inclined scientsts at
once confront it with the status quo and announce tiumphandy that 'it
is not in agreement with facts and received principles'. They are of
course right, and even tivially so, but not in the sense intended by
them. For at an early stage of development the contadicton only
indicates that the old and the new are di eet and out ofphae. lt does
not show which view is the bette one. A judgement of this kind
presupposes that the compettors confront each other on equal
terms. How shall we proeed in order to bring about such a fair
comparison?
The first step is clear: we must retain the new cosmology untl it has
been supplemented by the necessar auxliar sciences. We must
retain it in the face of plain and unambiguous reftng facts. We may,
of course, t to exlain our acton by saying that the critcal
obseratons are either not relevant or that they are illusor, but we
cannot support such an exlanaton by a single objectve reason.
Whatever exlanaton we give is nothing but a veral geture, a geode
invitaton to partcipate in the development of the new philosophy.
Nor can we reasonably remove the received theor of percepton
whic
h says that the obseratons are relevant, gives reasons for this
asserton, and is confirmed by independent evidence. Thus the new
view is arbitarily separated from data that suppored its predecessor
Science-loving philosophers, including those who call themselves 'critcal', are
quick to critcize thinker who do not share their pet idea. Bacon was ofen critcized
fo
r not
at once falling for Copericus. He wa critcized for this unspeakable crime by
p
hilo
soph
er whose own 'ratonalism' would never have aUowed Copricus to live.
An
exa
mple is K.R. Popper, The Oe Socet aad InEaic,Vol. 2, p. 1 6.
114 AGAI NST METHOD
and is made more 'metaphysical' : a new period in the histor of
science commences with a bakard mueet that returs us to an
earlier stage where theories were more vague and had smaller
empirical content. This backward movement is not just an accident;
it has a defnite fncton; it is essental if we want to overake the statu
qu, for it gives us the tme and the freedom that are needed for
developing the main vew in detail, and for fnding the necessar
auxliar sciences.
5
This backward movement is indeed essental - but how can we
persuade people to follow our lead? How can we lure them away from
a well-defned, sophistcated and empirically successful system and
make them tansfer their allegiance to an unfnished and absurd
hypothesis? To a hypothesis, moreover, that is contadicted by one
obseraton afer another if we only take the touble to compare it
with what is plainly shown to be the case by our senses? How can we
convnce them that the success of the statu quo is only apparent and is
bound to be shown as such in 500 years or more, when there is not a
single argument on our side (and remember that the illustatons I
used two paragraphs earlier derive their force from the successes of
classical physics and were not available to the Copemicans).
6
It is
clear that allegiance to the new ideas wlhave to be brought about by
means other than arguments. It will have to be brought about by
irational means such as propaganda, emoton, a hoc hypotheses, and
appeal to prejudices of all kinds. We need these 'irratonal means' in
order to uphold what is nothing but a blind faith untl we have found
the auxliar sciences, the facts, the arguments that tum the faith into
sound 'knowledge'.
It is in this context that the rise of a new secular class with a new
outlok and considerable contempt for the science of the schols, its
methods, its results, even for its language, becomes so imporant.
The barbaric Latn spoken by the scholars, the intellectual squalor of
academic science, its other-worldliness which is son interreted as
uselessness, its connecton with the Church - all these elements are
5. An example of a backward movement of this kind is Galileo's retur to the
kinematcs of the Cmtanelasand his disregard for the machiner of epiccles as
developed in the DcRnel. For an admirable ratieaal account of this step cf. Imre
Lakatos and Eli Zahar, 'Why Did Copericus' Research Programe Superede
Ptolemy's?', in ImrcLabtes,Fhiles@himlFapVol. I, Cambridge 1 978.
6. They were available to the sceptcs, especially to Aenesidemus, who pints out,
foUowing Philo, that no object appear as it is but is modifed by being combined with
air, light, humidit, heat, etc.; cf. DiegcLias,IX, 84. However, it seems that the
sceptcal view had only little influence on the development of moem atonomy, and
understndably s: one dos not star a movement by being reasonable.
ELEVEN 115
now lmued together with the Aristotelian cosmolog and the
contempt one feels for them is tansferred to ever single Aristotelian
argument.
7
This guilt-by-assoiaton does not make the arguments
less rational, or less conclusive, but it reduce their infuece on the
minds of those who are willing to follow Copericus. For Copericus
now stands for progress in other areas as well, he is a symbol for the
ideals of a new class that looks back to the classical tmes of Plato and
Cicero and forard to a free and pluralistc soiety. The associaton
of astonomical ideas and historical and class tendencies does not
produce new arguments either. But it engenders a frm commitent
to the heliocentic view- and this is all that is needed at this stage, as
we have seen. We have also seen how masterfully Galileo exploits the
situaton and how he amplifes it by ticks, jokes, and non-sequitur of
his own.
8
We are here dealing with a situaton that must be analysed and
understood if we want to adopt a more reasonable atttude towards
the issue between 'reason' and 'irratonality'. Reason grants that the
ideas which we intoduce in order to expand and to improve our
7. For these soial pressures cf. Olschki's magnifcent Ccchichtcdanprh-
lichmusscha/luh Litoatar. For the role of Puritanism cf. R.F. Jones, op. cit.,
Chapters V and V.
8. In a remarkable book, CHm Hoctic, Princeton, 1 987 (frst published, in
Italian, in 1 982), Pieto Redondi has described the groups both inside the Church (and
including the Pope himself) and outside of it who loked favourably upn new
scientfc developments, the views on perception, continuity, mater and moton that
had been explained by Galileo in hi>Zssoamong them. Being in direct confict with
the taditonal account of the Eucharist, the most imprant sacrament, these views
were considerably more dangerous than Copemicanism and could be tolerated only as
long as the groups and the Pope himself had the upper hand in the complex plitcal
developments of the tme (Thirt Years' War; French and Spanish plitcs; the French
alliance with the Pope). The politcal reversal of the Pope's forunes, the accustons of
leniency towards heretics that were raised against him on plitcal grooods cat a
shadow on his atttude towards scientifc maters as well (here, too, he seemed to
suppor heresy) and made protective measures necessar. Redondi ties to show (a)
that the physics of the time was connected with theological dotines such as the
dotrine of the Eucharist and that a histor of science that neglects the connecton
becomes incomprehensible and (b) that the attitude towards scientfc problems
caused by the connecton and thus the atttude towards innovaton changed with the
political climate. The second par of (b) may well b te but there is only weak
evid
ence to suppr the rest: what Galileo says about atomism in the Zssois much
to brief and indefnite to confict with tansubstantation (it is an aside almost, not an
elaborate statement) and with the exception of a rather problematc doument no such
conflict was perceived. (Cf. R.S. Westfall, Esss o thc 1nm e/Caldce, Vatcan
Obs
erator Publicatons, 1 989, pp. 84f.) What is valuable in Redondi's account is
that
he widens the domain of possible infuences and thus underines the
anachronistc) belief that then as now scientfc ratonality wa resticted to the
Int
eral problem situaton of a scientfc disipline.
116 AGAI NST METHOD
knowledge may arse in a very disorderly way and that the orgin of a
partcular point of view may depend on class prejudice, passion,
personal idiosyncrasies, questons of stle, and even on error, pure
and simple. But it also demands that in judging such ideas we follow
certain well-defined rules: our ealuation of ideas must not be
invaded by irratonal elements. Now, what our historical examples
seem to show is this: there are situatons when our most liberal
judgements and our most liberal rules would have eliminated a point
of view which we regard today as essental for science, and would not
have permitted it to prevail - and such situatons ocur quite
frequently. The ideas survived and they now are said to be in
agreement with reason. They survived because prejudice, passion,
conceit, errors, sheer pigheadedness, in short because all the
elements that characterize the context of discovery, op osed the
dictates of reason and becuse thee i"ational eleets were penitted to
hae their way. To express it diferently: Coeicnism and othe
'rational' views eist tody onl because reason was oerled at some time
in their past. (The opposite is also true: witchcraft and other
'irratonal' views have ceased to be influental only because reason was
overruled at some tme in therpast.)9
Now, assuming that Copercanism is a Good Thing, we must also
admit that its survival is a Good Thing. And, considering the
conditons of its survival, we must further admit that it was a Good
Thing that reason was overruled in the 16th, 1 7th and even the 1 8th
centuries. Moreover, the cosmologists of the 1 6th and 1 7th centuries
did not have the knowledge we have today, they did not know that
Copercanism was capable of giving rise to a scientfic system that is
acceptable from the point of view of'scientfic method'. They did not
know which of the many views that exsted at their tme would lead to
fture reason when defended in an 'irratonal' way. Being without
such guidance they had to make a guess and in making this guess they
could ony follow their inclinatons, as we have seen. Hence it is
advisable to let one's inclinatons go against reason in any circmstance,
for it makes life less constrained and science may profit from it.
9. These consideratons refte J. Dorling, who, in ntuheamal /rthcFhiles@
e/5n,Vol. 23, 1 972. 1 89f, presents my 'iratonalism' as a presuppsiton of my
research, not a a result. He contnues: ' . . . one would have thought that the
philosopher of science would be most interested in picking out and analysing in detail
those scientfic arguments which did seem to be ratonally reconstuctble.' One would
have thought that the philosopher of science would be most interested in picking out
and analysing in detail those moves which are necessar for the mamtof science.
Such moves, I have oed to show, ofen resist ratonal reconstucton.
ELEVEN 117
I t is clear that this argument, that advises us not t o let reason
overrule our inclinatons and occasionally to suspend reason
altogether, does not depend on the historical material which I have
presented. If my account of Galileo is historically correct, then the
argument stands as formulated. If it turs out to be a fair-tale, then
this fair-tale tells us that a conflict between reason and the
preconditons of progress is possible, it indicates how it might arise,
and it forces us to conclude that our chances to progress may be
obstructed by our desire to be ratonal. And note that progress is here
defined as a ratonalistc lover of science would defne it, i.e. as
entailing that Copercus is better than Aristotle and Einstein better
than Newton. Of course, there is no need to accept this defniton,
which is certainly quite narrow. I use it only to show that an idea of
reason accepted by the majority of ratonalists may prevent progress
as defined by the ver same majority. I now resume the discussion of
some details of the tansiton from Aristotle to Copericus.
The frst step on the way to a new cosmology, I have said, is a step
back: apparently relevant evidence is pushed aside, new data are
brought in by a hoc connectons, the empirical content of science is
drastcally reduced. Now the cosmology that happens to be at the
centre of attenton and whose adopton causes us to carry out the
changes just described differs from other views in one respect only: it
has features which at the tme in queston seem attractve to some
people. But there is hardly any idea that is totally without merit and
that might not also become the startng point of concentated effort.
No inventon is ever made in isolation, and no idea is, therefore,
completely without (abstract or empirical) support. Now if partal
support and partal plausibility sufce to start a new tend - and I
have suggested that they do - if startng a new tend means taking a
step back from the evidence, if any idea can become plausible and can
receive partal support, then the step back is in fact a step forard,
and away from the tyranny of tghtly-knit, highly corroborated, and
gracelessly presented theoretcal systems. 'Another diferent error',
writes Bacon on precisely this point,
1
0
'is the . . . peremptor
reducton of knowledge into ars and methods, from which tme the
sciences are seldom improved; for as young men rarely grow in
stature after their shape and limbs are fully formed, so knowledge,
whilst it lies in aphorisms and observatons, remains in a growng
10. Zdaact e/Lcamiag(1 65 editon), New York, 1 94, p. 21 . Cf. also the
NmamOqaaam,Aphorisms 79, 86, as well asJ.W.N. Watkins' splendid little book
Hebbc'5ste/Iq,London, 1965, p. 1 69.
118 AGAI NST METHOD
state; but when once fashioned into methods, though it may be
further polished, illustated and ftted for use, is no longer increased
in bulk and substance.'
The similarity with the arts which has often been asserted arises at
exactly this point. Once it has been realized that a close empirical ft is
no virtue and that it must be relaxed in tmes of change, then style,
elegance of expression, simplicity of presentaton, tension of plot and
narratve, and seductveness of content become important features of
our knowledge. They give life to what is said and help us to overcome
the resistance of the obseratonal material.
1 1
They ceate and
maintain interest in a theor that has been partly removed from the
obseratonal plane and would be inferior to its rivals when judged by
the customar standards. It is in this context that much of Galileo's
work should be seen. This work has often been likened to
pragand
1
2
-and propaganda it certainly is. But propaganda of this
kind is not a marginal affair that surrounds allegedly more substantal
means of defence, and that should perhaps be avoided by the
'professionally honest scientst'. In the circumstances we are
considering now, pragand is ofthe esece. It is of the essence
because interest must be created at a tme when the usual
methodological prescriptons have no point of attack; and because
this interest must be maintained, perhaps for centuries, untl new
reasons arrive. It is also clear that such reasons, i.e. the appropriate
auxliar sciences, need not at once tum up in full foral splendour.
They may at frst be quite inartculate, and may even confict with the
exstng evidence. Agreement, or partal agreement, with the
cosmolog is all that is needed in the beginning. The agreement
shows that they are at least releant and that they may some day
produce full-fedged positve evidence. Thus the idea that the
telescope shows the world as it really is leads to many diffcultes. But
the support it lends to, and receives from, Copericus is a hint that we
might be moving in the right directon.
We have here an extremely interestng relaton between a general
view and the partcular hypotheses which consttute its evidence. It is
often assumed that general views do not mean much unless the
relevant evidence can be fully specifed. Camap, for example, asserts
that 'there is no independent interretaton for [the language in
ters of which a certain theor or world-view is forulated] . The
system T[the axoms of the theor and the rules of derivaton] is itself
1 1 . 'What resttutes to scientfc phenomenon its life, is an' (1hcDmq e/Zaai
Nia,Vol !, p. 277).
12. Cf. A. Koyre, Etu&Calilac,Vol. Ill, Paris, 1939, pp. 53f.
ELEVEN 119
an uninterreted postulate system. [Its] terms obtain only an indirect
and incomplete interretaton by the fact that some of them are
con
nected by correspondence rules with obseratonal terms.'
1 3
'There is no independent interretaton,' says Carap and yet an idea
such as the idea of the moton of the earth, which was inconsistent
with the contemporar evidence, which was upheld by declaring this
evdence to be irrelevant and which was therefore cut from the most
important facts of contemporar astonomy, managed to become a
nucleus, a crstallizaton point for the aggregaton of other
inadequate views which gradually increased in artculaton and fnally
fused into a new cosmolog including new kinds of evdence. There
is no better account of this proess than the descripton which John
Stuart Mill has left us of the vcissitudes of his educaton. Referring
to the explanatons which his father gave him on logical matters he
writes: 'The explanatons did not make the matter at all clear to me at
the tme; but they were not therefore useless; they remained as a
nucleus for my obseratons and refectons to crstallize upon; the
import of his general remarks being interpreted to me, by the
partcular instances which came under my notce afearr.
1
4
In
exactly the same manner the Coperican vew, though devoid of
cognitve content from the point of vew of a stict empiricism or else
refuted, was needed in the constucton of the supplementar
sciences even before it became testable with their help and even
before it, in tur, provded them with supportng evdence of the
most forceful kind.
There is a further element in this tapesty of moves, infuences,
beliefs which is rather interestng and which received attenton only
recently - the role of patonage. Today most researchers gain a
reputaton, a salar and a pension by being assoiated with a
university and/or a research laborator. This involves certain
conditons such as an ability to work in teams, a willingness to
subordinate one's ideas to those of a team leader, a harmony between
one's ways of doing science and those of the rest of the profession, a
certain style, a way of presentng the evdence - and so on. Not
everyone fts conditons such as these; able people remain
unemployed because they fail to satsf some of them. Conversely the
reputaton of a university or a research laboratory rises with the
. 13. 'The Methodologcal Character of Theoretcal Concepts', Miaaceta5tadic
tathc
Fhiles@he/5occ,Vol. I,
Minneaplis, p. 47.
14. Zatebieap, quoted from Esstial erH e/eha5tmuMill, ed. Lerer,
New
York, 1 965, p. 21 .
120 AGAI NST METHOD
reputaton of its members. I n Galileo's tme patronage played a
similar role. There were certain ways of gaining a paton and of
keeping him. The paton in tum rose in estmaton only if he
succeeded to attact and to keep individuals of outstanding
achievement. According to Westfall, 1 5 the Church permitted the
publicaton of Galileo's Dialoge in the full kowledge of the
contoversial matters contained in it '[n]ot least because a Pope
[Urban VIII] who gloried in his reputaton as a Maecenas, was
unwilling to place it in jeopardy by saying no to the light of his tes',
and Galileo fell because he violated his side of the rules of
patronage.
1 6
Considering all these elements, the 'Rise of the Copercan
World-View' becomes a complicated matter indeed. Accepted
methodological rules are put aside because of social requirements
(atons need to be persuaded by means more efectve than
argument), instuments are used to redefine experience instead of
being tested by it, local results are extapolated into space despite
reasons to the contary, analogies abound - and yet all this turs out,
in retospect, to have been the correct way of circumventng the
restictons implied by the human conditon. This is the material that
should be used to get better insight into the complex process of
kowledge acquisiton and improvement.
To sum up the content of the last five chapters:
When the 'Pythagorean idea' of the moton of the earth was revived
by Copericus it met with difcultes which exceeded the difcultes
encountered by contemporary Ptolemaic astonomy. Stictly speak
ing, one had to regard it as refuted. Galileo, who was convinced of the
tuth of the Copercan view and who did not share the quite
common, though by no means universal, beliefin a stable experience,
looked for new kinds of fact which might support Copericus and
stll be acceptable to all. Such facts he obtained in two diferent ways.
First, by the inventon of his telecoe, which changed the sesor core
of everyday experience and replaced it by puzzling and unexplained
phenomena; and by his prnciple ofrelatiit and his dynamic, which
changed its concetual componets. Neither the telescopic phenomena
I S. op. cit., p. 73.
1 6. Funher details on these matters in Chapter 8, fotote 1 2 of the present essay,
Westfall, op. cit., and M. Biagioli, CahceCeanio.M. Finochiaro, Cahceaadthc Zn
e/Rcmeaiag,Dordrecht, 1 980, has commented on Galileo's use of rhetoric, while M.
Pera and W.R. Shea (eds), Fmamiag5occ- 1hcZne/5otcRhctenc, 1 991 , and
especialy Marcello Pera, 5occaadRhctenc, fonhcoming, comment on scientf
c
rhetoric in general.
ELEVEN 121
nor the new ideas of moton were acceptable to common sense (or to
the Aristotelians). Besides, the associated theories could be easily
shown to be false. Yet these false theories, these unacceptable
phenomena, were transformed by Galileo and converted into strong
support of Copericus. The whole rich reseroir of the everday
experience and of the intuition of his readers is utlized in the
argument, but the facts which they are invited to recall are arranged
in a new way, approxmations are made, known effects are omitted,
different conceptual lines are drawn, so that a new kind ofexerience
arises, manufacured almost out of thin air. This new experience is
then solidied by insinuatng that the reader has been familiar with it
all the tme. It is solidifed and soon accepted as gospel truth, despite
the fact that its conceptual components are vastly more speculatve
than are the conceptual components of common sense. Following
positvistc usage we may therefore say that Galileo's science rests on
an illustrated metaphysic. The distorton permits Galileo to advance,
but it prevents almost everone else from making his effort the basis
of a critcal philosophy (for a long tme emphasis was put either on his
mathematcs, or on his alleged experiments, or on his frequent appeal
to the 'tuth', and his propagandistc moves were altogether
neglected). I suggest that what Galileo did was to let refuted theories
support each other, that he built in this way a new world-view which
was only loosely (if at all!) connected with the preceding cosmology
(everyday experience included), that he established fake connectons
with the perceptual elements of this cosmolog which are only now
being replaced by genuine theories (physiological optcs, theor of
contnua), and that whenever possible he replaced old facts by a new
type of experience which he simply invented for the purose of
supportng Copericus. Remember, incidentally, that Galileo's
procedure drastcally reduces the content of dynamics: Aristotelian
dynamics was a general theor of change comprising loomoton,
qualitative change, generaton and corruption. Galileo's dynamics
and its successors deal with locomotion only, other kinds of moton
being pushed aside with the promissory note (due to Democritos)
that locomoton will eventually be capable of comprehending all
moton. Thus, a comprehensive empirical theory of moton is
repla
ced by a much narrower theory plus a metaphysics of moton,
iust as an 'empirical' experience is replaced by an experience that
con
tains speculatve elements. This, I suggest, was the actual
p
roc
edure followed by Galileo. Proceeding in this way he exhibited a
style, a sense ofhumour, an elastcity and elegance, and an awareness
of the valuable weaknesses of human thinkng, which has never been
equalled in the histor of science. Here is an almost inexhaustble
122 AGAI NST METHOD
source of material for methodological speculaton and, much more
importantly, for the recover of those features of knowledge which
not only inform, but which also delight us. 1 7
17. A few years ago Marn Gardner, the pitbull of scientsm, published a arcle
with the ttle 'Ant-Science, the Stange Cae of Paul Feyerabend' CntimIaqaiq,
Winter 1 98283. The valiant lghter seems to have overloked these and other
passages. I am not against science. I praise it foremost practtoners and (next chapter)
suggest that their proedures be adopted by philosophers. What I object to i narow
minded philosophical interference and a narow-minded extension of the latest
sientfc fahions to all areas of human endeavour - in shor what I object to is a
ratonalistc interretaton and defence of science.
1 2
Galleo 's method work in other feld a well. For eample, it can be used to
eliminate the eisting arments against materalism and to put an ed to
the philosophical mind/body problem. (he corrsponding scientfc
prbl
ems reain untouched, howrer.) It des not fllow that it should be
unie
rall applied.
Galileo made progress by changing familiar connectons between
words and words (he introduced new concepts), words and
impressions (he intoduced new natural interpretatons), by using
new and unfamiliar principles such as his law of inerta and his
principle of universal relatvity, and by altering the sensor core of his
obseraton statements. His motve was the wish to accommodate the
Coperican point of view. Copemicanism clashes with some obvious
facts, it is inconsistent with plausible, and apparently well
established, principles, and it does not ft in with the 'grammar' of a
commonly spoken idiom. It does not ft in with the 'form oflife' that
contains these facts, principles, and grammatcal rules. But neither
the rules, nor the principles, nor even the facts are sacrosanct. The
fault may lie with them and not with the idea that the earth moves. We
may therefore change them, create new facts and new grammatcal
rules, and see what happens once these rules are available and have
become familiar. Such an attempt may take considerable tme, and in
a sense the Galilean venture is not finished even today. But we can
already see that the changes were wise ones to make and that it would
hav
e been foolish to stick with the Aristotelian form of life to the
exclusion of everything else.
With the mind/body problem, the situaton is exactly the same. We
ha
ve
again obseratons, concepts, general principles, and gram
matcal rules which, taken together, consttute a 'form of life' that
app
aren
tly supports some views, such as dualism, and excludes
o
the
rs,
such as materialism. (I say 'apparently' for the situaton is
m
uc
h less
clear here than it was in the astronomical case.) And we
may
again
proceed in the Galilean manner, look for new natural
123
124 AGAI NST METHOD
interretatons, new facts, new grammatcal rules, new principles
which can accommodate materialism and then compare the total
systems - materialism and the new facts, rules, natural interreta
tons, and principles on the one side; dualism and the old 'forms of
life' on the other. Thus there is no need to try, like Smart, to show
that materialism is compatble with the ideology of common sense.
Nor is the suggested procedure as 'desperate' (Armstong) as it must
appear to those who are unfamiliar with conceptual change. The
procedure was commonplace in antquity and it occurs wherever
imaginatve researchers stike out in new directons (Einstein and
Bohr are recent examples).
1
So far the argument was purely intellectual. I tied to show that
neither logic nor experience can limit speculaton and that
outstanding researchers often transgressed widely accepted limits.
But concepts have not only a logical content; they also have
associatons, they give rise to emotons, they are connected with
images. These associatons, emotons and images are essental for
the way in which we relate to our fellow human beings. Removing
them or changing them in a fundamental way may perhaps make our
concepts more 'objectve', but it often violates important social
constaints. It was for this reason that Aristotle refused to abandon an
intuitve view of human beings simply because a more physiological
approach showed successes in a limited domain. For him a person
was a social entty and defined by his or her functon in the city no
matter what atomists or physicians involved in theory might say.
Similarly the Roman Church, being interested in souls and not only
in astonomical ticks, forbade Galileo to present his badly founded
guesses as tuths and punished him when he violated the prohibiton.
The tial of Galileo raises important questons about the role
products of specialists, such as abstact knowledge, are supposed to
play in society. It is for this reason that I shall now give a brief account
of this event.
I . For a more detailed disussion the reader is refered to my Fhihs@himFqm
,
Vol. I , Chapter 9 and 10.
1 3
Te Church at the time ofGalileo not onl ket closer to reason as dfned
the and, in par, ee now; it also considred the ethical and social
conse
queces ofCali/ eo s views. Its indictmet ofGali/eo was rational and
onl oporunism and a lack ofperpectie can dmand a reision.
There were many trials in the 1 7th centur. The proceedings started
either with accusatons made by private partes, with an offcial act by
a public ofcer, or with an inquir based on sometmes rather vague
suspicions. Depending on the locaton, the distibuton of jurisdic
ton and the balance of power at a partcular tme, crimes might be
examined by secular courts such as the courts of kings or of free
cites, by Church courts, such as the spiritual courts attached to ever
episcopate, or by the special courts of the Inquisiton. After the
middle of the 1 2th centur the episcopal courts were greatly aided by
the study of Roman law. Lawyers became so infuental that, even if
wholly untained in canon law and theolog, they had a much better
chance of high preferment than a theologian.
'
The inquisitorial
process removed safeguards provided by Roman law and led to some
well-publicized excesses. What has not been publicized to the same
extent is that the excesses of royal or secular courts often matched
those of the Inquisiton. It was a harsh and cruel age.
2 By 1 600 the
I . For this complaint (made by Roger Bacon) cf. H. Ch. Lea, Z Huteq e/thc
Ioqaisitieo e/thcMidlcZgc,Vol. I, p. 309. Chapter ixf explain the details of the
inquisitorial procedure, the ways in which they difered from other proedures and the
reas
ons for the diference. Cf. also G.G. Coulton, IoqauitieoaodLibcq, Boston,
1
959,
Chapters XI-XV.
2.
Charles Henr Lea, the great liberal historian, writes: 'On the whole we may
con
clude
that the secret prisons of the Inquisition were less intolerable places of abode
than
the
episcopal and public gaols. The general policy respectng them was more
hu
mane and enlightened than that of other jurisdictons, whether in Spain or
els
ewh
ere, although negligent superision allowed of abuses and there were ample
res
ourc
es of rigor in resere, when the obstinacy of the impenitent was to be broken
dow
n.'
Histeq
e/thcIoqaisitieo io 5paio,Vol. 2, New York, 1 906, p. 534. Prisoner
a
ccu
sed
before secular couns ocasionally commited crimes under the jurisdiction of
1 25
126 AGAI NST METHOD
Inquisiton had lost much of its power and aggressiveness. This was
tue especially in Italy and more partcularly in Venice.
3
The courts of the Inquisiton also examined and punished crimes
concering the producton and the use of knowledge. This can be
explained by their origin: they were supposed to exterminate heey,
i.e. complexes consistng of actons, assumptons and talk making
people inclined towards certain beliefs. The surprised reader who
asks what knowledge has to do with the law should remember the
many legal, social and financial obstacles knowledge-claims face
today. Galileowanted his ideas to replace the exstng cosmolog, but
he was forbidden to work towards that aim. Today the much more
modest wish of creatonists to have their views taught in schools side
by side with other and competnf views runs into laws settng up a
separaton of Church and State. Increasing amounts of theoretcal
and engineering informaton are kept secret for military reasons and
are thereby cut off from interatonal exchange. 5 Commercial
interests have the same restictve tendency. Thus the discovery of
superconductvit in ceramics at (relatvely) high temperatures,
which was the result of interatonal collaboraton, son led to
protectve measures by the Aerican goverment.6 Financial
arrangements can make or break a research programe and an entre
profession. There are many ways to silence people apart from
forbidding them to speak-and all of them are being used today. The
proess of knowledge producton and knowledge distibuton was
the Church so that they might be handed over to the Inquisiton: Henr Kamen, Die
5paaucheIaqauita,Munich, I 980, p. I 7.
3. In 1356 the secular ofcials ofVenice forbade the Inquisitor ofTreviso to t his
prisner, seized his inforant and torured them on the charge of pilfering the
proprty of the accused. Lea, IaqauitieaiatheMm Zgc,Vol. ii, p. 273.
4. A comprehensive repr of one of the tials that resulted from the conflict has
been published in 5n,Vol. 215, 1982, pp. 934f. Many other tials foUowed.
5. It seems that the need for secrec in nuclear matter was frt raised by te
sientsts themselves. Cf. the repr and the douments in Spncer R. Wear and
Gertude Weiss-Szilard (eds), Lw5zimrdHu Imieae/theF,Cambridge, Mas.,
1978, esp. Chapter 2f. Cf. als the material on the Oppnheimer cas. The inventor
of the telescop was forced to secrec as the militar imprance of the contivance was
son realized. Cf. Chapter 8, fotote 24.
Research teams become ver secretve when approaching what they think is a Big
Disover. Afer al, what is.at stake are patents, consultancies in indust, money and,
prhaps, the honour of a Nobel Prze. For a spcia case cf. R.M. Hae,
5qo en, London, 1988. The manipulaton of knowledge by the cour is
disussed, with many examples, by Peter W. Hubr, Cm: Rnge, New York,
1 991 .
6. 5cm, Vol. 237, 1987, pp. 476f and 593f. An imprant step toward
s
exclusivenes consisted in a igning par of the research to the militar.
THI RTEEN 127
never the free, 'objectve', and purely intellectual exchange
ratonalists make it out to be.
The tial of Galileo was one of many tials. It had no special
features except perhaps that Galileo was teated rather mildly,
despite his lies and attempts at decepton. 7 But a small clique of
intellectuals aided by scandal-hung writers succeeded in blowing it
up to enormous dimensions so that what was basically an altercaton
between an exert and an insttuton defending a wider view of thgs
now looks almost like a battle between heaven and hell. This is
childish and also ver unfair towards the many other victms of 1 7th
centur justce. It is especially unfair towards Giordano Bruno, who
was bured but whom scientfcally mded itellectuals prefer to
forget. It is not a concer for humanity but rather party interests
which play a major role i the Galileo hagiography. Let us therefore
take a closer look at the matter.
8
The so-called trial of Galileo consisted of two separate
proeedigs, or tials. The first occurred in 1 61 6. The Copercan
doctrine was examed and critcied. Galileo received an order, but
he was not punished. The second tial took place in 1 63233. Here
the Copercan doctine was no longer the point at issue. Rather,
what was considered was the queston of whether Galileo had obeyed
the order given him in the first tial, or whether he had deceived the
iquisitors ito believing that the order had never been issued. The
proceedigs of both tials were published by Antonio Favaro in Vol.
1 9 of the Natonal Editon of Galilean material. The suggeston,
rather popular i the 1 9th centur, that the proceedings contained
7. An example is Galileo's reply to the inquiries of 1 2 April 1 633: Maurice A.
Finochiaro, 1hcCalilceZ] air,Berkeley and Los Angeles, 1 989, p. 262, the firt two
lines. The reacton of an admirer is characteristc: 'This absurd pretence . . . '
Geymonat, op. cit., p. 1 49.
8. It
cannot be denied that pressure groups, peronal grievances, envy, the fact that
Galileo, 'being too infatated wth his own genius' was 'unsuferable' (Westall, op.
cit., pp. 52, 38) and the rles of patonage played an imporant role a they, or similar
circumstances, do at ever tial. However, the tensions between various groups of the
Church on the one side and the demands for scientfic autonomy on the other were
real enough; after all, their modem successors (should the sciences be given the rn of
our educatonal instttons and of soiety a a whole or should they be teated like any
other special interest group?) are stll wth us. Here the Church did the right thing: the
sciences do aethave the last word in humane matters, knowledge included.
The main doument peraining to the tial were assembled and tanslated wth
com
ments
and an intoducton by Finochiaro, op. cit. I shall use his tanslatons.
A
cco
unts
of the tials and their problems are found in G. de Santllana, 1hcCnmce/
C
a
g
e,
Chicago, 1 954, Geyonat, op. cit., Redondi, op. cit., and, most recendy, in
W
est
all.
128 AGAI NST METHOD
falsifed documents and that the second tial was therefore a farce,
seems no longer acceptable.
9
The frst trial was preceded by denunciatons and rumours, in
which greed and envy played a part, as in many other tials. The
Inquisiton started to examine the matter. Experts (qualicatores)
were ordered to give an opinion about two statements which
contained a more or less correct account of the Coperican
doctrine.
10
Their decision
1 1
concered two points: what would
today be called the scietic contet of the doctrine and its ethicl
(social implictions.
On the frst point the experts declared the doctine to be 'foolish
and absurd in philosophy' or, to use modem terms, they declared it to
be unscientfc. This judgement was made without reference to the
faith, or to Church dotine, but was based exclusively on the
scientfc situaton of the tme. It was shared by many outstanding
scientsts (Tycho Brahe having been one of them) and it wa
co"ea
12
when based on the facts, the theories and the standards of
9. One of the authors of the suggeston was the Galileo scholar Emil Wohlwill. His
reasons, rather impressive at the time, are given in his DoIaqauitiempmms&Cm
Cahh,Berlin, 1 870. According to Wohlwill two douments of the proeedings, dated
25 Feb. 1 61 6 and 26 Feb. 1616 (Finochiaro, op. cit., pp. l 47f) are mutually
contadictor. The frt advises Galileo to treat Copericus as a mathematcal moel;
should he reject the advice, then he is forbidden to menton Copericus in any for
whatsoever. In the second doument Galileo is advised as abve and immediately
forbidden (i.e. without waitng for his reacton) to menton Copericus. Wohlwill
thought the second doument to be a forger. This seems now refted. Cf. de
Santllana, Chapter 13. Stllman Drake (appendix to Geymonat) devised a ver
inhguing hypthesis to explain the discrepancy.
10. Some critcs used idiosyncrasies in the forulaton as prof of a lack of com
prehension on the pan of the expens. But there was no need for the inquisitors to stck
closely to the language of the author they examined. Their account ofCopemicanism
was clear enough without such textual puritanism.
I I . Finochiaro, op. cit., p. 146.
12. Note that in rendering my judgements I rely on standards subsribed to b
y
many moem scientst and philosopher of science. Retured to the early 1 7th
centur, these champions of ratonality would have judged Galileo as the Arstotelians
judged him then. Michelson, for example, would have been aghast at Galileo's attempt
to get knowledge out of an instment as little understo as the telesope and
Rutherford, who was never to happy abut the theor of relatvity, would have
prouced one of his characteristc rde remarks. Salvador Luria, an outtanding
microbiologist who favours theories decidable by 'clear-cut expermental step[s)'
,
would have relegated the debate to 'outfelds of science' like 'soiolog' and would
have stayed away from it ( 5letMhiac,amk1ct1abc,New York, 1 985, pp. 1 15,
1 19). For what Galileo suggested was no less than to regard as te a theor which had
only analogies in it favour and which suffered from numerous difcultes. And he
made this suggeston in public while even today it is a deadly sin for a scientst to
address the public before having consulted his peer (example in A. Pickering,
THI RTEEN 129
the tme. Compared with those facts, theories and standards the idea
of the moton of the earth was as absurd as were Velikovsky's ideas
when compared with the facts, theories and standards of the fftes. A
moder scientst really has no choice in this matter. He cannot cling
to his own very strict standards and at the same tme praise Galileo
for defending Copericus. He must either agree with the frst part of
the judgement of the Church experts, or admit that standards, facts
and laws never decide a case and that an unfounded, opaque and
incoherent doctine can be presented as a fundamental tuth. Only
fe
w admirers of Galileo have an inkling of this rather complex
situat
on.
The situaton becomes even more complex when we consider that
the Copericans changed not only views, but also standards for
judging views. Aristotelians, in this respect not at all unlike moder
epidemiologsts, molecular biologsts and 'empirical' sociologsts
who insist either on the examinaton oflarge statstcal samples or on
'clearcut experimental steps' in Luria's sense, demanded stong
empirical support while the Galileans were content with far
reaching, unsupported and partally refuted theories.
1 3
I do not
critcize them for that; on the contary, I favour Niels Bohr's 'this is
not crazy enough'. I merely want to reveal the contadicton in the
actons of those who praise Galileo and condemn the Church, but
become as stict as the Church was at Galileo's tme when turng to
the work of their contemporaries.
-
On the second point, the soial (ethical) implicatons, the experts
declared the Copercan doctine to be 'forally heretcal'. This
means it contadicted Holy Scripture as interpreted by the Church,
and it did so in full awareness of the situaton, not inadvertently (that
would be 'material' heresy).
'Co
nstaints on Contoversy: the Case of the Magnetc Monopole', 5enal5twice/
5occ, Vol. 1 1 , 1 981 , pp. 63f. All this is realized neither by 'progressive' (i.e.
scien
tfcally inclined) princes of the Church nor by scientsts, so the discussion of the
'ta
l of Galileo' ocurs in a dream world with only little relaton to the real world we
and
Galileo inhabit. Further arguments on that point are found in Chapter 9 of
FarcllteRcaseaand Chapter 19 below.
l 3. As indicated in Chapter 8, fotote 1 , Galileo's law of inerta was in conflict
with the Coperican as well as the Keplerian teatment of planetar moton. Galileo
hop
ed for future accommodatons. That was a sensible thing to do but not in
ageement with some standards ofhis tme and of today. Today a similar clash between
theoretcians and empiricist ocurs in the feld of epidemiolog. There are theoretcal
reasons to expect that X-rays and other fors of partculate radiaton consttute a
c.ncer-risk down to the smallest dose. Many epidemiologists demand empirical prof,
however, though it is clear that events, when occurring below a cerain treshold of
fr
eq
uen
cy,
cannot be detected in that way.
130 AGAI NST METHOD
The second point rests on a series of assumptons, among them the
assumpton that Scripture is an important boundar conditon of
human existence and, therefore, of research. The assumpton was
shared by all great scientsts, Copericus, Kepler and Newton among
them. According to Newton knowledge fows from two sources - the
word of God - the Bible - and the works of God - Nature; and he
postulated divine interentons in the planetar system, as we have
seen.
1
4
The Roman Church in additon claimed to possess the exclusive
rights of exploring, interpretng and applying Holy Scripture. Lay
people, according to the teaching of the Church, had neither the
knowledge nor the authority to tamper with Scripture and they were
forbidden to do so. This comment, whose rigidity was a result of the
new Tridentne Spirit,
1
5 should not surprise anyone familiar with
the habits of powerful insttutons. The atttude of the American
Medical Associaton towards lay practtoners is as rigid as the
atttude of the Church was towards lay interpreters - and it has the
blessing of the law. Experts, or ignoramuses having acquired the
formal insignia of expertse, always tried and often succeeded in
securing for themselves exclusive rights in special domains. Any
critcism of the rigidity of the Roman Church applies also to its
modem scientfic and science-connected successors.
Turing now from the form and the administratve backing of the
objecton to its content we notce that it deals with a subject that is
gaining increasing importance in our own tmes - the quality of
human exstence. Heresy, defined in a wide sense, meant a deviaton
from actons, atttudes and ideas that guarantee a well-rounded and
sanctfed life. Such a deviaton might be, and occasionally was,
encouraged by scientfic research. Hence, it became necessar to
examine the heretcal implicatons of scientfic developments.
Two ideas are contained in this atttude. First, it is assumed that
the quality oflife can be defined independently of science, that it may
clash with demands which scientsts regard as natural ingredients of
their actvity, and that science must be changed accordingly.
Secondly, it is assumed that Holy Scripture as interpreted by the
Holy Roman Church adumbrates a correct account of a well
rounded and sanctfied life.
1 4. Chapter 5, footote 4. See also the literature in fotote 6 to Chapter
4.
According to Galileo (letter to Grand Duchess Christna) the idea of the rwo source
s
gos back to Terullian m.Maroeac(E. Evans, ed.), I , 1 8.
1 5. For the exact wording see Denzinger-Schonmeter, Eachindiea5mblem
m,
36th editon, Freiburg, 1 976, pp. 365f.
THI RTEEN 131
The second assumpton can be rejected without denying that the
Bible is vastly richer in lessons for humanit than anything that might
ever come out of the sciences. Scientfc results and the scientfc
ethos (if there is such a thing) are simply to t a foundaton for a
life worth living. Many scientsts agree with this judgement.
1
6
They agree that the qualit of life can be defned independently of
science -which is the frst part of the frst assumpton. At the tme of
Galileo there existed an insttuton -the Roman Church -watching
over this qualit in its own partcular way. We must conclude that the
second point - Copericus being 'formally heretcal' - was
connected with ideas that are urgently needed today. The Church
was on the right tack.
But was it perhaps mistaken in rejectng scientfc opinions
inconsistent with its idea of a Good Life? In Chapter 3 I argued that
knowledge needs a pluralit of ideas, that well-established theories
16. Thus Konrad Lorenz in his interestng if somewhat suprcial bok DicZcht
1eda
q
at ZniluimMschhot,Munich, 1984 (frt published in 1 973), p. 70
writes: 'The eroneous belief that only what can be ratonally grasped or even what can
be proved in a scientfc way consttutes te slid knowledge of mankind ha disatous
consequences. It prompts the "sientficaly enlightened" younger generaton to
discard the immense teaures of knowledge and wisdom that are contained in the
taditons of every ancient culture and in the teachings of the great world religions.
Whoever thinks that al this is without sigifcance naturally succumbs to another,
equally pricious mistake, living in the conviction that sience is able, a a matter of
coure, to create from nothing, and in a ratonal way, an entre culture with all its
ingredients.' In a similar vein J. Needham, initator and par-author of a great history
of Chinese science and technolog, speaks of 'scientfc opium', meaning by it 'a
blindness to the sufering of other'. 1imc, thcRqrchiagRno,Nottngham, 1 986.
'Ratonalism', writes Peter Medawar (dicctea\eaag5citut,New York, 1979, p.
l 0 I), 'falls shor of answering the many simple and childlike questons pople like to
ask; questons about origins and purpses such as are often contemptuously dismissed
a non-questons, or pseudo-questons, although pople understand them clearly
enough and long to have an answer. These are intellectual pains that ratonalists -like
bad physicians confronted by ailments they cannot diagnose or cure - are apt to
dismiss as "imaginaton".'
The clearest and most perceptive statement is found in Jacques Monod, Chaaccaad
Ncccsi, New York, 1 972, p. 1 70 (text in brackets from p. 169): 'Cold and austere,'
writes Monod, 'propsing no explanaton but imposing an acetc renunciaton of all
other spiritual fare, [the idea that objectve knowledge is the only authentc source of
tuth ) was not of a kind to allay anxety but aggravated it instead. By a single stoke it
clai
med to sweep away the taditons of hundreds of thousands of years, which had
become one with human nature itself. It wrote an end to the ancient animist convenant
between man and nature, leaving nothing in place of that precious bond but an anxous
quest in a frozen universe of solitude. With nothing to recommend it but a cerain
puritan arrogance, how could such an idea win acceptance? It did not; it stll ha not. It
has however commanded recogiton; but that it did only because of it proigious
pwer of perforance.'
132 AGAI NST METHOD
are never strong enough to terminate the exstence of alteratve
approaches, and that a defence of such alteratves, being almost the
only way of discovering the errors of highly respected and
comprehensive points of view, is required even by a narrow
philosophy such as empiricism. Now ifit should tum out that it is also
required on ethical grounds, then we have two reasons instead of one
rather than a confict with 'science'.
Besides, the Church, and by this I mean its most outstanding
spokesmen, was much more modest than that. It did not say: what
contadicts the Bible as interpreted by us must go, no matter how
stong the scientfc reasons in its favour. A tuth supported by
scientfc reasoning was not pushed aside. It was used to revise the
interpretaton of Bible passages apparently inconsistent with it.
There are many Bible passages which seem to suggest a flat earth.
Yet Church doctne accepted the spherical earth as a matter of
course. On the other hand the Church was not ready to change just
because somebody had produced some vague guesses. It wanted
proof scientfc proof in scientfc matters. Here it acted no
differently from modem scientfc insttutons: universites, schools
and even research insttutes in various counties usually wait a long
tme before they incorporate new ideas into their curricula.
(rofessor Stanley Goldberg has described the situaton in the case
of the special theor of relatvity.) But there was as yet no convincing
proof of the Copercan dotine. Hence Galileo was advised to
teach Copercus as a hypothesis; he was forbidden to teach it as a
trth.
This distncton has survived untl today. But while the Church
was prepared to admit that some theories might be true and even that
Copercus' might be true, given suffcient evidence, 17 there are
1 7. In a widely discussed letter which Cardinal Robero Bellarino, master of
contoversial questons at the Collegio Romano, wrote on 1 2 April 1 61 5 to Paolo
Antonio Foscarini, a Carelite monk from Naples who had inquired about the reality
of the Coperican system, we fnd the following passage (Finochiaro, op. cit., p. 68):
. . . if there were a te demonstraton that the sun is at the center of the world and the
earh in the third heaven, and that the sun dos not circle the earh but the earh circles
the sun, then one would have to proeed with great care in explaining the Scriptures
that appear contar, aads rathcrthatmcd aetaadtaadth thaa thatmhatu
deastratcdis /ahc.But I will not believe that there is such a demonstraton, untl it is
shown me. Nor is it the same to demonstate that by supposing the sun to be at the
center and the earh in heaven one can save the appearances, and to demonstate that
in tth the sun is at the center and the earh in heaven; for I believe the frst
demonstaton may be available, but I have ver great doubts about the second, and in
case of doubt we must not abandon the Holy Scripture as interreted by the Holy
Fathers.' In his CsidatiseathcC@micaaQiaiea,Finochiaro, op. cit., pp. 70f,
THI RTEEN 133
now many scientsts, especially in high energy physics, who view all
theories as instuments of predicton and reject tuth-talk as being
metaphysical and speculatve. Their reason is that the devices they
use are so obviously designed for calculatng purposes and that
theoretcal approaches so clearly depend on consideratons of
elegance and easy applicability that the generalizaton seems to make
good sense. Besides, the foral propertes of'approxmatons' often
difer from those of the basic principles, many theories are first steps
towards a new point of view which at some future tme may yield them
as approximatons and a direct inference from theory to reality is
therefore rather naive
!
8
All ths was known to 1 6th- and 1 7th
century scientsts. Only a few astonomers thought of deferents and
epicycles as real roads in the sky; most regarded them as roads on
paper which might aid calculaton but which had no counterpart in
reality. The Coperican point of view was widely interpreted in the
same way - as an interestng, novel and rather efcient model. The
Church requested, both for scientfc and for ethical reasons, that
Galileo accept ths interpretaton. Considering the difcultes the
model faced when regarded as a descripton of reality, we must admit
that ' [l]ogic was on the side of . . . Bellarmine and not on the side of
Galileo,' as the historian of science and physical chemist Pierre
Duhem wrote in an interestng essay. 1 9
To sum up: the judgement of the Church experts was scientfcally
correct and had the right social intenton, v. to protect people from
the machinatons of specialists. It wanted to protect people from
being corrupted by a narrow ideology that might work in resticted
domains but was incapable of sustaining a haronious life. A revision
of the judgement might win the Church some friends among
scientsts but would severely impair its functon as a preserer of
important human and superhuman values.
20
esp. pp. 85f, Galileo addresses precisely these points. He agrees that if the Coperican
astonomers are 'not more than ninety percent right, they may be dismissed' but adds
that 'if all that is produced by philosophers and astonomers on the opposite side is
sho
wn to be mostly false and wholly inconsequential, then the other side should not be
disparaged, nor deemed paradoxcal, so as to think that it could never be clearly
prov
ed': research should be peritted even if demonstatons are not yet available.
This
does not conflict with Bellarino's suggestons; it did conflict and to a cerain
exte
nt stll does conflict with the atttude of many moem research insttutons.
1 8.
More on this pint in Nanc Canwright, HemthcLamse/FhsicLic,Oxford,
198
3.
1 9.
1e5mcthcFhmm,Chicago, 1 963, p. 78.
20.
After some apparent willingess to consider the matter (cf. the addres of Pope
Joh
n
Paul II on the centenar of Einstein's birth, published a an Epiloge in Paul
Car
dinal Poupard (ed.), Calih Calim: 1eard aRcelatm e/350 \cane/Dcbatc,
134 AGAI NST METHOD
Pittburgh, 1 987) Cardinal Joseph Ratinger, who holds a positon similar to that once
held by Bellarine, forulated the problem in a way that would make a revision of the
judgement anachronistc and poindess. Cf. his talk in Para of 1 5 March 1 990, pardy
reponed in I 5mate, 31 March 1 990, pp. SOf. As witesses the Cardinal quoted
Erst Bloh (being merely a matter of convenience the scientfc choice between
geoentism and helioentism cannot overle the practcal and religious centicit of
the eanh), C. F. von Weizsacker (Galileo leads direcdy to the atom bomb) and myself
(the chapter heading of the present chapter). I commented on the speech in to
interiews, I 5mate, 1 2 May 1 990, pp. 54f and LaRqabblica, 1 4 July 1 99, p. 20.
1 4
Galileo 's inquiries fred onl a small part ofthe so-called Coeican
Reolution. Ading the reaining eleents makes it still more dif clt to
reconcile the deloment with familiar principle oftheor ealuation.
Galileo was not the only scientst involved in the refor of physics,
astonomy and cosmolog. Neither did he deal with the whole area of
astonomy. For example, he never studied the moton of the planets
in as much detail as did Copericus and Kepler and he probably
never read the more technical parts of Copericus' great work. That
was not unusual. Then as now knowledge was subdivded into
specialites; an expert in one feld rarely was also an expert i another
and distant feld. And then as now scientsts with widely diverg
ing philosophies could and did comment on new suggestons
and developments. Tycho Brahe was an outstanding astonomer;
his observatons contributed to the downfall of generally accepted
vews. He notced the importance of Copericus' cosmolog -
yet he retained the unmoved earth, on physical as well a on
theological grounds. Copericus was a faithful Christan and a good
Aristotelian; he tied to restore cented circular moton to the
prominence it once had, postulated a movng earth, rearranged the
planetar orbits and gave absolute values for their diameters. The
astonomers surrounding Melanchthon and his educatonal reform
accepted and praised the frst part of his achievement, but (with a
single excepton - Rhetcus) either disregarded, or critcized, or
reint
erreted (Osiander!) the second. And they often tied to tansfer
C
opericus' mathematcal models to the Ptolemaic system.
1
Maestlin, Kepler's teacher, regarded comets as solid bodies and tied
calculatng the orbit of one of them. His (incorrect) result made h
accept the Coperican arrangement of the planetar orbits (it stll
I . Details and literarre in R.S. Westan, 'The Melanchthon Circle, Rhetcus,
a
nd
the
Wittenberg Interretaton of the Coprican Theor', Isis, Vol. 66, 1 975,
pp. 1 65f.
135
136 AGAI NST METHOD
influenced Kepler). Maestlin respected Aristotle but regarded
mathematcal correctess and harmony as signs of physical truth.
Galileo's approach had its own idiosyncrasies, it was more complex,
more conjectural, partly adapted to the greater role theological
consideratons played in Italy, partly determined by the laws of
rhetoric or patronage. Many different personalites, professions and
groups guided by different beliefs and subjected to different
constaints contibuted to the process that is now being described,
somewhat summarily, as the 'Coperican Revoluton'.
As I said at the beginning this process was not a simple thing
but consisted of developments in a variety of subjects, among
them the following: cosmology; physics; astronomy; the calculaton
of astronomical tables; optcs; epistemology; and theology.
I draw these distnctons not 'in order to be precise' but because
they reflect actually exstng subdivisions of research. Physics, for
example, was a general theory of motion that described change
without reference to the circumstances under which it ocurred. It
comprised locomoton, the growth of plants and animals a well as
the tansiton of knowledge from a wise teacher to an ignorant pupil.
Aristotle's Physic and the many mediaeval commentaries on it give us
an idea of the problems treated and the solutons proposed.
Cosmology described the structure of the universe and the special
motons that are found in it. A basic law of physic in the sense just
explained was that a moton without motor comes to a standstll -the
'natural' situaton of a body is rest (this includes lack of qualitatve
change). The 'natural' motons of cosmolog were those that occurred
without notceable interference; examples are the upward moton of
fire and the downward motion of stones. Aristotle's On the Heaes
and the many mediaeval commentaries on it give us an idea of the
problems and the view discussed in this domain.
The books I just mentoned were for advanced studies only.
Introductory texts omitted problems and alteratve suggestons and
concentated on the bare bones of the ideas then held. One of the
most popular introductory texts of cosmology, Sacrobosco's d spera,
contained a sketch of the world, and described the main SQheres
without giving the details of their motons - the rest is silence. 2 Stll,
2. Cf. Lynn Thordyke (ed.), 1hc 5phoc e/ 5ambesce aad Its Cemmtaten,
Chicago, 1 949. The elements and their motions are briefly mentioned in the frst
chapter together with a simple argument in favour of the unmoved eanh: the eanh is
situated in the centre (this is shown earlier by optical arguments, including the fact that
the constellations have the same size, no matter where the daily rotaton puts them)
and 'quicquid a medio movetur versus circumferentiam ascendit. Tera a medio
movetur, ergo ascendit, quod pro impossibile relinquitur' (p. 85). Equant, defere
nt
FOURTEEN 137
it was used as a basis for rather advanced critcal comments down to
Galileo's own tme.
Physics and cosmology claimed to make tue statements. Theology
which also claimed to make tue statements was regarded as a
boundar conditon for research in these fields though the strength of
this requirement and ofits insttutonal backing varied in tme and with
the locaton. It was never a necessar boundar conditon for
astonomy which dealt with the motons of the stars, but without
claiming tuth for its models. Astronomers were interested in models
that might correspond to the actual arrangement of the planets, but
they were not restricted to them. Handbooks of astronomy such as
Ptolemy's handbook and the various popularizatons based on it
contained detailed astonomical models preceded by sketchy
cosmological introductons. As far as these introductons were
concered, there exsted only one cosmology-Aristotle's. Some of the
handbooks also contained tables. Tables were a further step away from
'reality'. They not only used 'hypotheses', i.e. models that might not
reflect the stucture of reality, they also used approxmatons. But an
astonomer's approxmatons did not always correspond to the
excellence of his models. 'Advanced' (from our standpoint) models
might be combined with crude approxmatons and thus give worse
tables than their older counterarts.
3
The separaton between physics and cosmology on the one side
and astonomy on the other was not only a practical fact; it also had a
frm philosophical backing. According to Aristotle 4 mathematcs
does not deal with real things but contains abstractons. There exsts
therefore an essental difference between physical subjects such as
physics, cosmology, biolog and psycholog and mathematcal
subjects such as optics and astronomy. In the encyclopaedias of the
early Middle Ages the separaton was a matter of course.
Optical textbooks only rarely dealt with astonomical matters. 5
Astronomy used basic optcal laws such as the law of linear
propagaton, but the more complicated parts of optcal theor were
and epicycle are mentoned in the fourth chapter together with the miraculous nature
of the solar eclipse accompanying Christ's death.
3. The example of Ptolemy-opericus is teated by Stanley E. Babb Jr. in Isu,
Vol.
68, September 1977, especially p. +3Z.
4. Mct., Bok xii, Chapter Z,Fhsic, Bok ii, Chapter Z. For an account and
defe
nce of Aristotle's theor of mathematcs cf. Chapter 8 of my Farm cll teRcmea.
5. As an example I menton John Pecham's optcs (quoted from David Lindberg
(ed.),
ehaFccham aadthc5occ e/Qtic, Madison, 1970): astonomical matter
ocur here on pages 1 53 (mon illusion and the northward displacement of the sun
an
d the
fed star, explained by vapur near the horizon), Z0 (sintllaton of the
138 AGAI NST METHOD
not well known. The same is tue of epistemology. Galileo's
arguments (and the arguments of Copericus on which they are
based) brought epistemology back into science (the same happened
many years later, in connecton with the quantum theor).
Now is it to be expected that a collecton of relatvely independent
subjects, research stategies, arguments and opinions such as the one
just mentoned will develop in a uniform way? Can we really assume
that all the physicists, cosmologists, theologians and philosophers
who reacted to the Coperican doctine were guided by the same
motves and reasons and that these reasons were not only accepted by
them, but were also regarded as being binding for any scientst
entering the scene? The ideas of an individual scientst such as
Einstein may show a certain coherence
6
and this coherence may be
reflected in his standards and his theorizing. Coherence is to be
expected in totalitarian surroundings that guide research either by
laws, by peer pressure or by fnancial machinatons. But the
astonomers at the tme of Copericus and after did not live in such
surroundings; they lived at a tme of dissension, wars and general
upheaval, at a tme when one city (Venice, for example, and the cites
under its j urisdicton) would be safe for a progressive scientst while
another (such as Rome, or Florence) ofered considerable dangers,
and when the ideas of a single individual often faced groups of
scientsts not in agreement with his monomania. To show this, let us
look at two astonomers who partcipated in the development:
Copercus himself and Maestlin, Kepler's teacher.
Copercus wanted to reform astronomy. He explained his
misgivings and the ways in which he tied to overcome them. He
wrote:
7
star explained by unevennesses of their rotatng surfaces which reflect the sunligt),
21 8 (impossibility to deterine the size of the star frm their appearances), 233
(star appear to be smaller than they actually are), 225 (they are displaced towards
the norh at the horizon, and the more so, the greater their distance from the
meridian).
6. The case of Einstein shows that even this modest assumpton goes much to far.
Einstein recommended a loose opportunism as the best research stateg (cf. the
quotaton in the text to footote 6 of the Intoducton) and he wared that a good joke
(such as the consideratons leading to the special theor of relatvity) should not be
repeated too often: Philipp Frank, Eiascia,HuLtaadTime, London, 1 946, p. 261 .
7. Cmanelas, ed. E. Rosen, Orcc C@mima 1rcmuc, 3rd editon,
New
York, 1971 , tanslaton panly changed in accordance with F. Kraf, 'Copericus
Retoverus 1', CeheqaiaC@mimaaIl l and I, Proeedings of the Joint Symposium
of the !AU and the IUHPS, Torn, 1 973, p. 1 1 9. In what follows I shall also
use
Kraft, 'Copericus Retoversus II', Jo. cit.
FOURTEEN 139
The planetar theores of Ptolemaics and most other astonomer . . .
seemed . . . to present no small difculty. For these theores were not
adequate unless certain equants were also conceived; it then appeared
that a planet moved with unifor veloity neither along its own
deferent nor relatve to an actual cente . . . . Having become aware of
these defects I often considered whether there could perhaps be
found a more reasonable arangement of circles from which ever
apparent inequality could be derived and i which everg would
move uniformly about its proper cente as the rle of accomplished
moton requires . . . .
v
The critque of Copercus concers the following model that was
used for calculatng the longitudes ofMars,Jupiter and Satur. The
planet P moves on a small circle, the epicycle, whose cente is located
on a larger circle, the deferent. The cente of the epicycle proceeds
with constant angular velocity with respect to E, the equant point.
The planet is obsered from the earth T. E and T are on opposite
sides of the cente of the deferent, having the same distance from it.
Copericus does not queston the empirical adequacy of the
mo
del. On the contar, he admits that the planetar theories of the
P
tole
maeans and others are 'consistent with the numerical data'.
8
Nor does he believe that these data are in need of correcton. Instead
of introducing new obseratons he emphasies that
8. Rosen, op. cit., p. 59.
140 AGAI NST METHOD
we must follow in their [the ancient Greeks'] footsteps and hold fast to
their obseratons bequeathed to us like an inertance. And if anyone
on the contrar thinks that the ancients are untrstworthy in this
regard, surely the gates of this art are closed to him. 9
Neither new obseratons nor the inability of Ptolemy to take care of
what was known to him are the reason for Copericus' discomfor.
The difculty he perceives lies elsewhere.
In his account Copericus distnguishes between absolute
motons and apparent motons. The second inequality of planetar
moton, i.e. the fact that a planet may run ahead in its path and then
reverse its directon, is 'apparent' - it must be reduced to other
motons. According to Copericus these other motons are motons
on cented circles with a constant angular velocity around the cente.
Ptolemy violates the conditon; he uses equants. Equants explain
apparent motons not by tue motons but again by apparent motons
where the planet 'moves with uniform velocity neither along its own
deferent, nor relatve to an actual cente . . . .' For Copericus (and for
many other astonomers) real moton is a circular moton around a
cente with constant anglar veloity.
1
0
Copercus removes excente and equant and replaces them by
epicycles.
1
1
In the Ptolemaic scheme each planet has now three
epicycles: the old epicycle and two further epicycles for replacing the
ecentc and the equant.
In order to avoid this accumulaton of epicycles (which occasion
ally pushed the planets far out into space) Copericus looks for
a different explanaton of the second inequality. He is helped by the
fact that the second inequality agrees with the positon of the sun.
1
2
It can therefore be interreted as an apparent moton created by a
real (and, of course, circular) moton on par of the earth.
The argument as reconstucted so far (after Kraft) contains two
9. LctmZgaiastmo,in Rosen, op. cit., p. 99.
10. Eramus Reinhold wote on the tde page of his peronal copy of d
Rmlatieaibas: Zr~ Zstmaemimm: Metas chtu mqmlu ct n ormmm cl e
mqmlibm ct ormmnbas mpesitas. Quoted from Westan, 'Te Melanchthon
Circle', op. cit., p. 176.
l l . Tis is te of the Cemmtanelas. In his main work he again uses excentic
deferents. Only the equant i replaced by an epiccle. This 'liberaton from the equat'
(Eramus Reinhold), also in lunar theor, gready impressed sme of Copr
cus'
admirer who paid no atenton to his new cosmolog and the moton of the earh
.
Westan, 'The Melanchthon Circle, op. cit., pp. 175, 177.
1 2. The mean sun in Copericus. Kepler effects a reducton to the te sun a
nd
thereby stengthens te Coprican arngement.
FOURTEEN 141
elements; a purely formal element and a reality asserton. Formally it
is requested that any periodic moton be reduced to centred circular
motons. The request is connected with the assumpton that
inequalites are apparent while circular motons alone are real. Let us
call this the frt realit assumption. But Copericus also discovered
that his procedure allowed him to incorporate ever planetar path
into a system, containing the 'large circle', the circle of the earth, as
an absolute measure. 'All these phenomena', Copercus writes in
his main work,
1
3 'are connected with one another in a most noble
way, as if by a golden chain, and each planet with its positon and
order is a witess that the earth moves while we, who live on the
terrestrial globe, failing to recognize its moton, ascribe all sorts of
motons to the planets.' It is this inner connectedness of all parts of
the planetar system that convinced Copericus of the reality of the
moton of the earth. I call this the second realit assumption.
The frst reality assumpton was part of the Platonic taditon;
Aristotle gave it a physical basis. The second reality assumpton
conflicted with Aristotelian physics and cosmology. Aristotle had
already critcized an earlier (Pythagorean) version ofit: mathematcal
harmonies, which are abstactons, reflect truth only if they agree
with well-confrmed physical principles. This is a reasonable
request; it was used in our ow centur to reject Schroedinger's
interpretaton of wave mechanics. It is reasonable especially for those
thinkers who regard mathematcs as an auxliar science that may
describe but cannot consttute physical processes. It is unreasonable
for a Platonist or a Pythagorean. The resultng clash between two
interpretatons of the nature of mathematcal statements played an
important role in the ' Coperican Revoluton'.
Copericus strengthened the second reality assumpton by
referring to taditons such as the Hermetc taditon and the idea of
the exceptonal role of the sun
1 4
and by showing how it could be
reconciled with the phenomena. He made two assumptons. First,
that the moton of a body is appropriate to its shape. The earth is
spherical, hence its moton must be circular. Secondly, objects such
as stone stay with the hody (the earth) from which they were
13. DcRnel. , Preface to Pope Paul. Kraft assumes that Copericus discovered
this
harony in the course of his attempts to remove the equant and only later tured it
into a fundamental argument in favour of a real moton of the earh.
1 4. The phrase 'and in the middle stands the sun' was not new. In the older
astr
onomy the sun was indeed in the middle of the planets, with Mars, Jupiter and
Sat
ur
standing above it, and Venus, Mercur and the moon below. It also 'ruled' the
planets in the sense that its moton was mirored in the motons of all planets (the moon
excepted). Cf. e.g. Macrobius, 5emaiam5cipieais.
142 AGAI NST METHOD
separated - hence the falling stone stays close to the tower.
According to Aristotle the natural moton of objects, i.e. the upward
moton of fre and the downward moton of stones, was determined
by the stucture of space (cental symmety). According to Coper
nicus it is determined by the distibuton of matter.Copercus
'saves phenomena' such as the free fall of heavy bodies but provides
neither independent arguments nor stict laws that could lead to a
detailed comparison. His procedure is a hoc. This does not mean
that it is bad; it only means that it cannot be reconciled with the
leading methodologies of today.
My second example is Michael Maestlin, Kepler's teacher. Maestlin
was an exer astonomer and his judgement was generally
respected. He 'only reluctantly abandoned' the Ptolemaic distibu
ton of the spheres - but he was forced to do so by circumstances
beyond his conto1.
1 5
A far as we can see, the circumstances were,
frst, the nova of 1 572. Maestlin obsered it, measured its parallax
and put it beyond the sphere of the moon into the sphere of the fxed
stars. The frst par (beyond the moon) followed for Maestlin from
the missing parallax, the second par (fed stars) from the absence of
any proper moton. According to Copericus, whose ideas Maestlin
used at this point, a planet moves more slowly the greater its distance
from the sun. Observing the changes of colour and brightess
Maestlin (and Tycho who saw the new star on the way to hs
alchemical laborator) inferred that the region above the moon
cannot be without change, as Aristotle had assumed. However, it
would be rash to conclude that Maestlin (and Tycho) regarded the
nova as a 'blow against the peripatetc philosophy'.
1 6
Many Church
people, Theodore Beza among them, regrded the phenomenon as a
retur of the star of Bethlehem, i.e. as a superatural event.
17
Tycho
thought this comparison too modest; here, he said, is the greatest
miracle since the beginning of the world, comparable at least to
Joshua's stopping of the sun.
1 8
This means that as far as Tycho was
concered miracles refuted the idea of the autonom of the laws of
I 5. In what follows I am using the dissnaton by R.A. Jarell, OcLqaad5ntc
e/thc 1mbiag ZstmaoMichmlMtlia, Toronto, 1 972, a well a R.S.
Westan, 'Michael Maestlin's Adopton of the Coprican Theor', Celam
C@mimmI, Ossilineum, I 975, p. 53f.
I 6. Jarell, op. cit., p. 1 08.
I 7. Cf. the literature in P.H. Koher, 5mccaadRcliga iaElccthaaEagmad,
New York, 1969, pp. 1 74f, fototes 1 2 and 1 3. Cf. also Vol. v, Chapter xx i ofLynn
Thordike, ZHue/MaaadEmmtal5n,New York, 1 941 .
1 8. Prmnm, p. 548.
FOURTEEN 143
nature (which was an Aristotelian idea), they did not refute specfc
laws. Maestlin, on the other hand, being perhaps more sceptcal
abou
t miracles, may indeed have regarded the case as a 'blow against'
Ari
stotle.
The next queston is how serious a blow it was for him. The idea of
a permanent heaven was part of cosmolog and contained the special
hypothesis of a ffth element. The falsehood of this hypothesis
impaired neither the remaining laws of moton nor the tower
argu
ment. Both Clavus and Tycho accepted a changing heaven
1 9
but stll used the tower argument to exclude the moton of the earth.
If Maestlin's doubts reached further then this was due either to an
idiosyncratc interpretaton of the Aristotelian dotines, or to
personal inclinatons towards a non-Aristotelian world-vew. It
seems that we must assume the latter.
The next decisive event on Maestlin's jourey towards Copercus
was the comet of 1 577. Again Maestlin, prompted by 'numerous
observatons', puts the comet into the superlunar region.
2
0
The idea
that this region is free from change has now defnitely been dropped.
Maestlin also tied to determine the tajector of the comet. He
found it to be movng in the path of Venus as described in Book 6,
Chapter 12 of d Reo/utionibus. Somewhat hesitatngly he now
accepts the Copercan ordering of the spheres.
2
1
But, so he adds,
he was forced to do so 'by exteme necessity'.
2
2
This 'exteme necessity' arises only when geometical considera
tons are given the force of cosmological arguments. Many years later
Galileo cautoned against this way of reasoning: rainbows, he said,
cannot be caught by tiangulaton. Maestlin had no such doubts. He
accepted the traditonal distncton between physics and astonomy
and identfed astronomy with mathematcs: ' Copercus wrote his
entre book not as a physicist, but as an astonomer' is his comment
on the margin of his copy of d Reolutionibus.
23
He then interpreted
the results of mathematcal arguments by using the second reality
assumpton. This means that he did not uercome an Aristotelian
resistance against such an interpretaton, he acted as if such a
resi
stance did not eist. 'This argument', he wrote in his marginal
notes,
24
'is wholly in accord with reason. Such is the arrangement of
1
9
. For Clavius cf. his commentar on Sacrobosco's sphere, 1 593 editon, pp.
21 0
. Cf. also Westfall, op. cit., p. 4.
20. Jarrell, op. cit., p. 1 1 2.
21. Ibid., p. 1 1 7.
22. Ibid., p. 1 20.
23.
Westan, op. cit., p. 59
.
2
4. Ibid.
144 AGAI NST METHOD
this entre, immense mahina that it permits surer demonstatons,
indeed, the entre universe revolves in such a way that nothing can be
transposed without confusion of its [parts] and, hence, by means of
these [surer demonstratons] all the phenomena of moton can be
demonstrated most exactly, for nothing unfttng ocurs in the course
of their orbits. ' Kepler too became a Coperican because of this
harmony and because of the comet, the interestng fact being that
Maestlin's calculatons of the path of the comet contain serious
mistakes; it did not move in the orbit ofVenus.
Now let us compare these events and the situatons in which they
ocur with some once popular philosophies of science. We notce at
once that none of these philosophies considers all the disciplines that
contibuted to the debate. Astronomy is in the centre. A ratonal
reconstucton of the developments in this area is thought to be a
ratonal reconstucton of the Coperican Revoluton itself. The role
of physics (the tower argument), the fact that theolog occasionally
formed a stong boundary conditon (cf. Tycho's reacton to his nova
and to the idea of the moton of the earth) and the role of different
mathematcal philosophies shows that this cannot possibly be tue.
This fatal incompleteness is the frst and most fundamental objecton
against all reconstuctons that have been ofered. They stll depend
on the (positvistc) prejudice that obseratons alone decide a case
and that they can judge a theory all by themselves, without any help
(or hindrance) from alteratves, metaphysical alteratves included.
Moreover, they even fail in the narrow domain they have chosen for
reconstucton, v. astronomy. To show this, let us consider the
following accounts:
1 . Naie empirm: the Middle Ages read the Bible and never
looked at the sky. Then people suddenly looked upwards and found
that the world was diferent from the opinion of the schools.
This account has disappeared from astonomy - but its analogue
survives in other areas (for example, in some parts of the history of
medicine). The main argument against it is that Aristotle was an arch
empiricist and that Ptolemy used carefully collected data. 25
2. Sohisticated empircism: new obseratons forced astonomers
to modif an already empirical doctine.
25. 'Carefully' has been contested by R.R. Newon, cCnmce/CmadimFtek
,
Baltmore, 1 977. Newton shows that many of Ptolemy's 'data' mcmaa/madto
ft
his moel. For his optcs tis ha ben know for a long tme.
FOURTEEN 145
This certainly is not tue for Copericus and his followers in the
1 6th
centur. As we have seen, Copercus thought the Ptolemaic
system to be epircall aquate - he critcized it for theoretical
reasons. And his 'observatons' are essentally those of Ptolemy, as he
says
himself.
Modem comparisons of Coperican and Ptolemaic predictons
'with the facts', i.e. with 1 9th- and 20th-centur calculatons, show,
furthermore, that empirical predictons were not improved and
actually become worse when the competng systems are resticted to
the same number of parameters.
26
The only new observatons made were those ofTycho Brahe -but
they already led beyond Copericus to Kepler. Galileo's observa
tons belong to cosmolog, not to astonomy. They lend plausibilit to
some of Copercus' analoge. A compeling proof of the moton of the
earth did not emerge, however, for the Galilean observatons could
also be accommodated by the Tychonian system.
3. Falsicationism: new observatons refuted imporant assump
tons of the old astonomy and led to the inventon of a new one. This
is not correct for Copericus and the domain of astonomy (see
above, comments on 2). The 'refutaton' of the immutabilit of the
heavens was neither compelling nor decisive for the problem of the
moton of the earth. Besides, the idea of the moton of the earth was
in big touble or, if you will, 'refted'. lt could survive only if it was
teated wth kindness. But if i could be teated wth kindness, then so
could the older sstem.
We see here ver clearly how misguided it is to ty reducing the
process 'Copercan Revoluton' to a single principle, such as the
principle of falsifcaton. Falsifcatons played a role just as new
observatons played a role. But both were imbedded in a complex
patter of events which contained tendencies, atttudes, and
consideratons of an entrely different nature.
4. CoTetionalism: the old astonomy became more and more
complicated - so it was in the end replaced by a simpler theor. lt is
this assumpton that led to the mockng remark of the 'epicyclical
Z6.
Stanley E. Babb, 'Accuracy of Planetary Theories, Parcularly for Mars', Isu,
Sep. 1 977, pp. 426f. Cf. also the earlier arcle of Derek de Solla Price, 'Conta
Copericus', in M. Clagett (ed.), CntimlFrebls e/thcHuteqe/5occ, Madison,
1 9
59,
pp. 1 97f; N.R. Hanson,/si, No. 5 1 , 1 960, pp. 1 50f a well as Owen Gingerich,
'C
rsis
vs Aesthetcs in the Coperican Revoluton', in Beer (ed.), IistmiaZstmaep,
V
ol.
1 7,
1 974. Gingerich compares the whof Stoefer, Stadius, Maesdin, Magini
an
d
Organus and fnds al of them beset by erors of rougly the same magitude
(th
ough
not of the same distibuton along the ecliptc).
146 AGAI NST METHOD
degeneraton'. The theor overlook the fact that the Coperican
scheme has about as many circles as the Ptolemaic one
P
5. Te theor ofcse: astonomy was in a crisis. The crisis led to a
revoluton which brought about the tiumph of the Coperican
system.
The answer here is the same as under 2: epircll there was no
crisis and no crisis was resolved. A crisis did ocur in cosmolog, but
only aer the idea of the moton of the earth received a serious
hearing. The many complaints about the inexactess of astonomical
predictons that preceded Copericus (Regiomontanus, for example)
critcized the lack of precise inital conditons and accurate tables, not
basic theor, and such a critcism would have been quite unjust, as
the later examinaton of these theories shows.
28
27. The reader should consult the ver instrctve diagams in de Santllana's
editon of Galileo's Dialep,Chicago, 1 964.
28. Cf. fotote 26 above.
1 5
Te reults obtained so fr sugest abolihing the ditincion betwee a
cntet of dicuer and a contet ofjustication, nors and fas,
obseational tes and theoretical ters. None ofthee distincions plays a
rle in scietic praic. Anepts to enfrce the woul hae disastrus
consequece. Pop e's 'crticl' rationalism fils fr the same reaons.
Let us now use the material of the preceding sectons to throw light
on the following features of contemporar empiricism: (1) the
distnctons between a context of discover and a context of
justfcaton - nors and facts, obseratonal ters and theoretcal
ters; (2) Popper's 'critcal' ratonalism; (3) the problem of
incommensurability. The last problem will lead us back to the
problem of ratonalit and order vs anarchism, which is the main
topic of this essay.
One of the objectons which may be raised against my attempt to
draw methodological conclusions from historical examples is that it
confounds two contexts which are essentally distnct, v. a context of
discover, and a context of justfcaton. Discuer may be irratonal
and need not follow any recognized method. Justiction, on the
other hand, or - to use the Holy Word of a diferent school -c'ticism,
starts only afe the discoveries have been made, and it proceeds an
orderly way. 'It is one thing,' writes Herbert Feigl, 'to retace the
historical origins, the psychological genesis and development, the
socio-politcal-economic conditons for the acceptance o rejecton
of scientfc theories; and it is quite another thing to provide a logical
reconstructon of the conceptual stucture and of the testng of
scientfc theories.'
1
These are indeed two diferent thing, especially
as they are done by two different discipline (histor of science,
philosophy of science), which are quite jealous of their indepen
dence. But the queston is not what distctons a fertle mind can
I.
'
Te Orodox View of Teories, in Radner-Winokur (eds), Zmsc e/
Ihm
ncaadMcthede/FsmaadFqcheleg,Minneaplis, 1 970, p. 4.
147
148 AGAI NST METHOD
dream up when confronted with a complex process, or how some
homogeneous material may be subdivided; the queston is to what
extent the distncton drawn reflects a real diference and whether
science can advance without a stong interacton between the
separated domains. (A river may be subdivided by natonal
boundaries but this does not make it a discontnuous entty.) Now
there is, of course, a ver notceable diference between the rules of
testng as 'reconstructed' by philosophers of science and the
procedures which scientsts use in actual research. This diference is
apparent to the most superficial examinaton. On the other hand, a
most superficial examinaton also shows that a determined applica
ton of the methods of critcism and proof which are said to belong to
the context of justficaton would wipe out science as we know it -
and would never have permitted it to arise.
2
Conversely, the fact
that science exsts proves that these methods were frequently over
ruled. They were overruled by procedures which belong to the
context of discover. Thus the attempt 'to retace the historical
origins, the psychological genesis and development, the socio
politcal-economic conditons for the acceptance or rejecton of
scientfic theories', far from being irrelevant for the standards of test,
actually leads to a critcism of these standards - pr(idd the two
domains, historical research and discussion of test procedures, are
not kept apart by fiat.
In another paper Feigl repeats his arguments and adds some
frther points. He is 'astonished that . . . scholars such as N.R.
Hanson, Thomas Kuhn, Michael Polanyi, Paul Feyerabend,
Sigmund Koch et al. , consider the distncton as invalid or at least
misleading'.
3
And he points out that neither the psycholog of
inventon nor any similarity, however great, between the sciences and
the arts can show that it does not exst. In t he i cery right.
Even the most surprising stories about the manner in which scientsts
arrive at their theories cannot exclude the possibilit that they
proceed in an entrely diferent way once they have found them. But
thi possibilit is nee realied. Inventng theories and contemplatng
them in a relaxed and 'artstc' fashion, scientsts often make moves
that are forbidden by methodological rules. For example, they
interpret the evidence so that it fits their fanciful ideas, eliminate
diffcultes by a hoc proedures, push them aside, or simply refse to
take them seriously. The actvites which according to Feigl belong to
2. Cf. the exmples in Chapter 5.
3. 'Empiricism at Bay', MS, 1 972, p. 2.
FI FTEEN 149
the context of discover are, therefore, not just di eet from what
philosophers say about justficaton, the are in cnjia with it.
Scientfic practce does not contain two contexts moving sid b si, it
is a complicated miture of proedures and we are faced by the
queston if this mixture should be left as it is, or if it should be
replaced by a more 'orderly' arrangement. This is part one of the
argument. Now we have seen that science as we know it today could
not exst without a frequent overrling of the context of justficaton.
This is part two of the argument. The conclusion is clear. Pa one
shows that we do not have a diference, but a mixture. Pa two shows
that replacing the mixture by an order that contains discover on one
side and justficaton on the other would have ruined science: we are
dealing with a uniform practce all of whose ingredients are equally
important for the growth of science. This disposes of the distncton.
A similar argument applies to the ritual distncton between
methodological precptins and historical dcptions. Methodolog,
it is said, deals with what should be done and cannot be critcied by
reference to what is. But we must of coure make sure that our
prescriptons have a point ofattak in the historical material, and we
must also make sure that their determined applicaton leads to
desirable results. We make sure by considering (historical, soio
logical, physical, psychological, etc.) tednce and lws whch tell us
what is possible and what is not possible under the given circum
stances and thus separate feasible prescriptons from those which are
going to lead into dead ends. Again, progress can be made only if the
distncton between the ought and the i is regarded as a temporar
device rather than as a fundamental boundar line.
A distncton which once may have had a point but which has now
definitely lost it is the distncton between obseational terms and
thereticl terms. It is now generally admitted that t distncton is
not as sharp as it was thought to be only a few decades ago. It is also
admitted, in complete agreement with Neurath's original views, that
both theories and obseratons can be abandoned: theories may be
removed because of conflictng obseratons, obseratons may be
removed for theoretcal reasons. Finally, we have discovered that
learing does not go from obseraton to theor but always involves
both elements. Experience arises togehe with theoretcal assump
tons not before them, and an experience without theor is just as
incomprehensible as is (allegedly) a theor without experience:
eliminate part of the theoretcal knowledge of a sensing subject and
you have a peron who is completely disoriented and incapable of
c
arr
ng out the simplest acton. Eliminate frther knowledge and his
se
nsor world (his 'obseraton language') wl star disintegratng,
150 AGAI NST METHOD
colours and other simple sensatons will disappear untl he is in a
stage even more primitve than a small chld. A small chld, on the
other hand, does not possess a stable perceptual world which he uses
for making sense of the theories put before h. Quite the contary
he passes through various perceptual stages which are only losely
connected with each other (earlier stages disapear when new stages
take over - see Chapter 16) and which embody all the theoretcal
knowledge available at the tme. Moreover, the whole process stars
only because the child reacts correctly towards signals, interre the
corecl, because he possesses means of interpretaton even before he
has experienced his frst clear sensaton.
Al these discoveries cry out for a new terminolog that no longer
separates what is so intmately connected in the development both of
the individual and of science at large. Yet th distncton between
obseraton and theory is stll upheld. But what is its point? Nobody
will deny that the sentences of science can be classifed into long
sentences and shor sentences, or that its statements can be classifed
into those which are intuitvely obvious and others which are not.
Nobody wldeny that such distnctons can be mad. But nobody will
put great weight on them, or will even menton them, fr the d not
now pl a' dciie rle in the busines ofscec. (This was not always
so. Intuitve plausibility, for example, was once thought to be a most
imporant guide to the tuth; it disappeared from methodolog the
very moment intuiton was replaced by experience, and by foral
consideratons.) Does experience play such a role? It does not, as we
have seen. Yet the inference that the distncton between theory and
obseraton has now ceased to be relevant, is either not drawn or is
explicity rejected! Let us take a step forard and let us abandon
this last tce of dogmatsm in science!
Incommensurability, which I shall discuss next, is closely
connected with the queston of the ratonality of science. Indeed one
of the most general objectons not merely to the ue ofincommen
surable theories but even to the idea that thee are such theories to be
found in the history of science is the fear that they would severely
restict the efcacy of taditonal, non-dialectcal armet. Let us,
therefore, lok a little more closely at the critcal standrd which,
according to some, consttute the content of a 'ratonal' argument.
More especially, let us look at the stndards of the Popperian school,
4. 'Neurath fails to give . . . rles [which distnguish empirical statement from
others) and thus unwitngly throws empiricism overboard', K.R. Popper, DcLepc
e/
5otqcDie, New York and London, 1959, p. 97.
FI FTEEN 151
which are stll being taken seriously in the more backward regions of
knowledge. This will prepare us for the fal step in our discussion of
the issue between law-and-order methodologies and anarchism in
scie
nce.
Some readers of my arguments in the above text have pointed out
that Popper's 'critcal' ratonalism is sufciently liberal to accom
modate the developments I have described. Now critcal ratonalism
is either a meaningful idea or it is a collecton of slogans that can be
adapted to any situaton.
In the frst case it must be possible to produce rules, standards,
restictons which permit us to separate critcal behaviour (thinking,
singing, writng of plays) from other tpes ofbehaviour so that we can
discue irratonal actons and coTea them with the help of concrete
suggestons. It is not difcult to produce the standards of ratonality
defended by the Popperian school.
These standards are standards of cticsm: ratonal discussion
consists in the attempt to critcize, and not in the attempt to prove or
to make probable. Ever step that protects a view from critcism, that
makes it safe or 'well-founded', is a step away from ratonality.
Ever step that makes it more vulnerable is welcome. In additon,
it is recommended to abandon ideas which have been found want
ing and it is forbidden to retain them in the face of stong and
successful critcism unless one can present suitable counter
arguments. Develop your ideas so that they can be critcized;
attack them relentlessly; do not t to protect them, but exhibit
their weak spots; eliminate them a soon as such weak spots have
become manifest - these are some of the rules put forth by our
critcal ratonalists.
These rules become more defnite and more detailed when we
tur to the philosophy of science and, especially, to the philosophy of
the natural sciences.
Within the natural sciences, critcism is connected with experi
ment and obseratons. The content of a theor consists in the sum
total of those basic statements which contadict it; it is the class of its
potental falsifers. Increased content means increased vulnerability,
hence theories oflarge content are to be preferred to theories of small
content. Increase of content is welcome, decrease of content is to be
avoided. A theor that contradicts an accepted basic statement must
be given up. Ad hoc hypotheses are forbidden- and so on. A science,
however, that accepts the rules of a critcal empiricism of this kind
will develop in the following manner .
.
We start with a prble, such as the problem of the planets at the
te of Plato. This problem (which I shall discuss in a somewhat
152 AGAI NST METHOD
idealized form) is not merely the result of crosit, it is a theoreticl
reult. It is due to the fact that certain eecations have been
disappointed: on the one hand it seems to be clear that the stars must
be divine, hence one exects them to behave in an orderly and lawful
manner. On the other hand, one cannot fnd any easily discerble
regularit. The planets, to all intents and purposes, move in a quite
chaotc fashion. How can this fact be reconciled with the exectaton
and with the principles that underlie the exectaton? Does it show
that the exectaton is mistaken? Or have we failed in our analysis of
the facts? This is the problem.
It is important to see that the elements of the problem are not
simply ge. The 'fact' of irregularit, for example, is not accessible
without further ado. It cannot be discovered by just anyone who has
healthy eyes and a good mind. It is only through a certain expectaton
that it becomes an object of our attenton. Or, to be more accurate,
this fact of iregularit est because there is an expectaton of
regularit and because there are ideas which defne what it means to
be 'regular'. Aer al, the term 'irregularity' makes sense only if we
have a rle. In our case the rle that defnes regularity asser circular
moton with constant angular veloity. The fed stars agree with this
rule and so does the sun, if we tace its path relatve to the fed st.
The planet do not obey the rle, neither directly, with respect to the
earth, nor indirectly, with respect to the fed st.
(In the problem we are exg now the rule i formulated
explicitly and it can be discussed. This is not always the cae.
Recogg a colour as red i made possible by deep-lying
patters concerg the stcture of our suroundings and recog
nton dos not occur when these patters cease to exst.)
To sum up t par of the Popperian dotine: research stars with
a problem. The problem i the result of a conflict between an
expectaton and an obseraton which is consttuted by the
expectaton. It is clear that this doe difer from the dotine of
inductvism where objecte facts enter a passive mind and leave their
taces there. It was prepared by Kant, Mach, Poincare, Dingler, and
by Ml(On Liber).
Having formulated a problem, one ties to sole it. Solving a
problem means inventng a theor that is relevant, falsifable (to a
degree larger than any alteratve), but not yet falsified. In the case
mentoned above (planets at the tme of Plato), the problem i: to fnd
circular motons of constant angular veloity for the purpose of saving
the planetar phenomena. A firt soluton was provided by Eudoxos
and then by Heracleides ofPontos.
Next comes the cticm of the theor that has been put forth in the
FI FTEEN
new
theory
additional
predictions
success
of the old
theor
(par of
the trth
content
of the new
theor)
failure of the
old thery (par
of the falsity
content of the
new theory)
153
attempt to solve the problem. Successful critcism removes the
theor once and fr all and creates a new problem, v. to explain (a)
why the theor was successful so far; () why it failed. Trying to solve
this problem we need a new theor that reproduces the successful
consequences of the older theor, denies its mistakes and makes
additonal predictons not made before. These are some of the fral
conditions which a suitable successor ofa reuted theor must satsf.
Adoptng the conditons, one proceeds by conjecture and refutaton
from less general theories to more general theories and expands the
content of human knowledge.
More and more facts are discoered (or constucted with the help of
expectatons) and are then explained by theories. There is no
guarantee that scientsts will solve ever problem and replace ever
theor that has been refuted with a successor satsfing the formal
conditons. The inventon of theories depends on our talents and
other fortuitous circumstances such as a satsfactor sex life. But as
long as these talents hold out, the enclosed scheme is a correct
account of the growth of a knowledge that satsfes the rules of critcal
ratonalism.
Now at this point, one may raise two questons.
1 . Is it dirable to live in accordance with the rules of a critcal
ratonalism?
2. Is it possible to have both a science as we know it and these rules?
As far as I am concered, the frst queston is far more imporant
t
han
the second. True, science and related insttutons play an
im
pora
nt part in our culture, and they ocupy the cente of interest
for
many philosophers (most philosophers are opportunists). Thus
154 AGAI NST METHOD
the ideas of the Popperian school were obtained by generalizing
solutons for methodological and epistemological problems. Critcal
ratonalism arose from the attempt to understand the Einsteinian
revoluton, and it was then extended to politcs and even to the
conduct of one's private life. Such a procedure may satsf a school
philosohe, who looks at life through the spectacles of his own
technical problems and recognizes hated, love, happiness, only to
the extent to which they occur in these problems. But if we consider
human interests and, above all, the queston of human freedom
(freedom from hunger, despair, from the tyranny of constpated
systems of thought and not the academic 'freedom of the will'), then
we are proceeding in the worst possible fashion.
For is it not possible that science as we know it today, or a
'search for the truth' in the style of taditonal philosophy, will
create a monster? Is it not possible that an objectve approach that
frowns upon personal connectons between the enttes examined
will har people, tum them into miserable, unfriendly, self
righteous mechanisms without char and humour? 'Is it not
possible,' asks Kierkegaard, 'that my actvity as an objectve [or a
critco-ratonalJ obserer of nature will weaken my stength as a
human being?' I suspect the answer to many of these questons is
afratve and I believe that a refor of the sciences that makes
them more anarchic and more subjectve (in Kierkegaard's sense) i
urgently needed.
But these are not the problems I want to discuss now. In the
present essay I shall restict myself to the second queston and I shall
ask: is it possible to have both a science as we know it and the rules of
a critcal ratonalism as just described? And to this queston the
answer seems to be a fn and resounding NO.
To star with we have seen, though rather briefy, that te actual
development of insttutons, ideas, practces, and so on, ofen de not
star fom a prble but rather from some extaneous actvity, such as
playing, which, as a side efect, leads to developments which later on
can be interpreted as solutons to unrealized problems.
6
Are such
developments to be excluded? And, if we do exclude them, wlths
not considerably reduce the number of our aoaptve reactons and th
e
quality of our learng proess?
5. F@iro,ed. Heiberg, VI, Pt. I, sec. A, No. 1 82. Milties to show how scientfic
method can be undertod a par of a theor of man and thus gives a positve answer to
the queston raised by Kierkegaard; cf. fotote 2 to Chapter 4.
6. Cf. the brief comment on the relaton between idea and acton in Chapter
I.
For details cf. fototes 31 f of'Against Method_ Minneota5twic,Vol. 4, 1970.
FI FTEEN 155
Secondly, we have seen, in Chapters 8-1 4, that a strict pn"ncple of
fliction, or a 'naive falsificatonism' as Lakatos calls it, 7 would
wipe out science as we know it and would never have permitted it to
start.
The demand for inceaed contet is not satsfied either. Theories
which efect the overthrow of a comprehensive and well-entenched
point of view, and take over after its demise, are initally restricted to a
fairly
narrow domain of facts, to a series of paradigmatc phenomena
which lend them support, and they are only slowly extended to other
areas. This can be seen from historical examples (footote 1 2 of
Chapter 8), and it is also plausible on general grounds: ting to
develop a new theor, we must first take a ste bak from the evidence
and reconsider the problem of obseraton (this was discussed in
Chapter 1 1). Later on, of course, the theor is extended to other
domains; but the mode of extension is only rarely determined by the
elements that consttute the content of its predecessors. The slowly
emerging conceptual apparatus of the theory soon stars cning its
own prbles, and earlier problems, facts, and obseratons are either
forgotten or pushed aside as irrelevant. This is an entrely natural
development, and quite unobjectonable. For why should an ideology
be constained by older problems which, at any rate, make sense only
in the abandoned context and which look silly and unnatural now?
Wy should it even consid the 'facts' that gave rise to problems of
this kind or played a role in their solutons? Wy should it not rather
proceed in its own way, devising its own tasks and assembling its own
domain of 'facts'? A comprehensive theory, after all, is supposed to
contain also an ontolog that determines what exsts and thus delimits
the domain of possible facts and possible questons. The develop
ment of science agrees with these consideratons. New views soon
stike out in new directons and frown upon the older prbles (what
is the base upon which the earth rests? what is the specific weight of
phlogiston? what is the absolute veloit of the earth?) and the older
/as (most of the facts described in the Maleus Malecarm-Chapter
8, footote 2 - the facts of Voodoo - Chapter 4, footote 8 - the
propertes of phlogiston or those of the ether) which so much
exercised the minds of earlier thinkers. And where they d pay
attenton to preceding theories, they ty to accommodate their factual
core in the manner already described, with the help of a hoc
hyp
otheses, a hoc approxmatons, redefiniton of terms, or by simply
7. 'Falsifcaton and the Methodolog of Scientfc Research Progammes', in
Lak
atos
Musgrave (eds), Cntimm aadthcCmmthe/Kaemlcdgc, Cambridge, 1970,
pp
. 93f. ('Naive falsifcatonism' is here also called 'dogmatc'.)
156 AGAI NST METHOD
asering, without any more detailed study of the matter, that the core
'follows from' the new basic principles.
8
They are 'grafted on to
older programmes with which they [are] blatantly inconsistent'.
9
The result of all these proedures is an interestng eisteological
ilusion: the imagined content of the earlier theories (which is the
intersecton of the remembered consequences of these theories with
the newly recognized domain of problems and facts) shrnk and may
decrease to such an extent that it becomes smaller than the imagined
content of the new ideologes (which are the actual consequences of
these ideologes plus all those 'facts', laws, principles which are ted to
them by a hoc hypotheses, a hoc approxmatons or by the say-so of
some influental physicist or philosopher of science - and which
properly belong to the predecessor). Comparing the old and the new
it thus apear that the relaton of empirical contents is like this
new
or, perhaps, like this
while in actual fact it is much more like this
8. 'Einstein's teor is better than . . . Newton's theor anno 1916 . . . &me
it
explained evering that Newton's theor had successflly explained . . .', Lakato,
op. cit., p. 214.
9. Lakatos, disusing Copricus and Bohr, ibid., p. 143.
FI FTEEN 157
domain D representng the problems and facts of the old theor
which are stll remembered and which have been distorted so as to fit
into the new framework. It is this illusion which is responsible for the
p
ersist
ent survival of the demand for increased content.
10
Finally, we have by now seen quite distnctly the need for a hoc
hypothees: a hoc hypotheses and a hoc approxmatons create a
tentatve area of contact between 'facts' and those parts of a new view
which seem capable of explaining them, at some tme in the future
and afer additon of much frther material. They specif possible
exlananda and explanata, and thus detennine the directon of
future research. They may have to be retained forever if the new
framework i partly unfhed (this happened in the case of the
quantum theor, which needs the classical concepts to tum it into a
complete theor). Or they are incororated into the new theor as
theorems, leading to a redefton of the basic terms of the
proeeding ideolog (this happened in the cases of Galileo and of the
theor of relatvit). The demand that the tth-content of the earler
theor as conceied Jhile the earlier theor reiged supree be included
in the tuth-content of the successor i violated in either case.
To sum up: wherever we look, whatever examples we consider, we
see that the principles of critcal ratonalism (take falsificatons
seriously; increase content; avoid a hoc hypotheses; 'be honest' -
whatever that means; and so on) and, a frior, the principles of
logical empiricism (be precise; base your theories on measurements;
avoid vague and untestable ideas; and so on), though practsed
in special areas, give an inadequate account of the past development
of science as a whole and are lable to hinder it in the fture. They give
an inadequate account of science because science i much more
I 0. T illusion is te core of Elie Zhar's excellent papr on te development
fom Lorent to Einstein. According to Zahar, Einstein supreded Lorent with the
explanaton of the prhelion of Mercur (1915). But in 1915 noboy had a yet
suc
ceeded in giving a relattc account of cla ical praton theor to the degree
of approxmaton reached by Laplace ad Poincare, and the implicatons of Lorent on
the atomic level (electon theor of metls) were not accounted for either, but were
gradually replaced by the quatum teor: Lorent wa 'supreded' not by one, but b
at leat two diferent ad mutually incommensurable prm es. Lakatos, in
his
excellent reconstcton of the development of the researh progamme of
Copercus frm the Cmmmmlmto the DeRml., notes proresive chages but
o
because he omit the dcal and the optcal problem and concentates on
tnem
atcs,
pure and simple. Small wonder that bth Zhar and Latos are under the
!

pre
ssion
that the content conditon i stll satsfed. Cf. my shor note 'Zahar on
E
tnstein', in the ntuheum/rtheFhihsqqe/5nm,Marh 1974 a well a R.N.
N

g
aev, 'Spcial Relatvit a a Stage in te Development of Quantum Theor',
Hts
tem
5omnum,No. 34, 1988, pp. 57f.
158 AGAI NST METHOD
'sloppy' and 'irratonal' than its methodological image. And they are
liable to hinder it because the attempt to make science more 'ratonal'
and more precise is bound to wipe it out, as we have seen. The
diference between science and methodolog which is such an
obvious fact of histor, therefore, indicates a weakness of the latter,
and perhaps of the 'laws of reason' as well. For what appears as
'sloppiness', 'chaos' or 'opporunism' when compared with such laws
has a most imporant fncton in the development of those ver
theories which we today regard as essental parts of our knowledge of
nature. Tee 'diation', thee 'eror', are precnditions ofproges.
They permit knowledge to survive in the complex and difcult world
which we inhabit, they permit u to remain free and happy agents.
Without 'chaos', no knowledge. Without a frequent dismissal of
reason, no progress. Ideas which today form the ver basis of science
exst only because there were such things as prejudice, conceit,
passion; because these things op osed reaon; and because they wee
peitted to hae their way. We have to conclude, then, that ee within
science reason cannot and should not be allowed to be comprehen
sive and that it must ofen be overruled, or eliminated, in favour of
other agencies. There is not a single rule that remains valid under al
circumstances and not a single agency to which appeal can always be
made.
1 1
I I . Even Lakatos' ingenious methoolog dos not esap this indictent.
Lakato seems libral because he forbids ver little and he seems ratonal because he is
stU forbids smething. But the only thing he forbids is to &mbea 'degeneratng
research progamme', i.e. a research progamme lacking in novel predictons and
clutered with mhwadaptatons, a progesive. He dos not forbid it use. But this
meas tat his standards prmit a criminal to commit a many crmes a he want
provided he never lies about them. Details in my FhihsqhimF@m,Vol. 2, Chapter
10.
Appendi 1
Having listened to one of my anarchistc serons, Professor Wigner
exclaimed: 'But surely, you do not read all the manuscripts which
people send you, but you throw most of them into the wastepaper
basket.' I most certainly do. 'Anything goes' does not mean that I
shall read ever single paper that has been written - God forbid! -it
means that I make my selecton in a highly individual and
idiosyncratc way, partly because I can't be bothered to read what
doesn't interest me -and my interests change from week to week and
day to day - partly because I am convinced that humanit and even
Science will proft from everone doing his own thing: a physicist
might prefer a sloppy and partly incomprehensible paper full of
mistakes to a crstal-clear expositon because it is a natural extension
of his own, stll rather disorganized, research and he might achieve
success as well as clarit long before his rival who has vowed never to
read a single woolly line (one of the assets of the Copenhagen Schol
was its abilit to avoid premature precision). On other occasions he
mght look for the most perfect prof of a principle he is about to use
in order not to be sidetacked in the debate of what he considers to be
his main results. There are of course so-called 'thinkers' who
subdivide their mail in exactly the same way, come rain, come
sunshine, and who also imitate each other's principles of choice-but
we shall hardly admire them for their uniformit, and we shall
certainly not think their behaviour 'ratonal'. Science needs people
who are adaptable and inventve, not rigid imitators of 'established'
behavioural patters.
In the case of insttutons and organizatons such as the Natonal
Science Foundaton the situaton is exactly the same. The
physiognomy of an organizaton and its efciency depends on its
members and it improves with their mental and emotonal agilit.
ven Procter and Gamble realized that a bunch of yes-men is inferior
m
com
pettve potental to a group of people with unusual opinions
an
d
bus
iness has found ways of incorporatng the most amazing
no
nco
nformists into their machiner. Special problems arise with
1 59
160 AGAI NST METHOD
foundatons that distribute money and want to do this in a just and
reasonable way. Justce seems to demand that the alloaton of funds
be carried out on the basis of standards which do not change from
one applicant to the next and which refect the intellectual situaton in
the fields to be supported. The demand can be satsfied in an a hoc
manner without appeal to unieal 'standards of ratonality': any free
assoiaton of people must respect the illusions of its members and
must gve them insttutonal support. The illusion of rationalit
becomes especially stong when a scientfic insttuton opposes
politcal demands. In this case one class of standards is set against
another such class - and this is quite legitmate: each organizaton,
each party, each religous group has a right to defend its partcular
form of life and all the standards it contains. But scientits go much
furher. Like the defenders of The One True Religon before them
they insinuate that their standards are essental for arriving at the
Truth, or for gettng Results and they deny such authority to the
demands of the politcian. They oppose all politcal interference, and
they fall over each other tg to remind the listener, or the reader, of
the disastrous outcome of the Lysenko afair.
Now we have seen that the belief in a unique set of standards that
has always led to success and will always lead to success is nothing but
a chimera. The theoretical authority of science is much smaller than it
is supposed to be. Its social authority, on the other hand, has by now
become so overpowering that political intererence is necesar to retore a
balanced deloment. And to judge the ef as of such interference one
must study more than one unanalysed case. One must remember
those cases where science, left to itself, committed grievous blunders
and one must not for
9
et the instances when politcal interference did
improe the situaton. Such a balanced presentaton of the evidence
may even convince us that the tme is overdue for adding the
separaton of state and science to the by now quite customary
separaton of state and church. Science is only one of the many
instruments people invented to cope with their surroundings. It is not
the only one, it is not infallible and it has become too powerful, too
pushy, and too dangerous to be left on its own. Next, a word about the
praical aim ratonalists want to realize with the help of their
methodology.
Ratonalists are concered about intellectual polluton. I share this
concer. Illiterate and incompetent books food the market, empty
verbiage full of strange and esoteric terms claims to express profound
I . An example was disussed in the text to fototes 9I Zof Chapter 4.
APPENDI X I 161
insights, 'experts' without brains, character, and without even a
modicum of intellectual, stylistc, emotonal temperament tell us
about our 'conditon' and the means for improving it, and they do not
only preach to us who might be able to look through them, they are let
loose on our children and permitted to drag them down into their
own intellectual squalor.
2
'Teachers' using grades and the fear of
failure mould the brains of the young untl they have lost every ounce
of imaginaton they might once have possessed. This is a disastous
situaton, and one not easily mended. But I do not see how a
ratonalistc methodolog can help. As far as I am concered the frst
and the most pressing problem is to get educaton out of the hands of
the 'professional educators'. The constaints of grades, competton,
regular examinaton must be removed and we must alo searate the
prces oflearing frm the prearation fr a pariclar tra. I grant that
business, religions, special professions such as science or prosttu
ton, have a right to demand that their partcipants and/or
practtoners conform to standards they regard as imporant, and that
they should be able to ascerain their competence. I also admit that
this implies the need for special types of educaton that prepare a man
or a woman for the corresponding 'examinatons'. The standards
taught need not be 'ratonal' or 'reasonable' in any sense, though they
will be usually presented as such; it sufces that they are acted by
the groups one wants to join, be it now Science, or Big Business, or
The One Tre Religion. Afer all, in a democracy 'reason' has just as
much right to be heard and to be expressed as 'unreason' especially in
view of the fact that one man's 'reason' is the other man's insanity.
But one thing must be avoided at all costs: the special standards
which defne special subjects and special professions must not be
allowed to permeate geeal educaton and they must not be made the
defg propert of a 'well-educated person'. General educaton
should prepare citens to choose betwee the standards, or to fnd their
way
in a society that contains groups committed to various standards,
but it must undr no condition bed their mind so that the cnn to the
stand
rd ofone pariclar grup. The standards will be considred, they
will be discssed, children will be encouraged to get profciency in the
more imporant subjects, but onl a one get prfcec in a game, that
is, without serious commitent and without robbing the mind of its
ability to play other games as well. Having been prepared in this way a
young person may decide to devote the rest of his life to a partcular
pro
fession and he may start taking it seriously forthwith. This
2. Even the law now seems to suppon these tendences, as is show in Peter
Huber's Calilce1Rngc, New York, 1991.
162 AGAI NST METHOD
'commibent' should be the result of a conscious decision, on the
basis of a fairly complete knowledge of alteratves, and not a fregone
conclusion.
All this means, of course, that we must stop the scientsts from
taking over educaton and from teaching as ' fact' and as 'the one tue
method' whatever the myth of the day happens to be. Agreement with
science, decision to work in accordance with the canons of science
should be the result of examinaton and choice, and not of a partcular
way of bringing up children.
It seems to me that such a change in educaton and, as a result, in
perspectve will remove a great deal of the intellectual polluton
ratonalists deplore. The change of perspectve makes it clear that
there are many ways of ordering the world that surrounds us, that the
hated constaints of one set of standards may be broken by freely
acceptng standards of a diferent kind, and that there is no need to
reject all order and to allow oneself to be reduced to a whining steam
of consciousness. A society that is based on a set of well-defned and
restictve rules, so that being human becomes synonyous with
obeying these rules, frces the disseter into a no-man 's-land of no rle
at all and thus rbs him ofhis reason and his humanit. It is the paradox
of modem irratonalism that its proponents silently identf
ratonalism with order and artculate speech and thus see themselves
forced to promote stammering and absurdity - many forms of
'mystcism' and 'exstentalism' are impossible without a frm but
unrealized commibent to some principles of the despised ideolog
Gust remember the 'theor' that poety is nothing but emotons
colourfully expressed). Remove the principles, admit the possibility
of many diferent forms of life, and such phenomena will disappear
like a bad dream.
My diagnosis and my suggestons coincide with those of Lakatos
up to a point. Lakatos has identfied overly-rigid ratonality principles
as the source of some versions of irratonalism and he has urged us to
adopt new and more liberal standards. I have identfed overly-rigid
ratonality principles as well as a general respect for 'reason' as the
source of some forms of mystcism and irratonalism, and I also urge
the adopton of more liberal standards. But while Lakatos' great
'respect for great science'
3
makes him look for the standards within
the confines of modem science 'of the last two centuries',4 I
recommend to put science in its place as an interestng but by no
means exclusive form of knowledge that has many advantages but
3. 'Histor', p. I I3.
4. ibid., p. I I I.
APPENDI X I 163
also many drawbacks: 'Although science taken as a whole is a
nuisance, one can stll lear from it.'
5
Also I don't believe that
charlatans can be banned just by tghtening up rules.
Charlatans have exsted at all tes and in the most tghtly-knit
professions. Some of the examples whch Lakatos mentons6 seem
to indicate that the problem is created by too much contol and not by
too litde. 7 This is especially tue of the new 'revolutonaries' and
their 'refor' of the universites. Their fault is that they are Puritans
and not that they are libertnes.8 Besides, who would exect that
cowards will improve the intellectual climate more readily than will
libertnes? (Einstein saw this problem and he therefore advised
people not to connect their research with their profession: research
has to be free from the pressures which professions are likely to
impose.'1 We must also remember that those rare cases where liberal
methodologies d encourage empt verbiage and loose thinkng
('loose' from one point of view, though perhaps not from another)
may be inevitable in the sense that the guilt liberalism is alo a
preconditon of a free and humane life.
Finally, let me repeat that for me the chauvinism of science is a
much greater problem than the problem of intellectual polluton. It
may even be one ofits major causes. Scientsts are not content with
running their own playpens in accordance with what they regard as
the rules of scientfc method, they want to universalize these rules,
they want them to become par of societ at large and they use ever
means at their disposal - argument, propaganda, pressure tactcs,
intidaton, lobbying - to achieve their aims. The Chnese
Communsts recognized the dangers inherent in this chauvinism and
they proceeded to remove it. In the process they restored important
parts of the intellectual and emotonal heritage of the Chinese people
and they also improved the practce of medicine10 It would be of
advantage if other goverments followed suit.
5. Gottried Benn, leter to Gert Micha Simon of I I October 1 949, quoted from
Gott
fried Benn, LnkaadFmsa,ngaadDekamtc,Wiesbaden, 1962, p. 235.
6.
'Falsificaton', p. 1 76, fotote I.
7.
Cf. also his remarks on 'false consciousness' i n 'Histor', pp. 94, 1 08f.
8.
For
an older example, cf. the emEimtoa Lcttm,New York, 1 971 , p. I SO.
9.
ibid.
, pp. 1 05f.
1 0.
Cf. tex to fototes 91 2 of Chapter 4.
1 6
Finall, the kind ofcomparson that undrlie most methodlogie is possible
onl in some rather simple cses. It break dwn whe we tr to compare
non-scietic views with sciece or whe we considr the most adanced,
most geeal and therere most mythological pars ofsciece itel
I have much sympathy with the view, formulated clearly and elegantly
by Whorf (and antcipated by Bacon), that languages and the reacton
patters they involve are not merely instuments for dcbing events
(facts, states of affair), but that they are also shape of events (facts,
states of afairs), 1 that their 'grammar' contains a cosmolog, a
comprehensive view of the world, of soiety, of the situaton of man
2
which infuences thought, behaviour, percepton.
3
According to
Whorf the cosmolog of a language is expressed partly by the over
use of words, but it also rests on classifcatons 'which ha[ve] no over
mark . . . but which operate [] through an invisible "cental exchange"
of linkage bonds in such a way as to determine other words which
mark the class.'
4
Thus ' [t]he gender nouns, such as boy, girl, father,
wife, uncle, woman, lady, including thousands of given names like
George, Fred, Mar, Charlie, Isabel, Isadore, Jane, John, Alice,
Aloysius, Esther, Lester, bear no distnguishing mark of gender like
the Latn - u or - a within each motor process, but nevertheless each
of these thousands of words has an invariable linkage bond
connectng it with absolute precision either to the word "he" or to the
word "she" which, however, does not come into the overt behaviour
1 . According to Whorf 'the background linguistc system (in other words, t
e
grammar) of each language is not merely a reproducing system for voicing idea, but
rather is itself a shaper of idea, the programme and guide for the individual's mental
actvity, for his analysis of impressions, for his synthesis of his mental stok in tade'.
Laagage, 1heaghtaadReali,Cambridge, Mass., 1956, p. 1 21 . See also Appendix 2.
2. As an example cf. Whorfs analysis of Hopi Metaphysics in ibid., pp. 57f.
3. 'User of markedly diferent grammar are pointed by their grammar towards
diferent types of obseratons . . . ', ibid., p. 221 .
4. ibid., p. 69.
164
SI XTEEN 165
pic
ture
untl and unless special situatons of discourse require
it.'
5
Oe classifcatons (which, because of their subterranean nature,
are 'sensed rather than comprehended -awareness of[them] has an
intu
itve qualit'
6
- which 'are quite apt to be more ratonal than
over ones'
7
and which may be ver 'subtle' and not connected 'with
any grand dichotomy'
8
) create 'pattered resistances to widely
divergent points of view'. 9 If these resistances oppose not just the
tuth
of the resisted alteratves but the presumpton that an
alteratve has been presented, then we have an instance of
incommensurabilit.
I also believe that scientfc theories, such as Aristotle's theor of
moton, the theor of relatvit, the quantum theor, classical and
modem cosmolog are sufciently general, sufciently 'deep' and
have developed in sufciently complex ways to be considered along
the same lines as natural languages. The discussions that prepare the
tansiton to a new age in physics, or in astonomy, are hardly ever
resticted to the over features of the orthodox point of vew. They
often reveal hidden ideas, replace them by ideas of a diferent kd,
and change over as well as cover classifcatons. Galileo's analysis of
the tower argument led to a clearer formulaton of the Aristotelian
theor of space and it also revealed the diference between impetus (an
5. ibid., p. 68.
6. ibid., p. 70. Even '[a] phoneme may asume defnite smantc dutes a par of
its rapprt. In English the phoneme 3 ('thor'] (the voiced sound of th) ocur initally
only in the crptotyp [covert classifcaton not connected with any grand dichotomy
p
. 70] of demonstatve parcles (the, this, there, than, etc.). Hence, there is a pqchic
prcsurcagainst acceptng the voiced sound of th in new or imaginar words: thig, thay,
thob, thuzzle, etc., not having demonstatve meaning. Encountering such a new word
(e.g. thob) on a page, we will "instnctvely" give it the voiceles sund 9 of th in "think".
But it is not "instnct". Just our old frend linguistc rapprt again' (p. 76, my italics).
7.
ibid., p. 80. The passage contnues: ' . . . sme rather foral and not ver
meaningful linguistc group, marked by some overt feature, may happn to coincide
ver roughly with some concatenaton of phenomena in such a way a to sugest a
ratonalizaton of this parallelism. In the coure of phonetc change, the distnguishing
mark, ending, or what not is lost, and the class pases from a formal to a semantc one.
It reactance is now what distnguishes it a a class, and it idea is what unifes it. A
tme and use go on, it becomes increaingly organized around a ratonale, it attact

emantcally suitable words and loss forer membr that now are semantcally
mapproprate. Logic is now what holds it together.' Cf. a Mill's account of his
edu
caton
al development as desribd in text to fomote I 4 of Chapter 1 1 .
8 .
Whorf, op. cit., p. 70. Such subde classificatons are called crptotyps by
W
horf. A
crptotyp is 'a submerged, subde, and elusive meaning, corespnding to
no
act
ual
word, yet shown by linguistc analysis to be functonally imprnt in the
g
ra
m
mar'.
9.
ibid., p. 247.
166
AGAI NST METHOD
absolute magnitude that inheres in the object) and momentum (which
depends on the chosen reference system). Einstein's analysis of simul
taneity unearthed some features of the Newtonian cosmolog which,
though unknown, had influenced all arguments about space and tme,
while Niels Bohr found in additon that the physical world could not be
regarded as being entrely separated from the obserer and thus gave
content to the idea of independence that was part of classical physics.
Attending to cases such as these we realize that scientfc arguments
may indeed be subjected to 'pattered resistances' and we expect that
incommensurability will also occur among theories.
(As incommensurability depends on covert classifcatons and
involves major conceptual changes it is hardly ever possible to gve an
explicit defniton of it. Nor will the customary 'reconstuctons'
succeed in bringng it to the fore. The phenomenon must be shown,
the reader must be led up to it by being confronted with a great variety
of instances, and he must then judge for hiself. Ths wl be the
method adopted in the present chapter.)
Interestng cases of incommensurability occur already in the
domain of pection. Given appropriate stmuli, but diferent systems
of classifcaton (diferent 'mental sets'), our perceptual apparatus
may produce perceptual objects whch cannot be easily compared. 10
A direct judgement is impossible. We may compare the two objects in
our meor, but not while attending to the same piaure. The ft
drawig below goes one step frther. It gves rise to perceptual
objects which do not just neate other perceptual objects - thus
retg the basic categories - but prevent the formaton of any
object whatsoever (note that the cylinder in the middle fades into
nothingness a we approach the inside of the to-pronged stmulus).
Not even memory can now gve us a full view of the alteratves.
10. 'A master of intospcton, Kenneth Clark, has recendy described to us most
vividly how even he was defeated when he attempted to "stall" an illusion. Loking at a
great Velasquez, he wanted to obsere what went on when the brsh stokes and dabs of
pigent on the canvas tansfored themselves into a vision of tansfgured realit as he
stepped back. But t a he might, stepping backward and forard, he could never hold
both visions at the same tme . . .', E. Gombrich,Ar amIhmi,Princeton, 1956, p. 6.
S I XTEEN 167
Ever picture with only a modicum of perspectve exhibits this
phenomenon: we may decide to pay attenton to the piece of paper on
which the lines are drawn - but then there is no three-dimensional
patter; on the other hand we may decide to investgate the
propertes of this patter, but then the surface of the paper
disappears, or is integrated into what can only be called an illusion.
There is no way of 'catching' the tansiton from the one to the
other.
1 1
In all these cases the perceived iage depends on 'mental
sets' that can be changed at will, without the help of drugs, hypnosis,
reconditoning. But mental sets may become frozen by illness, as a
result of one's upbringing in a certain culture, or because of
physiological determinants not in our contol. (ot ever change of
language is accompanied by perceptual changes.) Our atttude
towards other races, or towards people of a diferent cultural
background, often depends on 'frozen' sets of the second kind:
having leared to 'read' faces in a standard way we make standard
judgements and are led astay.
A interestng example of physiologically determined sets leading
to incommensurability is provided by the delomet of human
percetion. As has been suggested by Piaget and his schol,
12
a
child's percepton proceeds through various stages before it reaches
its
relatvely stable adult form. In one stage, objects seem to behave
very much like after-images and are treated as such. The child
I I . Cf. R.L. Gregor, 1hcIotclligcotEc, London, 1970, Chapter 2. Cf. also the
distinction beteen eikon and phantasma in Plato, 5@hutc,23Sb8f: 'This "appear
ing" or "seeming" without really "being" . . . all these expressions have always been
and still are deeply involved in perplexity.' Plato talks about the distortons in statues of
col
ossa
l size which were introduced to make them q carwith the proper proprions.
'I cannot make use of an illusion and watch it,' says Gombrich in such cases, op. cit.,
p.
6.
12. ].
Piaget, 1hcCemt~ieae/RcaliiothcChiU,New York, 1954, pp. Sf.
168 AGAI NST METHOD
follows the object with his eyes untl i t disappears; he does not make
the slightest attempt to recover it, even if this should require but a
minimal physical (or intellectual) efort, an efort, moreover, that is
already within the child's reach. There is not even a tendency to
search - and this is quite appropriate, 'conceptually' speaking. For it
would indeed be nonsensical to 'look for' an after-image. Its
'concept' does not provide for such an operaton.
The arrival of the concept, and of the perceptual image, of material
objects, changes the situaton quite dramatcally. There occurs a
drastc reorientaton of behavioural patters and, so one may
conjecture, of thought. Afer-images, or things somewhat like them,
stll exst; but they are now difcult to fnd and must be discovered
by special methods (the earlier visual world therefore literally dis
apear).
13
Such methods proceed from a new conceptual scheme
(after-images occur in humans, they are not pars of the physical
world) and cannot lead back to the exact phenomena of the previous
stage. (These phenomena should therefore be called by a diferent
name, such as 'pseudo-after-images' - a very interestng perceptual
analogue to the tansiton from, say, Newtonian mechanics to special
relatvit: relatvit, too, does not give us Newtonian facts, but
relatvistc analogues of Newtonian facts.) Neither afer-images nor
pseudo-afer-images have a special positon in the new world. For
example, they are not teated as eidce on which the new noton of a
material object i supposed to rest. Nor can they be used to elin
t noton: afer-images arse togethe with it, they depend on it, and
are absent from the minds of those who do not yet recoge material
objects; and pseudo-afer-images disapear as soon as such
recogniton takes place. The perceptual feld never contains afer
images together with pseudo-after-images. It is to be admtted that
every stage possesses a kind of obseratonal 'basis' to which special
attenton is paid and from which a multtude of suggestons are
received. However, this basis (a) change from stage to stage, and (b) it
is par of the conceptual apparatus of a given stage, not its one and
only source of interpretaton as some empiricists would lke to make
us believe.
Considering developments such as these, we may suspect that the
family of concepts centing upon 'material object' and the famly of
concepts centng upon 'pseudo-after-image' are incommensurable
13. This seems to be a general feature of the acquisiton of new prceptual worlds:
'The older representatons for the most par have to be suppresed rather than
refored,' writes Statton i his epoh-maing essy 'Vision without Inverion of the
Retnal Image'. DcFqchelepmlRnimD,1 897, p. 471 .
SI XTEEN 169
in precisely the sense that is at issue here; these families cannot be
used simultaneously and neither logical nor perceptual connectons
can be established between them.
Now is it reasonable to expect that conceptual and perceptual
changes of this kind occur in childhood only? Should we welcome the
fact, if it is a fact, that an adult is stuck with a stable perceptual world
and an accompanying stable conceptual system, which he can modif
in many ways but whose general outlines have forever become
im
mobilized? Or is it not more realistc to assume that fundamental
changes, entailing incommensurabilit, are stll possible and that
they should be encouraged lest we remain forever excluded from
what might be a higher stage of knowledge and consciousness?
Besides, the queston of the mobilit of the adult stage is at any rate an
empirical queston that must be attacked by reearch, and cannot be
settled by methodological fat.
1
4 The attempt to break through the
boundaries of a given conceptual system is an essental part of
such research (it also should be an essental part of any interestng
life).
Such an attempt involves much more than a prolonged 'critcal
discussion'
1
5
as some relics of the enlightenment would have us
believe. One must be able to produce and to grap new perceptual and
conceptual relatons, including relatons which are not immediately
apparent (covert relatons -see above) and that cannot be achieved by
a critcal discussion alone (cf. also above, Chapters 1 and 2). The
orthodox accounts neglect the covert relatons that contribute to their
meaning, disregard perceptual changes and teat the rest in a rigidly
standardized way so that any debate of unusual ideas is at once
stopped by a series of routne responses. But now this whole array of
responses is in doubt. Ever concept that ocurs in it is suspect,
especially 'fundamental' concepts such as 'obseraton', 'test', and, of
course, the concept 'theor' itself. And as regards the word 'tuth', we
can at this stage only say that it certainly has people in a tzzy, but
has not achieved much else. The best way to proeed in such
circumstances is to use examples which are outside the range of the
routne responses. It is for this reason that I have decided to examine
means of representaton diferent from languages or theories and to
develop my terminolog in connecton with them. More especially, I
shall examine stles in paintng and drawing. It will emerge that there
1
4. As
Lakatos attempt to do: 'Falsifcaton', p. 1 79, footote 1 : 'Incommen
sur
able
theories are neither inconsistent wit each other, nor comparble for content.
B
ut
we
can mmcthem, by a dictionar, inconsistent and their content comparble.'
1
5. Popper in CntimmamthcCmmthe/Kaemlcdgc,p. 56.
1 70 AGAI NST METHOD
are no 'neutal' objects which can be represented in any style, and
which measure its closeness to 'reality'. The applicaton to languages
is obvious.
The 'archaic sle' as defined by Emanuel Loewy in his work on
ancient Greek art 6 has the following characteristcs.
(1) The stucture and the movement of the figures and of their
pars are limted to a few typical schemes; (2) the individual forms are
stylized, they tend to have a certain regularity and are 'executed with
. . . precise abstacton' ;17 (3) the representaton of a for depends
on the cntour which may retain the value of an independent line or
form the boundaries of a silhouette. 'The silhouettes could be gven a
number of postures: they could stand, march, row, drive, fight
i
die,
lament. . . . But always their essental stucture must be clear'; 8 (4)
c/our appear in one shade only, and gradatons of light and shadow
are missing; (5) as a rule the figures show their pars (and the larger
episodes their elements) in their most coplete apea - even if this
means awkwardness in compositon, and 'a certain disregard of
spatal relatonships'. The pars are gven their known value even
when this confcts with their seen relatonship to the whole;19 thus
(6) with a few well-determined exceptons the figures which form a
compositon are arranged in such a way that uelps are aoid and
objects situated behind each other are presented as being side by
side; (7) the mironmet of an acton (mountains, clouds, tees, etc.)
is either completely disregarded or it is omitted to a large extent. The
acton fors self-contained units of typical scenes (battes, fnerals,
etc.).20
These stylstc elements which are found, in various modifcatons,
in the drawings of children, i the 'frontal' art of the Egptans, in
early Greek a, a well a among so-called Primitves, are exlained
I 6. DieNmum ie&miadNtm CnechuchmKumt,Rome, 190 , Chapter I .
Lw uses 'achaic' a a gmmcter coverng phenomena i Egptan, Greek ad
Prmitve Art, in the drawing of children and of untutored obserer. In Greece his
remarks apply to the g~msle(100to 70 DC) dow t te archaupemd(70 to
50 DC) which teats the human figre in greater detail and involves it i lively
episoes. Cf. as F. Mat, CcchichtedCnemummKumt,Vol. I, 1 950, a weD a
Beazly and Ashmole, Cnek5mtureaadFaiatiag,Cambrdge, 1 966, Chapter II and
III.
1 7. Webster, Fm MmteHo, New York, 1 964, p. 292. Webster regards
this use of'simple and clear patter' i Greek geometic ar a 'the forerunner oflater
developments i a (ultmately the inventon of perpctve), mathematcs, and
philosophy'.
1 8. Webster, op. cit., p. 205.
19. ibid., p. 207.
20. Beazly and Ashmole, op. cit., p. 3.
SI XTEEN 171
by Loewy on the basis of psychological mechanisms: 'Side by side
with the images which reality presents to the physical eye there exsts
an entrely different world ofimages which live or, better, come to life
in our mind only and which, although suggested by reality, are totally
transformed. Ever primitve act of drawing . . . ties to reproduce
these images and them alone with the instnctve regularity of a
psychical functon. '21 The archaic style changes as a result of
'numerous planned obseratons of nature which modif the pure
mental images',22 initate the development towards realism and thus
start the histor of art. Natural, physiological reasons are given for the
archaic style and for its change.
Now it is not clear why it should be more 'natural' to copy memor
images than images of percepton which are better defned and more
permanent.23 We also fnd that realism ofen preced more
schematc forms of presentaton. This is tue of the Old Stone
Age, 24 of Egptan A, 25 of Attc Geometic Art. 26 In all these
cases the 'archaic style' is the result of a conciou efr (which may
of course be aided, or hindered, by unconscious tendencies and
physiological laws) rather than a natural reacton to interal deposits
21 . Lowy, op. cit., p. 4.
22. ibid., p. 6.
23. The facts of prpctve are notced hut they do not enter the pictoral
presentation; this is seen fom literar desriptons. Cf. H. Schafer, lZgtumn
Kamt,Wiesbaden, 1 963, pp. 88f, where the problem is frher disused.
24. Cf. Paolo Graiosi, FmlithicZu, New York, 1 960, and Andre Ler
Gourhan, 1remarce/FrehutencZu,New York, 1 967, both with exceUent illustatons.
These results were not ko to Lowy: Cartailhac's 'Mea culpa d'un septque', for
example, appared only in 192.
25. Cf. the change in the presentaton of animals in the coure of the tansiton
from predynatc tmes to te Firt Dnat. The Berlin lion (Berlin, Staatliches
Museum, Nr. 2240) is wild, threatening, quite diferent in exresion and executon
from the majestc animal of the Second and Third Dnates. The latter seems
to he more a representaton of the mcqt lion than of any individual lion. Cf.
also the diference between the falcon on the victor tablet of King Narer
(backside) and on the burial stone of King Wadji (Djet) of the First Dnat.
'Everywhere one advanced to pure clarit, the fors were stengthened and made
simp
le,' Schafer, op. cit., pp. 1 2f, espcially p. 1 5 where frher details are
given.
26.
'Attc geometc an should not he called primitve, although it ha not the kind
of
photo
graphic realism which literar scholar seem to demand in paintng. It is a
highly
sophistcated an with it own conventons which sere it own purses. As with
t
he
shap
es and the oramentaton, a revoluton separates it from late Mycenaean
pai
ntn
g. In this revoluton fgres were reduced to their minimum silhouettes, and ,
ou
t of
these minimum silhouettes the new an wa built up.' Webster, op. cit.,
p
. 20
5.
172 AGAI NST METHOD
of exteral stmuli.
27
Instead of looking for the psychological cause
of a 'style' we should therefore rather ty to discover its eleets,
analyse their function, compare them with other phenomena of the
same culture Oiterar style, sentence constructon, grammar,
ideology) and thus arrive at an outline of the underlying world-vie
including an account of the way in which this world-view influences
percepton, thought, argument, and of the limits it imposes on the
roaming about of the imaginaton. We shall see that such an analysis
of outlines provdes a better understanding of the proess of
conceptual change than either a naturalistc account which
recognizes only one 'reality' and orders arworks by their closeness to
it, or tite slogans such as 'a critcal discussion and a comparison of
. . . various frameworks is always possible'.
28
Of course, some knd of
comparison is always possible (for example, one physical theor may
sound more melodious when read aloud to the accompaniment of a
guitar than another physical theor). But lay down specfc rules for the
proess of comparison, such as the rules of logic as applied to the
relaton of content classes, or some simple rules of perspectve and
you will find exceptons, undue restrictons, and you will be forced to
talk your way out of touble at ever tur. It is much more interestng
and instuctve to examine what kinds of things can be said
(represented) and what kinds of things cannot be said (represented) i
the comparson has to take place within a ceain specfed and historcll
well-etreched frameork. For such an examinaton we must go
beyond generalites and study frameworks in detail. I start with an
account of some examples of the archaic style.
The human figure shows the following characteristcs: 'the men
are ver tall and thin, the tunk of a triangle tapering to the waist, the
head of a knob with a mere excrescence for a face: towards the end of
the style the head is lit up - the head knob is drawn in outline, and a
dot signifies the eye'.
29
All, or almost all, parts are shown in profile
and they are stung together like the limbs of a puppet or a rag doll.
They are not 'integrated' to form an organic whole. This 'additve'
feature of the archaic style becomes ver clear from the teatent of
the eye. The eye does not partcipate in the actons of the body, it
does not guide the body or establish contact with the surrounding
situaton, it does not 'look'. It is added on to the profile head like part
2 7. This thesis is funher suppned by the obseraton that so-called Primitves
often tm their back to the objects they want to draw; Schafer, op. cit., p. 102, afer
Conze.
28. Popper, in Cntimm,etc., p. 56.
29. Beazly and Ashmole, op. cit., p. 3.
S I XTEEN 173
of a notaton as if the artst wanted to say: 'and beside all these other
things such as legs, arms, feet, a man has also eyes, they are in the
head, one on each side'. Similarly, special states of the body (alive,
dead, sick) are not indicated by a special arrangement of its parts, but
by putng the same standard body into various standard positions.
Thus the body of the dead man in a funeral carriage is artculated in
exactly the same way as that of a standing man, but it is rotated
through 90 degrees and inserted in the space between the bottom of
the shroud and the top of the bier. 30 Being shaped like the body of a
live man it is in ad ition put into the death positon. Another instance
is the picture of a kid half swallowed by a lion. 3 1 The lion loks
feroious, the kid looks peacefl, and the act of swallowing is simply
tacked on to the presentaton of what a lion is and what a kid is. (We
have what is called a parataic agregate: the elements of such an
aggregate are all given equal importance, the only relaton beteen
them is sequental, there is no hierarchy, no part is presented as being
subordinate to and determined by others.) The picture rea:
feroious lion, peacefl kid, swallowing of kid by lion.
The need to show ever essental part of a situaton often leads to a
separaton of parts which are actually in contact. The picture
becomes a list. Thus a charioteer standing in a carriage i shown as
standing above the floor (which is presented in its fullest view) and
unencumbered by the rails so that his feet, the floor, the rails can all
be clearly seen. No touble arises if we regard the paintng as a viual
ctalogue of the part of an event rather than as an illusor rendering
of the event itself (no touble arises when we say: his fet touched the
foor which is reaangulr, and he was surrounded by a railing . + . )32
But such an interpretaton must be leared, it cannot be simply read
of the picture.
The amount of leag needed may be considerable. Some
Egptan drawings and paintngs can be decoded only with the help
of either the represented object itself or with the help of three
dimensional copies of it (statuar in the case of humans, animals,
etc.)
. Using such informaton we lear that the chair in Figre A
repr
esents the object of Figre C and not the object of Figure B and
that it must be read: 'chair with backrest and four legs, legs connected
30. Webster, op. cit., p. 204: 'The painter feels the need to say that he ha two
ars, two leg, and a manly chest.'
3 I . R. Hampl, DieClochauseHmamdwiUastsoaoZot,Tibingen, 1 952.
3 2. 'All geometc pictures of chariots show at least one of these distorons.'
Webster, op. cit., p. 204. Late Mycenaean potter, on the other hand, has the leg of
the
ocupants concealed by the side.
174 AGAI NST METHOD
by support' where it i s understood that the front legs are connected
with the back legs and not with each other.
33
The interpretaton of
groups is complicated and some cases are not yet understood.
3
4
FGURE A
FGURE B
FGURE C
(Being able to 'read' a certain style also includes knowledge of
what features are i"elrant. Not ever feature of an archaic list has
representatonal value just as not ever feature of a written sentence
plays a role i articulating its content. This was overlooked by the
Greeks who started inquiring into the reasons for the 'dignified
postures' of Egptan statues (already Plato commented on this).
Such a queston 'might have struck an Egptan artst as it would
strike us if someone inquired about the age or the mood of the king
on the chessboard'.
3
5
)
1
So far a brief account of some peculiarites of the 'archaic' style.
A style can be described and analysed in various ways. The
descriptons given so far paid attenton to fnnal fature: the archaic
style provides visible lists whose parts are arranged in roughly the
same way in which they occur in 'nature' except when such an
arrangement is liable to hide important elements. All parts are on the
33. Schafer, op. cit., p. 1 23.
34. ibid., pp. 223f.
35. Gombrich, op. cit., p. 1 34, wt literature.
S I XTEEN 175
same level, we are supposed to 'read' the lists rather than 'see' them
as
illusor accounts of the situaton. 36 The lists are not organized in
any way except sequentally, that is, the shape of an element does not
depend on the presence of other elements (adding a lion and the act
of swallowing does not make the kid look unhappy; adding the
process of dying does not make a man look weak). Archaic pictures
are paratacic agegates, not hypotactc systems. The elements of the
aggregate may be physical parts such as heads, arms, wheels, they
may be states of afair such as the fact that a body is dead, they may be
actons, such as the acton of swallowing.
Instead of describing the formal features of a style, we may
describe the ontologcal fatures of a world that consists of the elements
represented in the style, arranged in the appropriate way, and we may
also describe the impression such a world makes upon the viewer.
This is the proedure of the art critc who loves to dwell on the
peculiar behaviour of the characters which the artst puts on his
canvas and on the 'interal life' the behaviour seems to indicate.
Thus G.M.S. Hanfmann37 writes on the archaic fgure: 'No matter
how animated and agile archaic heroes may be, they do not appear to
move by their own will. Their gestures are explanator formulae
imposed upon the actors from without in order to explain what sort of
acton is going on. And the crucial obstacle to the convincing
portrayal of inner life was the curiously detached character of the
archaic eye. It shows that a person is alive, but it cannot adjust itself to
the demands of a specifc situaton. Even when the archaic arst
succeeds in denotng a humorous or tagic mood, these factors of
exteralized gesture and detached glance recall the exaggerated
animaton of a puppet play.'
A ontological descripton frequently adds just verbiage to the
36. 'We come closer to the factual content of frontal [goanuhigoj drawings of
objects, if we stan by remiag e{their partal contents in the form of naratve
declaratve sentences. The frontal moe of representaton gives us a "visual concept"
[5ehbq]| of the thing (the situaton) represented.' Schafer, op. cit., p. 1 1 8. Cf. als
Webster, op. cit., p. 202, about the 'naratve' and 'explanator' character of
Mycenaean and geometic an. But cf. H.A. Groenewegen-Frankfon, Znc aad
Mmt, London, 1951, pp. 33f: the scenes from daily life on the walls of Egptan
tom
bs 'should be "read": harestng entails ploughing, sowing, and reaping; care of
canle entails fording of streams and milking . . . the sequence of scenes is purely
conceptual, not naratve, nor is the writing which ocurs with the scenes dramatc in
character. The signs, remarks, names, songs and explanatons, which illuminate the
act
on
. . . do not link events or explain their development; they are typical sayings
belon
ging to typical situatons.'
37.
'Naraton in Greek An', Zmmcaaeamale/Zrchmle, Vol. 61 , Januar
195
7, p. 74.
1 76 AGAI NST METHOD
formal analysis; it is nothing but an exercise in 'sensitvity' and
cuteness. However, we must not disregard the possibility that a
partcular style ge a precise account ofthe world a it is perceied b the
arist and hi conteporaries and that every formal feature corresponds
to (hidden or explicit) assumptons inherent in the underlyng
cosmolog. (In the case of the 'archaic' style we must not disregard
the possibility that humans then actually flt themselves to be what
we today would call puppets guided by outside forces and that they
saw and treated others accordingly.) Such a realistic interretation of
styles would be in line with Whorfs thesis that in additon to being
instuments for dcrbing events (which may have other features, not
covered by any descripton) languages are also shape of events (so
that there is a linguistc limit to what can be said in a given language,
and this limit coincides with the limts of the thing itself) but it would
go beyond it by including non-linguistc means ofrepresentaton.
38
The realistc interpretaton is very plausible. But it must not be taken
for granted.
39
It must not be taken for granted, for there are technical failures,
special purposes (caricature) which may change a style without
changing the cosmolog. We must also remember that humans have
roughly the same neurophysiological equipment, so that percepton
cannot be bent in any directon one chooses.
40
And in some cases we
can indeed show that deviatons from a 'faithful rendering of nature'
ocur in the presence of a detailed knowledge of the object and side
by side with more 'realistc' presentatons: the workshop of the
sculptor Thutosis in Tel al-Amama (the ancient Achet-Aton)
contains masks directly taken from live models with all the details of
the foraton of the head (indentatons) and of the face intact, as well
as heads developed from such masks. Some of these heads presere
the details, others eliminate them and replace them by simple fors.
An extreme example is the completely smooth head of an Egptan
man. It proves that 'at least some artsts remained consciously
independent of nature'.
41
During the reign of Amenophis I (BC
1364-1347) the mode of representaton was changed twice; the frst
change, towards a more realistc style, occurred merely four years
38. Cf. fotote l and text of the present chapter.
39. For a sketch of the problems that arise in the case of phsimlthmnccf. my
'Reply to Critcism', estea5twiciatheFhilesqhe/5om,Vol. 2, 1965, sectons 5-
8, and especially the list of problems on p. 234. Hanson, Poppr and other take it for
granted that realism is corect.
4. It may be different with drg-induced states, espcially when they are made
pan of a systematc coure of educaton.
41 . Schafer, op. cit., p. 63.
S I XTEEN 177
after his ascension to the throne which shows that the technical ability
for realism exsted, was ready for use, but was intentonally left
undeveloped. An inferece from stle (or language) to cosmolog and mod
ofpercetion therere need special armet: it cannot be mad a a matter
ofcoure. (A similar remark applies to any inference from popular
theories in science, such as the theory of relatvity, or the idea of the
moton of the earth, to cosmology and modes of percepton.)
The argument (which can never be conclusive) consists in pointng
to characteristc features in distant felds. If the idiosyncrasies of a
partcular style of paintng are found also in statuary, in the grammar
of contemporary languages (and here especially in cover classifca
tons which cannot be easily twisted around), if it can be shown that
these languages are spoken by artsts and by the common folk alike, if
there are philosophical principles formulated in the languages which
declare the idiosyncrasies to be features of the world and not just
artfacts and which t to account for their orgin, if man and nature
have these features not only in paintngs, but also in poet, in
popular sayings, in common law, if the idea that the features are parts
of noral percepton is not contadicted by anything we know from
physiology, or from the psychology of percepton, if later thinkers
attack the idiosyncrasies as 'errors' resultng from an ignorance of the
'tue way', then we may assume that we are not just dealing with
technical failures and partcular puroses, but with a coheret way of
li, and we may exect that people involved in this way of life see the
world in the same way in which we now see their pictures. It seems
that all these conditons are satsfed in archaic Greece: the foral
stucture and the ideology of the Greek eic as reconstcted both
from the text and from later references to it repeat all the peculiartes
of the later geometc and the early archaic style!2
To star with, about nine-tenths of the Homeric epics consist of
frul
a which are prefabricated phrases extending in length from a
single word or two to several complete lines and which are repeated at
apropriate places!3 One-ffth of the poems consist of lines wholly
repeated from one place to another; in 28,000 Homerc lines there
are about 25,000 repeated phrases. Repettons ocur already in
Mycenaean cour poet and they can be taced to the poet of
42. Webster, op. cit., pp. 294f.
43. In the 20t centur the role of forulae was desribed and tested by Milman
P
arry
,
LJ
pithu trmitim ccchcHemrc, Paris, 1928; Hanard5tadic ia Clmsiml
F
hile
le
,Vols. 41 , 1 930, 43, 1932. For a brief account cf. D. l. Page, Huteqaadthc
H

mc
Iliad, Berkeley, 1 966, Chapter V, a weU as G.S. Kirk, HemoaadthcEpic,
Cambridge, 1965, Part I .
178 AGAI NST METHOD
easter courts: 'Titles of gods, kings, and men must be given
correctly, and in a courtly world the principle of correct expression
may be extended further. Royal correspondence is highly formal, and
this formalit is extended beyond the messenger scenes of poet to
the formulae used for intoducing speeches. Similarly, operatons are
reported in the terms of the operaton order, whether the operaton
order itself is given or not, and this technique is extended to other
descriptons, which have no such operaton orders behind them.
These compulsions all derve ultmately from the court of the king,
and it is reasonable to suppose that the court in tum enjoyed such
formalit in poet.'
44
The conditons of (Sumerian, Babylonian,
Hurrian, Hethitc, Phoenician, Mycenaean) courts also explain the
occurrence of standardized elements of contet (typical scenes;
the king and the nobles in war and peace; furiture; descripton
of beautful things) which, moving from cit to cit, and even
across natonal boundares are repeated, and adapted to local
circumstances.
The slowly arsing combinaton of constant and varable elements
that is the result of numerous adaptatons of this kind was utlized by
the illiterate poets of the 'Dark Age' of Greece who developed a
language and forms of expression that best sere the requirements of
oral composition. The requirement of meor demanded that there be
ready-made descrptons of events that can be used by a poet who
composes in his mind, and without the aid of writng. The
requirement of metre demanded that the basic descrptve phrases be
fit for use in the various parts of the line the poet is about to complete:
' Unlike the poet who writes out his lines . . . [the oral poet] cannot
think without hurry about his next word, nor change what he has
made nor, before going on, read over what he has just wrtten . . . . He
must have for his use word groups all made to fit his verse.'
45
Economy demanded that given a situaton and a cerain metical
constaint (beginning, middle or end of a line) there be only one way
of contnuing the narraton - and this demand is satsfied to a
surrsing extent: 'All the chief characters of the 1iad and the
Odysse, if their names can be fitted into the last half of the verse along
with an epithet, have a noun-epithet formula in the nominatve,
beginning with a simple consonant, which fills the verse between
the tochaic caesure of the third foot and the verse end: for in
stance, noA'Aw; bto 'ObuooEu. In a list of thirty-seven
characters who have formulae of this type, which includes all those
4. Webster, op. cit., pp. 75f.
45. M. Par, Hanard5twCl. Fhil.,41 , 1 930, p. 77.
S I XTEEN 179
havng any importance in the poems, there are only three names
which have a second forula which could replace the frst.'
4
'If you
take in the fve grammatc cases the singular of all the noun-epithet
formulae used for Achilles, you will fnd that you have forty-fve
diferent formulae of which none has, in the same case, the same
metrical value.'
47
Being provded for in this manner, the Homeric
poet 'has no interest in originali of expression, or in variety. He uses
or adapts inherited formulae.'
4
He does not have a 'choice, do[es]
not even think in terms of choice; for a given part of the line, whatever
declension case was needed, and whatever the subject matter might
be, the forular vocabular supplied at once a combinaton of words
ready-made'.
49
Using the forulae the Homeric poet gives an account of tpical
scee in which objects are occasional1 described by 'adding the parts
on in a strng ofword i appositon'. 0 Ideas we would today regard
as being logically subordinate to others are stated in separate,
grammatcally co-ordinate propositons. Example (Iliad, 9.556f:
Meleagros 'lay by his wedded wife, fair Cleopata, daughter of fair
ankled Marpessa, daughter of Euenos, and of Ides, who was the
stongest of men on earth at that tme - and he agaist lord Phoebus
Apollo took up his bow for the sake of the fair-ankled maid: her then
in their halls did her father and lady mother call by the name of
Alkyon because . . .' and so on, for ten more lies and two or three
more major themes before a major stop. This paratactic feature of
Homeric poetr which parallels the absence of elaborate systems of
subordinate clauses in early Greek
5 1
also makes it clear why
46. ibid., pp. 86f.
47. ibid., p. 89.
48. Page, op. cit., p. 230.
49. ibid., p. 242.
50. Webster, op. cit., pp. 9f; my italics.
51 . Cf. Raphael Kiihner,Ausfohrliche Grommotik dGrechiche Sprohe, 2. Teil,
reprin
ted Darmstadt, 1 966. In the 20t centur such a paratactc or 'simultanistc',
way of presentaton was used by the early expressionist, for example by Jacob von
Hoddis in his pom Welted:
Dem Birger fliegt vom spiten Kopf der Hut,
In allen Liften hallt es wie Geschrei.
Dachdecker stiren ab und gehn entwei,
Und an den Kisten - liest man-steigt die Flut.
Der Sturm ist da, die wilden Meere hupfen
A Land, ur dicke Damme zu zerdricken.
Die meisten Menschen haben einen Schnupfen.
Die Eisenbahnen fallen von den Brcken.
180 AGAI NST METHOD
Aphrodite is called 'sweetly laughing' when in fact she complains
tearfully (Iiad, 5. 3 7 5), or why Achilles is called 'swift footed' when he
is sittng talking to Priam (Iliad, 24.559). Just as in late geometic
pottery (in the 'archaic' style of Loewy) a dead body is a live body
brought into the positon of death (cf. above, text to footote 30) or an
eaten kid a live and peaceul kid brought into the appropriate relaton
to the mouth of a ferocious lion, in the very same way Aphrodite
complaining is simply Aphrodite -and that is the laughing goddess
insered into the situaton of complaining in which she parcipates
only exterally, without changing her nature.
The aditie treatmet of events becomes very clear in the case of
(human) moton. In Iia, 22.298, Achilles drags Hector along in the
dust 'and dust arose around him that was dragged, and his dark hair
flowed loose on either side, and in the dust lay his once fair head'
that is, the proces of dragging contains the state of lying as an
independent .art which together with other such parts consttutes
the moton.
5
Speaking more abstactly, we might say that for the
poet 'te is composed of moments'.
53
Many of the similes assume
that the parts of a complex entty have a life of their own and can be
separated with ease. Geometical man is a visible list of parts and
positons; Homeric man is put together from limbs, surfaces,
connectons which are isolated by comparing them with inanimate
objects of precisely defned shape: the tnk of Hippolohos rolls
through the battle feld like a log after Agamemnon has cut of his
Von Hoddis claims Homer as a precursor, explaining that simultaneit wa used by
Homer not in order to make an event more tansparent but in order to create a feeling
of immeasurable spaciousness. When Homer describes a battle and compares the
noise of the weapns with the beat of a woodcuter, he merely wants to show that while
there is battle there is also the quietess of wods, interupted only by the work of the
woodcuter. Catastophe cannot be thought without simultaneously thinking of some
utterly unimprant event. The Great is mixed up with the Small, the Imprant with
the Trivial. (For the repor cf.J.R. Becher in Eqrcsieaismm,ed. P. Raabe, Olten and
Freiburg, I 965, pp. 50f; this shor aricle also contains a descripton of the
temendous impression von Hodis' eight-liner made when it frst came out in I 91 1 .)
One cannot infer that the same impression was created in the listener of the Homeric
singers who did not pssess a complex and romantcizing medium that had
deteriorated into tearful sentmentalit as a background for comparison.
52. Cf. Gebhard Kur, Dantcllaag/m mschlicho mqag ia t llim,
Heidelberg, 1 966, p. 50.
53. This is the theor ascribed to Zeno by Aristotle, Fhsio,239b, 31 . The theor
comes forth most clearly in the argument of the arow: 'The arow at flight is at rest.
For, if everything is at rest when it occupies a space equal to itelf, and what is in fight
at any given moment always occupies a space equal to itself, it cannot move' (afer
Fhsio, 239b). We cannot say that the theor was held by Zeno himself, but we may
conjecture that it played a role in Zeno's tme.
S I XTEEN 181
arms and hs head (Ilia, 1 1 . 1 46 - OAJo, round stone of
cylindrical shape), the body of Hector spins like a top (Ilia, 1 4.41 2),
the head of Gorgythion drops to one side 'like a gardn /obeing
heavy with fruit and the showers of spring' (Ilia, 8.302); 4 and so on.
Also, the formulae of the epic, especially the noun-epithet combina
tons, are frequently used not according to content but according to
metrical convenience: 'Zeus changes from counsellor to storm
mountain god to pateral god not in connecton with what he is doing,
but at the dictates of mete. He is not neheleeata Ze when
he is gathering clouds, but when he is flling the metcal unit,
v v - v v - -',55 just as the geometical artst may distort spatal
relatons - intoduce contact where none exsts and break it where it
ocurs - in order to tell the visual stor in hs own partcular way.
Thus the poet repeats the formal features used by the geometc and
the early archaic artsts. Neither seems to be aware of a 'underlying
substance' that keeps the objects together and shapes their parts so
that they reflect the 'higher unity' to which they belong.
Nor is such a 'hgher unity' found in the concepts of the language.
For example, there is no expression that could be used to describe the
human body as a single entty.56 Soma is the corse, d is
accusatve of specifcaton, it means 'in stucture', or 'as regards
shape', reference to limbs occurs where we today speak of the body
(yuia, limbs as moved by the joints; JtAm, limbs in their
bodily stength; MAuvo yia, hs whole body tembled;
tOQo f JAtwv EQQV, his body was flled with stength).
All we get is a puppet put together from more or less artculated
parts.
The puppet does not have a soul in our sense. The 'body' is an
aggregate of limbs, tunk, moton, the 'soul' is an aggregate of
'mental' events which are not necessarily private and which may
belong to a diferent individual altogether. 'Never does Homer in his
descripton of ideas or emotons, go beyond a purely spatal, or
quanttatve defton; never does he attempt to sound their special,
54.
Kur, lo. cit.
55. R. Lattmore, 1heIlime/Hemo,Chicago, 1 951 , pp. 39f.
56. For the following cf. B. Snell, 1heDuce/theMiad, Harr Torchboks,
I 960, Chapter 1 . Snell's views have been critcized but seem to surive the critcism.
Cf. the repon in F. Kraf, leqhcb Uatmachaagm za Hemo aad Hcie
H
@a~ata, Heft 6, Gittngen, 1963, pp. 25f. In his Ccammelte 5chn/m,
G
ittngen, 1 966, p. 1 8, Snell also arges that 'in Homer we never find a prnal
de
cisio
n, a conscious choice made by an actng human being. A human being who is
faced with various possibilites never thinks: "now it depnds on me, it depnds on
what I decide to do".'
182 AGAI NST METHOD
non-physical nature.'
57
Actons are initated not by an 'autonomous
I', but by further actons, events, occurrences, including divine
interference. And this is precisely how mental events are e
perieced.
58
Dreams, unusual psychological feats such as sudden
remembering, sudden acts of recogniton, sudden increase of vital
energ, during battle, during a stenuous escape, sudden fts of anger
are not only elained by reference to gods and demons, they are also
flt as such. Agamemnon's dream 'listened to his [Zeus'] words and
descended' (Iiad, 2. 1 6) -the dream descends, not a fgure i it -'and
it stood then beside his [Agamemnon's] head in the likeness of
Nestor' (Iiad, 2.20). One does not hae a dream (a dream is not a
'subjectve' event), one sees it (it is an 'objectve' event) and one also
sees how it approaches and moves away.
59
Sudden anger, fts of
strength are described and flt to be divine acts:
60
'Zeus builds up
and Zeus diminishes stength in man the way he pleases, since his
power is beyond all others' (Iiad, 20.241) is not just an objectve
descripton (that may be extended to include the behaviour of
animals) it also expresses the feling that the change has entered from
the outside, that one has been 'flled . . . with stong courage' (Iiad,
1 3.60). Today such events are either forgotten or regarded as purely
accidental.
61
'But for Homer
6
or for early thought i general, there is
no such thing as accident.' 2 Every event is accounted for. This
makes the events clearer, strengthens their objectve features,
moulds them into the shape of known gods and demons and thus
turs them into powerful evidence for the divine apparatus that is
57. Snell, Ccammcltc5chn/, .1 8.
58. Cf Dodds, 1hcCrccHaadthcInatieaal,Boston, I 957, Chapter I .
59. With some effor this experience can be repeated even today. Step 1 : lie dow,
close your eyes, and attend to your hypnagogic hallucinatons. Step 2: perit the
hallucinatons to proeed on their own and according to their own tendencies. They
will then change from events in front of the eyes into events that gradually suround the
viewer but without yet making him an actve participant of an acton in a three
dimensional dream-space. Step 3: switch over from ricmiagthe hallucinator event to
boagpanof a complex of real events which act on the viewer and can be acted upon by
him. Step 3 can be revered either by the act of an almost non-exstent will or by an
outside noise. The three-dimensional scener becomes two-dimensional, rns
together into an area in front of the eyes, and moves away. It would be interestng to see
how such /~alelements change from culture to culture.
60. Today we say that somebody is 'overcome' by emotons and he may feel his
anger as an alien thing that invades him against his will. The daemonic ontolog of the
Greeks contains objectve terminolog for describing this feature of our emotons aad
thocqstmilucit.
6 I . Psychoanalysis and related ideologies now again contbute to making such
events pan of a wider context and thereby lend them substantality.
62. Dodds, op. cit., p. 6.
S I XTEEN 183
used for exlaining them: 'The gods are present. To recogne this a
a given fact for the Greeks is the first conditon for comprehending
their
religion and their culture. Our knowledge of their presence
rests upon an (inner or outer) experience of either the Gods
them
selves or of an acton of the Gods. '
63
To sum up: the archaic world is much less compact than the world
that surrounds us, and it is also experienced as being less compact.
Archaic man lacks 'physical' unity, his 'body' consists of a multtude
of parts, limbs, surfaces, connectons; and he lacks 'mental' unity, his
'mind' is composed of a variety of events, some of them not even
' mental' in our sense, which either inhabit the body-puppet a
additonal consttuents or are brought into it from the outside. Events
are not shaped by the individual, they are complex arrangements of
parts into which the body-puppet is insered at the appropriate
place.
64
This is the world-view that emerges from an analysis of the
fral features of 'archaic' art and Homeric poet, taken in
conjunction with an analysis of the conct which the Homeric poet
used for describing what he perceives. Its main features are
63. Wilamowit-Mollendorf, Do Claabc d Hcll, 1 , 1 955, p. 1 7. Our
conceptons of the world subdivide an otherise unifor material and create
differences in perceived brightess where objectve brightess has no gradient. The
same proess is responsible for the ordering of the rather chaotc impressions of our
inner life, leading to an (inner) percepton of divine interference, and it may even
intouce daemons, gos, sprites into the domain of outer perceptons. At any rate -
there is a sufcient number of daemonic experiences not to reject this conjecture out
of hand.
64. This means that success is not the result of an effort on the part of the
individual but the forunate ftting together of circumstances. This shows itself even in
words like :Qdnnv, which seem to designate actnitic. In Homer such words
emphasize not so much the effect of the agent as the fact that the result comes about in
the right way, that the proess that brings it about does not encounter to many
disturbances; it fts into the other proesses that surround it (in the Attic dialect
Eu:anw stll means 'I am doing well'). Similarly ttuxnv emphasizes
not so much a personal achievement as the fact that things go well, that they ft into
their
suroundings. The same is tre of the acquisiton of knowledge. 'Odysseus has
seen a lot and experienced much, moreover, he is the :ol. ufixavo who can
always help himself in new ways, and, fnally, he is the man who listens to his goddess
Athena. The part ofknowledge that is based on seeing is not really the result of his own
actvity and research, it rather happened to him while he was driven around by exteral
circumstances. He is ver different from Solon who, as Herodotus tells us, was the first
to tavel for theoretcal reaons, because he wa interested in research. In Odysseus
e knowledge of many things is stangely separated from his actvity in the field of the
E(oao6m: this actvity is resticted to fnding means for reaching a cerain
aim
, in
order to save his life and the life of his assoiates.' B. Snell, DicmtCncch
aadir,Gottngen, 1962, p. 48. In this place also a more detailed analysis of pertnent
ters. Cf. also fotote 56 on the apparent non-exstence of personal decisions.
184 AGAI NST METHOD
eereced by the individuals using the concepts. Tese indiiul lie
inded in the same kind ofJrld that is constred b their arits.
Further evidence for the conjecture can be obtained from an
examinaton of 'meta-atttudes' such as general religious atttudes
and 'theories' of (atttudes to) knowledge.
For the lack of compactess just described reappears in the field of
ideology. There is a toleance in religious matter which later
generatons found morally and theoretcally unacceptable and which
even today is regarded as a manifestaton of frivolous and simple
minds.
65
Archaic man is a religious eclectc, he dos not object to
foreign gods and myths, he adds them to the exstng furture of the
world without any attempt at synthesis, or a removal of contadic
tons. There are no priests, there is no dogma, there are no
categorical statements about the gods, humans, the world.
6
(This
tolerance can stll be found with the Ionian philosopher of nature
who develop their ideas side by side with myth without tying to
eliminate the latter.) There i no religious 'moralit' in our sense, nor
are the gods abstact embodiments of eteral principles.
67
This they
became later, during the archaic age and as a result they 'lost [their]
humanit. Hence Olympianism in its moralized for tended to
become a religion of fear, a tendency which is reflected in the
religious voabulary. There is no word for "god-fearing" in the
Iia.'
6
8 This is how life was dehumanized by what some people are
pleased to call 'moral progress' or 'scientfic progress'.
Similar remarks apply to the 'theory of knowledge' that is implcit
in this early world view. The Muses in Ila, 2. 284f, have knowledge
because they are cose to things -they do not have to rely on rmours
and because they know all the many things that are of interest to the
writer, one after the other. 'Quantt, not intensit is Homer's
standard of judgement' and of knowledge,
69
as becomes clear
from such words as JolqQWV and Jolic, 'much
pondering' and 'much thinking', as well as from later critcisms such
as 'Learin of many things [JoAUJa9(f] does not teach
intelligence'. 0 A interest in, and a wish to undertand, many
amaing thing (such as earthquakes, eclipses of the sun and the
moon, the paradoxcal rising and falling of the Nile), each of them
65. Example: F. Schacherayer, Dic / hcKmsikdCncm,Stutgart, 1 966.
66. Cf. Wilamowt-MoUendorf, op. cit.
67. M.P. Nilssn, ZHuteqe/CrcckRcligi,Oxord, 1 949, p. 1 52.
68. Dods, op. cit., p. 35.
69. Snell, 1hcDumpe/thcMiad, . l8.
70. Heraclitus, afer Diogenes Laerus, IX, I.
SI XTEEN 185
explained in its own partcular way and without the use of universal
principles, persists in the coastal descriptons of the 8th and 7th (and
later) centuries (which simply enumerate the tribes, tibal habits, and
coa
stal formatons that are successively met during the jourey), and
even a thinker such as Thales is satisfed with making many
interestng observatons and providing many explanatons without
trying to te them together in a system.
71
(The frst thinker to
construct a 'system' was Aaxmander, who followed Hesiod.)
Knowledge so conceived is not obtained by trying to grasp an essence
behind the reports of the senses, but by (1) puttng the observer in the
right positon relatve to the object (process, aggregate), by insertng
him into the appropriate place in the complex patter that consttutes
the world, and (2) by adding up the elements which are noted under
these circumstances. It is the result of a complex survey carried out
from suitable vantage points. One may doubt a vague report, or a
ffth-hand account, but it is not possible to doubt what one can clearly
see with one's own eyes. The objec depicted or described is the
proper arrangements of the elements which may include foreshort
enings and other perspectoid phenomena.
7
2
The fact that an oar
looks broken in water lacks here the sceptcal force it assumes in
another ideology.
73
Just as Achilles sittng does not make us doubt
that he is swift-footed - as a matter of fact, we would start doubtng
his swiftess if it tured out that he is in principle incapable of sittng
- in the very same way the bent oar does not make us doubt that it is
perfectly straight in air - as a matter of fact, we would start doubtng
its straightess if it did not look bent in water.
7
4
The bent oar is not
71. The idea that Thales used a principle expressing an underlying unity of
natural phenomena and that he identifed this principle with water is frt found in
Aristotle, Mctaphsic,983b61 2 and 26f. A closer lok at this and other passages and
consultation of Herodotus suggests that he still belongs to the group of those thinkers
who deal with numerous extraordinar phenomena, without tying them together in a
system. Cf. the vivid presentation in F. Krafft, CcchichtcdNatamissscha/,l,
Freiburg, 1971 , Chapter 3.
72. Perspectoid phenomena are sometimes treated as if they were special
properties of the objects depicted. For example, a container of the Old Kingdom
(Ancient Egypt) has an indentaton on top, indicating prpective, but the indentaton
is presented as a feature of the object itself, Schafer, op. cit., p. 266. Some Greek
artists tr to fnd situatons where perpectve does not need to be considered. Thus
the peculiarity of the so-called red-fgure style that arises in about 530 BC 'dos not so
much consist in the fact that foreshortenings are drawn, but in the new and highly
varied
ways to circumvent them', E. Pfuhl, MalooaadZochaaagdCncch,Vol. I,
Munich, 1923, p. 378.
73.
Cf. the disussion in Chapter I of A.. Ayer's Feaa&tieas e/ Empincal
Kae
m/cq
c.The example wa familiar in antquity.
14.
This
is also the way in which J .L. Austn takes care of the case. Cf. 5scaad
186 AGAI NST METHOD
an apec that denies what another apec says about the nature of the
oar, it is a partcular par (situaton) of the real oar that is not only
compatible with its staightess, but that demands it: the objects of
knowledge are as additve as the visible lists of the archaic artst and
the situatons described by the archaic poet.
Nor is there any unifor concepton of knowledge.75 A great
variety of words is used for expressig what we today regard as
diferent form of knowledge, or as diferent ways of acquing
knowledge. ooc(a 76 means expertse i a cerai profession
(carpenter, singer, general, physician, charioteer, wrestler) icluding
the ars (where it praises the artst not as an outstandig creator but as
a master of his craft); ELOEVat, literally 'having seen', refers
to knowledge gaied from ispecton; 0V(TJ, especially i
the Iia, though often tanslated as 'listenig' or 'understanding', is
stonger, it contains the idea of following and obeying, one absorbs
something and acts i accordance with it (hearing may play an
imporant role). Ad so on. Many of these expressions entail a
receptve atttude on the par of the knower, he, as it were
7
acts out the
behaviour of the things around h, he follows them, 7 he act as
befts an entty that is insered at the place he ocupies.
To repeat and to conclude: the modes of representaton used
during the early archaic period in Greece are not just reflectons of
icompetence or of special artstc interests, they give a faithful
account of what are felt, seen, thought to be fundamental features of
the world of archaic man. This world is an open world. Its elements
are not fored or held together by an 'underlying substance', they
are not appearances from which this substance may be inferred with
difculty. They ocasionally coalesce to for assemblages. The
relaton of a single element to the assemblage to which it belongs is
like the relaton of a par to an aggregate of pars and not like the
relaton of a par to an overpowering whole. The partcular aggregate
called 'man' i visited, and occasionally inhabited by 'mental events'.
Sensibilia, New York, I 962. It is clear that problems such a the 'problem of the
exstence of theoretcal enttes' cannot arse under these circumstances either. Al
these problems are neatedbythe new approach that supereded the additve ideolog
of archaic and pre-archaic tmes.
75. B. Snell, Die Zusd~ cb /r d Bfd usm ia d wmteauch
Fhiks@hie,Berlin, I 924. A shor account is given in Snell, DiealtmCnedmumm ir,
pp. 4 I f. Cf. also von Frt, Fhihs@hieumsrhluho ZusdmckboD~eHt,Fwe, um
Zmteteh,Leipzig-ParsLondon, 1 938.
76. Only ocurence in Homer, Ilm, 15, 42, concering the ooa of a
carenter (an 'exper carenter' tanslates Lattmore).
77. Cf. 5ncII,Zusd~ cke,p. 50.
SI XTEEN 187
Such events may reside in him, they may also enter from the outside.
Like ever other object man is an exchange staton of influences
rather than a unique source of acton, an 'I' (Descartes' 'cogito' has
no point of attack in this world, and his argument cannot even start).
There is a great similarity between this view and Mach's cosmolog
except that the elements of the archaic world are recognizable
physical and mental shapes and events while Mach's elements are
more abstact, they are as yet unknown aim of research, not its obec.
In sum, the representatonal units of the archaic world view admit of a
realistc interpretaton, they express a coherent ontolog, and
Worfs obseratons apply.
At this point I interupt my argument in order to make some
comments which connect the preceding obseratons with the
problems of scientfic method.
1 . It may be objected that foreshortenings and other indicatons of
perspectve are such obvious features of our perceptual world that
they cannot have been absent from the perceptual world of the
Ancients. The archaic manner of presentaton is therefore incom
plete, and its realistc interpretaton incorrect.
Reply: Foreshortenings are not an obvious feature of our percep
tual world unless special attenton is drawn to them (in an age of
photography and f this is rather frequently the case). Unless we
are professional photographers, flm-makers, painters we perceive
thing, not apecs. Moving swftly among complex objects we notce
much less change than a percepton of aspects would permit.
Aspects, foreshortenings, if they enter our consciousness at all, are
usually suppressed just as after-images are suppressed when the
appropriate stage of perceptual development is completed
78
and
they are notced in special situatons only
?
9
In ancient Greece such
special situatons arose in the theate, for the frst-row viewers of the
impressive productons of Aeschylus and Agatharchos, and there is
ind
eed a school that ascribes to the theate a decisive influence on the
development of perspectve.
80
Besides, why should the perceptual
world of the ancient Greeks coincide with ours? It needs more
argument than reference to a non-exstent form of percepton to
consolidate the objecton.
78. Cf. fomote I 2f and tex of the present chapter.
79. Cf. foomote 13.
80. Cf. Par II of Hedwig Kenner, Dm1hcaoaaddRcalummiadCnechuch
Kaast,Vienna, l 954, especially pp. l 2l f.
188 AGAI NST METHOD
2. The procedure used for establishing the peculiarites of the
archaic cosmology has much in common with the method of an
anthropologist who examines the world-view of an associaton of
tribes. The diferences are due to the scarcity of the evidence and to
the partcular circumstances of its origin (written sources; works of
art; no personal contact). Let us take a closer look at this procedure!
A anthropologist trying to discover the cosmology of his chosen
tibe and the way in which it is mirrored in language, in the ars, in
daily life, first lears the language and the basic social habits; he
inquires how they are related to other activites, including such prma
fae unimportant actvites as milking cows and cooking meals;
81
he
tries to identf key ideas.
82
His attenton to minutae is not the
result of a misguided urge for completeness but of the realizaton that
what looks insignificant to one way of thinking (and perceiving) may
play a most important role in another. (The diferences between the
paper-and-pencil operatons of a Lorentzian and those of an
Einsteinian are ofen minute, if discerible at all; yet they reflect a
major clash of ideologies.)
Having found the key ideas the anthropologist tries to undtand
them. This he does in the same way in which he originally gained an
understanding of his own language, including the language of the
special profession that provides him with an income. He intealie
the ideas so that their connectons are firmly engraved in his memor
and his reactons, and can be produced at will. 'The natve society has
to be in the anthropologist himself and not merely in his notebooks if
he is to understand it.'
83
This process must be ket fee fom eteal
intererece. For example, the researcher must not ty to get a better
hold on the ideas of the tribe by likening them to ideas he already
knows, or finds more comprehensible or more precise. On no
account must he attempt a 'logical reconstructon'. Such a procedure
would te him to the known, or to what is preferred by certain groups,
and would forever prevent him from grasping the unknown world
view he is examining.
Having completed his study, the anthropologist carries within
himself both the natve society and his own background, and he may
now start comparing the two. The comparison decides whether the
natve way of thinking can be reproduced in European terms
(provided there is a unique set of' European terms'), or whether it has
a 'logic' of its own, not found in any Wester language. In the course
81 . Evans-Pritchard, 5mmZathpel@,New York, 1 965, p. 80.
82. ibid., p. 80.
83. ibid., p. 82.
SI XTEEN 189
of the comparison the anthropologist may rephrase certain natve
ideas in English. This does not mean that English a spoke
indedtl ofthe comparson already contains natve ideas. It means
that langages can be bet in many directons and that understanding
does
not depend on any partcular set of rules.
3. The examinaton of key ideas passes through various stages,
none
of which leads to a complete clarificaton. Here the researcher
must exercise firm contol over his urge for instant clarit and logical
perfecton. He must never t to make a concept clearer than is
suggested by the material (except as a temporary aid for further
research). It is this material and not his logical intuiton that
determines the content of the concepts. To take an example. The
Nuer, a Nilotc tibe which has been examined by Evans-Pritchard,
have some interestng spato-temporal concepts. 84 The researcher
who is not too familiar with Nuer thought wl find the concepts
'unclear and insufciently precise'. To improve matters he might t
explicatng them, using modem logical notons. That might create
clear concepts, but they would no longer be Nuer concepts. If, on the
other hand, he wants to get concepts which are both clear and Nuer,
then he must keep his key notons vage and incomplete until the rght
infration come along, i.e. untl field study turs up the missing
elements which, taken by themselves, are just as unclear as the
elements he has already found.
Each item of informaton is a building blok of understanding,
which means that it has to be clarified by the discovery of further
bloks from the language and ideology of the tribe rather than by
premature definitons. Statements such as ' . . . the Nuer . . . cannot
speak of tme as though it was something actual, which passes, can be
waited for, can be saved, and so forth. I do not think that they ever
experience the same feeling of fightng against tme, or of having to
co-ordinate actvites with an abstract passage of tme, because their
points of reference are mainly the actvites themselves, which are
generally of a leisurely character e e '85 are either building blocks -in
this case their own content is incomplete and not fully understood -
or
they are preliminary attempts to antcipate the arrangement of the
tota
lity of all bloks. They must then be tested, and elucidated by the
discover of further blocks rather than by logical clarificatons (a
child lears the meaning of a word not by logical clarificaton but by
84.
Evans-
Pritchard, 1hcNue, Oxford, 1940, Par Ill; cf. also the brief account in
5em
l Zathr@ele,pp. 1 02f.
85.
Oc
Nue, p. 103.
190 AGAI NST METHOD
realizing how it goes together with things and other words). Lack of
clarity of any partcular anthropological statement reflects the
scarcity of the material rather than the vagueness of the logical
intuitons of the anthropologist, or of his tibe.
4. Exactly the same remarks apply to any attempt to explore
important modem notons such as the noton ofincommensurability.
Within the sciences incommensurability is closely connected with
meaning. A study of incommensurability in the sciences will
therefore produce statements that contain meaning-terms - but
these ters will be only incompletely understood, just as the ter
'tme' is incompletely understood in the quotaton of the preceding
paragraph. Thus the remark that such statements should be made
only afe producton of a clear theor of meaning86 is as sensible as
the remark that statements about Nuer tme, which are the material
that lea to an understanding ofNuer tme, should be written down
only after such an understanding has been achieved.
5. Logicians are liable to object. They point out that an examina
ton of meanings and of the relaton between terms is the task of logc,
not of anthropology. Now by 'logic' one may mean at least two
different things. 'Logic' may mean the study of, or results of the study
of, the stuctures inherent in a certain type of discourse. And it may
mean a partcular logical system, or set of systems.
A study of the frst kind belongs to anthropology. For in order to
see, for example, whether AB v AB = A is part of the 'logic of
quantum theor' we shall have to study quantum theor. And as
quantum theor is not a divine emanaton but a human product, we
shall have to study it in the for in which human products usually are
available, that is, we shall have to study historical records -textbooks,
original papers, records of meetngs and private conversatons,
letters, and the like. (In the case of quantum theor our positon is
improved by the fact that the tibe of quantum theoretcians has not
yet died out. Thus we can supplement historical study with
anthropological feld work such as the work of Kuhn and his
collaborators. 87)
86. Achinstein, Miaaceta5tadic iathcFhiles@he/5occ,Vol. 4, Minneaplis,
1 970, p. 224, says that 'Feyerabend owe[s] us a theor of meaning' and Hempel is
prepared to accept incommensurability only a/othe noton of meaning involved in it
has been made clear, op. cit., p. 156.
87. Repor in T.S. Kuhn, J.L. Heilbron, P. Foran and L. Allen, 5earcc/rthc
Huteqe/QmatamFsic,Amercan Philosphical Soiety, Philadelphia, 1967. The
materal asembled under the programme desrbed in this repr can b consulted at
various univerites, the University of Califoria in Berkeley among them.
S I XTEEN 191
It is to be admitted that these records do not, by themselves,
produce a unique soluton to our problems. But who has ever
assumed that they do? Historical records do not produce a unique
soluton for historical problems either, and yet nobody suggests that
they be neglected. There is no doubt that the records are necesar for
a logical study in the sense examined now. The queston is how they
should be used.
We want to discover the stucture of the feld of discourse, of
which the records give an incomplete account. We want to lear
about it without changing it in any way. In our example we are not
interested in whether a peraed quantum mechanics of the fture
employs AB v A
B
= A or whether an inetion of our own, whether a
little bit of 'reconstucton' which changes the theor so that it
conforms to some preconceived principles of moder logic and
readily provides the answer employs that principle. We want to know
whether quantum theor a aauall praaised b physicists employs the
principle. For it is the work of the physicists and not the work of the
reconstuctonists we want to examine. And this work may well be fll
of contadictons and lacunae. Its 'logic' (in the sense in which I am
now using the term) may well be 'illogical' when judged from the
point of view of a partcular system of formal logic.
Puttng the queston in this way we realize that it may not admit of
any answer. There may not exst a single theor, one 'quantum
theor', that is used in the same way by all physicists. The diference
between Bohr, Dirac, Feynman and von Neumann suggests that this
is more than a distant possibility. To test the possibility, i.e. to either
eliminate it or to give it shape, we must examne concrete cases. Such
an examinaton of concrete cases may then lead to the result that
quantum theoretcians difer from each other as widely as do
Catholics and the various types of Protestants: they may use the same
texts (though even that is doubtul - just compare Dirac with von
Neumann), but they sure are doing diferent things wth them.
The need for anthropological case studies in a feld that initally
seemed to be dominated by a single myth, always the same, always
used in the same manner, indicates that our common knowledge of
science may be severely defectve. It may be entrely mistaken (some
mistakes have been hinted at in the preceding chapters). In these
circ
umstances, the only safe way is to confess ignorance, to abandon
reconstuctons, and to start studying science from scratch. We must
approach science like an anthropologist approaches the mental
contortons of the medicine-men of a newly discovered associaton of
tibes. And we must be prepared for the discover that these
contortons are wildly illogical (when judged from the point of view of
192 AGAI NST METHOD
a partcular system of formal logic) and have to be wildly illogical in
order to functon as they do.
6. Only a few philosophers of science interret 'logic' in this
sense, however. Only few philosophers are prepared to concede that
the basic structures that underlie some newly discovered idiom might
differ radically from the basic structures of the more familiar systems
of formal logic and absolutely nobody is ready to admit that this might
be tue of science as well. Most of the tme the 'logic' (in the sense
discussed so far) of a partcular language, or of a theory, is
imediately identfed with the features of a partcular logical system
without considering the need for an inquiry concering the adequacy
of such an identfcaton. Professor Giedymin, for example, 88 means
by 'logic' a favourite system of his which is fairly comprehensive, but
by no means all-embracing. (For example, it does not contain, nor
could it be used to formulate, Hegel's ideas. And there have been
mathematcians who have doubted that it can be used for expressing
informal mathematcs.) A logical study of science as Giedymin and
his fellow logicians understand it is a study of sets of formulae of this
system, of their structure, the propertes of their ultmate con
sttuents (intension, extension, etc.), of their consequences and of
possible models. If this study does not repeat the features an
anthropologist has found in, say, science then this either shows that
science has some faults, or that the anthropologist does not know any
logic. It does not make the slightest difference to the logician in this
second sense that his formulae d not look like scientfc statements,
that they are not used like scientfc statements and that science could
not possibly grow in the simple ways his brain is capable of
understanding (and therefore regards as the only permissible ways).
He either does not notce the discrepancy or he regards it as being
due to imperfectons that cannot enter a satsfactory account. Not
once does it occur to him that the 'imperfectons' might have a
positve funaion, and that scientfc progress might be impossible
once they are removed. For him science is axomatcs plus model
theory plus correspondence rules plus obseraton language.
Such a procedure assumes (without notcing that there is an
assumpton involved) that an anthropological study which familiar
izes us with the overt and the hidden classifcatons of science has
been completed, and that it has decided in favour of the axomatc
(etc. , etc.) approach. No such study has ever been carried out. And
the bits and pieces of field work available today, mainly as the result of
88. ntuh eamal/r thc Fhiles@ e/5oc, August 1 970, pp. 257f and
Febrar 1 971 , pp. 39f.
S I XTEEN 193
the work of Hanson, Kuhn, Lakatos and the numerous historians
who remained untouched by positvistc prejudices, show that the
logician's approach removes not just some inessental embroideries
of science, but those ver features which make scientfc progress and
thereby science possible.
7. The discussions of meaning I have alluded to are another
illustaton of the defciencies of the logician's approach. For
Giedymin, this term and its derivatves, such a the term 'incommen
surabilit', are 'unclear and insufciently precise'. I agree. Giedymin
wants to make the terms clearer, he wants to understand them better.
Again agreement. He ties to obtain the clarit he feels is lacking by
explicaton in terms of a partcular kind of formal logic and of the
double language model, restictng the discussion to 'intension' and
'extension' as explained in the chosen logic. It is here that the
disagreement starts. For the queston is not how 'meaning' and
'incommensurabilit' occur within a partcular logical system. The
queston is what role they play in (actual, i.e. non-reconstucted)
science. Clarifcaton must come from a more detailed study of this
role, and lacunae must be flled with the results of such study. And as
the flling takes tme the key terms will be 'unclear and insufciently
precise' for years and perhaps decades. (See also items 3 and 4
above.)
8. Logicians and philosophers of science do not see the situaton
in this way. Being both unwilling and unable to carr out an informal
discussion, they demand that the main terms of the discussion be
'clarified'. And to 'clarif' the terms of a discussion does not mean to
study the ad itional and as yet unknown propertes of the domain in
queston which one needs to make them fully understood, it means to
fll them with eisting notons from the entrely diferent domain of
logic and common sense, preferably obseratonal ideas, untl they
sound common themselves, and to take care that the proess of flling
obeys the accepted laws of logic. The discussion is permitted to
p
roeed only afe its inital steps have been modifed in this manner.
So
the course of an investgaton is deflected into the narrow
channels of things already understood and the possibilit of fun
damental conceptual discover (or of fundamental conceptual
change) is considerably reduced. Fundamental conceptual change,
on the other hand, presupposes new world-views and new languages
capable of expressing them. Now, building a new world-view, and a
corresponding new language, is a process that takes tme, in science
as well as in meta-science. The terms of the new language become
clear only when the process is fairly advanced, so that each single
word is the cente of numerous lines connectng it with other words,
194 AGAI NST METHOD
sentences, bits of reasoning, gestures which sound absurd at frst but
which become perfectly reasonable once the connectons are made.
Argments, theories, terms, points of view and debates can therefore
be clarifed in at least two diferent ways: (a) in the manner already
described, which leads back to the familiar ideas and teats the new a
a special case of things already understood, and (b) by incorporaton
into a langage of the future, which means that one must lear to are
with unelained ters and to use seteces fr which no cear rle ofusage
are a yet aailable. Just as a child who starts using words without yet
understanding them, who adds more and more uncomprehended
lingistc fragments to his playful actvity, discovers the sense-giving
principle only after he has been actve in this way for a long tme-the
actvity being a necessar presuppositon of the fnal blossoming forth
of sense -in the ver same way the inventor of a new world-view (and
the philosopher of science who tries to understand his proedure)
must be able to talk nonsense untl the amount of nonsense created
by him and his friends is big enough to give sense to all its parts.
There is again no better account of this process than the descripton
which John Stuar Mill has left us of the vicissitudes of his educaton.
Referring to the explanatons which his father gave him on logical
matters, he wrote: 'The explanatons did not make the matter at all
clear to me at the tme; but they were not therefore useless; they
remained as a nucleus for my obseratons and reflectons to
crstallise upon; the impor of his general remarks being interpreted
to me, by the partcular instances which came under my notce
afteard.'
89
Building a new langage (for understanding the world,
or knowledge) is a process of exactly the same kind ecet that the
inital 'nuclei' are not given, but must be invented. We see here how
essental it is to lear talking in riddles, and how disastrous an efect
the drive for instant clarity must have on our understanding. (In
additon, such a drive betrays a rather narrow and barbaric mentality:
'To use words and phrases in an easy going way without scrutnizing
them too curiously is not, in general, a mark of il breeding; on the
contrar, there is something low bred in being too precise . . . .'
All these remarks are rather tivial and can be illustated
89. There is much more randomness in this proess than a rationalist would ever
perit, or suspect, or even notice. Cf. von Kleist, 'Uber die allmahliche Verfertgun
g
der Gedanken beim Reden', inMmtmokcDmuchoLitoatarkntik,ed. Hans Meyer,
Stutgar, I 96Z,pp. 7+I-7.Hegel had an inkling of the situaton. Cf. K. Lowith and]
.
Riedel (eds), Hqcl, 5tadiaasgmcI.Frankfr, I 968,p. 5+.For Mill cf. Chapter
I I ,
footote I 3.
9. Plato, 1hcaitctes, I 8+c.Cf. also I. Dnng, Zmtetclc,Heidelberg, l 966,p.
379,
critcizing Aristotle's demand for instant precision.
SI XTEEN 195
by
obvious examples. Classical logic arrived on the scene only
when there was sufcient argumentatve material (in mathematcs,
rhe
toric, politcs) to sere as a startng point and a a testng ground.
Arithmetc developed without any clear understanding of the
concept of number; such understanding arose only when there
existed a sufcient amount of arithmetcal 'facts' to give it substance.
In the same way a proper theory of meaning (and of incommen
surabilit) can arise only after a sufcient number of'facts' has been
assembled to make such a theory more than an exercise in concept
pushing. This is the reason for the examples in the present secton.
9. There is stll another dogma to be considered before returng
to the main narraton. It is the dogma that all subjects, however
assembled, quite automatcally obey the laws of logic, or ought to
obey the laws of logic. If this is so, then anthropological feld work
would seem to be superfluous. 'What is tue in logic is tue in
psycholog . . . in scientfc method, and in the history of science,'
writes Popper.
91
This dogmatc asserton is neither clear nor is it (in one of its main
interretatons) true. To start with, assume that the expressions
'psycholog', 'history of science', 'anthropolog' refer to certain
domains of facts and regularites (of nature, of percepton, of the
human mind, of societ). Then the asserton is not cear as there is not
a single subject - LOGIC - that underlies all these domains. There
is Hegel, there is Brouwer, there are the many logical systems
considered by modem constuctvists. They ofer not just diferent
interretatons of one and the same bulk of logical 'facts', but
different 'facts' altogether. And the asserton is not tre as there exst
legitmate scientfc statements which violate simple logical rules. For
example, there are statements which play an important role in
established scientfc disciplines and which are obseratonally
adequate only if they are self-contradictory: fte a moving patter
that has just come to a standstll, and you will see it move in the
opposite directon, but without changing its positon. The only
phenomenologically adequate descripton is 'it moves, in space, but it
does
not change place' -and this descripton is self-contadictory.
92
9I
O
qcuncKaemlcdgc, Oxford, I 97Z,p. 6. Antcipated e.g. by Comte, Ceanc,SZO
Lecon
and, of course, Aristotle.
9Z. It has been objected (Ayer, G.E.L. Owen) that we are dealing with appearances,
n
?
t with actual event, and that the corect description is 'it appears to move . . . . ' But the
dif
culty
remains. For if we intoduce the 'appear', we must put it at the beginning of the
se
nten
ce,
which will read 'it appears that it moves and does not change place'. And a
ap

earan
ces belong to the domain of phenomenological psycholog we have made our
pmt, v
z. that this domain contains self-inconsistent elements.
196 AGAI NST METHOD
There are examples from geomet:
93
thus the enclosed figure
(which need not appear in the same way to ever person) is seen as an
isosceles tiangle whose base is not halved by the perpendicular. And
there are examples with a b & b c & a l c as the only
phenomenologically adequate descripton.
94
Moreover, there is not
a single science, or other for oflife that is usefl, progressive as well
as in agreement with logical demands. Ever science contains
theories which are inconsistent both with fact and with other
theories and which reveal contadictons when analysed in detail.
Only a dogmatc belief in the principles of an allegedly unifor
discipline 'Logic' will make us disregard this situaton. And the
objecton that logical principles and principles of, say, arithmetc
difer from empirical principles by not being accessible to the method
of conjecture and reftatons (or, for that matter, any other
'empirical' method) has been defused by more recent research in this
field.
95
Secondly, let us assume that the exressions 'psychology', 'anthro
pology', 'histor of science', 'physics' do not refer to fact and laws,
but to certain method of assembling facts including certain ways
of connectng obseraton with theor and hyothesis. That is, let
us conider the aiit ' science' and it various subdivisions. Then
we may lay down idal dmand of knowledge and knowledge
acquisiton, and we may t to constuct a (soial) machiner that
obeys these demands. Almost all epistemologists and philosophers of
science proeed in this way. Occasionally they succeed in fding a
93. E. Rubin, 'Visual Figres Apparendy Incompatble with Geomet', Zm
Fqchelepm,VII, 1 950, pp. 365f. Cf also the drawings on pages 1 667.
94. E. Trnekjaer-Rasmussen, 'Perpectoid Distances', Zm Fqchegim, X,
1 955, p. 297.
95. Mainly by the work of lmre Lakatos, 'Profs and Refutaton', ntuhjea~ /
r
theFhihs@e/5o, 1 962/63.
S I XTEEN 197
machinery that might work in certain ideal conditons, but they never
inquire, or even fnd it worth inquiring, whether the conditons are
satsfed in this real world of ours. Such an inquiry, on the other
hand, will have to explore the way in which scientsts auall deal
with their surroundings, it will have to examine the actual shape of
their product, viz. 'knowledge', and the way in which this product
changes as a result of decisions and actons in complex soial and
material conditons. In a word, such an inquiry will have to be
an
thropological.
There is no way of predictng what an anthropological inquiry will
bring to light. In the preceding chapters, which are rough sketches of
an anthropological study of partcular episodes, it has emerged that
science is full of lacunae and contadictons, that ignorance,
pigheadedness, reliance on prejudice, lying, far from impeding the
forard march of knowledge may actually aid it and that the
traditonal virtues of precision, consistency, 'honesty', respect for
facts, maximum knowledge under given circumstances, if practsed
with determinaton, may bring it to a standstll. It has also emerged
that logical principles not only play a much smaller role in the
(argumentatve and non-argumentatve) moves that advance science,
but that the attempt to enforce them would seriously impede science.
(One cannot say that von Neumann has advanced the quantum
theory. But he certainly made the discussion of its basis more long
winded and cumbersome.
9
6
)
Now a scientst engaged in a certain piece of research has not yet
completed all the steps that lead to defnite results. His future is stll
open. Will he follow the barren and illiterate logician who preaches to
him about the virtues of clarity, consistency, experimental support (or
experimental falsifcaton), tghtess of argument, 'honesty', and so
on, or will he imitate his predecessors in his own feld who advanced
by breaking most of the rules logicians want to lay on him? Will he rely
on abstract injunctons or on the results of a study of concrete
episodes? I think the answer is clear and with it the relevance of
anthropological feld work not just for the anthropologists but also for
the members of the societes he examines. I now contnue my
narraton and proceed to describing the tansiton from the paratactc
universe of the archaic Greeks to the substance-appearance universe
of their followers.
.
96. Besides, the imprecisions which he removes from the foralism now reappear
In
the
relat
on beteen theor and fact. Here the corespndence principle stU reigs
sup
reme. Cf. fotote 25 of Chapter 5.
198 AGAI NST METHOD
The archaic cosmology (which from now on I shall call cosmology A
contains things, events, their pars; it does not contain appearances. 9
Complete knowledge of an object is complete enumeraton ofits pars
and peculiarites. Humans cannot have complete knowledge. There
are too many things, too many events, too many situatons (Iia,
2.488), and they can be close to only a few of them (Ii, 2.485). But
although humans cannot have complete knowledge, they can have a
sizeable amount of it. The wider their experience, the greater the
number of adventures, of things seen, heard, read, the greater their
knowledge. 98
The new cosmolog (cosmology B) that arises in the 7th to 5th
centuries BC distngu!shes between much-knowing, 3OAUJa8(T,
and tue knowledge,99 and it wars against tustn
o
'custom
bor of manifold experience', eo 3OAU3ELQOV. 1 Such a
distncton and such a warg make sense only in a world whose
stcture difers from the stucture of A. In one version which played a
large role in the development of Wester civton and which
underlies such problems a the problem of the exstence of theoretcal
enttes and the problem of alienaton the new events for what one
mght call a Tr Worl, while the events of everday life are now
apearance that are but its dim and msleading reflecton. 101 The
Tre World is simple and coherent, and it can be described in a
unifor way. So can ever act by which its elements are
comprehended: a few abstact notons replace the numerous concepts
that were used in cosmolog A for describing how humans mght be
'insered' into their surroundings and for expressing the equally
numerous types ofinformaton thus gained. From now on there is only
one imporant type of informaton, and that is: kowledge.
97. Snell, Zmdm ke, p. 28 (referring to Homer), speaks of a 'knowledge that
proeeds from appearances and draws their multtude together in a unit which is ten
posited as their te essence'. T may apply to the Presoratcs, it dos not apply to
Homer. In the case of Homer 'the world is comprehended as the sum of things, visible
in space, and not as reaon actng intensively' (ibid., p. 67, disussing Empedokles;
cf
also the lines following the quotaton for a furher elaboraton of the theme).
98. Snell, Diemt Cnemamir,p. 48.
99. Cf Heraclitus, fr. 4 (DielsKran).
10. Paenides, fr. 7, 3. 'Here for the frt tme sense and reason are contasted';
W.K. Guthrie, ZHuteqe/CreekFhiles@,Vol. II, Cambridge, I 965, p. 25.
I 01. This distncton i characteristc of cerin mythological views as well. Hom
er
thus differ both from the preceeding mythologies and from the succeeding
philosophies. His point of view is of great originality. In the 20 centur J .L. Autn
has developed similar ideas. And he has critcized the development from Thales via
Plato to the present essentalism. Cf. the ft chapter of 5seam5sibilm.Chapter 3
of Farm eht Remeacontains details.
S I XTEEN 199
The conceptual totalitarianism that arises as a result of the slow
arri
val of world B has interestng consequences, not all of them
desirable. Situatons which made sense when ted to a parcular type
of
cogniton now become isolated, unreasonable, apparendy inconsis
ten
t with other situatons: we have a 'chaos of appearances'. The
'chaos' is a direct consequence of the simplificaton oflanguage that
acco
mpanies the belief in a True World.
1
0
2
Moreover, all the
manifold abilites of the obserers are now directed towards this True
World, they are adapted to a unir aim, shaped for one pariclar
purose, they become more similar to each other which means that
humans become impoverished toger1er with their language. They
become impoverished at precisely the moment they discover an
autonomous 'I' and proceed to what some have been pleased to call a
'more advanced noton of God' (allegedly found in Xenophanes),
which is a noton of God lacking the rich variety of typically human
features.
1
03
'Mental' events which before were teated in analo
with events of the body and which wee eerenced acrdingl
1
become more 'subjectve', they become modificatons, actons,
revelatons of a spontaneous soul: the distncton between appearance
(first impression, mere opinion) and reality (te knowledge) spreads
everwhere. Even the task of the arst now consists in arranging his
shapes in such a manner that the underlying essence can be grasped
with ease. In paintng this leads to the development of what one can
only call systematc methods for deceiving the eye: the archaic arst
teats the surface on which he paints as a writer might teat a piece of
papyrus; it i a real surface, it is supposed to be see as a real surface
(though attenton is not always directed to it) and the marks he draws
on it are comparable to the lines of a blueprint or the letters of a word.
They are symbols that infr the reader of the straureoftheobjea, ofits
parts, of the way in which the parts are related to each other. The
simple drawing overleaf, for example, may represent three paths
meetng at a point. The artst using perspectve on the other hand,
regards the surface and the marks he puts on it as stimuli that tigger the
ilusion of an arrangement of three-dimensional objects. The illusion
occurs because the human mind is capable of producing illusor
experiences when properly stmulated. The drawing is now seen
102. Snell, Zmdm ,pp. 80f; von Frt, Fhil@hicaadqrhlmoZmdmbo
D
mHt, FmteawZmteulc,Leipzig-ParisLondon, 1938, p. I I .
103 . . . in becoming te emboiment of cosmic justce Zeus lot h humanit.
Hence Olympianism in its moralized for tended to become a religion of fear . . . ',
Dds, CrccH,p. 35. For Xenophanes cf. Chapter 2 of Farmcll teRcmea.
1
0. Snell, Diae, p. 69.
20 AGAI NST METHOD
either a the comer of a cube that extends towards the viewer, or as the
comer of a cube that points away from him (and is seen from below), or
else as a plane floatng above the surface of the paper carrying a two
dimensional drawing of three paths meetng.
Combining this new way of seeing with the new concept of
knowledge that has just been described, we obtain new enttes, v.
physical objects a they are understood by most contemporary
philosophers. To explain, let me agin take the case of the oar.
In the archaic view 'the oar' is a complex consistng of parts some of
which are objects, some situatons, some events. It is possible to say
'the staight oar is broken' (not 'appears to be broken') just as it is
possible to say 'swift-footed Achilles is walking slowly', for the
elements are not set against each other. They are part of a paratactc
aggregate. Just as a taveller explores all parts of a stange count and
describes them in a 'periegesis' that enumerates its peculiarites, one
by one, in the same way the student of simple objects such a oars,
boats, horses, people inserts himself into the 'major oar-situatons',
apprehends them in the appropriate way, and reports them in a list of
propertes, events, relatons. And just as a detailed periegesis exhausts
what can be said about a count, in the same way a detailed list
exhausts what can be said about an object.
105
'Broken in water'
belongs to the oar as does 'staight to the hand'; it is 'equally real'. In
cosmology B, however, 'broken in water' is a 'semblance' that
contraias what is suggested by the 'semblance' of staightess and
thus shows the basic untustworthiness of all semblances.
106
The
concept of an object has changed from the concept of an aggregate of
equi-imporant perceptble parts to the concept of an imperceptble
I 05. The idea that knowledge consists in lucreaches back far into the Sumeran
past. Cf. von Soden, LmaagaadCrm5ammch-@leauchom mcha/, new
edn, Darstadt, I 965. The difference between Babylonian and Greek mathematcs
and astonomy lies precisely in this. The one develops methods for the presentaton of
what we toay caD 'phenomena' and which were interestng and relevant events in the
sk, while the other ties to develop astonomy, 'while leaving the heavens alone'
(lato, Re., 530bf; Lg., SI Sa).
I 06. Xenophanes, fr. 34.
S I XTEEN 201
esse
nce underlying a multtude of deceptve phenomena. (We may
guess that the appearance of an object has changed in a similar way,
that objects now look less 'flat' than before.)
Considering these changes and peculiarites, it is plausible to
assume that the comparison of A and Ba interreted b the paricpants
(rather than as 'reconstructed' by logically well-trained but otherwise
illi
terate outsiders) will raise various problems. In the remainder of
this chapter only some aspects of some of these problems will be
discussed. Thus I shall barely menton the psychological changes that
accompany the tansiton from A to Band which are not just a matter of
conjecture, but can be established by independent research. Here is
rich material for the detailed study of the role of frameworks (mental
sets, languages, modes of representaton) and the limits of ratonalism.
To start with, cosmos A and cosmos B are built from diferent
elements.
The elements of A are relatvely independent parts of objects which
enter into exteral relatons. They partcipate in aggregates without
changing their intinsic propertes. The 'nature' of a partcular
aggregate is determined by its parts and by the way in which the parts
are related to each other. Enumerate the pars in the prer ordr, and you
hae the objea. This applies to physical aggregates, to humans (minds
and bodies), to animals, but it also applies to social aggregates such a&
the honour of a warrior.
The elements of B fall into two classes: essences (objects) and
appearances (of objects - what follows is tue only of some rather
streamlined versions of B). Objects (events, etc.) may again combine.
They may for haronious totalites where each part gives meaning
to the whole and receives meaning from it (an extreme case is
Parmenides where isolated parts are not only unrecognizable, but
altogether unthinkable). Aspects properly combined do not produce
objeas, but psychological conditons for the apprehension of phantoms
which are but other aspects, and partcularly misleading ones at that
(they look so convincing). No enumeration ofapeas i idnticl with the
objea (problem of inducton).
The transiton from A to B thus intoduces new enttes and new
relatons between enttes (this is seen ver clearly in paintng and
statu
ar). It also changes the concept and the self-experience of
humans. An archaic individual is an assemblage oflimbs, connectons,
trunk, neck, head, 107 (s)he is a puppet set in moton by outside forces
107. 'To be precise, Homer does not even have any words for the ars and the
legs; he
speaks of hands, lower ars, uppr ars, feet, calves, and thigs. Nor is there
a
com
prehensive ter for the tnk.' Snell, Dice, Chapter I , fotote 7.
202 AGAI NST METHOD
such as enemies, social circumstances, feelings (which are described
and perceived as objectve agencies - see above):
10
8
'Man is an open
target of a reat many forces which impinge on him, and penetate his
very core. '
0
9
He is an exchange staton of material and spiritual, but
always objectve, causes. Ad this is not just a 'theoretcal' idea, it is a
social fact. Man is not only dscrbed in this way, he is picured in this way,
and he fel himself to be consttuted in this manner. He does not
possess a cental agency of acton, a spontaneous 'I' that produces its
own ideas, feelings, intentons, and difers from behaviour, social
situatons, 'mental' events of type A. Such an I is neither mentoned
nor is it notced. It is nowhere to be found within A. But it plays a very
decisive role within B. Indeed, it is not implausible to assume that some
outstandiny;eculiarites ofB such as aspects, semblances, ambiguit
of feeling
1
enter the stage as a result of a sieable inceae ofsel
cnsciousnes.
Now one might be inclined to explain the tansiton as follows:
archaic man has a limited cosmology; he discovered some things, he
missed others. His universe lack important objects, his language lack
important concepts, his percepton lack important stctures. Add
the missing elements to cosmos A, the missing terms to language A, the
missing stctures to the perceptual world of A, and you obtain cosmos
B, language B, percepton B.
Some tme ago I called the theory underlying such an explanaton
the 'hole theory' or the 'Swiss cheese theory' oflanguage (and other
means of representaton). According to the hole theory every
cosmology (every language, every mode of percepton) has sizeable
lacunae which can be flled, leaing eerthing ele unchanged. The hole
theory is beset by numerous difcultes. In the present case there is the
difcult that cosmos B does not contain a single element of cosmos A.
Neither common-sense terms, nor philosophical theories; neither
paintg and statuary, nor artstc conceptons; neither religion, nor
theological speculaton contain a single element of A once the
tansiton to B has been completed. This is a historclfa
!
1
1
Is this
1 08. 'Emotons do not spring spontaneously from man, but are bestowed on him
by the gods,' ibid., p. 52. See also the account earlier in the present chapter.
1 09. ibid., p. 20.
1 10. Cf. Sappho's 'bitter-sweet Eros', ibid, p. 60.
I l l . The fact is not eay to establish. Many presentatons of A, including some
ver detailed and sophistcated ones, are infected by B-concept. A example is
quoted in footote 97 to the present chapter. Here a elsewhere only te
anthropological method can lead to knowledge that is more than a reflecton of wishfl
thinking. A similar situaton in the coure of individual development is described in the
text to fn. 1 2 of the present chapter.
S I XTEEN 203
fact an accident, or has A some stuctural propertes that prevent the
co-exstence of A-situatons and B-situatons? Let us see!
I have already mentoned an example that might give us an inkling of
a reason as to why B does not have room for A-facts: the drawing
below may be the intersecton of three paths as presented in
accordance with the principles of A-pictures (which are visual lists).
Perspectve having been intoduced (either as an objectve method or
as a mental set), it can no longer be seen in this manner. Instead oflies
on paper we have the illusion of depth and a three-dimensional
panorama, though of a rather simple kind. There is no way of
incororatng the A-picture into the B-picture except as par of this
illusion. But an illusion of a visual list is not a visual list.
The situaton becomes more tansparent when we tum to concepts.
I have said above that the 'nature' of an object (=aggregate) in A is
determined by the elements of the aggregate and the relaton between
the elements. One should add that t determinaton is 'closed' in the
sense that the elements and their relatons cnstitute the object; when
they are given, then the object is given as well. For example, the
'elements' described by Odysseus in his speech in Ilia, 9.225f
constitute honour, grace, respect. A-concepts are thus very similar to
notons such as 'checkmate' : given a certai arrangement of pieces on
the board, there is no way of 'discovering' that the game can stll be
contnued. Such a 'discovery' would not fla gap, it would not add to
our knowledge of possible chess positons, it would put an end to the
game. And so would the 'discovery' of'real meanings' behind other
moves and other constellatons.
Precisely the same remarks apply to the 'discovery' of an individual
I that is diferent from faces, behaviour, objectve 'mental states' of
the type that occur in A, to the 'discovery' of a substance behind
'appearances' (forerly elements of A), or to the 'discovery' that
honour may be lacking despite the presence of al its outer
manifestatons. A statement such as Heraclitus' 'you could not fnd
the limits of the soul though you are tavelling every way, so deep is its
logos' (Diels, B 45) does not just ad to cosmos A, it undcts
204 AGAI NST METHOD
the principles which are needed in the constucton of A-type
'mental states' while Heraclitus' rejecton of JOAUJaSCT and
Parmenides' rejecton of an eo JOA tJELQOV undercuts
rules that gover the constructon of ee single fa of A. An
entre world-view, an entre universe of thought, speech, percepton
is dissolved.
It is interestng to see how this process of dissolving manifests itself
in partcular cases. In his long speech in Ilia, 9 .308f, Achilles wants
to say that honour may be absent even though all its outer
manifestatons are present. The terms of the language he uses are so
intmately ted to defnite soial situatons that he 'has no language to
express his disillusionment. Yet he expresses it, and in a remarkable
way. He does it by misusing the language he disposes of. He ak
questons that cannot be answered and makes demands that cannot
be met.'112 He acts in a most 'irratonal' way.
The same irratonality is found in the writngs of all other early
authors. Compared with A the Presoratcs speak stangely indeed.
So do the lyrical poets who explore the new possibilites of seltood
they have 'discovered'. Freed from the fetters of a well-constucted
and unambiguous mode of expression and thinking, the elements of
A lose their familiar functon and start floatng around aimlessly-the
'chaos of sensatons' arises. Freed from frm and unambiguous soial
situatons feelings become fleetng, ambivalent, contadictory: 'I
love, and I love not; I rave, and I do not rave,' writes Anakreon. 11
3
Freed from the rules of late geometc paintng the artsts produce
stange mixtures of perspectve and blueprint. 114 Separated from
well-determined psychological sets and freed of their realistc
import, concepts may now be used 'hypothetcally' without any
odium of lyng and the arts may begin exploring possible worlds in an
imaginatve way. 115 This is the same 'step back' which was earlier
I 1 2. A. Par, 'The Language of Achilles', 1raas. C Fmc.Zmo. Fhil. Zssec. ,87,
I 956, p. 6. Cf. the discussion of the case in FarmcllteRcmea,Chapter I 0.
I 13. Dich|,ZathelegiaLnca,fr. 79.
I 14. Pfuhl, op. cit., cf. also J. White, FmcunciaZaotDramiagaadFaiatiag,
London, 1 965.
I I 5. Plutarch reports the following stor in his Lqe/5elea:'When the company of
Thespis began to exhibit tagedy, and it novelty was attactng the ppulace but had
not yet got as far as public compettons, Solon, being fond of listening and learng
and being rather given in his old age to leisure and amusement, and indeed to drinking
pares and music, went to see Thespis act in his ow play, a wa te practce in
ancient tmes. Solon approached him afer the prforance and aked h if he w
not ahamed to tell so many lies to so many pople. When Thespis sid there wa
nothing dreadfl in representng such works and actons in fn, Solon stuck the
ground violently with his walking stck: "If we aplaud these things in fn," he sid,
S I XTEEN 205
seen to be a necessary presuppositon of change and, possibly,
progress
1 16-
only it now does not just discard obseratons, it discards
some imporant standards of ratonality as well. Seen from A (and also
from the point of view of some later ideologies) all these thkers,
poets, artsts, are raving maniacs.
Remember the circumstances which are responsible for this
situaton. We have a point of view (theory, framework, cosmos, mode
of representaton) whose elements (concepts, 'facts', pictures) are
built up in accordance with certain principles of constcton. The
principles involve something like a 'closure': there ar things that
cannot be said, or 'discovered', without violatng the principles
(which does not mean contadictng them). Say the things, make the
discovery, and the principles are suspended. Now take those
constctve principles that underlie every element of the cosmos (of
the theory), every fact (every concept). Let us call such principles
unieal prncple of the theory in queston. Suspending universal
principles means suspending all fact and all concepts. Finally, let us
call a discovery, or a statement, or an atttude incommeurabl with
the cosmos (the theory, the framework) if it suspends some of its
universal principles. Heraclitus B 45 is incommensurable with the
psychological part of A: it suspends the rules that are needed for
consttutng individuals and puts an end to all A-facts about
individuals (phenomena corresponding to such facts may of course
persist for a considerable tme a not all conceptual changes lead to
changes in percepton and a there exst conceptual changes that
never leave a tace in the appearances; however, such phenomena
can no longer be dcbed in the customary way and cannot therefore
count a obseratons of the customary 'objectve facts').
Note the tentatve and vague nature of this explanaton of
'incommensurable' and the absence of logical terminology. The
reason for the vagueness has already been explained (items 3 and 4
above). The absence of logic is due to the fact that we deal with
phenomena outside of it domain. My purpose is to fnd terminology
"we shaU son fnd ourelves honourng them in earest".' The stor seems
historcally impossible yet elucidates a widespread atttude (for this atttude cf.
Chapter 8 of John Fordyke Crcc bqrcHo, New York, 1964). Solon himself
seems to have been somewhat les impressed by taditonal for of thought and he
may
have
been one of the frst drmatc actor (of the politcal varety): G. Else, Dc
Ongia aadEarFe~e/1ragc@, Cambrdge, 1 965, pp. 4f. The opposite atttude,
which reveals the secure and already somewhat conceited citen of B, is expresed by
Simonides who answered the queston why the Thessalian were not deceied by him
b saying 'Because they ae to stupid'. Plutrch, Dcamp.,l SD.
1 1 6. Chapter I I , text to fotote 5.
206 AGAI NST METHOD
for describing certain complex historical-anthropological phenomena
which are only imperfectly understood rather than defning
propertes oflogical systems that are specifed in detail. Terms, such
as 'universal principles' and 'suspend', are supposed to summarize
anthropological informaton much in the same way in which Evans
Pritchard's account of Nuer tme (text to footote 85) summarizes
the anthropological informaton at his disposal (cf. also the brief
discussion in item 3 above). The vageness of the explanaton
reflects the incompleteness and complexty of the material and invites
artculaton by further research. The explanation has to have some
content - otherwise it would be useless. But it must not have too much
content, or else we have to revise it ever second line.
Note, also, that by a 'principle' I do not simply mean a stateent
such as 'concepts apply when a fnite number of conditons is
satsfed', or 'knowledge is enumeraton of discrete elements which
form paratactc aggregates' but the gammatical habit corresponding
to the statement. The two statements just quoted describe the habit
of regarding an object as given when the list of its parts has been fully
presented. This habit is suspended (though not contadicted) by the
conjeaure that even the most complete list does not exhaust an object;
it is alo suspended (but again not contadicted) by any unceasing
search for new aspects and new propertes. (It is therefore not
feasible to defne 'incommensurability' by reference to state
ments. 1 1 7) If the habit is suspended, then A -objects are suspended
with it: one cannot examine A-objects by a method of conjectures
and refutatons that knows no end.
How is the 'irratonality' of the tansiton period overcome? It is
overcome in the usual way (cf. item 8 above), i.e. by the determined
producton of nonsense untl the material produced is rich enough to
permit the rebels to reveal, and everone else to recognize, new
universal principles. (Such revealing need not consist in writng the
principles down in the form of clear and precise statements.)
Madness turs into sanity provided it is suffciently rich and
sufciently reglar to functon as the basis of a new world-view. And
when that happens, then we have a new problem: how can the old
view be compared with the new view?
From what has been said it is obvious that we cannot compare the
contents of A and B. A-facts and B-facts cannot be put side by side,
not even in memor: presentng B-facts means suspending principles
1 1 7. This takes care of a critcism in fomote 63 of Shapere's aricle in Miadam
Cesmes,Pittsbur, 1 966. The classifcations achieved by the principles are 'cover'
in
the sense ofWhorf: cf. above, fomote 4 and text dow to fomote 9.
S I XTEEN 207
assumed in the constucton of A-facts. All we can do is draw B
pictures of A-facts in B, or intoduce B-statements of A-facts into B.
We cannot use A-statements of A-facts in B. Nor is it possible to
translate language A into language B. This does not mean that we
cannot discss the two views - but the discussion wllead to sizeable
changes of both views (and of the languages in which they are
exp
ressed).
Now it seems to me that the relaton between, say, classical
mechanics (interpreted realistcally) and quantum mechanics (inter
preted in accordance with the views of Niels Bohr), or between
Newtonian mechanics (interreted realistcally) and the general
theor of relatvity (also interpreted realistcally) is in many respects
similar to the relaton between cosmolog A and cosmolog B. Thus
ever fact of Newton's mechanics presumes that shapes, masses,
periods are changed only by physical interactons and this
presumpton is suspended by the theor of relatvity. Similarly the
quantum theor consttutes facts in accordance with the uncertainty
relatons which are suspended by the classical approach.
At this point it is imporant to interpret the situaton in a
sensible manner of else scientfc (cultural) change becomes an
inexplicable miracle. The idea that comprehensive ways of tg,
actng, perceiving such as cosmolog A (and, in a much more narrow
domain, classical physics) and cosmolog B (relatvity or quantum
mechanics) are closed frameworks with fed rules creates an
unbridgeable gulf between situatons which, though diferent in
surprising ways, are yet connected by arguments, allusions,
borrowings, analogies, general principles of the kind exlained in the
text above. Logicians who confne the ter 'argument' to chains of
reasoning involving stable and precise concepts and who reconstuct
theories and world-views using equally precise and unambiguous
terms are forced to call such connectons 'irratonal' while their
opponents can report the 'discover' that science, that alleged
stonghold of reason, often violates reason in a decisive way. Both are
talking about chimaeras, not about science and culture as they really
are. Things change when we use scientfc practce or cultural reality
and not logic as our informants, in other words, when we engage in
sociological research, not in reconstucton. We then discover
that scientfc concepts (and concepts, shapes, percepts, styles in
general) are ambiguous in the sense that decisive events can afect
their appearance, their perceived implicatons and, with them, the
'log
ic'
they obey. Achilles (see the text to footote 1 1 2 above)
'
msus
es' the language he has at his disposal by assertng a diference
be
twe
en
'real' honour and its social manifestatons. Assertng
208 AGAI NST METHOD
diferences is not in confict with view A; for example, there i s a great
diference between the knowledge, the power, the actions of the gods
on the one side and the knowledge, the power and the actons of
humans on the other. Assuming that honour is in the hands of gods
who don't give a damn about the aspiratons of humans devalues the
social manifestatons of honour, makes them secondar. The
assumpton fts well into the general outlines of view A but Achilles is
the frst to make it. Wy? Because his anger, his sufering makes him
see connectons which, because of a widespread optmism, are not
part of the general views about honour and do not contibute to its
'defniton'. He seems to violate basic soial rules but viewed with the
anxety caused by Agamemnon's actons such rules give way to a
diferent idea that is regarded as being implicit in the exstng material
but as not having surfaced so far. Generalizing, we can say that
concepts have potentalites over and above the usages that seem to
defne them; it is this feature that makes them capable of connectng
entrely diferent conceptual systems. More about this in my (I
promise!) last bok, Te Conquet ofAbundnce.
Appendi 2
Whorf speaks of 'Ideas', not of 'events' or of 'facts', and it is not
always clear whether he would approve of my extension of his views.
On the one hand he says that 'tme, velocity, and matter are not
essental to the constructon of a consistent picture of the universe, 1
and he asserts that 'we cut up nature, organize it into concepts, and
ascribe signifcances as we do, largely because we are partal to an
agreement to organize it in this way', 2 which would seem to imply
that widely diferent languages posit not just diferent ideas for the
ordering of the same facts, but that they posit also diferent facts. The
'linguistc relatvity principle' seems to point in the same directon. It
says, 'in informal terms, that users of markedly diferent grammars
are pointed by their grammars towards diferent tpes of obseratons
and diferent evaluatons of exterally similar acts of obseraton, and
hence are not equivalent obserers, but must arrive at somewhat
diferent views of the world'.
3
But the 'more formal statements'
4
of
the principle already contains a diferent element, for here we are
told that 'all obserers are not led by the same physical eidce to the
same picture of the universe, unless their linguistc backgrounds are
similar, or can in some way be calibrated',
5
which can either mean
that obserers using widely diferent languages will posit di eret faas
under the same physical circumstances in the same physical world, or
it can mean that they will arange similar faas in di eret ways. The
second interpretaton fnds some support in the examples given,
where diferent isolates of meaning in English and Shawnee are said
to be 'used in reportng the same eerece'
6
and where we read that
'languages classif items of experience diferently';
7
experience is
l . Wor, op. cit., p. 216.
2.
ibid., p. 213.
3.
ibid., p. 221.
4. ibid., p. 221.
5.
ibid., p. 21 4, my italics.
6.
ibid., p. 208.
7.
ibid., p. 209.
209
210 AGAI NST METHOD
regarded as a uniform reseroir offacts which are cassied diferently
by diferent languages. It fnds further support in Whorls
descripton of the tansiton from the horr-vai account of
barometric phenomena to the modem theor: 'If once these
sentences [Why does water rise in a pump? Because Nature abhors a
vacuum.] seemed satsfing to logic, but today seem idiosyncrasies of
a partcular jargon, the change did not come about because science
has discovered new facts. Science has adopted new linguistc
formulatons of the old facts, and now that we have become at home
in the new dialect, certain traits of the old one are no longer binding
on us'.
8
However, I regard these more conseratve statements as
secondar when compared with the great infuence ascribed to
grammatcal categories and especially to the more hidden 'rapport
systems' of a language.
9
Whorf and those who follow him regard language as the main and
perhaps as the only 'shaper of events'. That" is much too narrow a
point of view. Animals have no langage in the sense ofWhorf, yet
they do not live in a shapeless world. Planets, at least as conceived
today, are not even alive, but they afect their surroundings and react
to them in a lawful manner. In humans rituals, music, the arts,
adaptve behaviour that occurs without the interpositon of words
make important contibutons to the way in which the world apear
and, to those living accordingly, i. In the sciences we have not only
statements (the old idea that science is a system of statements has by
now been thoroughly discredited), but obseratons, experimental
equipment, an intuitve relaton between obserers and their
equipment that has to be leared in a practcal way and cannot be
written down, the work of experimentalists which has much in
common with the work of artsts - what they want are not merely
results, but results that emerge in a simple, compelling and
aesthetcally pleasing way - and so on. A concentaton on language
alone, or on 'texts', can easily lead into absurdit, as is shown
by Austn and by the practce of deconstucton: on the one
hand philosophers produce texts, like poets; on the other hand
they take it for granted that their texts reveal a realit beyond the
thoughts, impressions, memories, fgres of speech, etc., etc. from
which they arose. (Scientfc realists to a certain extent share in t
predicament.)
Finally, some comments on what I think about incommensurabilit
and how I arrived at the idea.
8. ibid., p. 222.
9. ibid., pp. 68f.
APPENDI X 2 211
I think that incommensurability tur up when we sharen our
concepts in the manner demanded by the logcal positvists and their
ofspring and that it undine their ideas on explanaton, reducton
and progress. Incommensurability disapear when we use concepts
as scientsts use them, in an open, ambiguous and often counter
intuitve manner. Incommensurability is a problem for philosophers
not for scientsts, though the latter may become pschologcall
confsed by unusual things. I arrived at the phenomenon while
studying the early literature on basic statements and by considering
the possibility of perceptons radically diferent from our own. In my
thesis
10
I exned the meaning of obseratonal statements. I
considered the idea that such statements describe 'what is gven' and
tied to identf this 'gven'. Pheomeologcall this did not seem to be
possible; we notce objects, their propertes, their relatons, not 'the
given'. It is of course true that we can give quick repors on the
propertes of everday objects but this does not change them into
non-objects but only shows that we have a special relaton to them.
Phenomenologically what is given consists of the same things which
can also exst unobsered - it is not a new kind of object. Special
arrangements such as the reducton screen intoduce new con
ditons, they do not reveal ingredient in objects we already know.
Result: the given cannot be isolated by obseraton.
The second possibility was to isolate it by logical means: what is
given can be ascertained with ceraint, hence I obtain the given
contained in the table before me by removing from the statement
'there is a table' all the consequences that make fture correctons
possible. This shows that the given is the result of an unreasonable
decision: untestable statements cannot sere as a basis for science.
Following this argument I intoduced the assumpton that the
meaning of obseraton statement depends on the nature of the
objects described and, as this nature depends on the most advanced
theories, on the content of these theories. Or as I formulated it in my
first English paper on the topic: the interretaton of an obseraton
language is determined by the theories which we use to explain what
we obsere, and it changes as soon as these theories change.
1 1
In a
word: obseraton statements are not just theor-/a (the views of
Toulmin, Hanson and apparently also Kuhn) but full theoreticl
and the distncton between obseraton statements ('protocol
l 0. Vienna, 195 1 wten afer to year of extensive discussion in the Kraf
Circ
le and
superised by Professor Victor Kraft of the Univerity of Vienna.
I I . '
A
Atempt at a Realistc Interretaton of Exrence', Fm.Zmt.5w. 1958,
rep
rnted
in Fhihs@himFapm,Vol. I . Te pasage (in italics) i on p. 31 .
212 AGAI NST METHOD
statements' in the terminology of the Vienna Circle) and theoretcal
statements is a pragmatc distncton, not a semantc distncton;
there are no special 'obseratonal meanings'. Thus in the same year
as Hanson (Hanson'sPaters ofDiscuer appeared in 1 958) and four
years before Kuhn I formulated a thesis a weaker form of which
became ver popular later on. Moreover, my thesis not only was
stonger than the thesis of theor-ladenness, it also came from a
diferent source. For while Toulmin and Hanson were inspired by
Wittgenstein's Philosohicl betigations I started from and retured
to ideas that had been developed in the Vienna Circle - and I said
so.
1 2
Quine, whose philosophy shows close connectons to the
philosophy of the Vienna Circle,
13
also used a criterion of
obserability that is rather similar to mine.
14
Now when Feigl heard of these ideas he pointed out that
interpretng obseratons in terms of the theories they are
obseratons of makes nonsense of crucial experiments; for how can
an experiment decide between two theories when its interpretaton
already depends on these theories and when the theories themselves
have no common elements, such as a common obseraton language?
In the paper just mentoned and in 'Explanaton, Reducton and
Empiricism', published in 1 962, I took up the challenge. I frst
increased it by constuctng cases where important terms of one
theor cannot in any way be defned in another which, moreover, tes
to do its job. My example which I found in Anneliese Maier's Die
Vor/i ufe Gali/eis im 14. Jahrund was the relaton of the terms
'impetus' and 'momentum'. I also developed a theor of test to
answer the challenge. In 1 962 I called theories such a those
containing 'impetus' and 'momentum' incommensurable theories,
said that only a special class of theories, so-called non-instantal
theories could be (but need not be) incommensurable and added that
successive incommensurable theories are related to each other by
replacement, not by subsumpton. The year 1 962 is also that of
Kuhn's great book-but Kuhn used a diferent approach to apply the
same term to a similar (not an identcal) situaton. His approach was
historical, while mine was abstact.
In 1 960 I started the studies described in chapters 8, 9 and 1 6.
They revealed tht percepton and experimentaton obey laws of their
I 2. Fhiles@himlFam,Vol. I, pp. 49, I 25.
13. Details in Dirk Koppelberg, Dic Za)bcbaag d Zaatuch Fhiles@hic,
Frankfurt, 1 987.
14. Fhiles@himlFam,Vol. l, pp. l7f.
APPENDI X 2 213
own which cannot be reduced to theoretcal assumptons and are
therefore beyond the grasp of theory-bound epistemologies.
I also joined Kuhn in demanding a historical as opposed to
an epistemological grounding of science but I stll difer from
him by opposing the politcal autonomy of science. Apar from
that our views (i.e. my published views and Kuhn's as yet
unpublished recent philosophy) by now seem to be almost
identcal,
1
5 except that I have little sympathy for Kuhn's atempt to
te up history with philosophical or linguistc, but at any rte with
theoretcal ropes: a connecton with theory just brings us back to what
I at least want to escape from - the rigid, though chimaerical
(deconstucton!) boundaries of a 'conceptual system'.
15. Cf. my 'Realism and te Historcit of Knowledge', De eu~e/Philoh,
Vol.
lx
, 1 989, pp. 353f, esp. fotote 26 ad te ptpt to te prsnt essy.
1 7
Nether sciece nor rationalit are unieal measure ofecellece. Te are
pariclar traditions, unaware ofther historical grounding.
So far I have tied to show that reason, at least in the form in which it
is defended by logicians, philosophers and some scientsts, does not
ft science and could not have contributed to its growth. This is a
good argument against those who admire science and are also slaves
of reason. They must now make a choice. They can keep science;
they can keep reason; they cannot keep both.
But science is not sacrosanct. The mere fact that it exsts, is
admired, has results is not sufcient for makg it a measure of
excellence. Modem science arose from global objectons against
earlier views and ratonalism itself, the idea that there are general
rules and standards for conductng our affairs, affairs of knowledge
included, arose from global objectons to common sense (example:
Xenophanes against Homer). Are we to refrain from engaging in
those actvites that gave rise to science and ratonalism in the frst
place? Are we to rest content with their results? Are we to assume that
everg that happened after Newton (or afer Hilbert) is
perfecton? Or shall we admit that modem science may have basic
faults and may be in need of global change? And, having made the
admission, how shall we proceed? How shall we localize faults and
carr out changes? Don't we need a measure that is independent of
science and conflicts with it in order to prepare the change we want to
bring about? And will not the rejecton of rules and standards that
conflict with science forever prevent us from fnding such a measure?
On the other hand - have not some of the case studies shown that a
blunt applicaton of 'ratonal' proedures would not have given us a
better science, or a better world but nothing at all? And how are we to
judge the results themselves? Obviously there is no simple way of
guiding a practce by rules or of critcizing standards of ratonality by
a practce.
The problems I have sketched are old ones and much more
21 4
S EVENTEEN 215
general than the problem of the relaton between science and
rationality. They occur whenever a rich, well-artculated and familiar
practce - a practce of composing, of paitng pictures, of stage
producton, of selecting people for public ofce, of keeping order and
punishing criminals, a practce of worship, of organizing society - is
confronted by a practce of a diferent kind that can interact with it.
The inteaions and their results depend on historical conditons and
var from one case to the next. A powerful tibe invadig a county
may impose its laws and change the indigenous taditons by force
only to be changed itself by the remnants of the subdued culture. A
ruler may decide, for reasons of convenience, to use a popular and
stabilizing religion as the basic ideology of his empire and may
thereby contribute to the transformation both of his empire and of
the religion chosen. An individual, repelled by the theatre of his tme
and in search of something better, may study foreign plays, ancient
and modem theories of drama and, using the actors of a friendly
company to put his ideas into practce, change the theate of a whole
naton. A group of painters, desirous of adding the reputaton of
being scientsts to their already enormous reputaton as skilled
crafsmen, may intoduce scientfc ingredients such as geometr
into paintng and thereby create a new style and new problems for
painters, sculptors, architects. A astronomer, critcal of the
diference between classical principles of astronomy and the exstng
practce and desirous to restore astronomy to its former splendour,
may fnd a way to achieve his aim and so initate the removal of the
classical principles themselves.
In all these cases we have a practce, or a traditon, we have certain
influences upon it, emerging from another practce or taditon and
we obsere a change. The change may lead to a slight modifcaton of
the original practce, it may eliminate it, it may result in a traditon
that barely resembles either of the interactng elements.
Interactons such as those just described are accompanied by
changing degrees of awarees on part of the partcipants. Copericus
knew very well what he wanted and so did Constantne the Great (I
am now speaking about the inital impulse, not about the
transformaton that followed). The intusion of geometr into
paintng is less easily accounted for in terms of awareness. We have
no
idea why Giotto tried to achieve a compromise between the
surface of the paintng and the cororeality of the things painted,
especially as pictures were not yet regarded as studies of a material
reality. We can surmise that Brunelleschi arrived at his constucton
by a natural extension of the architects' method of representng
three-dimensional objects and that his contacts with contemporar
216 AGAI NST METHOD
scientsts were not without consequence. I t is stll more diffcult to
understand the gradually rising claims of artsans to make
contributons to the same kind of knowledge whose principles were
explained at universites in very diferent terms. Here we have not a
critcal study of alteratve taditons as we have in Copercus, or in
Constantne, but an impression of the uselessness of academic science
when compared with the fascinatng consequences of the joureys of
Columbus, Magellan and their successors. There arose then the idea
of an 'America of Knowledge', of an entrely new and as yet
unforeseen contnent of knowledge that could be discovered, just as
the real America had been discovered: by a combinaton of skill and
abstact study. Marxst have been fond of confounding insufcient
inforaton concering the awareness that accompanies such
proesses with irrelevance and they have ascribed only a secondar
role to individual consciousness. In this they were right - but not in
the way they thought. For new ida, though often necessar, were not
sufcient for explaining the change that now ocurred and that
depended also on the (often unknown and unrealized) circmstance
under which the ideas were applied. Revolutons have tansfored
not only the practces their initators wanted to change but the ver
principles by means of which, intentonally or unintentonally, they
carried out the change.
Now considering any interacton of taditons we may ask two
kds of questons which I shall call obserer quetions and paricipant
quetions respectvely.
Obserer quetions are concered with the details of an interacton.
They want to give a historical account of the interacton and, perhaps,
forulate laws, or rules of thumb, that apply to all interactons.
Hegel's tiad: positon, negaton, synthesis (negaton of the negaton)
is such a rule.
Paricipant quetions deal with the atttude the members of a
practce or a taditon are supposed to take towards the (possible)
intusion of another. The obserer asks: what happens and what is
going to happen? The partcipant asks: what shall I do? Shall I
support the interacton? Shall I oppose it? Or shall I simply forget
about it?
In the case of the Copercan Revoluton, for example, the
obserer asks: what impact did Copercus have on Wittenberg
astonomers at about 1 560? How did they react to his work? Did they
change some of their beliefs and if so, why? Did their change of
opinion have an efect on other astonomers, or were they an isolated
group, not taken seriously by the rest of the profession?
The questons of a partcipant are: this is a stange book ideed -
S EVENTEEN 217
should I take it seriously? Should I study 1t m detail or only
superficially or should I simply contnue as before? The main theses
seem absurd at first sight - but, maybe, there is something in them?
How shall I fnd out? And so on.
It is clear that obserer questons must take the questons of the
partcipants into account and partcipants will also listen most
carefully (if they are inclined that way, that is) to what obserers have
to say on the matter - but the intenton is diferent in both cases.
Obserers want to know what is going on, partcipants what to do. An
obserer describes a life he does not lead (except accidentally), a
partcipant wants to arange his own life and asks himself what
atttude to take towards the things that may influence it.
Partcipant can be op orunists and act in a straightforard and
practcal way. In the late 1 6th century many princes became
Protestants because this furthered their interests and some of their
subjects became Protestants in order to be left in peace. Wen
Britsh colonial ofcials replaced the laws and habits of foreign tibes
and cultures by their own 'civilized' laws the latter were ofen
accepted because they were the laws of the king, or because one had
no way to oppose them, and not because of any intinsic excellence.
The source of their power and 'validity' was clearly understod, both
by the ofcials and by the more astute of their unforunate subjects.
In the sciences and especially in pure mathematcs one often pursues
a partcular line of research not because it is regarded as intinsically
perfect, but because one wants to see where it leads. I shall call the
philosophy underlying such an atttude of a partcipant a pragatic
philosohy.
A pragmatc philosophy can flourish only if the taditons to be
judged and the developments to be influenced are seen as temporar
makeshifts and not as lastng consttuents of thoughts and acton. A
partcipant with a pragmatc philosophy views practces and
traditons much as a taveller views foreign counties. Each count
has features he likes and things he abhors. In deciding to settle down
a taveller will have to compare climate, landscape, language,
temperament of the inhabitants, possibilites of change, privacy,
looks of male and female populaton, theate, opportunites for
advancement, quality of vices and so on. He wlalso remember that
his
inital demands and expectatons may not be very sensible and so
permit the proess of choice to afect and change his 'nature' a well
which
, after all is just another (and mor) practce or taditon
entering the proess. So a pragmatst must be both a partcipant and
an obserer even in those exteme cases where he decides to live in
accordance with his momentary whims entrely.
218 AGAI NST METHOD
Few individuals and groups are pragmatsts in the sense just
described and one can see why: it is ver difcult to see one's own
most cherished ideas in perspectve, as parts of a changing and,
perhaps, absurd taditon. Moreover this inability not only eists, it is
also ecuraged as an atttude proper to those engaged in the study and
the improvement of man, society, knowledge. Hardly any religon has
ever presented itself just as something worth trying. The claim is
much stonger: the religion is the truth, everything else is error and
those who know it, understand it but stll reject it are rotten to the
core (or hopeless idiots).
Two elements are contained in such a claim. First, one
distnguishes beteen taditons, practces and other results of
individual and/or collectve human actvity on the one side and a
diferent domain that may act on the taditons without being one.
Secondly, one explains the stucture of this special domain in detail.
Thus the word of God is powerful and must be obeyed not because
the taditon that carries it has much force, but because it is outside
all taditons and provides a way of improving them. The word of
God can start a taditon, its meaning can be handed on from one
generaton to the next, but it is itself outside all taditons.
The first element - the belief that some demands are 'objectve'
and taditon-independent - plays an important role in rationalism
which is a secularized form of the belief in the power of the word of
God. And this is how the oppositon reason-practce obtains its
polemical stng. For the to agencies are not seen as to practces
which, while perhaps of unequal value, are yet both imperfect and
changing human products but as one such product on the one side
and lastng measures of excellence on the other. Early Greek
ratonalism already contains this version of the conflict. Let us
examine what circumstances, assumptons, proedures - what
features of the historical process - are responsible for it!
To start with the taditons that oppose each other - Homeric
common sense and the various forms of ratonalism that arise in the
6th to 4th centuries - have di eret inteal straure.
1
On the one
hand we have complex ideas that cannot be easily explained, they
'work' but one does not know how, they are 'adequate', but one does
not know why, they apply in special circumstances only, are rich in
content but poor in similarites and, therefore, in deductve
connectons. On the other side there are relatvely clear and simple
concepts which, having just been intoduced, reveal a good deal of
1 . For details see Chapter 1 6.
S EVENTEEN 219
their stucture and which can be linked in many ways. They are poor
in content, but rich in deductve connectons. The diference
becomes especially stiking in the case of mathematcs. In geomet,
for example, we start with rules of thumb applying to physical object
and their shapes under a great variety of circumstances. Later on it
can be pred why a given rule applies to a given case -but the proofs
make use of new enttes that are nowhere found in nature.
In antquity the relaton between the new enttes and the familiar
world of common sense gave rise to various theories. One of them
which one might call Platonism assumes that the new enttes are real
while the enttes of common sense are but their imperfect copies.
Another theor, due to the Sohists, regards natural object as real
and the objects of mathematcs (the objects of 'reason') as
simpleminded and unrealistc images of them. These two theories
were also applied to the diference between the new and fairly
abstact idea of knowledge propagated by Plato (but found already
before) and the common-sense knowledge of the tme (Plato wisely
uses a distorted image of the latter to give substance to the forer).
Again it was either said that there exsted only one tue knowledge
and that human opinion was but a pale shadow ofit or human opinion
was regarded as the only substantal knowledge in exstence and the
abstact knowledge of the philosophers as a useless dream ('I can see
horses, Plato,' said Antsthenes, 'but I nowhere see your ideal
horse').
It would be interestng to follow this ancient conflict through
histor down to the present. One would then lear that the conflict
turs up in many places and has many shapes. Two examples must
sufce to illustate the great variety of its manifestatons.
When Gottsched wanted to reform the Geran theate he looked
for plays worth imitatng. That is, he looked for taditons more
orderly, more dignifed, more respectable than what he found on the
stage of his tme. He was attacted by the French theate and here
mainly by Coreille. Being convinced that 'such a complex edifce of
poety (as tagedy) could hardly exst without rules'
2
he loked for
the rules and found Aristotle. For him the rules of Aristotle were not
a partcular way of viewing the theate, they were the reason for
excellence where excellence was found and guides to improvement
where improvement seemed necessar. God theate was an
emb
odiment of the rules of Aristotle. Lessing gradually prepared a
diff
erent view. First he restored what he thought to be the real
2. 'Vorrede zum "Sterbenden Cato" ' quoted from). Chr. Gottsched, 5chn/zar
Litoo
tar,Stuttgan, 1972, p. 20.
220 AGAI NST METHOD
Aristotle as opposed to the Aristotle of Comeille and Gottsched.
Next he permitted violatons of the letter of Aristotle's rules provided
such violatons did not lose sight of their aim. And, fnally he
suggested a diferent paradigm and emphasized that a mind inventve
enough to constuct it need not be restricted by rules. If such a mind
succeeds in his eforts 'then let us forget the textbok!'
3
In a different (and much less interestng) domain we have the
oppositon between those who suggest that languages be constucted
and reconstructed in accordance with simple and clear rules and who
favourably compare such ida/ langage with the sloppy and opaque
natural idioms and other philosophers who assert that natural
languages, being adapted to a wide variety of circumstances, could
never be adequately replaced by their anaemic logical compettors.
This tendency to view diferences in the stucture of taditons
(complex and opaque vs simple and clear) as diferences in kind (real
vs imperfect realizaton of it) is reinforced by the fact that the critcs
of a practce take an obserer's positon with respect to it but remain
partcipants of the practce that provides them with their objectons.
Speaking the language and using the standards of this practce they
'discover' limitatons, faults, errors when all that really happens is
that the two practces - the one that is being critcized and the one
that does the critcizing-don't ft each other. Many armets against
an out-and-out matealim are of this kind. They notce that
materialism changes the use of 'mental' terms, they illustrate the
consequences of the change with amusing absurdites (thoughts
having weight and the like) and then they stop. The absurdites show
that materialism clashes with our usual ways of speaking about
minds, they do not show what is better - materialism or these ways.
But taking the partcipants' point of view with respect to common
sense turs the absurdites into arguments against materialism. It is
as if Americans were to object to foreign currencies because they
cannot be brought into simple relatons ( 1 : 1 or 1 : 1 0 or 1 : 1 00) to the
dollar.
The tendency to adopt a partcipant's view with respect to the
positon that does the judging and so to create an Archimedian point
for critcism is reinforced by certain distnctons that are the pride
and joy of armchair philosophers. I refer to the distncton between
an evaluaton and the fact that an evaluaton has been made, a
3. Hambaqo Dramataqc, Stck 48. Cf., however, Lessing's critcism of te
claims of the 'original geniuses' of his time in Sti ck 96. Lessing's account of the
relaton between 'reason' and practce is quite complex and in agreement with the view
developed furher below.
SEVENTEEN 221
proposal and the fact that the proposal has been accepted, and the
related distncton between subjectve wishes and objectve standards
of excellence. When speaking as obserers we often say that certain
groups accept certain standards, or think highly of these standards.
Speaking as partcipants we equally ofen use the standards without
any reference to their origin or to the wishes of those using them. We
say 'theories ought to be falsifable and contadicton free' and not 'I
want theories to be falsifable and contadicton free' or 'scientsts
become very unhappy unless their theories are falsifable and
contadicton free'. Now it is quite correct that statements of the frst
kind (proposals, rules, standards) (a) contain no reference to the
wishes of individual human beings or to the habits of a tbe and (b)
cannot be derived from, or contadicted by, statements concerg
such wishes, or habits, or any other facts. But that does not make
them 'objectve' and independent of taditons. To infer from the
absence of terms concerng subjects or groups in 'there ought to
be . . .' that the demand made is 'objectve' would be just as erroneous
as to claim 'objectvity' i.e. independence from personal or group
idiosyncrasies, for optcal illusions and mass hallucinatons on the
grounds that the subject, or the group, nowhere occurs in them.
There are many statements that are frulated 'objectvely', i.e.
without reference to taditons or practces, but are stll meant to be
undtood in relaton to a practce. Examples are dates, co-ordinates,
statements concering the value of a currency, statements of logic
(after the discovery of altertve logics), statements of geomet
(after the discovery of Non-Euclidean geometes) and so on. The
fact that the retort to 'you ought to do X' can be 'that's what you
think!' shows that the same is tue of value statements. And those
cases where the reply is not allowed can be easily rectfed by using
discoveries in value theory that correspond to the discovery of
alteratve geometes, or alteratve logical systems: we confront
'objectve' value judgement from diferent cultures or diferent
practces and ask the objectvist how he is going to resolve the
conflict.4 Reducton to shared principles is not always possible and
so we must admit that the demands or the formulae expressing them
4. In the play DcRaliagCms(later tured into a somewhat vapid fm with Peter
O'Tole) two madmen claiming to be God are confronted with each other. This
marellous idea so confuses the playwright that he uses fre and brmstone instead of
dialogue to get over the problem. His final soluton, however, is quite interestng. The
one
madman turs into a good, uprght, noral Brtsh Citen who plays Jack the
Ripper on the side. Did the playwright mean to say that our modem 'objectvist' who
have
been through the fre of relativism can retr to noralc only if they are
Peritted to annihilate all distrbing elements?
222 AGAI NST METHOD
are incomplete as used and have to be revised. Contnued insistence
on the 'objectvity' of value judgements however would be as illiterate
as contnued insistence on the 'absolute' use of the pair 'u)down'
after discovery of the spherical shape of the earth. And an argument
such as 'it is one thing to utter a demand and quite a different thing to
assert that a demand has been made - therefore a multplicity of
cultures does not mean relatvism' has much in common with the
argument that antpodes cannot exst because they would fall 'down'.
Both cases rest on antediluvian concepts (and inadequate distnc
tons). Small wonder our 'ratonalists' are fascinated by them.
With this we have also our answer to (b). It is tue that statng a
demand and describing a practce may be two different thing and
that logcal connectons cannot be established between them. This
does not mean that the interacton between demands and practces
cannot be teated and evaluated as an interacton of practces. For the
difference is due, first, to a difference between obserer-atttude and
partcipant-atttude: one side, the side defending the 'objectvity' of
its values, ue its taditon instead of eamining it - which does not
tum the taditon into an objectve measure of validity. And secondly,
the difference is due to concepts that have been adapted to such one
sidedness. The colonial ofcial who proclaims new laws and a new
order in the name of the king has a much better grasp of the situaton
than the ratonalist who recites the mere letter of the law without any
reference to the circumstances of its applicaton and who regards this
fatal incompleteness as proof of the 'objectvity' of the laws recited.
After this preparaton let us now look at what has been called 'the
relaton between reason and practce'.
Simplifing matters somewhat we can say that there exst three
views on the matter.
A. Reason guides practce. Its authority is independent of the
authority of practces and taditons and it shapes the practce in
accordance with its demands. This we may call the idalistic veion of
the relaton.
B. Reason receives both its content and its authority from practce.
It describes the way in which practce works and formulates its
underlying principles. This version has been called naturalim and it
has occasionally been attibuted to Hegel (though erroneously so).
Both idealism and naturalism have difcultes.
The difcultes of idealism are that the idealist does not only want
to 'act ratonally' he also wants his ratonal actons to have results.
And he wants these results to occur not only among the idealizatons
he uses but in the real world he inhabits. For example, he wants real
S EVENTEEN 223
human beings to build up and maintain the society of his dreams, he
wants to understand the motions and the nature of real stars and real
stones. Though he may advise us to 'put aside (all obseraton of) the
heavens'
5
and to concentrate on ideas only he eventually returs to
nature in order to see to what extent he has grasped its laws.
6
It then
often turs out and it often has tured out that actng ratonally in the
sense preferred by him does not produce the expected results. This
conflict between ratonality and expectations was one of the main
reasons for the constant reform of the canons of ratonality and much
encouraged naturalism.
But naturalism is not satsfactory either. Having chosen a popular
and successful practce the naturalist has the advantage of'being on
the right side', at least for the time being. But a practce may
deteriorate; or it may be popular for the wrong reasons. (Much of the
popularity of modem scientfic medicine is due to the fact that sick
people have nowhere else to go and that television, rumours, the
technical circus of well equipped hospitals convince them that they
could not possibly do better.) Basing standards on a practce and
leaving it at that may forever peretuate the shortcomings of this
practce.
The difficulties of naturalism and idealism have certain elements
in common. The inadequacy of standards often becomes clear from
the barrenness of the practce they engender, the shortcomings of
practces often are very obvious when practces based on diferent
standards flourish. This suggests that reason and practce are not two
different kinds of enttes but pars ofa single dialeaical proces.
The suggeston can be illustrated by the relaton between a map
and the adventures of a person using it or by the relaton between an
artsan and his instuments. Originally maps were constucted as
images of and guides to reality and so, presumably, was reason. But
maps like reason contain idealizations (Hecataeus of Miletus, for
example, imposed the general outlines of Anaxmander's cosmolog
on his account of the occupied world and represented contnents by
geometrical figures). The wanderer uses the map to fnd his way but
he also corrects it as he proceeds, removing old idealizatons and
intr
oducing new ones. Using the map no matter what will soon get
him into trouble. But it is better to have maps than to proceed without
them. In the same way, the example says, reason without the
guidance of a practce will lead us astray while a practce is vastly
improved by the additon of reason.
5. Plato, Rtublic, 530bf.
6
. Epinomis.
224 AGAI NST METHOD
This account, though better than naturalism and idealism and
much more realistic, is still not entrely satsfactor. It replaces a one
sided acton (of reason upon practce or practce upon reason) by an
interacton but it retains (certain aspects oO the old views of the
interactng agencies: reason and practce are stll regarded as enttes
of diferent kinds. They are both needed but reason can exst without
a practce and practce can exst without reason. Shall we accept this
account of the matter?
To answer the queston we need only remember that the
diference between 'reason' and something 'unreasonable' that must
be formed by it or can be used to put it in its place arose from turing
stuctural diferences of practces into diferences ofkind. Even the
most perfect standards or rules are not independent of the materal
on which they act (how else could they find a point of attack in it?) and
we would hardly understand them or know how to use them were
they not well-integrated parts of a rather complex and in places quite
opaque practce or taditon, viz. the language in which the desor
rationi exresses his stem commands. 7 On the other hand even the
most disorderly practce is not without its regulartes, as emerges
from our atttude towards non-partcipants.
8
Wat is caled 'reason '
and 'raaice' are therere two di eret tpe ofpraaice, the diference
being that the one clearly exhibits some simple and easily producible
formal aspects, thus making us forget the complex and hardly
understood propertes that guarantee the simplicity and producibility,
while the other drowns the formal aspects under a great variety of
accidental propertes. But complex and implicit reason is stll reason
and a practce with simple formal features hovering above a perasive
but unnotced background of linguistc habits is stll a practce.
Disregarding (or, rather, not even noticing) the sense-giving and
applicaton-guaranteeing mechanism in the first case and the implicit
regularites in the second a ratonalist perceives law and order here
and material yet in need of being shaped there. The habit, also
commented upon in an earlier part of this secton, to take a
partcipant's point of view with respect to the former and an
7. This pint has been made with great force and with the help of many examples
by Wittgenstein (cf. my essay 'Wittgenstein's Fhiles@himlh ctigatieas',FhilRn.,
1955). Wat have ratonalists replied? Russell (coldly): 'I don't undertand.' Sir Karl
Popper (breathlessly): 'He is rght, he is rght-I don't undertand it either!' In a word:
the pint is irelevant because leading rationalist don't undertand it. I, on the other
hand, would stan doubtng the intelligence (and perhaps also the inte.Uectual honest)
of ratonalists who don't undertand (or pretend not to undertand) such a simple
point.
8. Cf. my shon comment on 'coven classifcations' in Chapter 1 6.
SEVENTEEN 225
observer's attitude towards the latter further separates what is so
intmately connected in reality. And so we have finally two agencies,
stem and orderly reason on the one side, a malleable but not entrely
yielding material on the other, and with this all the 'problems of
ratonality' that have provided philosophers with intellectual (and, let
us not forget, also with fnancial) nourishment ever since the 'Rise of
Ration
alism in the West'. One cannot help notcing that the
arguments that are stll used to support this magnifcent result are
indistnguishable from those of the theologian who infers a creator
wherever he sees some kind of order: obviously order is not inherent
in matter and so must have been imposed from the outside.
The interacton view must therefore be supplemented with a
satsfactory account of the interactng agencies. Presented in this way
it becomes a tiviality. For there is no traditon no matter how hard
headed its scholars and how hard-limbed its warriors that will remain
unaffected by what occurs around it. At any rate -what changes, and
how, is now a matter either for historical research or for politicl aion
carried out by those who partcipate in the interactng traditons.
I shall now state the implicatons of these results in a series of theses
with corresponding explanatons.
We have seen that ratonal standards and the arguments supportng
them are visible parts of special traditons consistng of clear and
explicit principles and an unnotced and largely unknown but
ahsolutely necessary background of dispositons for acton and
juu6ement. The standards become 'objectve' measures of excel
lence when adopted by partcipants of taditons of this kind. We have
then 'objectve' ratonal standards and arguments for their validity.
We have further seen that there are other traditons that also lead to
judgements though not on the basis of explicit standards and
principles. These value judgements have a more 'immediate'
character, but they are stll evaluations, just like those of the
ratonalist. In both cases judgements are made by individuals who
partcipate in traditons and use them to separate ' Good' from ' Evil'.
We can therefore state:
i. Traditions are neither good nor bad, the simpl are. 'Objectvely
speaking', i.e. independently of partcipaton in a taditon, there is
not
much to choose between humanitarianism and ant-Semitsm.
Coroll
ary: ratonality is not an arbiter of taditons, it is itself a
trad
iton
or an aspect of a traditon. It is therefore neither good nor
bad, it simply is.
226 AGAI NST METHOD
ii. A tradition assumes dsirable or undirable proeies onl whe
compared with some tradition, i.e. only when viewed by partcipants
who see the world in terms of its values. The projectons of these
partcipants appear objeaie and statements describing them sound
objeaie because the partcipants and the traditon they project are
nowhere mentoned in them. They are subjeaie because they depend
on the taditon chosen and on the use the partcipants make of it.
The subjectvity is notced as soon as partcipants realize that
different taditons give rise to diferent judgements. They will then
have to revise the content of their value statements just as physicists
revised the content of even the simplest statement concering length
when it was discovered that length depends on reference systems and
just as everybody revised the content of 'down' when it was
discovered that the earth is spherical. Those who don't carry out the
revision cannot pride themselves on forming a special school of
especially astute philosophers who have overcome moral relatvism,
just as those who stll cling to absolute lengths cannot pride
themselves on forming a special school of especially astute physicists
who have overcome relatvity. They are just pig-headed, or badly
informed, or both.
iii. i. and ii. impl a relatiism ofprecisel the kind that seems to hae
bee dedd b Prtagoras. Protagorean relatvism is reasonable
because it pays attenton to the pluralism of traditons and values.
And it is ciilized for it does not assume that one's own village and the
strange customs it contains are the navel of the world. 9
iv. Ever tradition has special ways of gaining fllower. Some
traditons reflect about these ways and change them from one group
to the next. Others take it for granted that there is only one way of
making people accept their views. Depending on the taditon
adopted this way will look acceptable, laughable, ratonal, foolish, or
will be pushed aside as 'mere propaganda'. Argument is propaganda
for one obserer, the essence of human discourse for another.
v. We have seen that individuals or groups participatng in the
interacton of traditons may adopt a pragmatic philosophy when
judging the events and structures that arise. The principles of their
philosophy often emerge only during the interacton (people change
while obsering change or partcipatng in it and the taditons they
use may change with them). This means that judging a historcal proces
one may use an as yet unspecifed and unspecifable praice. One may base
9. Protagoras is discussed in detail in Chapter l, sectons 3f of FarmchteRcmea.
S EVENTEEN 227
judgements and actons on standards that cannot be specifed in
advance but are intoduced by the ver judgements (actons) they are
supposed to guide and one may even act without any standards,
simply following some natural inclinaton. The ferce warrior who
cures his wounded enemy instead of killing him has no idea why he
acts as he does and gives an entrely erroneous account of his reasons.
But his acton intoduces an age of collaboraton and peaceful
competton instead of permanent hostlity and so creates a new
taditon of commerce between natons. The queston- how will you
decide what path to choose? How will you know what pleases you and
what you want to reject? has therefore at least two answers, v. (1)
there i s no decision but a natural development leading to taditons
which in retospect give reasons for the acton had it been a decision
in accordance with standards or (2) to ask how one will judge and
choose in as yet unknown surroundings makes as much sense as to
ask what measuring instuments one will use in as yet unexlored
domains. Standards which are intellectual measuring instrments
often have to be ireted to make sense of new historical situatons
just as measuring instuments have constantly to be invented to make
sense of new physical situatons.
v. There are therefore at least two di eet ways ofcolleaiel
dciding an issue which I shall call a gidd echange and an oe
echange respectvely.
In the frst case some or all partcipants adopt a well-specifed
taditon and accept only those responses that correspond to its
standards. If one party has not yet become a partcipant of the chosen
traditon he will be badgered, persuaded, 'educated' untl he does
and then the exchange begins. Educaton is separated from decisive
debates, it occurs at an early stage and guarantees that the grown-ups
will behave properly. A rational dbate is a special case of a guided
exchange. If the partcipants are ratonalists then all is well and the
debate can start right away. If only some partcipants are ratonalists
and if they have power (an important consideraton! ) then they will
not take their collaborators seriously untl they have also become
ratonalists: a society based on ratonality is not entrely free; one has
to play the game of the intellectuals.
1 0
A open exchange, on the other hand, i s guided by a pragmatc
philo
sophy. The taditon adopted by the partes is unspecifed in the
10. 'It is perhaps hardly necessar to say', says John Star Mill, 'that this dotine
(pluralism of ideas and insttutons) i meant to apply only to human beings in the

atrity of their facultes' - i.e. to fellow intellectals and their pupils. 'On Libery',
m 1hcFhihsqe/eha5tmnMih, ed. M. Cohen, New York, 1 961 , p. 197.
228 AGAI NST METHOD
beginning and develops as the exchange proceeds. The partcipants
get immersed into each other's ways of thinking, feeling, perceiving
to such an extent that their ideas, perceptons, world-views may be
entrely changed - they become different people partcipatng in a
new and different traditon. An open exchange respects the parter
whether he is an individual or an entre culture, while a ratonal
exchange promises respect only within the framework of a ratonal
debate. An open exchange has no organon though it may invent one,
there is no logic though new forms oflogic may emerge in its course.
An open exchange establishes connectons between diferent
taditons and tanscends the relatvism of points iii and iv. However,
it tanscends it in a way that cannot be made objectve but depends in
an unforeseeable manner on the (historical, psychological, material)
conditons in which it occurs. (Cf. also the last paragraph of Chapter
1 6.)
vi. A free societ is a societ in which all traditions are gien equal
rghts, equal aces to education and other positions ofpower. This is an
obvious consequence of i, ii and iii. If taditons have advantages only
from the point of view of other traditons then chosing one taditon
as a basis of a free society is an arbitary act that can be justfed only
by resortng to power. A free society thus cannot be based on any
partcular creed; for example, it cannot be based on ratonalism or on
humanitarian consideratons. The basic stucture of a free society is a
prtecie strcure, not an ideolog, it functons like an iron railing not
like a convicton. But how is this stucture to be conceived? Is it not
necessary to dbate the matter or should the stucture be simply
imposed? And if it is necessar to debate the matter then should this
debate not be kept free from subjectve influences and based on
'objectve' consideratons only? This is how intellectuals ty to
convince their fellow citens that the money paid to them is well
spent and that their ideolog should contnue to assume the cental
positon it now has. I have already exposed the errors-cum
deceptons behind the phrase of the 'objectvity of a ratonal debate':
the standards of such a debate are not 'objectve' they only apear to
be 'objectve' because reference to the group that profts from their
use has been omitted. They are like the invitatons of a clever tyrant
who instead of saying 'I want you to do . . . ' or 'I and my wife want you
to do . . . ' says 'What all of us want is . . . ' or 'what the gods want of us
is . . . ' or, even better, 'it is ratonal to do . . .' and so seems to leave out
his own person entrely. It is somewhat depressing to see how many
intelligent people have fallen for such a shallow tck. We remove it by
observing:
SEVENTEEN 229
vll. that a free socet will not be imposed but will eere onl whee
peole egagng in an oe echange (cf. vi above) introduce protectie
str
cture ofthe kind aludd to. Citzen initatves on a small scale,
collabo
ration between natons on a larger scale are the developments
I have in mind. The United States are not a free society in the sense
described here.
i. The dbate setling the strure ofa free societ are oe dbate not
gidd dbate. This does not mean that the concrete developments
described under the last thesis alreay use open debates, it means that
they
could use them and that ratonalism is not a necessar ingredient
of the basic stucture of a free soiety.
The results for science are obvious. Here we have a partcular
taditon, 'objectvely' on par with all other taditons (theses i and
vi ). Its results will appear magnifcent to some taditons, execrable
to others, barely worth a yawn to stll further taditons. Of course,
our well-conditoned materialistc contemporaries are liable to burst
with excitement over events such as the moonshots, the double helix,
non-equilibrium therodyamics. But let us look at the matter from
a diferent point of view, and it becomes a ridiculous exercise in
futlity. It needed billions of dollars, thousands of well-tained
assistants, years of hard work to enable some inartculate and rather
limited contemporaries1 1 to perform a few graceless hops in a place
nobody in his right mind would think of visitg-a dried out, airless,
hot stone. But mystcs, using only their minds, tavelled across the
celestal spheres to God himself, whom they viewed in all his
splendour, receivng stength for contuing their lives and
enlightenment for themselves and their fellow men. It is only the
illiteracy of the general public and of their stem tainers, the
intellectuals, and their amazing lack of imaginaton that makes them
reject such comparisons without further ado. A free society does not
object to such an atttude but it will not permit it to become a basic
ideolog either.
x. A free societ insists on the searation ofsciec and societ. More
about this topic in Chapter 1 9.
I I .
Cf.
Noran Maer, O/a Freo thcMm,London, 1 970.
1 8
Yet it i possible to ealuate standrd ofrationalit and to imprle them.
The princple of impruement are neithe able tradition nor beond
change and it is impossible to nail them dwn.
I shall now illustate some of these results by showing how standards
are and have been critcized in physics and astonomy and how this
procedure can be extended to other fields.
Chapter 1 7 started with the general problem of the relaton
between reason and practce. In the illustaton reason becomes
scientfc ratonality, practce the practce of scientfc research, and
the problem is the relaton between scientfic ratonality and
research. I shall discuss the answers given by idealism, naturalism
and by a third positon, not yet mentoned, which I shall call naive
anarchism.
According to idalism it is ratonal (proper, in accordance with the
will of the gods - or whatever other encouragng words are being
used to befuddle the natves) to do certain things-come what ma. It is
ratonal (proper, etc.) to kill the enemies of the faith, to avoid ad hoc
hypotheses, to despise the desires of the body, to remove
inconsistencies, to support progressive research programmes and so
on. Ratonality Gustce, the Divine Law) are universal, independent
of mood, context, historical circumstances and give rise to equally
universal rules and standards.
There is a version of idealism that seems to be somewhat more
sophistcated but actually is not. Ratonality (the law, etc.) is no longer
said to be universal, but there are universally valid conditonal
statements assertng what is ratonal in what context and there are
corresponding conditonal rules.
Some reviewers have classified me as an idealist in the sense just
described with the proviso that I try to replace familiar rules and
standards by more 'revolutonary' rules such as proliferaton and
counterinducton and almost everyone has ascribed to me a
'methodology' with 'anything goes' as its one 'basic principle'. But in
230
EI GHTEEN 231
Chapter 2 I say quite explicitly that 'my intenton is not to replace one
set of general rules by another such set: my intenton is, rather, to
convince the reader that, al methodlogies, ee the most obious ones,
hae their limits' or, to express it in terms just explained, my intenton
is to show that idealism, whether of the simple or of the context
dependent kind, is the wrong soluton for the problems of scientfc
ratonality. These problems are not solved by a change of standards
but by taking a diferent view of standards altogether.
Idealism can be dogmatc and it can be critcal. In the frst case the
rules proposed are regarded as fnal and unchangeable; in the second
case there is the possibility of discussion and change. But the
discussion does not take practces into account -it remains resticted
to an abstact domain of standards, rules and logic.
The limitaton of all rules and standards is recognized by naie
anarchism. A naive anarchist says (a) that both absolute rules and
context-dependent rules have their limits and infers (b) that all rules
and standards are worthless and should be given up. Most reviewers
regard me as a naive anarchist in this sense, overlooking the many
passages where I show how certain procedures aidd scientsts in
their research. For in my studies of Galileo, of Brownian moton,
of the Presoratcs I not only demonstate the failures of familiar
standards, I also t to show what not so familiar proedures
did actually succeed. Thus while I agree with (a) I do not agree
with (b). I argue that all rules have their limits and that there
is no comprehensive 'ratonality', I do not argue that we should
proeed without rules and standards. I also argue for a context
ual account but again the contextual rules are not to relae the
absolute rules, they are to supleet them. Moreover, I suggest a
new relation between rules and practces. It is this relaton and not any
partcular rule-content that characterizes the positon I wish to
defend.
This positon adopts some elements of naturalism but it rejects the
naturalist philosophy. According to naturalism rules and standards
are obtained by an analysis of traditons. As we have seen the problem
is which traditon to choose. Philosophers of science will of course
opt for science as their basic taditon. But science is not one taditon,
it is many, and so it gives rise to many and partly incompatble
standards (I have explained this difculty in my discussion of
Lakatos).
1
Besides, the procedure makes it impossible for the
philosopher to give reasons for his choice of science over myth
I .
Fhil@himlF@m,Vol. 2, Chapter I 0. Cf. as Chapter 19.
232 AGAI NST METHOD
or Aristotle. Naturalism cannot solve the problem of scientifc
ratonality.
As in Chapter 1 7 we can now compare the drawbacks of
naturalism and idealism and arrive at a more satsfactory view.
Naturalism says that reason is completely dterined by research. Of
this we retain the idea that research can change reason. Idealism says
that reason completely guers research. Of this we retain the idea
that reason can change research. Combining the two elements we
arrive at the idea of a guid who is par ofthe aiit guidd and
is changed b it. This corresponds to the interactonist view of
reason and practce formulated in Chapter 1 7 and illustated by
the example of the map. Now the interactonist view assumes
two diferent enttes, a disembodied guide on the one side and a
well-endowed practce on the other. But the guide seems dis
embodied only because its 'body', i.e. the very substantal practce
that underlies it, is not notced and the 'practce' seems crude and in
need of a guide only because one is not aware of the complex and
rather sophistcated laws it contains. Thus the problem is not the
interacton of a practce with something diferent and exteral, but the
deloment ofone tradition undr the impac ofother. A look at the way
in which science treats its problems and revises its 'standards'
confrms this picture.
In physics theories are used both as descriptons of facts and as
standards of speculaton and factual accuracy. Measurng instrments
are constucted in accordance with laws and their readings are tested
under the assumpton that these laws are correct. In a similar way
theories giving rise to physical principles provide standards to judge
other theore by: theories that are relativistcally invariant are better
than theories that are not. Such standards are of course not
untouchable. The standard of relatvistc invariance, for example,
may be removed when one discovers that the theory of relatvit has
serious shortcomings. Shortcomings are occasionally found by a
direct examinaton of the theory, for example by an examinaton ofits
mathematcs, or its predictve success. They may also be found by the
development of alteratives (cf. Chapter 3) - i.e. by research that
violates the standards to be examined.
The idea that nature is infnitely rich both qualitatvely and
quanttatvely leads to the desire to make new discoveries and thus to
a principle of content increase which gives us another standard to
judge theories by: theories that have excess content over what is
already kown are preferable to theories that have not. Again the
standard is not untouchable. It is in trouble the moment we discover
that we inhabit a fnite world. The discovery is prepared by the
EI GHTEEN 233
development of 'Aristotelian' theories which refrain from going
beyond a given set of propertes -it is again prepared by research that
violates the standard.
The procedure used in both cases contains a variety of elements
and so there are diferent ways of describing it, or reactng to it.
One element and to my mind the most important one is
csmological. The standards we use and the rules we recommend
make sense only in a world that has a certain stucture. They become
inapplicable, or start running idle in a domain that does not exhibit
this stucture. When people heard of the new discoveries of
Columbus, Magellan, Diaz they realized that there were contnents,
climates, races not enumerated in the ancient accounts and they
conjectured there might be new contnents of knowledge as well, that
there might be an 'America of Knowledge' just as there was a new
geographical entty called 'America', and they ted to discover it by
venturing beyond the limits of the received ideas. The demand for
content increase now became ver plausible. It arose from the wish to
discover more and more of a nature that seemed to be infinitely rich
in extent and quality. The demand has no point in a finite world that
is composed of a finite number of basic qualites.
How do we find the cosmology that supports or suspends our
standards? The reply introduces the second element that enters the
revision of standards, viz. theorzing in a general sense, including
myth and metaphysical speculaton. The idea of a finite world
becomes acceptable when we have theories describing such a world
and when these theories tum out to be better than their infinitst
rivals. The world is not directly given to us, we have to catch it
through the medium of traditons which means that even the
cosmological argument refers to a certain stage of competton
between world-views, theories of ratonality included.
Now when scientsts become accustomed to treatng theories in a
certain way, when they forget the reasons for this teatent but
simply regard it as the 'essence of science' or as an 'important part of
what it means to be scientfic', when philosophers aid them in their
forgetfulness by systematzing the familiar procedures and showing
how they flow from an abstact theor of ratonality then the theories
needed to show the shortcomngs of the underlying standards will not
be introduced or, if they are introduced, will not be taken seriously.
They will not be taken seriously because they clash with customar
habits and systematzatons thereof.
For example, a good way of examining the idea that the world is
finite both qualitatvely and quanttatvely is to develop an
Aris
totelian cosmology. Such a cosmology provides means of
234 AGAI NST METHOD
descripton adapted to a finite world while the corresponding
methodology replaces the demand for content increase by the
demand for adequate descriptons of this kind. Assume we introduce
theories that correspond to the cosmology and develop them in
accordance with the new rules. What will happen? Scientsts will be
unhappy for the theories have unfamiliar propertes. Philosophers of
science will be unhappy because they intoduce standards unheard of
in their profession. Being fond of surrounding their unhappiness
with arias called 'reasons' they will go a little further. They will say
that they are not merely unhappy, but have 'arguments' for their
unhappiness. The arguments in most cases are elaborate repettons
and variatons of the standards they grew up with and so their
cognitve content is that of 'But the theor is ad hoc! ' or 'But the
theories are developed without content increase! ' And all one hears
when asking the further question why that is so bad is either that
science has proceeded diferently for at least 200 years or that
content increase solves some problems of confirmaton theor. Yet
the queston was not what science does but how it can be improved
and whether adoptng some confirmaton theories is a good way of
learing about the world. No answer is forthcoming. And so
interestng possibilites are removed by firmly insistng on the status
quo. It is amusing to see that such insistence becomes the more
determined the more 'critcal' the philosophy that is faced with the
problem. We, on the other hand, retain the lesson that the validit,
useulnes, adquac ofpoular standrd can be checked onl b reearch
that violate them.
A further example, to illustrate the point. The idea that
informaton concering the exteral world tavels undisturbed via
the senses into the mind leads to the standard that all kowledge
must be checked by obseraton: theories that agree with obseraton
are preferable to theories that do not. This simple standard is in need
of replacement the moment we discover that sensor informaton is
distorted in many ways. We make the discover when developing
theories that confict with obseraton and finding that they excel in
many other respects (Chapters 5 to 1 1 describe how Galileo
contributed to the discover).
Finally, the idea that things are well defined and that we do not live
in a paradoxcal world leads to the standard that our kowledge must
be self-consistent. Theories that contain contradictions cannot be
part of science. This apparently quite fundamental standard which
many philosophers accept as unhesitatingly as Catholics once
accepted the dogma of the immaculate concepton of the Virgin loses
its authority the moment we find that there are facts whose only
EI GHTEEN 235
adequate descripton is inconsistent and that inconsistent theories
may be fruitul and easy to handle while the attempt to make them
conform to the demands of consistency creates useless and unwieldy
monsters.2
The last example raises further questons which are usually
formulated as objectons against it (and against the critcism of other
standards as well, standards of content increase included).
One objecton is that non-contadicton is a necessary conditon of
research. A procedure not in agreement with this standard is not
research - it is chaos. It is therefore not possible to examine non
contadicton in the manner described in the last example.
The main part of the objecton is the second statement and it is
usually supported by the remark that a contadicton implies every
statement. This it does - but ony in rather simple logical systems.
Now it is clear that changing standards or basic theories has
repercussions that must be taken care of. Admttng velocites larger
than the velocity of light into relatvit and leaving everything else
unchanged gives us some rather puzling results such as imaginary
masses . and velocites. Admittng well-defed positons and
momenta into the quantum theory and leaving everything else
unchanged creates havoc with the laws of interference. Admttng
contadictons into a system of ideas allegedly connected by the laws
of standard logic and leaving everg else unchanged makes us
assert every statement. Obviously we shall have to make some further
changes, for example we shall have to change some rules of
derivaton in the last case. Carrying out the change removes the
problems and research can proceed as planned. (Scientfc practce
containing inconsistencies is already arranged in the right way.)
But - says an objecton that is frequently raised at this point: how
will the results of the research be evaluated if fundamental standards
have been removed? For example, what standards show that research
in violaton of content increase produces theories which are 'bete
than their infnitst rivals' as I said a few paragraphs ago? Or what
standards show that theories in conflict with observatons have
something to offer while their observatonally ipeccable rivals have
not? Does not a decision to accept unusual theories and to reject
fa
miliar ones assume standards and is it not clear, therefore, that
cosmological investgatons cannot t to provide alteratves to all
standards? These are some of the questons one hears with tring
regularit in the discussion of 'fundamental principles' such as
2. Cf. Chapter I 6, text to foototes 9 I f.
236 AGAI NST METHOD
consistency, content increase, observational adequacy, falsifability,
and so on. It is not diffcult to answer them.
It is asked how research leading to the revision of standards is to be
evaluated. For example, when and on what grounds shall we be
satsfed that research containing inconsistencies has revealed a fatal
shortcoming of the standard of non-contradicton? The queston
makes as little sense as the queston what measuring instuments will
help us to explore an as yet unspecifed region of the universe. We
don't know the region, we cannot say what will work in it. To advance
we must either enter the region, or start making conjectures about it.
We enter the region by artculatng unusual intellectual, soial,
emotonal tendencies, no matter how strange they may seem when
viewed through the spectacles of established theories or standards. It
would certainly be silly to disregard physical fature that do not agree
with deeply ingrained spiritual notons. But it is equally shortsighted
to curtailfntaie that do not seem to ft into the physical universe.
Fantasies and, in fact, the entre subjectvity of human beings are just
as much a part of the world as fleas, stones and quarks and there is no
reason why we should change them to protect the latter.
Similar consideratons apply to the standards that are supposed to
gide our thoughts and actons. They are not stable and they cannot
be stabilized by tying them to a partcular point of view. For Aistotle
knowledge was qualitatve and observatonal. Today knowledge is
quanttatve and theoretcal, at least as far as our leading natural
scientsts are concered. Who is right? That depends on what kind of
informaton has privileged status and this in tum depends on the
culture, or the 'cultural leaders' who use the informaton. Many
people, without much thought, prefer technology to harmony with
Nature; hence, quanttatve and theoretcal informaton is regarded
as 'real' and qualites as 'apparent' and secondar. But a culture that
centes on humans, prefers personal acquaintance to abstract
relatons (intelligence quotents; efciency statstcs) and a naturalists'
approach to that of molecular biologists will say that knowledge is
qualitatve and will interret quanttatve laws as bookkeeping
devices, not as elements of reality.
Combining the consideratons of the last two paragraphs we see
that even the apparently hardest scientfc 'fact' can be dissolved by
decisions undermining the values that make it a fact and/or by
research that replaces it by facts of a diferent kind. This is not
a new procedure. Philosophers from Parmenides to 20th-centur
(undialectical) materialists and scientsts from Galileo and Descartes
to Monod used it to devalue, and to declare as mere appearance, the
qualitatve features of human life. But what can be used to support
EI GHTEEN 237
science can also be used against it. The (cultural) measuring
instruments that separate 'reality' from 'appearance' change and
must change when we move from one culture to another and from
one historical stage to the next, just as our physical measuring
instruments change and must change when we leave one physical
region (one historical period) and enter another.
1 9
Scece i neither a single tradition, nor the bet tradition there is, ecet fr
peole who hae become acstomed to its preece, its beets and it
disadantage. In a dmocac it should be searated from the state just a
churche are now searated fom the state.
I shall now sum arize the arguments of the precedig chapters by
tng to answer the followig three questons.
1 . Wat is scece? How do scientsts proceed, how do their
standards difer from the standards of other enterprises?
2. Wat's so great about scece? What are the reasons that might
compel us to prefer the sciences to other fors of life and ways of
gathering knowledge?
3. How are we to use the scece and who dcid the matter?
My answer to the frst queston is that the wide divergence of
individuals, schools, historical periods, entre sciences makes it
extemely difcult to identf comprehensive principles either of
method, or of fact. The word 'science' may be a sigle word - but
there is no single entty that corresponds to that word.
In the domain of method we have scientsts like Salvador Luria who
want to te research to events permittng 'stong inferences',
'predictons that will be stonfly supported and sharply rejected by a
clear-cut experimental step'.
According to Luria the experiments (Luria and Delbrueck, 1 943)
which showed that the resistance of bacteria to phage invasion is a
result of environment-independent mutatons and not of an
adaptaton to the environment had precisely this character. There
was a siple predicton: fluctuatons, from one culture to the next, of
surviving colonies of bacteria on an agar contaiig an excess of
bacteriophages would be small i the frst case, but would contain
I . S.E. Lura, A 5letMhiac,a mk1ct1mc,New York, 1 985, p. l i S.
238
NI NETEEN 239
avalanches in the second. The predicton could be tested in a simple
and staightforard way and there was a decisive result. (The result
refuted Lamarckism, which was popular among bacteriologists but
practcally extnct elsewhere - a frst indicaton of the complexity of
scie
nce.)
Scientsts inclined in the manner of Luria show a considerable
'lack of enthusiasm in the "big problems" of the Universe or of the
early Earth or in the concentaton of carbon dioxde in the upper
atosphere',
2
al subjects that are 'loaded with weak inferences'.
3
In a way they are contnuing the Aristotelian approach which
demands close contact with e'erience and objects to following a
plausible idea to the bitter end.
However, this was precisely the procedure adopted by Einstein, by
researchers in celestal mechanics between Newton and Poincare, by
the proponents of atomism and, later, the kinetc theor, by
Heisenberg during the inital stages of matix mechanics and by
almost all cosmologists. Einstein's frst cosmological paper is a purely
theoretcal exercise containing not a single astonomical constant.
The subject of cosmolog itself for a long tme found few supporters
among physicists. Hubble the obserer was respected, the rest had a
hard tme:
Jourals accepted papers from obserers, giving them only the most
cursor refereeing whereas our own paper always had a stf passage,
to a point where one became quite wor out with exlaining points of
mathematcs, physics, fact and logic to the obtuse minds who
consttute the mysterous anonymous class of referees, doing their
work, like owls, i the darkness of the night. 5
'Is it not really stange', asks Einstein, 'that human beings are norally
deaf to the stongest argument while they are always inclined
to overestmate measuring accuracies?'
6
- but just such an 'over
estmatng of measuring accuracies' is the rule in epidemiolog,
demography, genetcs, spectoscopy and in other subjects. The
variety increases when we move into sciences like cultural
2.
ibid., p. 1 1 9.
3.
ibid.
4. DcCecle293a24ff.
5. F. Hoyle in Y. Terian and E.M. Bilson (eds), CesmeleaadZstphsic,Ithaca
and
London, 1982, p. 21 .
6.
Letter
t o Max Bor, quoted from the em-Eiastcia Lm m, New York, 1971 ,
p. 192.
240 AGAI NST METHOD
anthropology where a compromise has to be found between the
effects of personal contact and the idea of an objectve approach on
the one side and the practcal needs for quick acton and theoretcal
thoroughness on the other. 'To hear a seminar at a university about
modes of producton in the moring', writes Robert Chambers,
and then attend a meetng in a goverent ofce about agricultural
extension in the afteron leaves a shizoid feeling. One might not
know that bth refered to the same small farmer and might doubt
whether either disussion has anything to contibute to the other. 7
But is it not tue that scientsts proeed i a methodical way,
avoid accidents and pay attenton to obseraton and exeriment?
Not always. Some scientsts propose theories and calculate cases
which have little or no connecton with reality. 'The great growth
i technical achievements which began i the nineteenth centur',
we read i L. Prandtl's lectures Fundmetal of Hydr- and
Aeomechanic,
left sientfic knowledge far behind. The multtudinous problems of
practce could not be answered by the hydroynamics of Euler; they
could not even b discussed. This was chiefly because, starng from
Euler's equatons of moton the sience had become more and more a
purely academic analysis of the hypothetcal frictonless 'ideal fluid'.
This theoretcal development is assoiated with the names of
Helmholt, Kelvin, Lamb and Rayleigh.
The analytcal results obtained by means of this so called 'classical
hydroynamics' virtually do not agree at all with the practcal
phenomena . . . . Therefore the engineers . . . put their trust in a mass
of empircal data collectvely known as the 'science of hydraulics',
a branch of knowledge which gew more and more unlike hydro
dynamics.
8
Accordig to Prandtl we have a disorderly collecton of facts on the
one side, sets of theories startng from simple but counterfactual
assumptons on the other and no connecton between the two. More
recently the axomatc approach i quantum mechanics and
especially i quantum feld theor was compared by cynical obserers
to the shakers, 'a religious sect of New England who built solid bas
7. RarmDeelmml, London, 1983, p. 29.
8. E. O.G. Tietens, Ne York, 1954, p. 3.
NI NETEEN 241
and led celibate lives, a non -scientfc equivalent of proving rigorous
theorems and calculatng no cross sectons' .
9
Yet in quantum mechanics this apparently useless actvity has led
to a more coherent and far more satsfactor codifcaton of the facts
than had been achieved before, while in hydrodynamics 'physical
commonsense' occasionally tured out to be less accurate than the
results of rigorous proofs based on wildly unrealistc assumptons. A
early example is Maxwell's calculaton of the viscosity of gases. For
Maxwell this was an exercise in theoretcal mechanics, an extension
of his work on the rings of Satur. Neither he nor his contemporaries
believed the outcome - that viscosity remains constant over a wide
range of density - and there was contar evidence. Yet more precise
measurements confrmed the predicton.
1
Few people were pre
pared for such a tum of events. Mathematcal curiosity had started
the work, cross-fertlizaton, not general principles, had brought it
to a conclusion.
Meanwhile the situaton has changed in favour of theor. In the
sies and seventes, when science was stll in public favour, theor
got the upper hand, at universites, where it increasingly replaced
professional skills, even in medicine, and in special subjects such as
biology or chemist where earlier morhological and substance
related research was replaced by a study of molecules. In cosmolog a
firm belief in the Big Bang now tends to devalue observatons that
clash with it. 'Such obseratons', writes C. Burbidge,
are delayed at the r.fereeing stage as long as possible with the hope
that the author will give up. If this does not ocur and they are
published the second line of defence is to ignore them. If they give rise
to some comment, the best approach is to arge simply that they are
hopelessly wrong and then, if all else fails, an obserer may be
threatened with loss of telescope tme untl he changes his program.
1 1
9. R.F. Steater and A.S. Wightan, FC1,5pia,5tatuticaadZllDm,New York,
1964, p. l .
I 0. For quantm mechanics cf. sectons 4. 1 and 4.2 of Hans Primas, Chistq,
QaaatamMcchaaicaadRcdauieaum,Berlin-New York, 1981 . Maxwell's calculatons
are reprouced in 17c5citcFapme/amcClokMmch, ed. W.D. Niven, New
York, 1965 (frst published in 1 890), pp. 377f. The conclusion is stated on p. 391 : 'A
remarkable result here presented to us . . . is that if this explanaton of gaseous frcton
be tue, the coefcient of frcton is independent of the density. Such a consequence of
a mathematcal theory is ver stanling, and the only experiment I have met with on the
s

bject does not seem to confr it.' For examples from hydrodynamics cf. G.
Brk
hof,
Hdmqaamic,New York, 1955, sectons 20 and 21 .
I I .
'Proble
ms of Cosmogony and Cosmology', in F. Berola, j.W. Sulentc and
D
.F.
Madore (eds), NmIdmiaZstmae,Cambrdge, 1988, p. 229.
242 AGAI NST METHOD
Thus all we can say i s that scientsts proceed in many diferent ways,
that rules of method, if mentoned explicitly, are either not obeyed at
all, or functon at most like rules of thumb and that imporant results
come from the confuence of achievements produced by separate and
often confictng tends. The idea that ' "scientfc" knowledge is in
some way peculiarly positve and free from diferences of opinion'1
2
is nothing but a chimaera.
The situaton in the arts is quite similar - as a matter of fact, it
ocurs in all areas of human actvity. Cennino Cennini's Libo
di're of 1390 contains practcal advice based on a rich experience
and complex skills. Leon Battsta Aberts DellPitura of1 435/6 i a
theoretcal teatse closely ted to cental perspectve and academc
optcal theor. Perspectve soon became a mania among artsts.
Leonardo and Raphael then pointed out, the one in words, the other
practcally (cf. the sphere on the right hand side of his School ofAthes
in the Stanza della Segnatura of the Vatcan), that a picture that is to
be viewed under normal circumstances, from a comfortable but not
well-defned distance and with both eyes wide open cannot obey the
rules of cental perspectve. They thereby clarifed the diference
between physiological optcs and geometical optcs which Kepler,
more than a centur later, stll tied to bridge by an easily refuted
hypothesis (cf. Ch. 9, text to footote 50). But cental perspectve
remained a basis on which various changes were superimposed.
So far I have been talking about proedure, or method. Now
methods that are not used as a matter of habit, without any thought
about the reasons behind them, are often ted to metaphysical beliefs.
For example, a radical form of empiricism assumes either that
humans are the measure of things or that they are in harmony with
them. Applied consistently methodological rules may produce results
which agree with the corresponding metaphysics. Luria's procedure
is an example. It did not fail; it helped to build a subject which today is
at the forefront of research. Einstein's approach did not end in
disaster; it led to one of the most fascinatng moder theories -
general relatvity. But methods are not resticted to the area where
they scored their first tiumphs. Luria's requirements, for example,
also tured up in cosmolog; they had been used by Heber Curts, in
his 'grand debate' with Harlow Shapely; by Ambarzumjan, who
opposed empiricism to abstact principles; and they are now being
applied by Halton Arp, Margaret Geller and their collaborators.
Whatever the results, a world built up in the manner of Luria has
1 2. N.R. Camb,Feamie/Scee, New York, 1957, p. 21 .
NI NETEEN 243
little in common with the world of Einstein and this world again
differs considerably from the world of Bohr. Johann Theodore
Mer describes in detail how abstact world-views using cor
responding methods produced results which slowly flled them with
empirical content. 13 He discusses the following views. First, the
astronomical view, which rested on mathematcal refnements of
acton at a distance laws and was extended (by Coulomb, Neumann,
Ampere and others) to electicity and magnetsm. Laplace's theor of
capillarity was an outstanding achievement of this approach.
Secondly, the atomic view, which played an important role in chemical
research (example: stereohemistry) but was also opposed by
chemists. Thirdly, the kinetic and mechanical view, which employed
atoms in the area ofheat and electic phenomena. For some scientsts
atomism was the foundaton of everything. Fourthly, the physicl view,
which tried to achieve universality in a different way, on the basis of
general notons such as the noton of energ. It could be connected
with the knetc view, but often was not. Physicians, physiologists and
chemists like Mayer, Helmholtz, du Bois Reymond and, in the
practcal area, Liebig were outstanding representatves of this view in
the second half of the 1 9th centur while Ostwald, Mach and Duhem
extended it into the 20th. Startng his descrpton of the morhologcl
view, Merz writes:
The different aspects of nature which I have reviewed in the foregig
chapters and the various sciences which have been elaborated by their
aid, comprise what may appropriately be tered the abstact study of
natural objects and phenomena. Though all the methods of reasoning
with which we have so far become acquainted originated primary
through obseraton and the refecton over things natural, they have
this in common that they - for the purose of examinaton - remove
their objects out of the positon and suroundings which nature has
assigned to them: that they astrat them. This proess of abstacton
is either literally a proess of removal from one place to another, from
the great work - and storehouse of nature herself to the smal
workrom, the laborator of the expermenter; or - where such
removal is not possible -the proess is carried out merely in the realm
of contemplaton; one or two special properes are noted and
described, whilst the number of collateral data are for the moment
disregarded. [A third method, not developed at the tme, is the
creaton of 'unnatural' conditons and, thereby, the producton of
'unnatural' phenomena.]
1 3.
Huteqe/Eupeaa1heughtiatheI9thCmtuq(frt published 1 91 2).
244 AGAI NST METHOD
. . . There is, moreover, i n additon to the aspect of convenience,
one ver powerful inducement for scientfc workers to persevere in
their proess of abstacton. . . . This is the practcal usefulness of
such researches in the arts and industres. . . . The wants and
creatons of artfcial life have thus proved the greatest incentves to
the abstract and artfcial treannent of natural objects and proesses
for which the chemical and electical laboratories with the calculatng
rom of the mathematcian on the one side and the workshop and
factor of the other, have in the course of the centur become so
renowned . . . .
There is, however, in the human mind an opposite interest which
fortunately counteracts to a considerable extent the one-sided
working of the spirit of abstracton i science . . . . This is the genuine
love of nature, the consciousness that we lose all power if, to any great
extent, we sever or weaken that connecton which tes us to the world
as it is -to things real and natural: it finds its expression i the ancient
legend of the mighty giant who derived all his strength from his
mother earth and collapsed if severed from her . . . . In the study of
natural objects we meet [therefore] with a class of students who are
attacted by things as they . . . . [Their] sciences are the tuly
descriptve sciences, in oppositon to the abstract ones. 14
I have quoted this descripton at length for it shows how diferent
procedures rest on and provide evidence for, diferent world-views.
Finally, Merz mentons the geetic view, the pschohysicl view, the
vitalistic view, the statitical view together with their procedures and
their findings.
What can a single comprehensive 'world-view of science' or a
single comprehensive idea of science offer under such circumstances?
It can ofer a surey, a list similar to the list given by Mer,
enumeratng the achievements and drawbacks of the various
approaches as well as the clashes between them and it can identf
science with this complex and somewhat scattered wars on many
fronts. Alteratvely it can put one view on top and subordinate the
others to it, either by pseudo-derivatons, or b declaring them to be
meaningless. Reductonists love to play that game. Or it can
disregard the diferences and present a paste job where each
partcular view and the results it has achieved is smoothly connected
with the rest thus producing an impressive and coherent edifice -
'the' scientfic world-view.
1 4. Ibid., Vol. Z, NewYork, 1 96S, pp. 20f.
NI NETEEN 245
Expressing it diferently we may say that the assumpton of a single
coherent world-view that underlies all of science is either a
metaphysical hypothesis ting to antcipate a future unity, or a
pedagogical fake; or it is an attempt to show, by a judicious up- and
downgrading of disciplines, that a synthesis has already been
achieved. This is how fans of uniformity proeeded in the past (cf.
Plato's list of subjects in Chapter vii of his Reublic), these are the
ways that are stll being used today. A more realistc account,
however, would point out that ' [t]here is no simple "scientfc" map of
reality - or if there were, it would be much too complicated and
unwieldy to be grasped or used by anyone. But there are many
diferent maps of reality, from a variety of scientfc viewpoints'.
15
It may be objected that we live in the 20th centur, not in the 1 9th,
and that many unifcatons which seemed impossible then have been
achieved by now. Examples are statstcal therodynamics, molecu
lar biolog, quantum chemist and superstings. These are indeed
flourishing subjects, but they have not produced the unity the phrase
'the' scientfc view of the world insinuates. Actually, the situaton is
not ver diferent from that which Mer had notced in the 1 9th
centur. Truesdell and others contnue the physical approach:
Prandtl maligned Euler, Truesdell praises h for having provided
rigorous concepts for research. Morpholog, though given a low
status by some and declared to be dead by others, has been revived by
ecologists and by Lorenz's study of animal behaviour (whch added
frs ofmotion to the older static frs) and it has always been of
importance in galactc research (Hubble's classifcaton). Having
been in the doghouse, cosmolog is now being coured by high
energ physicists but clashes with the philosophy of complementarit
accepted by the same group. Comentng on the problem M.
Kafatos and R. Nadeu write:
The essental requirement of the Copenhagen interretaton that the
experimental setup must be taken into account when making
obseratons is seldom met in obseratons with cosmological import
[though such obseratons rely on light, the pardig case of
complementarity] .
1 6
Moreover, the obseratons of Ar, M. Geller and others have
I S. John Ziman, 1chiagaadLcamiagZbat5omaad5mc, Cambridge, 1 980,
p. l 9.
16.
'Complementty and Cosmolog', in M. Kafato (ed.), ck's 1hmr~,
Qmatam1hceqaadthcCeaqtieme/thcUame,Dordrecht, 1 980, p. 263.
246 AGAI NST METHOD
thrown considerable doubt on the homogeneit assumpton which
plays a cental role in it. Extended to I ,000 megaparsec, Geller's
research may blow up the entre subject. We have a rabid materialism
in some parts (molecular biolog, for example), a modest to radical
subjectvism in others (some versions of quantum measurement,
anthropic principle). There are many fascinatng results, specula
tons, attempts at interretaton and it is certainly worth knowing
them. But pastng them together into a single coherent 'scientfc'
world-view, a procedure which has the blessings even of the Pope
1 7
- this i s going too far. After all, who can say that the world which so
stenuously resists unifcaton really is as educators and meta
physicians want it to be - tdy, uniform, the same everwhere?
Besides, as was shown in Chapters 3f, a paste job eliminates
precisely those conficts that kept science going in the past and will
contnue inspiring its practtoners if presered.
At this point some defenders of uniformit rise to a higher level.
Science may be complex, they say, but it is stll 'ratonal'. Now the
word 'ratonal' can either be used as a collectng bag for a variet of
proedures - this would be its nominalist interretaton - or it
describes a general feature found in ever single scientfc acton. I
accept the frst defniton, but I reject the second. In the second case
ratonalit is either defned in a narrow way that excludes, say, the
arts; then it also excludes large sectons of the sciences. Or it is
defned in a way that lets all of science survve; then it also applies to
love-making, comedy and dogfghts. There is no way of delitng
'science' by something stonger and more coherent than a list.
I come to the second queston-what's so great about science? There
are various measures of greatess. Poulart, i.e. familiarit with
some results and the belief that they are important, is one of them.
Now it is tue that despite periodic swings towards the sciences and
away from them they are stll in high repute with the general public
or, rather, not the sciences, but a mythical monster 'science' (in the
singular - in German it sounds even more impressive: Die
Wisseschaf). For what the general public seems to assume is that the
achievements they read about in the educatonal pages of their
newspapers and the threats they seem to perceive come from a single
source and are produced by a uniform procedure. They know that
biolog is different from physics which is different from geolog. But
1 7. Cf. his message on the ocasion of the 30h anniverar of Newon's
Fnacipia, published in ehaFaal ea5ciccaadRcligiea, Note Dame, 1990, esp.
M6f.
NI NETEEN 247
these disciplines, it is assumed, arise when 'the scientfc way' is
applied to different topics; the scientfc way itself, however, remains
the same. I have tried to argue that scientfc practce is much more
diverse. Adding that scientsts keep complaining about the scientfc
illiteracy of the general public and that by the 'general public' they
mean the Wester middle class, not Bolivian peasants (for example),
we have to conclude that the popularity of science is a ver doubtful
matter indeed.
What about praical adantage? The answer is that 'science'
sometmes works and sometmes doesn't. Some sciences (economic
theor, for example) are in a pretty sorr shape. Others are
sufciently mobile to tum disaster into tiumph. The cn d so becuse
the are not tied to any pariclar mehod or world-view. The fact that an
approach is 'scientfc' according to some clearly formulated criterion
therefore is no guarantee that it will succeed. Eah ce must be judged
searatel, especially today, when the fear ofindustial espionage, the
wish to overtake compettors on the way to a Nobel Prize, the uneven
distributon of funds, national rivalries, fear of accusatons (of
malpractce, plagiarism, waste of funds, etc.) put restictons on what
some dreamers, many philoso
R
hers among them, stll regard as a
'free intellectual adventure'. 8 The queston of trth, fnally,
remains unresolved. Love of tuth is one of the stongest motves for
replacing what really happens by a steamlined account, or, to
express it in a less polite manner, love of tuth is one of the stongest
motves for lying to oneself and to others. Besides, the quantum
theor seems to show, in the precise manner so much beloved by the
admirers of science, that reality is either one, which means there are
no obserers and no things obsered, or it is many, in which case what
is found does not exst in itself but depends on the approach chosen.
What are the views that are being compared with science when it is
declared to be superior? E.O. Wilson, the ' father' of sociobiology,
writes:
religion . . . will endure for a long tme as a vital force in soiety. Like
the mythical giant Antaeus who drew energ from his mother, the
earth, religion cannot be defeated by those who may cast it down. The
spiritual weakness of scientfc naturalism is due to the fact that it has
no such primal source of power . . . . So the tme has come to ask: does
I 8. This was realized by goverment advisers after the postwar euphoria had wor
off. See Joseph Ben-David, 5otcCremth,Berkeley, I 991 , p. 525, quoted above.
248 AGAI NST METHOD
a way exst to divert the power of religion into the serices of the great
new enterrise?
19
For Wilson the main feature of the alteratves is that they have
power. I regard this as a somewhat narrow characterizaton. World
views also answer questons about origins and puroses which soner
or later arise in almost ever human being. Answers to these
questons were available to Kepler and Newton and were used by
them in their research; they are no longer available today, at least not
within the sciences. They are par of non-scientfc world-views
which therefore have much to ofer, also to scientsts. When Wester
Civilizaton invaded what is now called the Third World it imposed
its own ideas of a proper environment and a rewarding life. It thereby
disrupted delicate patters of adaptaton and created problems that
had not exsted before. Both human decency and some appreciaton
of the many ways in which humans can live with nature prompted
agents of development and public health to think in more complex or,
as some would say, more 'relatvistc' ways. There exst approaches,
the approach called 'Primar environmental Care' among them,
which ofer legal, politcal and scientfc informaton but modifed in
accordance with the needs, the wishes and, what is most important,
the skll and the kowledge of local populatons.
20
Similarly, the
movement called liberaton theology has modifed Church dotne
to bring it closer to the spiritual needs of the poor and disadvantaged,
especially in South America.
Let me point out, incidentally, that not all ideas which seem
repulsive to the prophets of a New Age come from science. The idea
of a world machie and the related idea that nature is material to be
shaped by man should not be blamed on modem, i.e. post-Caresian,
science. It is older and stonger than a purely philosophical dotine
could ever be. The expression 'world machine' is found in Pseudo
Dionysius Areopagita, a mystc of unknown identt who wrote about
500 AD and had temendous influence. Oresme, who died in 1382
as bishop of Lisieux, compares the universe to a vast mechanical
clok set running by God so that 'all the wheels move as
harmoniously as possible'. The sentment can be easily understod:
I9. OHuma Nature, Cambridge, Ms., I 972, pp+ I 92f
20. LcsLeamedia Cemmuai-medEmireamtalMaaag~t, Proeedings
of the I9WPrimar Environmental Care Workhop, ed. Graia Borrini, Interatonal
Coure for Primar Health Care Manager at Distict Level in Developing Counties,
lsttuto Supriore di Saniti, Rome, I 99l . For a more popular presentaton cf. Graia
Brrini, 'Prmar Environmental Care: For Environmental Advoates and Polic
Maker', UCOCeuno,forhcoming.
NI NETEEN 249
this was the time when mechanical clocks 'of astounding intricacy
and elaboraton' were constructed all over Europe - ever town was
supposed to have one. Lynn White Jr., from whose book I have taken
this informaton, also describes the change of atttude that occurred
in the Carolingian Age:
The old Roman Calendars had occasionally shown genre scenes of
human actvity but the dominant traditon (which contnued in
Byzantium) was to depict the months as passive personifcatons being
symbols of attributes. The new Carolingian calendar which set the
pattern for the Middle Ages . . . shows a coercive attude towards
natural resources . . . . The pictures [are about) scenes of ploughing,
haresting, woodchopping; people knockng down acors for the pigs,
pig slaughtering. Man and Nature are now two things and man is the
master.
21
To sum up: there is no 'scientfic world-view' just as there is no
uniform enterprise 'science' -except in the minds of metaphysicians,
schoolmasters and politcians tying to make their naton compet
tve. Stll, there are many things we can lear from the sciences. But
we can also lear from the humanites, from religion and from the
remnants of ancient traditons that survived the onslaught ofWestem
Civilizaton. No area is unified and perfect, few areas are repulsive
and completely without merit. There is no objectve principle that
could direct us away from the supermarket 'religion' or the
supermarket 'art' towards the more modem, and much more
expensive supermarket 'science'. Besides, there are large areas of
knowledge and acton in which we use procedures without any idea as
to their comparatve excellence. A example is medicine, which,
though not a science, has increasingly been connected with scientfic
research. There are many fashions and schools in medicine j ust as
there are many fashions and schools in psycholog. It follows, frst,
the idea of a comparison of 'Wester medicine' with other medical
procedures does not make sense. Secondly, such a comparison is
often against the law, even if there should be volunteers: a test is
legally impossible. Adding to this that health and sickness are
culture-dependent concepts we see that there are domains, such as
medicine with no scientfc answer to queston 2. This is not really a
drawback. The search for objectve guidance is in confict with the
idea of individual responsibilit which allegedly is an important
21 .
Mcdimal1cchaeleaad5eoalChaagc,Oxord, 1 960, pp. 56 f.
250 AGAI NST METHOD
ingredient of a 'ratonal' or scientfc age. It shows fear, indecision, a
yearg for authority and a disregard for the new opportunites that
now exst: we can build world-views on the basis of a personal choice
and thus unite, for ourselves and for our friends, what was once
separated by a series ofhistorical accidents.22
On the other hand, we can agree that in a world full of scientfc
products scientsts may be given a special status just as henchmen
had a special status at tmes of social disorder or priests had when
being a citzen coincided with being the member of a single universal
Church. We can also agree that appealing to a chimaera (such as that
of a uniform and coherent 'scientfc world-view') can have important
politcal consequences. In 1 854 Commander Perr, using force,
opened the pors of Hakodate and Shimoda to American ships for
supply and tade. This event demonstated the military inferiority of
Japan. The members of the Japanese enlightenment of the early
1 870s, Fukuzawa among them, now reasoned as follows. Japan can
keep its independence only if it becomes stonger. It can become
stonger only with the help of science. It will use science efectvely
only if it does not just practse science but also believes in the
underlying ideology. To many taditonal Japanese this ideology
'the' scientfc world-view - was barbaric. But, so the followers of
Fukuzwa argued, it was necessary to adopt barbaric ways, to regard
them as advanced, to intoduce the whole of Wester Civilizaton in
order to suriveP Having been thus prepared Japanese scientsts
son branched out as their Wester colleagues had done before and
falsifed the uniform ideology that had stared the development. The
lesson I draw from this sequence of events is that a uniform 'scientfc
view of the world' may be useful fr peole ding scec-it gives them
motvaton without tying them down. It is like a fag. Though
presentg a single patter it makes people do many diferent things.
However, it is a disaste fr outid (philosophers, fy-by-night
mystcs, prophets of a New Age). It suggests to them the most
narowminded religious commitent and encourages a similar
narowmindedness on their part.
22. Wolfgang Pauli, who was deeply concered about the intellectual situaton of
the tme, demanded that science and religion again be united: letter to M. Fier, 8
August 1948. I agee but would add, entrely in te spirit of Pauli, that the unifcaton
should be a pronal matter; it should not be prepared by philosophical-sientfc
alchemist of the mind and impsed by teir minion in educaton. (It is diferent in the
Third World where a stong faith stll surives.)
23. Details in Caren Blacker, Dc@aaccEalightmt,Cambridge, 1969. For
the plitcal backgound cf. Chapter 3 and 4 of Rchard Storr, A Huteqe/M
@aa,Harondsworh, 1 982.
NI NETEEN 251
What I have said so far already contains my answer to queston 3: a
com
munity wil use science and scientsts in a way that agrees with its
values and aims and it will correct the scientfc insttutons in its
midst to bring them closer to these aims. The objecton that science is
self-correcting and thus needs no outside interference overlooks,
first, that ever enterrise is self-correctng (look at what happened to
the Catholic Church after Vatcan II) and, secondly, that i a
democracy the self-correction of the whole which tries to achieve
more humane ways ofliving overrules the self-correcton of the parts
which has a more narrow aim - unless the parts are given temporar
independence. Hence in a democracy local populatons not only will,
but also should, use the sciences in ways most suitable to them. The
objection that citzens do not have the expertse to judge scientfic
matters overlooks that important problems often lie across the
boundaries of various sciences so that scientsts within these sciences
don't have the needed expertse either. Moreover, doubtful cases
always produce experts for the one side, experts for the other side,
and experts in between. But the competence of the general public
could be vastly improved by an educaton that exposes expert
fallibility instead of actng as if it did not exist.
20
The point ofview undrling this book i not the reult ofa well-planned train
ofthought but ofarmets prmpted b accidtal ecounter. Anger at the
wanton dtrcion ofcltural achieeet from which we all could hae
leared, a the conceited assurance with which some intellecual interere
with the lie ofpeole, and contept fr the treacl phrase the use to
ebellih their mided wa and still is the motie frce behind my work.
The problem of knowledge and educaton in a free society frst stuck
me during my tenure of a state fellowship at the Weimar Insttut zur
Methodologischen Emeuerung des Deutschen Theaters ( 1 946),
which was a contnuaton of the Deutsches Theater Moskau under
the directorship of Maxm Vallentn. Stf and students of the Insttut
periodically visited theates in Easter Germany. 1 A special tain
l . Like many people of my generaton I was involved in the Second World War. This
event had little influence on my thinking. For me the war was a nuisance, not a moral
problem. Before the war I had intended to study astonomy, actng and singing and to
practse these professions simultaneously. I had excellent teacher (Adolf Vogel, my
singing teacher, had an interatonal reputaton and taught outstanding opera singer
such as Noran Bayley) and had just overcome some major vocal difcultes when I
received my draft notce (l was eighteen at the tme). How inconvenient, I thought. Why
the hell should I parcipate in the war games of a bunch ofidiots? How do I get out ofit?
Varous attempts misfred and I became a soldier. I applied for ofcers' taining to avoid
bullets as long as possible. The atempt was not entirely successfl; I was a lieutenant
before the war had come to an end and found myselfin the middle of the Gern reteat
in Poland and then in East Germany, surounded by fleeing civilians, infant units,
tanks, Polish auxiliares whom I suddenly commanded (the higher ofcer quickly
disappeared when matter became stcky). The whole colourfl chaos then appeared to
me like a stage and I became careless. A bullet hit me on my right hand, a second bullet
grazed my face, a third got stuck in my spine, I fell to the gound, unable to rise, but with
the happy thought 'the war is over for me, now at last I can retur to singing and my
beloved astonomy books'. It was only much later that I became aware of the moral
problems of the entre age. It seems to me that these problems are stll with us. They arise
whenever an individual or a goup objectvizes it ow peronal conceptons of a go
life and act accordingly. Cf. FarmehteRemm ,pp. 309f. This explains the ocasional
violence of my argument.
252
TWENTY 253
brought us from city to city. We arrived, dined, talked to the actors,
watched two or three plays. After each performance the public was
asked to remain seated while we started a discussion of what we had
just seen. There were classical plays, but there were also new plays
which tried to analyse recent events. Most of the tme they dealt with
the work of the resistance in Nazi Germany. They were indistn
guishable from earlier Nazi plays eulogizing the actvity of the Nazi
underground in democratc countries. In both cases there were
ideological speeches, outbursts of sincerity and dangerous situatons
in the cops and robbers traditon. This puzzled me and I commented
on it in the debates: how should a play be structured so that one
recognizes it a presenting the 'good side'? What has to be added to
the acton to make the struggle of the resistance fighter appear
morally superior to the struggle of an illegal Nazi in Austria before
1 938? It is not sufficient to give him the 'right slogans' for then we
take his superiority for granted, we do not show wherein it consists.
Nor can his nobility, his 'humanity' be the distnguishing mark; ever
movement has scoundrels as well as noble people among its
followers. A playwright may of course decide that sophistcaton is
luxur in moral battles and give a black-white account. He may lead
his followers to victor but at the expense of turing them into
barbarians. What, then, is the soluton? At the tme I opted for
Eisenstein and ruthless propaganda for the 'right cause'. I don't kow
whether this was because of any deep convicton of mine, or because I
was carried along by events, or because of the magnificent art of
Eisenstein. Today I would say that the choice must be left to the
audience. The playwright presents characters and tells a stor. If he
errs it should be on the side of sympathy for his scoundrels, for
circumstances and sufering play as large a role in the creaton of evil
and evil intentions as do those intentons themselves, and the general
tendency is to emphasize the latter. The playwright (and his
colleague, the teacher) must not tr to anticipate the decision
of the audience (of the pupils) or replace it by a decision of his
own if they should tum out to be incapable of makng up their
own minds. Undr no circumstance must he tr to be a 'moral frce'.
A moral force, whether for good or for evil, turs people into slaves
and slaver, even slaver in the service of The Good, or of God
Himself, is the most abject condition of all. This is how I see the
situ
ation today. However, it took me a long tme before I arrived at
this view.
Afte
r a year in Weimar I wanted to add the sciences and the
hum
anites to the arts, and the theate. I left Weimar and became a
student (histor, auxliar sciences) at the famous lnsttut fur
254 AGAI NST METHOD
Osterreichische Geschichtforschung which is part of the Universit
of Vienna. Later on I added physics and astronomy and so finally
retured to the subjects I had decided to pursue before the
interruptons of the Second World War.
There were the following 'influences'.
(l ) The Kraf Circe. Many of us science and engineering students
were interested in the foundatons of science and in broader
philosophical problems. We visited philosophy lectures. The lectures
bored us and we were soon thrown out because we asked questons
and made sarcastc remarks. I still remember Professor Heintel
advising me with raised arms: 'Herr Feyerabend, entweder sie halte
n
das Maul, oder sie verlassen den Vorlesungsaal! ' We did not give up
and founded a philosophy club of our own. Victor Kraft, one of my
teachers, became our chairman. The members of the club were
mostly students,
2
but there were also visits by facult members
and foreign dignitaries. Juhos, Heintel, Hollitscher, von Wright,
Anscombe, Wittgenstein came to our meetings and debated with
us. Wittgenstein, who took a long tme to make up his mind and
then appeared over an hour late, gave a spirited performance
and seemed to prefer our disrespectful attitude to the fawning
admiraton he encountered elsewhere. Our discussions started
in 1 949 and proceeded with interruptons up to 1952 (or 1 953).
Almost the whole of my thesis was presented and analysed at the
meetngs and some of my early papers are a direct outcome of these
debates.
(2) The Kraft Circle was part of an organizaton called the Autran
Colege Societ. The Society had been founded in 1945 by Austian
resistance fighters
3
to provide a forum for the exchange of scholars
and ideas and so to prepare the politcal unificaton of Europe. There
were seminars, like the Kraft Circle, during the academic year and
interatonal meetngs during the summer. The meetings took place
(and stll take place) in Alpbach, a small mountain village in Tirol.
Here I met outstanding scholars, artists, politcians and I owe my
academic career to the friendly help of some of them. I also began
suspectng that what counts in a public debate are not arguments but
2. Many of them have now become scientists or engineers. Johnny Sagan is
Professor of Mathematics at the University of Illinois, Henrich Eichhor director of
New Haven obseratory, Goldberger-de Buda adviser to electronic frs, while Erich
Jantsch, who died much too soon, met members of our circle at the astronomical
obseratory and later became a gr of dissident or pseudo-dissident scientists, trying
to use old traditions for new purposes.
3. Otto Molden, brother of Fritz Molden of the Molden publishing house, was for
many years the dynamic leader and organizer.
TWENTY 255
certain ways of presentng one's case. To test the suspicion I
interened in the debates defending absurd views with great
assurance. I was consumed by fear - after all, I was just a student
surrounded by bigshots -but having once attended an actng school I
proved the case to my satsfacton. The difcultes of scietic
ratonality were made ver clear by
(3) Feli Ehrehaf, who arrived in Vienna in 1 947. We, the
students of physics, mathematcs, astonomy, had heard a lot about
h. We knew that he was an excellent experimenter and that his
lectures were performances on a grand scale which his assistants had
to prepare for hours in advance. We knew that he had taught
theoretcal physics which was as exceptonal for an experimentalist
then as it is now. We were also familiar with the persistent rumours
that denounced him as a charlatan. Regarding ourselves as defenders
of the purity of physics we looked forard to exposing him in public.
At any rate our curiosity was aroused-and we were not disappointed.
Ehrenhaft was a mountain of a man, full of vitality and unusul
ideas. His lectures compared favourably (or unfavourably, depend
ing on the point of vew) with the more refined performances of his
colleagues. 'Are you dumb? Are you stupid? Do you really agree with
everg I say?' he shouted at us who had intended to expose him
but sat in silent astonishment at his performance. The queston was
more than justfied for there were large chunks to swallow. Relatvity
and quantum theor were rejected at once, and almost as a matter of
course, for being idle speculaton. In this respect Ehrenhaft' s atttude
was ver close to that of Stark and Lenard both of whom he
mentoned with approval. But he went further and critcized the
foundatons of classical physics as well. The first thing to be removed
was the law of inerta: undisturbed objects instead of going in a
straight line were supposed to move in a helix. Then came a sustained
attack on the principles of electomagnetc theor and especially
on the equaton div B 0. Many years before the fundamental
debate he produced convincing evidence for mesoscopic magnetc
monopoles. Then new and surprising propertes of light were
demonstated - and so on and so forth. Each demonstaton was
accompanied by a few gently ironical remarks on 'school physics' and
the 'theoretcians' who built castles in the air without considering the
experiments which Ehrenhaft devised and contnued devising in all
fields and which produced a plethora of inexplicable results.
We had soon an opportunity to witess the atttude of orthodox
phy
sicists. In 1 949 Ehrenhaft came to Alpbach. In that year Popper
con
ducted a seminar on philosophy, Rosenfeld and M.H.L. Prce
taug
ht
physics and phlosophy of physics (mainly from Bohr's
256 AGAI NST METHOD
comments on Einstein which had just appeared), Max Hartmann
biology, Duncan Sandys talked on problems of Britsh politcs,
Hayek on economics and so on. There was Hans Thirring, the senior
theoretical physicist from Vienna, a superb teacher who constantly
tried to impress on us that there were more important things than
science, who had taught physics to Feigl, Popper as well as the
present author and was an early and very actve member of the peace
movement. His son Walter Thirring, now Professor of Theoretcal
Physics in Vienna, was also present - a very distnguished audience
and a very critical one.
Ehrenhaft came well prepared. He set up a few of his simple
experiments in one of the country houses of Apbach and invited
everyone he could lay hands on to have a look. Every day from two or
three in the afteroon partcipants went by in an atttude of wonder
and left the building (if they were theoretcal physicists, that is) as if
they had seen something obscene. Apart from these physical
preparations Ehrenhaft also carried out, as was his habit, a beautful
piece of advertsing. The day before his lecture he attended a fairly
technical talk by von Hayek on 'The Sensory Order' (now available,
in expanded form, as a book). During the discussion he rose,
bewilderment and respect in his face, and started in a most innocent
voice: 'Dear Professor Hayek. This was a marellous, an admirable, a
most leared lecture. I did not understand a single word . . . .' Next
day his lecture had an overflow audience.
In this lecture Ehrenhaft gave a brief account of his discoveries,
adding general obserations on the state of physics. 'Now,
gentlemen,' he concluded triumphantly, turing to Rosenfeld and
Pryce who sat in the front row, 'what can you say?' And he answered
immediately. 'There is nothing at all you can say with all your fne
theories. Sitzen missen sie bleiben! Stll missen sie sein! '
The discussion, as was to be expected, was quite turbulent and it
was contnued for days with Thirring and Popper taking Ehrenhaft's
side against Rosenfeld and Pryce. Confronted with the experiments
the latter occasionally acted as some of Galileo's opponents must
have acted when confronted with the telescope. They pointed out
that no conclusions could be drawn from complex phenomena and
that a detailed analysis was needed. In short, the phenomena were a
Dreckef ct-a word that was heard quite frequently in the arguments.
What was our atttude in the face of all this commoton?
None of us was prepared to give up theory or to deny its excellence.
We founded a Club for the Salvaton of Theoretcal Physics and
started discussing simple experiments. It tured out that the relaton
between theory and experiment was much more complex than is
TWENTY 257
shown in textbooks and even in research papers. There are a few
paradigmatc cases where the theor can be applied without major
adjustents but the rest must be dealt with by occasionally rather
doubtful approxmatons and auxliar assumptons.
4
I fnd it
interestng to remember how little efect all this had on us at the tme.
We contnued to prefer abstactons as if the diffcultes we had
found had not been an expression of the nature of things but could be
removed by some ingenious device, yet to be discovered. Only much
later did Ehrenhaft's lesson sink in and our atttude at the tme as well
as the atttude of the entre profession provided me then with an
excellent illustaton of the nature of scientfc ratonalit.
(4) Philip Frank came to Alpbach a few years afer Ehrenhaft. He
undermined common ideas of ratonalit in a diferent way by
showing that the arguments against Copericus had been perfectly
sound and in agreement with experience while Gailieo's proedures
were 'unscientfc' when viewed from a modem standpoint. His
observatons fascinated me and I examined the matter further.
Chapters 8 to 1 1 are a late result of this study (I am a slow worker).
Frank's work has been teated quite unfairly by philosophers like
Putam who prefer simplistc models to the analysis of complex
historical events. Aso his ideas are now commonplace. But it was he
who announced them when almost everone thought diferently.
(5) In Vienna I became acquainted with some of the foremost
Marxst intellectuals. This was the result of an ingenious PR job by
Marxst students. They tured up - as did we - at all major
discussions whether the subject was science, religion, politcs, the
theate, or free love. They talked to those of us who used science to
ridicule the rest -which was then my favourite occupaton - invited
us to discussions of their own and introduced us to Marxst thinkers
from all felds. I came to know Berthold Viertel, the director of the
Burgtheater, Hanns Eisler, the composer and music theoretcian, a
n
d
Walter Hollitscher, who became a teacher and, later on, one of my best
friends. When startng to discuss with Hollitscher I was a raving
positvist, I favoured strict rules of research and had only a pitying
smile for the three basic principles of dialectcs which I had read in
Stalin's little pamphlet on dialectical and historical materialism. I was
interested in the realist positon, I had tied to read ever book on
realism I could lay hands on (including Kilpe's excellent Realisierng
and, of course, Materalism and Empirocticism) but I found that the
arguments for realism worked only when the realist assumpton had
4.
Cf. Chapter 5 on mhwapproxmatons.
258 AGAI NST METHOD
already been introduced. Kilpe, for example, emphasized the
distnction between impression and the thing the impression is about.
The distncton gives us realism only if it characterizes real features
of the world -which is the point at issue. Nor was I convinced by the
remark that science is an essentally realistic enterrise. Why should
science be chosen as an authority? And were there not positvistc
interretatons of science? The so-called 'paradoxes' of positvism,
however, which Lenin exposed with such consummate skill, did not
impress me at all. They arose only if the positvist and the realist
mode of speech were mixed and they exposed their diference. They
did not show that realism was better though the fact that realism came
with common sense gave the impression that it was.
Hollitscher never presented an argument that would lead, step by
step, from positvism into realism and he would have regarded the
attempt to produce such an argument as philosophical folly. He
rather developed the realist positon itself, illustrated it by examples
from science and common sense, showed how closely it was
connected with scientifc research and everyday acton and so
revealed its strength. It was of course always possible to tur a
realistc procedure into a positvistc procedure by a judicious use of
ad hoc hypotheses and ad hoc meaning changes and I did this
frequently, and without shame (in the Kraft Circle we had developed
such evasions into a fne art). Hollitscher did not raise semantc
points, or points of method, as a critcal ratonalist might have done,
he contnued to discuss concrete cases untl I felt rather foolish with
my abstact objectons. For I saw now how closely realism was
connected with facts, procedures, principles I valued and that it had
helped to brng the about while positivism merely dscribed the results
in a rather complicated way after they had been found: realism had
fruits, positvism had none. This at least is how I would speak today,
long afer my realist conversion. At the tme I became a realist not
because I was convinced by any particular argument, but because the
sum total of realism plus the arguments in favour of it plus the ease
with which it could be applied to science and many other things I
vaguely felt but could not lay a fnger on
5
fnally looked better to me
5. I remember that Reichenbach's answer to Dingler's account of relativity played
an imponant pan: Dingler extrapolated from what could be achieved by simple
mechanical operatons (manufacture of a Euclidean plain surface, for example) while
Reichenbach pointed out how the actual strcture of the world would modif the
results of these operations in the large. It is of course tre that Reichenbach's account
can be interpreted as a more efcient predictive machine and that it seemed impressive
to me only because I did not slide into such an interpretation. Which shows to what
extent the force of arguments depends on irratonal changes of attitude.
TWENTY 259
than the sum total of positvism plus the arguments one could ofer
for it plus . . . etc., etc. The comparison and the final decision had
much in common with the comparison of life in diferent counties
(weather, character of people, melodiousness of language, fod,
laws, insttutons, weather, etc., etc.) and the final decision to take a
job and to start life in one of them. Experiences such as these have
played a decisive role in my atttude towards ratonalism.
While I accepted realism I did not accept dialectcs and historical
materialism - my predilecton for abstact arguments (another
positvist hangover) was stll too strong for that. Today Stalin's rules
seem to me preferable by far to the complicated and epicycle-ridden
standards of our modem friends of reason.
From the ver begnning of our discussion Hollitscher made it
clear that he was a communist and that he would ty to convince me of
the intellectual and social advantages of dialectcal and historical
materialism. There was none of the mealy-mouthed 'I may be wrong,
you may be right-but together we shall find the tuth' talk with which
'critcal' ratonalists embroider their attempts at indoctinaton but
which they forget the moment their positon is seriously endangered.
Nor did Hollitscher use unfair emotonal or intellectual pressures.
Of course, he critcized my atttude but our personal relatons have
not suffered from my reluctance to follow him in ever respect. This
is why Walter Hollitscher is a teacher while Popper, whom I also
came to know quite well, is a mere propagandist.
At some point of our acquaintance Hollitscher asked me whether I
would like to become a producton assistant of Brecht - apparently
there was a positon available and I was being considered for it. I
declined. For a while I thought that this was one of the biggest
mistakes of my life. Enriching and changing knowledge, emotons,
atttudes through the ars now seems to me a much more fruitful
enterprise and also much more humane than the attempt to infuence
minds (and nothing else) by words (and nothing else). Reading about
the tensions inside the Brecht Circle, the almost religous atttude of
some of its members, I now think that I escaped just in tme.
(6) During a lecture (on Descartes) I gave at the Austian College
Society I met Elibeth Anscombe, a powerful and, to some people,
forbidding Britsh philosopher who had come to Vienna to lear
German for her tanslaton of Wittgenstein's works. She gave me
manuscripts ofWittgenstein's later writngs and discussed them with
me. The discussions extended over months and occasionally
proeeded from moring over lunch untl late into the evening. They
had a profound infuence upon me though it is not at all easy to
specif partculars. On one occasion which I remember vividly
260 AGAI NST METI I OD
Anscombe, by a series of skilful questons, made me see how our
concepton (and even our perceptons) of well-defned and apparently
self-contained facts may depend on circumstances not apparent in
them. There are enttes such as physical objects which obey a
'conseraton principle' in the sense that they retain their identty
through a variety of manifestatons and even when they are not present
at all while other enttes such as pains and after-images are
'annihilated' with their disappearance. The conseraton principles
may change from one developmental stage of the human organism to
another
6
and they may be different for different langages (cf. Whorfs
'covert classifcatons' as described in Chapter 1 6). I conjectured that
such principles would play an important role in science, that they might
change during revolutons and that deductive relatons between pre
revolutonar and post-revolutonartheories might be broken of as a
result. I explained this early version ofincommensurability in Popper's
seminar (1 952) and to a small group of people in Anscom be's flat in
Oxford (also in 1 952 with Geach, von Wright and L.L. Hart present)
but I was not able to arouse much enthusiasm on either occasion.
Wittgenstein's emphasis on the need for concrete research and his
objectons to abstact reasoning ('Look, don't think!') somewhat
clashed with my own tendency towards abstractess and the papers in
which his infuence is notceable are therefore mixtures of concrete
examples and sweeping principles.
7
Wittgenstein was prepared to take
me on as a student in Cambridge but he died before I arrived. Popper
became my supervisor instead.
(7) I had met Poper in Alpbach in 1
9
48. I admired his freedom of
manners, his cheek, his disrespectful atttude towards the German
philosophers who gave the proceedings weight in more senses than
one, his sense of humour (yes, the relatvely unknown Karl Popper of
1 948 was ver different from the established Sir Karl of later years)
and I also admired his ability to restate ponderous problems in simple
and jouralistc langage. Here was a free mind, joyfully puttng
forth his ideas, unconcered about the reacton of the 'professionals. '
Things were different as regards these ideas themselves. The
members of our circle knew deductvism from Kraft who had written
about it before Popper,
8
the falsifcatonist philosophy was taken for
6. Cf. Chapter 16, text to footnotes 1 2f.
7. For details cf. my comments on these papers in De ittscha/tthwrctitchc
RcalitmataaddicZaterittduttcha/,Vieweg Wiesbaden, 1 978.
8. Cf. my review of Kraft's Erkataulchrcin F5,Vol. 1 3, 1 963, pp. 319f. and
esp. p. 3 21 , second paragraph. Cf. also the references in Popper, Legc e/5otc
Ducq. Mill's 5tt e/Legc,Vol. 2, London, 1879, Chapter 14, gives a detailed
account of the proedure.
TWENTY 261
granted in the physics semnar of the conference under the
chairmanship of Arthur March and so we did not understand what all
the fuss was about. 'Philosophy must be in a desperate state', we said,
'if tivialites such as these can count as major discoveries.' Popper
himself did not seem to think too much ofhis philosophy of science at
the tme for when asked to send us a list of publicatons he included
the Oe Socet but not the Logic ofScetic Dice.
While in London I read Wittgenstein's Philsohical Iretigations
in detail. Being of a rather pedantc tum of mnd I rewrote the
book so that it looked more like a teatse with a contnuous
argument. Part of this teatse was tanslated by Anscombe into
English and published as a review by Philosohicl Reie in 1955. I
also visited Popper's semnar at the LSE. Popper's ideas were simlar
to those of Wittgenstein but they were more abstact and anaemic.
This did not deter me but increased my own tendencies to
abstacton and dogmatsm. At the end of my stay in London Popper
invited me to become his assistant. I declined despite the fact that
I was broke and did not know where my next meal was going to
come from. My decision was not based on any clearly recogniz
able tain of thought but I guess that having no fxed philosophy I
preferred stumbling around in the world of ideas at my own speed
to being guided by the ritual of a 'ratonal debate'. Again I was
lucky. Joseph Agassi who got the job did not have much privacy.
Two years later Popper, Schrodinger and my own big mouth got
me a job in Bristol where I started lecturing on the phiosophy of
science.
(8) I had studied theate, history, mathematcs, physics and
astonomy; I had never studied philosophy. The prospect ofhaving to
address a large audience of eager young people did not exactly fll my
heart with joy. One week before the lectures started I sat down and
wrote everg I knew on a piece of paper. It hardly flled a page.
Agassi came up with some excellent advice: 'Lok, Paul,' he said, 'the
first line, this is your frst lecture; the second line, this is your second
lecture - and so on.' I tok his advice and fared rather well except
that my lectures became a stale collecton of wisecracks from
Wittgenstein, Bohr, Popper, Dingler, Eddington and others. Wile
in Bristol I contnued my studies of the quantum theory. I found that
important physical principles rested on methodological assumptons
that are violated whenever physics advances: physics gets authority
from ideas it propagates but never obeys in actual research,
methodologists play the role of publicity agents whom physicists hire
to praise their results but whom they would not permit access to the
enterprise itself. That falsifcatonism is not a soluton became ver
262 AGAI NST METHOD
clear in discussions with David Bohm who gave a Hegelian account
of the relaton between theories, their evidence, and their
successors.
9
The material of Chapter 3 is the result of these
discussions (I frst published it in 1961).
10
Kuhn's remarks on the
omnipresence of anomalies ftted these difcultes rather nicely
1 1
but I stll tried to find general rules that would cover all cases
1 2
and
non-scientfc developments as well.
1 3
Two events made me realize
the futlity of such attempts. One was a discussion with Professor
C. F. von Weizsicker in Hamburg (1 965) on the foundatons of the
quantum theor. Von Weizsicker showed how quantum mechanics
arose from concrete research while I complained, on general
methodological grounds, that important alteratves had been
omitted. The argments supportng my complaint were quite good -
they are the argments summarized in Chapter 3 - but it was
suddenly clear to me that imposed without regard to circumstances
they were a hindrance rather than a help: a person tring to solve a
problem whether in science or elsewhere mut be gen complete
feedm and cannot be resticted by any demands, norms, however
plausible they may seem to the logician or the philosopher who has
thought them out in the privacy of his study. Norms and demands
must be checked hi research, not by appeal to theories of rationality.
In a lengthy artcle 4 I explained how Bohr had used this philosophy
and how it differs from more abstract procedures. Thus Professor
von W eizsicker has prime responsibility for my change to
'anarchism' - though he was not at all pleased when I told him so
in 1 977.
9. I have explained the Hegelianism of Bohm in the essay ' Against Method' which
appeared in Vol. 4 of the Miaaceta5tadic/rthcFhiles@he/5oce, 1970.
I 0. Popper once remarked (in a discussion at the Minnesota Center for the
Philosophy of Science in the year 1962) that the example of Brownian motion is just
another version of Duhem's example (conflict between specifc laws such as Kepler's
laws and general theories such as Newton's theor). But there is a most impnant
diference. The deviations from Kepler's laws are in principle obserable ('in
principle' meaning 'given the known laws of nature') while the microscopic deviations
from the second law of thermodynamics are not (measuring instruments are subjected
to the same fluctuatons as the things they are supposed to measure). Here we maaet
do without an alterative theor. Cf. Chapter 4, fn. 2.
I I . I read Kuhn's book in manuscript in 1960 and discussed it extensively with
Kuhn.
1 2. Cf. the account in 'Reply to Criticism', estea5tadic,Vol. 2, 1965.
13. Cf. 'On the Improvement of the Sciences and the Ans and the Possible
Identty of the Two' in estea5tadic,Vol. 3, 1 967.
14. 'On a Recent Critque of Complementarity', Fhiles@h e/5oce 1968/69
(two pans).
TWENTY 263
(9) The second event that prompted me to move away from
ratonalism and to become suspicious of all intellectual pretensions
was quite diferent. To explain it, let me start with some general
obseratons. The way in which social problems, problems of energ
distributon, ecolog, educaton, care for the old and so on are
'solved' in First World societes can be roughly described in the
following way. A problem arises. Nothing is done about it. People get
concered. Politcians broadcast this concer. Experts are called in.
They develop theories and plans based on them. Power-groups with
experts of their own efect various modifcatons untl a watered down
version is accepted and realized. The role of experts in this proess
has gradually increased. We have now a situaton where soial and
psychological theore of human thought and acton have taken the
place of this thought and acton itself. Instead of asking the people
involved in a problematc situaton, developers, educators, tech
nologists and soiologists get their informaton about 'what these
people really want and need' from theoretcal studies carried out by
their esteemed colleagues in what they think are the relevant felds.
Not live human beings, but abstract models are consulted; not the
target populaton decides, but the producers of the models.
Intellectuals all over the world take it for granted that their models
will be more intelligent, make better suggestons, have a better grasp
of the reality ofhumans than these humans themselves. What has this
situaton got to do with me?
From 1 958 to 1 990 I was a Professor of Philosophy at the
University of Califoria in Berkeley. My functon was to carr out the
educatonal policies of the State of Califora which means I had to
teach people what a small group of white intellectuals had decided
was knowledge. I hardly ever thought about this functon and I would
not have taken it ver seriously had I been informed. I told the
students what I had leared, I arranged the material in a way that
seemed plausible and interestng to me - and that was all I did. Of
course, I had also some 'ideas of my own' -but these ideas moved in a
fairly narrow domain (though some of my friends said even then that I
was going batt).
In the years around 1 964 Mexcans, blacks, Indians entered the
university as a result of new educatonal policies. There they sat,
partly curous, partly disdainful, partly simply confused hoping to get
an 'educaton'. What an opportunity for a prophet in search of a
following! What an opportunity, my ratonalist frends told me, to
contibute to the spreading of reason and the improvement of
mankind! What a marellous opportunity for a new wave of
enlightenment! I felt ver diferently. For it now dawned on me that
264 AGAI NST METHOD
the intricate arguments and the wonderful stories I had so far told to
my more or less sophistcated audience might just be dreams,
refectons of the conceit of a small group who had succeeded in
enslaving everyone else with their ideas. Who was I to tell these
people what and how to think? I did not know their problems though I
knew they had many. I was not familiar with their interests, their
feelings, their fears though I knew that they were eager to lear. Were
the arid sophistcatons which philosophers had managed to
accumulate over the ages and which liberals had surrounded with
schmaltzy phrases to make them palatable the right thing to ofer to
people who had been robbed of their land, their culture, their dignity
and who were now supposed first to absorb and then to repeat the
anaemic ideas of the mouthpieces of their oh so human captors?
They wanted to know, they wanted to lear, they wanted to
understand the strange world around them - did they not deserve
better nourishment? Their ancestors had developed cultures of their
own, colourful languages, harmonious views of the relaton between
people, and between people and nature whose remnants are a living
critcism of the tendencies of separaton, analysis, self-centredness
inherent in Wester thought. These cultures have important
achievements in what is today called sociolog, psycholog,
medicine, they express ideals of life and possibilites of human
existence. Yet the were neer eamined with the respea the deed
except by a small number of outsiders; they were ridiculed and
replaced as a matter of course first by the religion of brotherly love
and then by the religion of science or else they were defused by a
variety of'interretatons'. Now there was much talk ofliberaton, of
racial equality - but what did it mean? Did it mean the equality of
these traditons and the traditons of the white man? It did not.
Equality meant that the members of diferent races and cultures now
had the wonderful chance to partcipate in the white man's manias,
they had the chance to participate in his science, his technolog, his
medicine, his politcs. These were the thoughts that went through my
head as I looked at my audience and they made me recoil in revulsion
and terror from the task I was supposed to perform. For the task-this
now became clear to me - was that of a very refined, very
sophistcated slavedriver. And a slavedriver I did not want to be.
Experiences such as these convinced me that intellectual
procedures which approach a problem through concepts are on the
wrong tack and I became interested in the reasons for the
temendous power this error has now over minds. I started examining
the rise of intellectualism in Ancient Greece and the causes that
brought it about. I wanted to know what it is that makes people who
TWENTY 265
have a rich and complex culture fall for dr abstactons and multlate
their traditons, their thought, their language so that they can
accommodate the abstractons. I wanted to know how intellectuals
manage to get away with murder - for it is murder, murder of minds
and cultures that is committed year in year out at schools,
universites, educatonal missions in foreign counties. The tend
must be reversed, I thought, we must start learing from those we
have enslaved for they have much to ofer and, at any rate, they have
the right to live as they see ft even if they are not as pushy about their
rights and their views as their Wester conquerors have always been.
In
1 9
6
4-5 when these ideas frst ocurred to me I tied to fnd an
intel/eaual soluton to my misgivings, that is, I took it for granted that
it was up to me and the likes of me to devise educatonal policies for
other people. I envisaged a new kind of educaton that would live
from a rich reseroir of diferent points of view permitng the choice
of taditons most advantageous to the individual. The teacher's task
would consist in facilitatng the choice, not in replacing it by some
'tuth' of his own. Such a reseroir, I thought, would have much in
common with a theatre of ideas as imagined by Piscator and Brecht
and it would lead to the development of a great variety of means of
presentaton. The 'objectve' scientfc account would be one way of
presentng a case, a play another way (remember that for Arstotle
tagedy is 'more philosophical' than histor because it reveals the
straure of the historical proess and not only its accidental details), a
novel stll another way. Why should knowledge be shown in the
garment of academic prose and reasoning? Had not Plato obsered
that written sentences in a book are but tansitor stages of a complex
process of growth that contains gestures, jokes, asides, emotons and
had he not tied to catch this proess by means of the dialogue? And
were there not diferent forms of knowledge, some much more
detailed and realistc than what arose as 'ratonalism' in the 7th and
6
th centur in Greece? Then there was Daism. I had studied
Dadaism after the Second World War. What attracted me to this
movement was the style its inventors used when not engaged in
Dadaistc actvites. It was clear, luminous, simple without being
banal, precise without being narrow; it was a style adapted to the
expression of thought as well as of emoton. I connected this style
with the Dadaistc exercises themselves. Assume you tear language
apart, you live for days and weeks in a world of cacophonic sounds,
jumbled words, nonsensical events. Then, after this preparaton, you
sit
down and write: 'the cat is on the mat'. This simple sentence
which we usually utter without thought, like talking machines (and
much of our talk is indeed routne), now seems like the creaton of an
266 AGAI NST METHOD
entre world: God said let there be light, and there was light. Nobody
in modem tes has understood the miracle oflanguage and thought
as well as the Dadaists for nobody has been able to imagine, let alone
create, a world in which they play no role. Having discovered the
nature of a living ordr, of a reason that is not merely mechanical, the
Dadaists soon notced the deterioraton of such an order into routne.
They diagnosed the deterioraton oflanguage that preceded the First
World War and created the mentality that made it possible. Afer the
diagnosis their exercises assumed another, more sinister meaning.
They revealed the frightening similarity between the language of the
foremost commercial tavellers in 'imporance', the language of
philosophers, politcians, theologians, and brute inartculaton. The
praise of honour, patiotsm, truth, ratonality, honesty that fills our
schols, pulpits, politcal meetngs impercetibl mere into inaricl
tion no matter how much it has been wrapped into literar language
and no matter how hard its authors t to copy the style of the classics,
and the authors themselves are in the end hardly distguishable
from a pack of gruntng pigs. Is there a way to prevent such
deterioraton? I thought there was. I thought that regarding all
achievements as tansitor, resticted and peonal and ever tuth as
ceated by our love for it and not as 'found' would prevent the
deterioraton of once promising fair-tales and I also thought that it
was necessar to develop a new philosophy or a new religion to give
substance to this unsystematc conjecture.
I now realize that these consideratons were just another example of
intellectualistc conceit and folly. It is conceited to assume that one
has solutons for people whose lives one does not share and whose
problems one does not know. It is foolish to assume that such an
exercise in distant humanitarianism will have efects pleasing to the
people concered. From the ver beginning of Wester Ratonalism
intellectuals have regarded themselves as teachers, the world as a
school and 'people' as obedient pupils. In Plato this is ver clear. The
same phenomenon occurs among Christans, Ratonalists, Fascists,
Marxsts. Marxists did not tr to lear from those they wanted to
liberate; they attacked each other about interpretatons, viewpoints,
evidence and took it for granted that the resultg intellectual hash
would make fine food for the natves (Bakunin was aware of the
doctinarian tendencies of contemporar Marxsm and he intended
to retur all power - power over ideas included - to the people
immediately concered). My own view difered from those just
mentoned but it was stll a view, an abstact fancy I had invented and
now tied to sell without having shared even an ounce of the lives of
TWENTY 267
the receivers. This I now regard as insufferable conceit. So - what
remains?
Two things remain. I could follow my own advce to address and
tr to influence only those people whom I think I understand on a
personal basis. This includes some of my friends; it may include
philosophers I have not met but who seem to be interested in similar
problems and who are not too upset by my style and my general
approach. It may also include people from different cultures who are
attracted, even fascinated by Wester science and Wester
intellectual life, who have started partcipatng in it but who stll
remember, in thought as well as in feeling the life of the culture they
left behind. My account might lessen the emotonal tension they are
liable to feel and make them see a way of unitng, rather than
opposing to each other, the various stages of their lives.
Another possibility is a change of subject. I started my career as a
student of actng, theatre producton and singing at the Insttute for
the Methodological Reformaton of the German Theatre in the
German Democratc Republic. This appealed to my intellectualism
and my dramatc propensites. My intellectualism told me that
problems had to be solved by thought. My dramatc propensites
made me think that hamming it up was better than going through an
abstract argument. There is of course no conflict here for argument
without illustaton leads away from the human elements which affect
the most abstract problems. The arts, as I see them today, are not a
domain separated from abstract thought, but complementar to it
and needed to fully realize its potental. Examining this functon of
the arts and trying to establish a mode of research that unites their
power with that of science and religion seems to be a fascinatng
enterrise and one to which I might devote a year (or two, or
three . . . ).
Postscript on Relativism
In a critcal notce of my book Farewel to Reaon Andrew Lugg
suggests 'that Feyerabend and likeminded social critcs should teat
relatvism with the disdain that they normally resere for ratonal
ism'.
1
This I have now done, in Three Dialog ofKnorledge,
Z
where I say that relatvsm gives an excellent account of the relaton
beteen dogmatc world-vews but is only a frst step towards an
understanding of live taditons, and in Beond Reaon: Essas on the
Philsohy ofPaul K Feeabend, where I write that ' relatvism is as
much of a chimaera as absolutsm [the idea that there exst an
objectve tuth], its cantankerous t'.3 In the same book I call my
earlier advice to keep hands of taditons an 'idioy'.
4
In both cases I
raise objectons against relatvsm, indicate why I changed my mind.
and menton some of the remaining difcultes.
Adrew Lugg adds that my 'commitent to relatvism as a general
theor (or principled outlook) is considerably less than total and [that
I] can plausibly be read as arguing that the touble with taditonal
versions of relatvism is that they are pitched at to high a level of
abstacton'.
5
This is certainly tue of what I say in Farewel - but
antcipatons (which I notce only now, as a result of Lugg's
comments) occur already in Sciece in a Free Societ.
6
There I
distnguish between partcipants and exteral obserers of taditons,
describe objectvism as an illusion created by the special positon of
the former and summarize my arguments in a series of theses, all of
them printed in italics. Thesis i reads: Traditons are neither good
nor bad, they simply are. Thesis ii: A taditon assumes desirable or
I . Caa. mme/Fhih@,Vol. 21 , 199l, p. l l6 - received 1 989.
2. Oford, 1 991 , pp. 1 5 1 f. (MS finished 1 989/90.)
3. Drecht, 1 991 , p. 51 5. (MS finished 1 989.)
4. ibid., p. 50.
5. Jo. cit.
6. London, 1978, par I , secton 2, pp. 27f- reprnted without change in Chapter
1 7 of the scond editon of ZgaimtMnhm,London, 1988, and with added comment in
Chapter 1 7, p. 225f of te presnt editon.
268
POSTSCRI PT ON RELATI VI SM 269
undesirable properties only when compared with some traditon, i.e.
only when viewed by participants who see the world in terms of their
own values. And so on. This sounds like Protagoras, and I say so, in
thesis iii. However, I then describe (theses v and vi) how traditons
interact. I discuss two possibilites, a guided exchange and an open
exchange. A guided exchange adopts 'a well-specifed traditon and
accept[ s] only those responses that correspond to its standards. If one
party has not yet become a partcipant . . . he will be badgered,
persuaded, 'educated' untl he does - and then the exchange begins.'
'A rational debate', I contnue, 'is a special case of a guided exchange.'
In the case of an open exchange 'the participants get immersed into
each other's ways of thinking, feeling, perceiving to such an extent
that their ideas, perceptons, world-views may be entrely changed
they become diferent people partcipatng in a new and different
traditon. An open exchange respects the parter whether he is an
individual or an entre culture, while a ratonal exchange promises
respect only within the framework of a ratonal debate. An open
exchange has no organon though it may invent one; there is no
logic though new forms of logic may emerge in its course.' In sum,
an open exchange is part of an as yet unspecifed and unspecifiable
practce.
These comments imply, first, that traditons are rarely well defined
(open exchanges are going on all the tme) and, secondly, that their
interactions cannot be understood in general terms. Keeping
traditons alive i the face of exteral influences we act i an only
partly conscious way. We can describe results after they have
ocurred, we cannot incorporate them into a lastng theoretcal
structure (such as relatvism). In other words, there cannot be any
theor of knowledge (except as part of a special and fairly stable
traditon), there can at most be a (rather incomplete) histor of the
ways in which knowledge has changed in the past. In my next book I
shall discuss some episodes of such a histor.
In the meantme I have started using the term 'relatvism' again,
but in a new sense. In the second editon of the present book I
explained this sense by saying that 'Scientists [and, for that matter, all
members of relatvely uniform cultures] are sculptors of reality.'7
That sounds like the stong programme of the sociolog of science
except that sculptors are resticted by the propertes of the material
they use. Similarly individuals, professional groups, cultures can
create a wide variety of surroundings, or 'realites' - but not all
7.
op. cit., p. 270. Cf. also the more detailed account in 'Realism and the
Hist
oricity of Knowledge' ,eamle/Fhiles@h,1 989.
270 AGAI NST METHOD
approaches succeed: some cultures thrive, others linger for a while
and then decay. Even an 'objectve' enterprise like science which
apparently reveals Nature As She Is In Herself intervenes,
eliminates, enlarges, produces and codifes the results in a severely
standardied way - but again there is no guarantee that the results
will congeal into a unifed world. Thus all we apprehed when
experimentng, or interfering in less systematc ways, or simply livng
as part of a well-developed culture is how what surrounds us repond
to our actons (thoughts, observatons, etc.); we d not aprehed these
surounding thesele: Culture and Nature (or Being, to use
a more general term) are always entangled in a fashion that can be
explored only by entering into further and even more complicated
entanglements.
Now, considering that scientsts use diferent and often contradic
tor methods of research (I describe some of them in Chapter 1 9 of
the present editon), that most of these methods are successful and
that numerous non-scientfc ways of life not only survived but
protected and enriched their inhabitants we have to conclude that
Being responds diferently, and positiel, to many diferent
approaches. Being is like a person who shows a friendly face to a
friendly visitor, becomes angr at an angr gesture, remains unmoved
by a bore without giving any hint as to the principles that make Him
(Her? It? Them?) act the way they do in the diferent circumstances.
What we fnd when livng, experimentng, doing research is therefore
not a single scenario called 'the world' or 'being' or 'reality' but a
variety of responses, each of them consttutng a special (and not
always well-defned) reality for those who have called it forth. This is
relatvism because the type of reality encountered depends on the
approach taken. However, it difers from the philosophical dotrine
by admitng failure: not ever approach succeeds. In my reply to
critcs
8
I called this form of relatvism 'cosmological' relatvism, in
an artcle published in lrd9 I spoke of an 'ontological' relatvsm, in
'Nature as a Work of Art'
1 0
I argued that the world of modem
science (and not only the descripton of this world) is an artwork
constucted by generatons of artsan/ scientsts while in 'Realism
and the Historicity of Knowledge'
1 1
I indicated how such views are
related to the ideas of Niels Bohr. In the last artcle I also mentoned
8. In Gonzlo Munevar (ed.), 8qdRemea,Dorecht-BostonLondon, 1991, p.
570.
9. No. 8, n.s., Jan.-Apr. 192.
I 0. CemmKaemledge,Vol. I, No. 3, 1993.
I I . op. cit., fotote 7 abve.
POSTSCRI PT ON RELATI VI SM
271
that ontological relatvsm might be similar to Thomas Kuhn's more
recent philosophy.
Havng before me a copy of Kuhn's Robert and Maurine
Rothschild Distnguished Lecture of 1 9 November 1991 I can now
describe the similarites and the differences in greater detail.
We both oppose the stong programme in the sociolog of science.
As a matter of fact I would say, exactly as Kuhn does, that 'the claims
of the stong programme' are 'absurd: an example of deconstucton
gone mad'. I also agree that it is not enough to undermine the
authority of the sciences by historical arguments: why should the
authority of history be greater than that of, say, physics? All we can
show historically is that a geeal appeal to scientfc authority rns
into contadictons. That undermines any such appeal; however, it
does not tell us how science should now be interpreted or used.
(Such questons, I would say, have to be answered by the interested
partes themselves, according to their standards, conceptons,
cultural commitents.)
Kuhn says that 'the difcultes that have seemed to undermine the
authority of science should not be simply seen as obsered facts about
its practce. Rather they are necessary characteristcs of any
developmental or evolutonary process.' But how do we know that
science is an evolutonary process rather than a statc way of fnding
more facts and better laws? Either from 'obsered facts about its
practce' or from interretatons that are imposed from the outside.
In the frst case we are back at the situaton Kuhn wants to overcome
while the second case means that science is being incorporated into a
wider (cultural) context -a context that values developments -and is
interpreted accordingly (the procedure I mentoned in parentheses
above). It seems that is what Kuhn really wants, i.e. he wants to settle
the queston phlosophically, not by appealing to facts. I would agree
if I knew that for him this is one way among many and not the only
possible procedure.
Summarizing his argument Kuhn makes three assertons. 'First,
the Archimedian platform, outside history, outside of tme and space,
is gone beyond recall.' Yes, and no. It is gone as a stucture that can
be described and yet shown to be independent of any descripton. It is
not gone as an unknown background of our exstence which affects
us but in a way which forever hides its essence. Nor is
Archimedianism gone as a possible approach. It would be the
politcally correct approach in a theocracy, for example.
Secondly, Kuhn says that in the absence of an Archimedian
platorm 'comparatve evaluaton is all there is'. That is of course tue
-and tvally so. Thirdly, he challenges the taditonal noton of tuth
272 AGAI NST METHOD
a correspondence to reality. ' I am not suggestng, let me emphasize,
that there is a reality which science fails to get at. My point is, rather,
that no sense can be made of the noton of a reality a it has ordinarily
functoned in the philosophy of science.' Here I agree with the
proviso that more metaphysical notons of reality (such a those
proposed by Pseudo Dionysius Areopagita) have not yet been
disposed of.
Let me repeat that the cultures that call forth a certain reality and
these realites themselves are never well defned. Cultures change,
they interact with other cultures and the indefniteness resultng
therefrom is reflected in their worlds. This is what makes
intercultural understanding and scientfc change possible: poten
tally ever culture is all cultures. We can of course imagine a world
where cultures are well defned and stictly separated and where
scientfc ters have fnally been nailed down. I such a world only
miracles or revelaton could reform our cosmolog.
Index
Albert, Leon Battsta 242
anamneis 71-3, 105
anarchism 49
anything goes 14, 231
epistemological 9
methodolog and 9-10, 1 2-13,
1 9, 262
naive 4 7-9, 231
politcal 1 2-13
Anaxmander 95, 1 85
Anscombe, Elizabeth 259-60
anthropolog 1 8890, 1 97
case study of quantum
tibe 1 901
anything gos see anarchism
appearances 1 95n, l 97-8, 203
see alo natural
interretatons
argument
emotons and 1617
incommensurability and I SO
value of 1 5-17, 64
see alo anarchism;
incommensurability
Aristotle
Copercan theor and 72-3,
79
cosmolog 44, 1 09-12, 135, 141,
233-4
dynamics and motion 34n, 1 21 ,
135
intuitve view ofhumans 1 24
knowledge & percepton 89
A, Halton 242, 245
art
archaic style and percepton
169-186
perspectve 185n, 1 87, 1 99-201,
203, 215
TheAssaye(Galileo) 39-40n, 80
astronomy
ancient 36n, 1 35, 136
Ptolemaic 1 35-6
mediaeval 135, 136
see alo Copercus; Galileo;
Newton; telescope
auxiliar sciences 51-2, 1 10,
1 1 3-14
Bacon, Francis 32n, 6, 1 1 2-13n,
1 1 7-18
Barow, Isaac 45-6
Bohm, David 262
Bohr, Niels 1 5n, 31-2, 40
Boltzmann, Ludwig 46
Brahe, Tycho 55, 135, 142
Brecht, Bertolt 9, 259
Brownian moton (Dr Robert
Brown) 27-8, 28n, 262n
Brno, Giordano 1 27
Burbidge, C. 241
Carap, Rudolph 1 19
Chambers, Robert 240
children
development of 15-17
perceptual stages 150, 1 67-8
Chalmers, Alan 76n
China
medicine 367
Church & Christanity 248
accepts proven science 132-3
atttudes 1 25-6, 1301
Galileo and 1 24, 1 25-34
273
274 I NDEX
clashing with accepted theories
see counterinducton
classical mechanics 467, 72-3,
1 13
clasifcatons 1 64-6
common sense
Copercanism and 103-4, 1 10,
1 12-13
see alo natural interretatons;
obseraton
communicaton
guidedvs open 45-6, 229
concept
archaicparatactc 1 706,
177-85, 1 99-203, 21 8-19
changes of 155
integal nature of 601, 205
totlitrianism of 1 65-9,
1 99-207
see alo ideas; counterinducton;
natural interretatons
consistenc conditon 24-7, 28-30,
50,234-5
Copercus, Nichola
acceptance of 96, 1 45-6
defes evidence 1 7, 35, 39,
51 , 60, 61
efectof'Revoluton' 135-6
Galileo's defence of 72, 77
methodolog 138-2
symbol of progress 1 1 621
theor detached from exerience
103-4, 1 1 0, 1 1 2-13
see alo Galilei, Galileo
cosmolog 238, 241 , 242
archaic 1 84-6, 1 88, 1 98-9, 239
Aristode and 4, 1 09-1 2, 135,
1 41
cultural perceptons and 1 7 5-6
fmite 233-4
need for re-evaluaton 1 13-15,
1 20
se alo astonomy; concepts
counterinducton
Copercan theor and 51-3,
77-80
tol for research 203, 61-4
crisis, theor of 145-6
critcal ratonalism 14 7, 1 5 1-8
see alo Popper, Karl; ratonalism
culture 3-4
Dadaism 265-6
Descares, Rene 54
DicureonMelhod 49n
dialectc 1 8
Diogenes ofSinope 62
discover 1-2, 467, 10, 1 1 7
see alo science; theories of science
Duhem, Piere 24
educaton 263-5
sience 2, 1 1-12, 1 61-2
Ehrenhaft, Felix 28, 4 7
Einstein, Alben
Brownian moton and 28
methodolog 101 1, 42-3n,
138n, 163, 239
theor 18, 4, 107, 242
electoynamics 4 7
empiricism 18, 20, 29, 72, 109-10,
1 17-1 8, 1 47
Aristode and 109-10
autonomyprnciple 29
epistemological illusion 73
epistemolog 9, 1 1, 1 2, 1 36
illusions 73, 1 38, 156
prejudices 1 967
essentalism see paratctc
aggregates
evaluaton of theories see
methoolog
Evans-Pritchard, E.E. 189
events, efects on sience 1617, 138
exerience see natural
interretatons; obseraton
exeriments see empircism;
facts
facts
autonomy principle 267
collecton & disover 77, 1445
selecton & suppression 201,
501 , 155
I NDEX 275
theoretcal nature 1 1 , 22,
39-0, 63
see alo counternducton;
empircism; natural
interretatons
fair-tales see myths & fairy-tales
faith see religion
f&sificaton 23, 145, 154, 221
science sufoated by 50-I, 1 54
see alo crtcal ratonalism;
Popper, Karl
Favaro, Antonio 1 27
Feigl, Herber 147-8, 212
Feller, Margaret 242, 245-
Frank, Philipp 257
foralism see logic
freedom 1 7, 228- 9
Galilei, Galileo 1 7-18, 39-0, 49,
54, 1 23
TeAssae 390n, 80
boat argment 65-7
counterinducton and 77-80
dynamics & mechanics 24 5
moonand 91-9
propaganda 65, 70-3, 76n,
104-5, 1 1 8
SieNunu 81-2
telescopeandoptcs 81-5,
88- 90, 99102
toer argment 540, 68,
74, 107
ti&s of 125-34
see alo Church; Copericus,
Nicholas
Giedymin,J. 192-3
Gottsched,j ohann 21 9
goverent and sience 163
grammar see language &
lingistcs
Greeks, ancient 264-5
ar & culture shape views 1 704,
1 77-87, 198- 202
astonomy & cosmolog 36n,
1 84 7, 197-202
see alo Aristotle; Plato;
Parenides
Hanson, N.R. 21 2
Hanfmann, G.M.S. 1 75
Heraclitus 95, 203-4
heretcwrtng 35, 141
Hesse, Mar 34
histor
conceptu& changes 1 72, 201
educaton of sience 1 1-12
evaluaton of theores 1-2, 32,
33-7, 107-8, 212
events of, efects on sience
1617
mater&ism 259
methodolog 910, 51-3, 147-8
taditons 225-, 228
world views 243-5
Hollitscher, W&ter 257-9
Homer 1 77-82, 203, 204
humanitarianism 3-4, 1 62, 228
Hubble, E.P. 239
Hume, David 50
hypnagogic h&lucinaton 1 82n
hypotheses 10
ahoc 141 5, 44, 49, 75
contadictor 1415, 20-3
see alo ideas; theores
idealism 222-3, 230-1
ideas
and acton 1 7, 149, 216, 232
comparison to other ideas 21
see alo concepts; hypotheses
ideolog 1 7-18, 62, 1 84
Ilia (Homer) 1 77-82, 203, 204
incommensurability 150, 1 65, 190
author's arval at thesis 21 1-13,
262-3
suspension of univer&s 205-7
inconsistency see consistency
theor; fact
individu&, development of 1 2
see alo freedom; humanitarianism
inductve logic see logic
instuments 1 7, 1 10, 232
see alo telesope
intuiton 1 1 , ISO
iratonalism 158, 204
276 I NDEX
Japan 250
justficaton 147-9
Kant, Immanuel 51 , 55-6, 56n
Kaufmann, W. 4
Kepler,Johann 24, 94n, 242
Otic 45, 82, 88, 9l n, 98n,
9910
polyopia of 88n
Kierkegaard, Seren 17, 1 54
knowledge 21
'America of 21 6
enumeraton vs undertanding
184-6
see alo epistemolog
Kraft Circle 2545
Kropotkin, Peter A. 13
Kuhn, Thomas 3 1 , 21 1 , 212-13
Lagella,J ulius Caesar 84
Lakatos, lmre 34n, 158n, 1 62-3
language & linguistcs
classificaton 1 645
deterioraton of 266
effectof 1 64, 172, 1 99
inventon of scientfic 1 1 , 18,
63-4, 71, 1 50, 1 93-4
obseraton 57, 59, 64, 71, 1 1 2
philosophy 206-7, 220
relatvity principle 208-9
tanslatng ideas & concepts
1889, 206-7, 209-10
law & order see und methoolog
learing see educaton
Lessing, Gotthold 219 20
Lenin, Vadimir I. 9n, 10
libert see freedom
linguistcs see language & linguistcs
Loewy, Emmanuel 171
logic
anthropolog and 1 904
contadicton of 1 5n, 194-6
falsifcaton 51
inductve 1 5, 152
limits speculaton 1 1 , 1 06, 192,
195
Lorent, Hendrik 46-7
Loren, Konrad 1 3l n, 245
Luria, S.E. 238-9, 242
Lysenko, Tromfim D. 37, 160
McMullin, Eran 68-9n, 91 n
Maestlin, Michael 135-, 142-4
Mar, Karl, & Marxsm 1 07n
mathematcs 50, 136, 1 95
Maell,] ames Clerk 46, 241
Medawar, Peter 13l n
medicine 36-8, 249
Melanchthon, Philip 135
mental sets 1 99, 202
Mer,Johann T. 243-4
metaphysics, science as 76, 78, 1 14,
1 21 , 154, 245
methodolog
as the actvity of science 239-42
anarchistc 910, 12-13, 19, 231
counterinducton 23, 5 1-3,
61-4
critcal ratonalism 14 7, 1 51-8
empiricism 18, 20, 29, 72,
10910, 1 1 7-1 8
falsifcaton 23, 51, 145, 155
law&order 13, 19
logic and IO, 1 1 , 1 5n, 106, 1 52,
1 97, 221
patence needed with status
quo 1 1415
pluralistc 23, 28, 32, 33, 108
preconceptons in 1-2, 53
problemsof 187-97
ratonal 12-13, 17, 147
scientfic progress and 1419,
157-8
see alo hypotheses; ratonalism;
science; theories of science
Mill,JohnStuart 1 19, 1 94
describes acceptance of theories
29 30, 3 1
OLibe 34n, 38
mindody problem 64, 123-4
model theor see theories of
science
morality 1 6, 253
see alo humanitarianism
I NDEX
277
moton
Galileo and 55-0, 65-76,
1 201
peretual 27-8, 28n
relatvit of 69-70n, 701
myth & fair-tales 21 , 36n, 53
natural interretatons 22, 59,
605, 74
see alo obseraton
naturalism 222-3, 231-2
Neumann, John von 49n
Neurath, Otto 149
Newton, Isaac 35, 4, 168
acceptance of theor 245
optcs 4-
Nuerpeoples 1 89-90
objectvity 21 7, 221-2, 225-, 228
see alo facts; mental sets
obseraton
astonomical vs terestal 845,
8692
Copercus moves sience
fom 103-5, 1 10, 1 1 2
historical changes 52
langage 57, 59, 64, 71 , 1 1 2
natural interretatons 22, 59,
60-5, 74
sensor 567, 234
theories and 39, 149, 1 51
see alo methoolog; percepton;
telescope
Odssq (Homer) 1 78
ontolog 1 55, 1 75-
Otic (Kepler) 45, 82, 88, 9l n,
98n, 99-10
optcs 24, 4-, 1 37-8, 242
natural obseraton 867
telescope and 82, 8698, 10-5
Oresme, Nicole 248
para tactc aggegates 169-7 5,
1 77-86, 199-202
Parenides 43-4, 62, 20
parcipant vs obserer 21 5-21
passion (Kierkegaardian) 1 7
percepton 95, 1 089, 1 14, 1 52
of archaic cultures 169-75
conceptual fraework and
1 648, 1 75-7
see alo natural interretatons;
obseraton; perpectve
peretual moton 27-8, 28n, 262n
Perr, Commander atthew 250
perpectve 1 85n, 1 87, 1 99-20,
202-3, 21 5, 242
Philosohi betigatins
(Wittgenstein) 261
philosophy of sience 1-4, 9, 148,
1 51 , 1 54, 1 967
'prgmatc' 21 7-18, 227-8
standards 233
see alo Bacon, Francis; Ku,
Thomas; theories of sience
physics 47, 73, 1 13
see alo moton; optcs
Piaget,Jean 167
Plato 71 , 15 1 , 245
pluralism 203, 29, 32, 33
Plutarch 96, 98, 20-5n
Popper, Sir Karl R. 34n, 42n, 147,
1 51-8
author's acquaintance with
260-1
Prandtl, L. 24
prejudice, recogniton 22-3, 58
presuppsitons 58
see alo facts: theoretcal nature
propaganda 161 7, 1 1 4
Galileo and 65, 703, 76n,
10-5, 1 1 8
Protagoras 226
Pseudo Dionysius Apropagita 248
psycholog see mental set;
percepton
Ptolemy (Claudius
Ptolemaeus) 135, 137-4
public parcipaton in sience 2, 21
Pythagoreans 35
quantum theor 18, 4
case study of 191
Quine, Willard van Oran 21 2
278 I NDEX
ratonalism 2, 62, 64, 1 1 4,
1 61. 246, 265
critcal 147, 15 1-8
falsifcaton 23, 51, 145, 1 54
'LawofReason' 12-13
nature of 15, 1 1 5-16, 2245
objectvity 217, 221-2, 225-6,
228
reconstucton 191
vs science 21415
standards within 230
realism 141-2, 144
naive 59-60, 71
reality assumptons 141-2
relatvity
Einstein's 1 8-19, 1 , 48-9
Galileo's 73-4, 1 1 8
religion 1 1 , 53, 137, 218
co-exstence with science
see alo Church & Christanity
research
see facts; empiricism;
methodolog
Sacrobosco,Johannes 1 36
Schwarchild soluton 48
science
anarchism 9-10, 12-13, 14,
19
auxliar 51-2, 1 10, 1 1 3-14
chauvinismin 2-4, 37
coherence vs subdivision 1368,
144
conseratvism in 1 1 , 106, 192
cultural influence 3-4
discover & success 2-4, 467,
106, 1 16, 148
educaton and 2, 1 1-12, 1 61-2
moneyand 37
standards 1, 12, 1 61-3, 214,
227-8, 231 , 232-3, 23842
world views 243-5
see alo histor; methodolog;
philosophy of science; theories
of science
senses see und natural
interretatons; obseraton
Sies Nuncius (Galileo) 81-2
speculaton see hypotheses; ideas;
metaphysics
Stalin,Joseph 257
state see goverent
status quo 1 1415
Te Straure ofScietic Reolutions
Quhn) 31 , 212-13
Swiss cheese theor 262
technolog 263
telescope
celestal dificultes 8691
Galileo and 81-5, 120
senses and 93-102, 103-4
terestrial success 83-4
see alo optcs
tests see empiricism; methodolog
Thales 1 85n
theolog see Church & Christanity;
religion
theories of science
acceptance of 17-18, 32, 63,
77
changing of perceptons 23,
52-3, 1 23-4
consistenc conditons 247,
28-30, 50, 234-5
historical view 1-2, 32, 33-7
incommensurability 150, 1 65,
190, 205-7, 21 1-13
justfcaton 147-9
as metaphysics 76, 78, 1 14, 1 21 ,
154
numerical disagreement 39-42
qualitatve disagreement 39,
42-50
philosophical 1-4, 9, 3 1 , 148,
1 51 , 1 54, 196, 212-13
separaton from obseraton
.
14 7,
1 51-8
terinolog 1 1 , 18, 63-4, 71,
1 50, 202-3
world views 243-5
see alo concepts; histor;
hypotheses; methoolog;
science
traditon 225-6, 228
see alo histor
translaton 207
undertanding
see knowledge; percepton
USA, unfree soiety 229
value judgements 221 , 227
Vodoo and witchcraf 35-6, 78n
Weizacker, C.F. von 262
I NDEX
279
White, Lynn 249
Whorf, Benjamin L. 1 645, 1 76,
209, 21 0
Wilson, E.O. 247-8
Wittgenstein, Ludwig 224, 260
Philsohic betigations 261
world views, abstact 243-5
Xenophanes 95
Zahar, Elie 157n
Zeno 62