SCIENCE: CONJECTURES AND REFUTATIONS

KARL POPPER

This essay was first published as part of the book Conjectures and Refutations. See below for a short description of the book. Conjectures and Refutations: The Growth of Scientific Knowledge is a book written by philosopher Karl Popper. Published in 1963 by Routledge this book is a collection of his lectures and papers that su!!arised his thoughts on the philosophy of science. Popper suggested that all scientific theories are by nature con"ectures and inherently fallible and that refutation to old theory is the para!ount process of scientific disco#ery. Should any new theory sur#i#e !ore of such refutations it would ha#e a higher verisi ilitude and therefore Popper concluded closer to truth. $%ikipedia&

Page 1 of '(

)There could be no fairer destiny for any. . . theory than that it should point the way to a !ore co!prehensi#e theory in which it li#es on as a li!iting case.* +,-.RT ./0ST./0

)1r. Turnbull had predicted e#il conse2uences . . . and was now doing the best in his power to bring about the #erification of his own prophecies.* +0T3405 TR4,,4P. 66666666666666666666666666666 + lecture gi#en at Peterhouse 7a!bridge in Su!!er 19(3 as part of a course on 8e#elop!ents and trends in conte!porary -ritish philosophy organi9ed by the -ritish 7ouncil: originally published under the title ;Philosophy of Science< a Personal Report; in -ritish Philosophy in 1id=7entury ed. 7. +. 1ace 19(>. 66666666666666666666666666666

I
%3.0 / recei#ed the list of participants in this course and reali9ed that / had been asked to speak to philosophical colleagues / thought after so!e hesitation and consultation that you would probably prefer !e to speak about those proble!s which interest !e !ost and about those de#elop!ents with which / a! !ost inti!ately ac2uainted. / therefore decided to do what / ha#e ne#er done before< to gi#e you a report on !y own work in the philosophy of science since the autu!n of 1919 when / first began to grapple with the proble! ;%hen should a theory be ranked as scientific?; or ;/s there a criterion for the scientific character or status of a theory?;

The proble! which troubled !e at the ti!e was neither ;%hen is a theory true?; nor ;%hen is a theory acceptable?; 1y proble! was different. / wished to distinguish between science and pseudo=science: knowing #ery well that science often errs and that pseudo=science !ay happen to stu!ble on the truth.

Page @ of '(

/ knew of course the !ost widely accepted answer to !y proble!< that science is distinguished fro! pseudo=science==or fro! ;!etaphysics;==by its e!piricaa1 !ethod which is essentially inducti#e proceeding fro! obser#ation or eAperi!ent. -ut this did not satisfy !e. 4n the contrary / often for!ulated !y proble! as one of distinguishing between a genuinely e!pirical !ethod and a non=e!pirical or e#en a pseudo=e!pirical !ethod==that is to say a !ethod which although it appeals to obser#ation and eAperi!ent ne#ertheless does not co!e up to scientific standards. The latter !ethod !ay be eAe!plified by astrology with its stupendous !ass of e!pirical e#idence based on obser#ation==on horoscopes and on biographies.

-ut as it was not the eAa!ple of astrology which led !e to !y proble! / should perhaps briefly describe the at!osphere in which !y proble! arose and the eAa!ples by which it was sti!ulated. +fter the collapse of the +ustrian .!pire there had been a re#olution in +ustria< the air was full of re#olutionary slogans and ideas and new and often wild theories. +!ong the theories which interested !e .instein;s theory of relati#ity was no doubt by far the !ost i!portant. Three others were 1arA;s theory of history Breud;s psycho=analysis and +lfred +dler;s so= called ;indi#idual psychology;.

There was a lot of popular nonsense talked about these theories and especially about relati#ity $as still happens e#en today& but / was fortunate in those who introduced !e to the study of this theory. %e all==the s!all circle of students to which / belonged==were thrilled with the result of .ddington;s eclipse obser#ations which in 1919 brought the first i!portant confir!ation of .instein;s theory of gra#itation. /t was a great eAperience for us and one which had a lasting influence on !y intellectual de#elop!ent.

The three other theories / ha#e !entioned were also widely discussed a!ong students at that ti!e. / !yself happened to co!e into personal contact with +lfred +dler and e#en to co=operate with hi! in his social work a!ong the children and young people in the working=class districts

Page 3 of '(

of Cienna where he had established social guidance clinics.

/t was during the su!!er of 1919 that / began to feel !ore and !ore dissatisfied with these three theories=the 1arAist theory of history psychoanalysis and indi#idual psychology: and / began to feel dubious about their clai!s to scientific status. 1y proble! perhaps first took the si!ple for! ;%hat is wrong with 1arAis! Psycho=analysis and indi#idual psychology? %hy are they so different fro! physical theories fro! 0ewton;s theory and especially fro! the theory of relati#ity ?

To !ake this contrast clear / should eAplain that few of us at the ti!e would ha#e said that we belie#ed in the truth of .instein;s theory of gra#itation. This shows that it was not !y doubting the truth of those other three theories which bothered !e but so!ething else. 5et neither was it that / !erely felt !athe!atical physics to be !ore eAact than the sociological or psychological type of theory. Thus what worried !e was neither the proble! of truth at that stage at least nor the proble! of eAactness or !easurability. /t was rather that / felt that these other three theories though posing as sciences had in fact !ore in co!!on with pri!iti#e !yths than with science: that they rese!bled astrology rather than astrono!y.

/ found that those of !y friends who were ad!irers of 1arA Breud and +dler were i!pressed by a nu!ber of points co!!on to these theories and especially by their apparent eAplanatory power. These theories appeared to be able to eAplain practically e#erything that happened within the fields to which they referred. The study of any of the! see!ed to ha#e the effect of an intellectual con#ersion or re#elation opening your eyes to a new truth hidden fro! those not yet initiated. 4nce your eyes were thus opened you saw confir!ing instances e#erywhere< the world was full of #erifications of the theory. %hate#er happened always confir!ed it. Thus its truth appeared !anifest: and unbelie#ers were clearly people who did not want to see the

Page ' of '(

!anifest truth: who refused to see it either because it was against their class interest or because of their repressions which were still ;un=analysed; and crying aloud for treat!ent.

The !ost characteristic ele!ent in this situation see!ed to !e the incessant strea! of confir!ations of obser#ations which ;#erified; the theories in 2uestion: and this point was constantly e!phasi9ed by their adherents. + 1arAist could not open a newspaper without finding on e#ery page confir!ing e#idence for his interpretation of history: not only in the news but also in its presentation==which re#ealed the class bias of the paper==and especially of course in what the paper did not say. The Breudian analysts e!phasi9ed that their theories were constantly #erified by their ;clinical obser#ations;. +s for +dler / was !uch i!pressed by a personal eAperience. 4nce in 1919 / reported to hi! a case which to !e did not see! particularly +dlerian but which he found no difficulty in analysing in ter!s of his theory of inferiority feelings although he had not e#en seen the child. Slightly shocked / asked hi! how he could be so sure. ;-ecause of !y thousandfold eAperience ; he replied: whereupon / could not help saying< ;+nd with this new case / suppose your eAperience has beco!e thousand=and=one= fold.;

%hat / had in !ind was that his pre#ious obser#ations !ay not ha#e been !uch sounder than this new one: that each in its turn had been interpreted in the light of ;pre#ious eAperience; and at the sa!e ti!e counted as additional confir!ation. %hat / asked !yself did it confir!? 0o !ore than that a case could be interpreted in the light of the theory. -ut this !eant #ery little / reflected since e#ery concei#able case could be interpreted in the light of +dler;s theory or e2ually of Breud;s. / !ay illustrate this by two #ery different eAa!ples of hu!an beha#iour< that of a !an who pushes a child into the water with the intention of drowning it: and that of a !an who sacrifices his life in an atte!pt to sa#e the child. .ach of these two cases can be eAplained with e2ual ease in Breudian and in +dlerian ter!s. +ccording to Breud the

Page ( of '(

first !an suffered fro! repression $say of so!e co!ponent of his 4edipus co!pleA& while the second !an had achie#ed subli!ation. +ccording to +dler the first !an suffered fro! feelings of inferiority $producing perhaps the need to pro#e to hi!self that he dared to co!!it so!e cri!e& and so did the second !an $whose need was to pro#e to hi!self that he dared to rescue the child&. / could not think of any hu!an beha#iour which could not be interpreted in ter!s of either theory. /t was precisely this fact==that they always fitted that they were always confir!ed==which in the eyes of their ad!irers constituted the strongest argu!ent in fa#our of these theories. /t / began to dawn on !e that this apparent strength was in fact their weakness.

%ith .instein;s theory the situation was strikingly different. Take one typical instance== .instein;s prediction "ust then confir!ed by the findings of .ddington;s eApedition. .instein;s gra#itational theory had led to the result that light !ust be attracted by hea#y bodies $such as the sun& precisely as !aterial bodies were attracted. +s a conse2uence it could be calculated that light fro! a distant fiAed star whose apparent position was close to the sun would reach the earth fro! such a direction that the star would see! to be slightly shifted away fro! the sun: or in other words that stars close to the sun would look as if they had !o#ed a little away fro! the sun and fro! one another. This is a thing which cannot nor!ally be obser#ed since such stars are rendered in#isible in dayti!e by the sun;s o#erwhel!ing brightness: but during an eclipse it is possible to take photographs of the!. /f the sa!e constellation is photographed at night one can !easure the distances on the two photographs and check the predicted effect. 0ow the i!pressi#e thing about this case is the risk in#ol#ed in a prediction of this kind. /f obser#ation shows that the predicted effect is definitely absent then the theory is si!ply refuted. The theory is inco!patible with certain possible results of obser#ation==in fact with results which e#erybody before .instein would ha#e eApected.1 This is 2uite different fro! the situation / ha#e pre#iously described when it turned out that the theories in 2uestion were co!patible with the !ost di#ergent hu!an beha#iour so that it was practically i!possible to

Page 6 of '(

describe any hu!an beha#iour that !ight not be clai!ed to be a #erification of these theories. These considerations led !e in the winter of 1919=@D to conclusions which / !ay now refor!ulate as follows.

$1& /t is easy to obtain confir!ations or #erifications for nearly e#ery theory=if we look for confir!ations.

$@& 7onfir!ations should count only if they are the result of risky predictions: that is to say if unenlightened by the theory in 2uestion we should ha#e eApected an e#ent which was inco!patible with the theory==an e#ent which would ha#e refuted the theory.

$3& .#ery ;good; scientific theory is a prohibition< it forbids certain things to happen. The !ore a theory forbids the better it is.

$'& + theory which is not refutable by any concei#able e#ent is nonscientific. /rrefutability is not a #irtue of a theory $as people often think& but a #ice.

$(& .#ery genuine test of a theory is an atte!pt to falsify it or to refute it. Testability is falsifiability: but there are degrees of testability< so!e theories are !ore testable !ore eAposed to refutation than others: they take as it were greater risks.

$6& 7onfir!ing e#idence should not count eAcept when it is the result of a genuine test of the theory: and this !eans that it can be presented as a serious but unsuccessful atte!pt to falsify the theory. $/ now speak in such cases of ;corroborating e#idence;.&

$>& So!e genuinely testable theories when found to be false are still upheld by their

Page > of '(

ad!irers==for eAa!ple by introducing ad hoc so!e auAiliary assu!ption or by re=interpreting the theory ad hoc in such a way that it escapes refutation. Such a procedure is always possible but it rescues the theory fro! refutation only at the price of destroying or at least lowering its scientific status. $/ later described such a rescuing operation as a ;con#entionalist twist; or a ;con#entionalist stratage!;.&

4ne can su! up all this by saying that the criterion of the scientific status of a theory is its falsifiability or refutability or testability.

II
/ !ay perhaps eAe!plify this with the help of the #arious theories so far !entioned. .instein;s theory of gra#itation clearly satisfied the criterion of falsifiability. .#en if our !easuring instru!ents at the ti!e did not allow us to pronounce on the results of the tests with co!plete assurance there was clearly a possibility of refuting the theory.

+strology did not pass the test. +strologers were greatly i!pressed and !isled by what they belie#ed to be confir!ing e#idence==so !uch so that they were 2uite uni!pressed by any unfa#ourable e#idence. 1oreo#er by !aking their interpretations and prophecies sufficiently #ague they were able to eAplain away anything that !ight ha#e been a refutation of the theory had the theory and the prophecies been !ore precise. /n order to escape falsification they destroyed the testability of their theory. /t is a typical soothsayer;s trick to predict things so #aguely that the predictions can hardly fail< that they beco!e irrefutable.

The 1arAist theory of history in spite of the serious efforts of so!e of its founders and followers ulti!ately adopted this soothsaying practice. /n so!e of its earlier for!ulations $for eAa!ple in 1arA;s analysis of the character of the ;co!ing social re#olution;& their Page E of '(

predictions were testable and in fact falsified.@ 5et instead of accepting the refutations the followers of 1arA re=interpreted both the theory and the e#idence in order to !ake the! agree. /n this way they rescued the theory fro! refutation: but they did so at the price of adopting a de#ice which !ade it irrefutable. They thus ga#e a ;con#entionalist twist; to the theory: and by this stratage! they destroyed its !uch ad#ertised clai! to scientific status.

The two psycho=analytic theories were in a different class. They were si!ply non=testable irrefutable. There was no concei#able hu!an beha#iour which could contradict the!. This does not !ean that Breud and +dler were not seeing certain things correctly< / personally do not doubt that !uch of what they say is of considerable i!portance and !ay well play its part one day in a psychological science which is testable. -ut it does !ean that those ;clinical obser#ations; which analysts nai#ely belie#e confir! their theory cannot do this any !ore than the daily confir!ations which astrologers find in their practice& +nd as for Breud;s epic of the .go the Super=ego and the /d no substantially stronger clai! to scientific status can be !ade for it than for 3o!er;s collected stories fro! D1y!pus. These theories describe so!e facts but in the !anner of !yths. They contain !ost interesting psychological suggestions but not in a testable for!.

+t the sa!e ti!e / reali9ed that such !yths !ay be de#eloped and beco!e testable: that historically speaking all==or #ery nearly all==scientific theories originate fro! !yths and that a !yth !ay contain i!portant anticipations of scientific theories. .Aa!ples are .!pedocles; theory of e#olution by trial and error or Par!enides; !yth of the unchanging block uni#erse in which nothing e#er happens and which if we add another di!ension beco!es .instein;s block uni#erse $in which too nothing e#er happens since e#erything is four=di!ensionally speaking deter!ined and laid down fro! the beginning&. / thus felt that if a theory is found to be non= scientific or ;!etaphysical; $as we !ight say& it is not thereby found to be uni!portant or

Page 9 of '(

insignificant or ;!eaningless; or ;nonsensical;.' -ut it cannot clai! to be backed by e!pirical e#idence in the scientific sense==although it !ay easily be in so!e genetic sense the ;result of obser#ation;.

$There were a great !any other theories of this pre=scientific or pseudo=scientific character so!e of the! unfortunately as influential as the 1arAist interpretation of history: for eAa!ple the racialist interpretation of history another of those i!pressi#e and all=eAplanatory theories which act upon weak !inds like re#elations.&

Thus the proble! which / tried to sol#e by proposing the criterion of falsifiability was neither a proble! of !eaningfulness or significance nor a proble! of truth or acceptability. /t was the proble! of drawing a line $as well as this can be done& between the state!ents or syste!s of state!ents of the e!pirical sciences and all other state!ents==whether they are of a religious or of a !etaphysical character or si!ply pseudo=scientific. 5ears later==it !ust ha#e been in 19@E or 19@9==/ called this first proble! of !ine the ;proble! of de!arcation;. The criterion of falsifiability is a solution to this proble! of de!arcation for it says that state!ents or syste!s of state!ents in order to be ranked as scientific !ust be capable of conflicting with possible or concei#able obser#ations.

III
Today / know of course that this criterion of de!arcation==the criterion of testability or falsifiability or refutability==is far fro! ob#ious: for e#en now its significance is seldo! reali9ed. +t that ti!e in 19@D it see!ed to !e al!ost tri#ial although it sol#ed for !e an intellectual proble! which had worried !e deeply and one which also had ob#ious practical conse2uences $for eAa!ple political ones&. -ut / did not yet reali9e its full i!plications or its philosophical significance. %hen / eAplained it to a fellow student of the 1athe!atics Page 1D of '(

8epart!ent $now a distinguished !athe!atician in Freat -ritain& he suggested that / should publish it. +t the ti!e / thought this absurd: for / was con#inced that !y proble! since it was so i!portant for !e !ust ha#e agitated !any scientists and philosophers who would surely ha#e reached !y rather ob#ious solution. That this was not the case< / learnt fro! %ittgenstein;s work and fro! its reception: and so / published !y results thirteen years later in the for! of a criticis! of %ittgenstein;s criterion of !eaningfulness.

%ittgenstein as you all know tried to show in the Tractatus $see for eAa!ple his propositions 6.(3: 6.(': and (& that all so=called philosophical or !etaphysical propositions were actually non=propositions or pseudo=propositions< that they were senseless or !eaningless. +ll genuine $or !eaningful& propositions were truth functions of the ele!entary or ato!ic propositions which described ;ato!ic facts; i.e.==facts which can in principle be ascertained by obser#ation. /n other words !eaningful propositions were fully reducible to ele!entary or ato!ic propositions which were si!ple state!ents describing possible states of affairs and which could in principle be established or re"ected by obser#ation. /f we call a state!ent an ;obser#ation state!ent; not only if it states an actual obser#ation but also if it states anything that !ay be obser#ed we shall ha#e to say $according to the Tractatus ( and '.(@& that e#ery genuine proposition !ust be a truth=function of and therefore deducible fro! obser#ation state!ents. +ll other apparent propositions will be !eaningless pseudo=propositions: in fact they will be nothing but nonsensical gibberish.

This idea was used by %ittgenstein for a characteri9ation of science as opposed to philosophy. %e read $for eAa!ple in '.11 where natural science is taken to stand in opposition to philosophy&< ;The totality of true propositions is the total natural science $or the totality of the natural sciences&.; This !eans that the propositions which belong to science are those deducible fro! true obser#ation state!ents: they are those propositions which can be #erified by

Page 11 of '(

true obser#ation state!ents. 7ould we know all true obser#ation state!ents we should also know all that !ay be asserted by natural science.

This a!ounts to a crude #erifiability criterion of de!arcation. To !ake it slightly less crude it could be a!ended thus< ;The state!ents which !ay possibly fall within the pro#ince of science are those which !ay possibly be #erified by obser#ation state!ents: and these state!ents again coincide with the class of all genuine or !eaningful state!ents.; Bor this approach then #erifiability !eaningfulness and scientific character all coincide.

/ personally was ne#er interested in the so=called proble! of !eaning: on the contrary it appeared to !e a #erbal proble! a typical pseudo=proble!. / was interested only in the proble! of de!arcation i.e. in finding a criterion of the scientific character of theories. /t was "ust this interest which !ade !e see at once that %ittgenstein;s #erifiability criterion of !eaning was intended to play the part of a criterion of de!arcation as well: and which !ade !e see that as such it was totally inade2uate e#en if all !isgi#ings about the dubious concept of !eaning were set aside. Bor %ittgenstein;s criterion of de!arcation==to use !y own ter!inology in this conteAt==is #erifiability or deducibility fro! obser#ation state!ents. -ut this criterion is too narrow $and too wide&< it eAcludes fro! science practically e#erything that is in fact characteristic of it $while failing in effect to eAclude astrology&. 0o scientific theory can e#er be deduced fro! obser#ation state!ents or be described as a truth=function of obser#ation state!ents.

+ll this / pointed out on #arious occasions to %ittgensteinians and !e!bers of the Cienna 7ircle. /n 1931=@ / su!!ari9ed !y ideas in a largish book $read by se#eral !e!bers of the 7ircle but ne#er published: although part of it was incorporated in !y ,ogic of Scientific 8isco#ery&: and in 1933 / published a letter to the .ditor of .rkenntnis in which / tried to co!press into

Page 1@ of '(

two pages !y ideas on the proble!s of de!arcation and induction.( /n this letter and elsewhere / described the proble! of !eaning as a pseudo=proble! in contrast to the proble! of de!arcation. -ut !y contribution was classified by !e!bers of the 7ircle as a proposal to replace the #erifiability criterion of !eaning by a falsifiability criterion of !eaning==which effecti#ely !ade nonsense of !y #iews.6 1y protests that / was trying to sol#e not their pseudo=proble! of !eaning but the proble! of de!arcation were of no a#ail.

1y attacks upon #erification had so!e effect howe#er. They soon led to co!plete confusion in the ca!p of the #erificationist philosophers of sense and nonsense. The original proposal of #erifiability as the criterion of !eaning was at least clear si!ple and forceful. The !odifications and shifts which were now introduced were the #ery opposite.> This / should say is now seen e#en by the participants. -ut since / a! usually 2uoted as one of the! / wish to repeat that although / created this confusion / ne#er participated in it. 0either falsifiability nor testability were proposed by !e as criteria of !eaning: and although / !ay plead guilty to ha#ing introduced both ter!s into the discussion it was not / who introduced the! into the theory of !eaning.

7riticis! of !y alleged #iews was widespread and highly successful. / ha#e yet to !eet a criticis! of !y #iews.E 1eanwhile testability is being widely accepted as a criterion of de!arcation.

IV
/ ha#e discussed the proble! of de!arcation in so!e detail because / belie#e that its solution is the key to !ost of the funda!ental proble!s of the philosophy of science. / a! going to gi#e you later a list of so!e of these other proble!s but only one of the!==the proble! of induction==can be discussed here at any length. Page 13 of '(

/ had beco!e interested in the proble! of induction in 19@3. +lthough this proble! is #ery closely connected with the proble! of de!arcation / did not fully appreciate the connection for about fi#e years.

/ approached the proble! of induction through 3u!e. 3u!e / felt was perfectly right in pointing out that induction cannot be logically "ustified. 3e held that there can be no #alid logical9 argu!ents allowing us to establish ;that those instances of which we ha#e had no eAperience rese!ble those of which we ha#e had eAperience;. 7onse2uently ;e#en after the obser#ation of the fre2uent or constant con"unction of ob"ects we ha#e no reason to draw any inference concerning any ob"ect beyond those of which we ha#e had eAperience;. Bor ;shou;d it be said that we ha#e eAperience;1D===eAperience teaching us that ob"ects constantly con"oined with certain other ob"ects continue to be so con"oined===then 3u!e says ;/ wou;d renew !y 2uestion why fro! this eAperience we for! any conclusion beyond those past instances of which we ha#e had eAperience;. /n other words an atte!pt to "ustify the practice of induction by an appeal to eAperience !ust lead to an infinite regress. +s a result we can say that theories can ne#er be inferred fro! obser#ation state!ents or rationally "ustified by the!.

/ found 3u!e;s refutation of inducti#e inference clear and conclusi#e. -ut / felt co!pletely dissatisfied with his psychological eAplanation of induction in ter!s of custo! or habit. /t has often been noticed that this eAplanation of 3u!e;s is philosophically not #ery satisfactory. /t is howe#er without doubt intended as a psychological rather than a philosophical theory: for it tries to gi#e a causal eAplanation of a psychological fact==the fact that we belie#e in laws in state!ents asserting regularities or constantly con"oined kinds of e#ents==by asserting that this fact is due to $i.e. constantly con"oined with& custo! or habit. -ut e#en this refor!ulation of 3u!e;s theory is still unsatisfactory: for what / ha#e "ust / called a ;psychological fact; !ay itself be described as a custo! or habit==the custo! or habit of

Page 1' of '(

belie#ing in laws or regularities: and it is neither #ery surprising nor #ery enlightening to hear that such a custo! or habit !ust be eAplained as due to or con"oined with a custo! or habit $e#en though a different one&. 4nly when we re!e!ber that the words ;custo!; and ;habit; are used by 3u!e as they are in ordinary language not !erely to describe regular beha#iour but rather to theori9e about its origin $ascribed to fre2uent repetition& can we refor!ulate his psychological theory in a !ore satisfactory way. %e can then say that like other habits our habit of belie#ing in laws is the product of fre2uent repetition=of the repeated obser#ation that things of a certain kind are constantly con"oined with things of another kind.

This genetico=psychological theory is as indicated incorporated in ordinary language and it is therefore hardly as re#olutionary as 3u!e thought. /t is no doubt an eAtre!ely popular psychological theory==part of ;co!!on sense; one !ight say. -ut in spite of !y lo#e of both co!!on sense and 3u!e / felt con#inced that this psychological theory was !istaken: and that it was in fact refutable on purely logical grounds.

3u!e;s psychology which is the popular psychology was !istaken / felt about at least three different things< $a& the typical result of repetition: $b& the genesis of habits: and especially $c& the character of those eAperiences or !odes of beha#iour which !ay be described as ;belie#ing in a law; or ;eApecting a law=like succession of e#ents;.

$a& The typical result of repetition==say of repeating a difficult passage on the piano==is that !o#e!ents which at first needed attention are in the end eAecuted without attention. %e !ight say that the process beco!es radically abbre#iated and ceases to be conscious< it beco!es ;physiological;. Such a process far fro! creating a conscious eApectation of law=like succession or a belief in a law !ay on the contrary begin with a conscious belief and destroy it by !aking it superfluous. /n learning to ride a bicycle we !ay start with the belief that we

Page 1( of '(

can a#oid falling if we steer in the direction in which we threaten to fall and this belief !ay be useful for guiding our !o#e!ents. +fter sufficient practice we !ay forget the rule: in any case we do not need it any longer. 4n the other hand e#en if it is true that repetition !ay create unconscious eApectations these beco!e conscious only if so!ething goes wrong $we !ay not ha#e heard the clock tick but we !ay hear that it has slopped&.

$b& 3abits or custo!s do not as a rule originate in repetition. .#en the habit of walking or of speaking or of feeding at certain hours begins before repetition can play any part whate#er. %e !ay say if we like that they deser#e to be called ;habits; or ;custo!s; only after repetition has played its typical part: but we !ust not say that the practices in 2uestion originated as the result of !any repetitions. $c& -elief in a law is not 2uite the sa!e thing as beha#iour which betrays an eApectation of a law=like succession of e#ents: but these two are sufficiently closely connected to be treated together. They !ay perhaps in eAceptional cases result fro! a !ere repetition of sense i!pressions $as in the case of the stopping clock&. / was prepared to concede this but / contended that nor!ally and in !ost cases of any interest they cannot be so eAplained. +s 3u!e ad!its eyen a single striking obser#ation !ay be sufficient to create a belief or an eApectation===a fact which he tries to eAplain as due to an inducti#e habit for!ed as the result of a #ast nu!ber of long repetiti#e se2uences which had been eAperienced at an earlier period of life.11 -ut this / contended was !erely his atte!pt to eAplain away unfa#ourable facts which threatened his theory: an unsuccessful atte!pt since these unfa#ourable facts could be obser#ed in #ery young ani!als and babies==as early indeed as we like. ;+ lighted cigarette was held near the noses of the young puppies; reports B. -age. ;They sniffed at it once turned tail and nothing would induce the!; to co!e back to the source of the s!ell and to sniff again. + few days later they reacted to the !ere sight of a cigarette or e#en of a rolled piece of

Page 16 of '(

white paper by bounding away and snee9ing.;1@ /f we try to eAplain cases like this by postulating a #ast nu!ber of long repetiti#e se2uences at a still earlier age we are not only ro!ancing but forgetting that in the cle#er puppies; short li#es there !ust be roo! not only for repetition but also for a great deal of no#elty and conse2uently of non=repetition.

-ut it is not only that certain e!pirical facts do not support 3u!e: there are decisi#e argu!ents of a purely logical nature against his psychological theory.

The central idea of 3u!e;s theory is that of repetition based upon si!ilarity $or ;rese!blance;&. This idea is used in a #ery uncritical way. %e are led to think of the water= drop that hollows the stone< of se2uences of un2uestionably like e#ents slowly forcing the!sel#es upon us as does the tick of the clock. -ut we ought to reali9e that in a psychological theory such as 3u!e;s only repetition=for=us based upon si!ilarity=for=us can be allowed to ha#e any effect upon us. %e !ust respond to situations as if they were e2ui#alent: take the! as si!ilar: interpret the! as repetitions. The cle#er puppies we !ay assu!e showed by their response their way of acting or of reacting that they recogni9ed or interpreted the second situation as a repetition of the first< that they eApected its !ain ele!ent the ob"ectionable s!ell to be present. The situation was a repetition=for=the! because they responded to it by anticipating its si!ilarity to the pre#ious one.

This apparently psychological criticis! has a purely logical basis which !ay be su!!ed up in the following si!ple argu!ent. $/t happens to be the one fro! which / originally started !y criticis!.& The kind of repetition en#isaged by 3u!e can ne#er be perfect: the cases he has in !ind cannot be cases of perfect sa!eness: they can only be cases of si!ilarity. Thus they are repetitions only fro! a certain point of #iew. $%hat has the effect upon !e of a repetition !ay not ha#e this effect upon a spider.& -ut this !eans that for logical reasons there !ust always

Page 1> of '(

be a point of #iew==such as a syste! of eApectations anticipations assu!ptions or interests= before there can be any repetition: which point of #iew conse2uently cannot be !erely the result of repetition. $See now also appendiA GA $1& to !y ,.Sc.8.&

%e !ust thus replace for the purposes of a psychological theory of the origin of our beliefs the nai#e idea of e#ents which are si!ilar by the idea of e#ents to which we react by interpreting the! as being si!ilar. -ut if this is so $and / can see no escape fro! it& then 3u!e;s psychological theory of induction leads to an infinite regress precisely analogous to that other infinite regress which was disco#ered by 3u!e hi!self and used by ru! to eAplode the logical theory of induction. Bor what do we wish to eAplain? /n the eAa!ple of the puppies we wish to eAplain beha#iour which !ay be described as recogni9ing or interpreting a situation as a repetition of another. 7learly we cannot hope to eAplain this by an appeal to earlier repetitions once we reali9e that the earlier repetitions !ust also ha#e been repetitions=for= the! so that precisely the sa!e proble! arises again< that of recogni9ing or interpreting a situation as a repetition of another.

To put it !ore concisely si!ilarity=for=us is the product of a response in#ol#ing interpretations $which !ay be inade2uate& and anticipations or eApectations $which !ay ne#er be fulfilled&. /t is therefore i!possible to eAplain anticipations or eApectations as resulting fro! !any repetitions as suggested by 3u!e. Bor e#en the first repetition=for=us !ust be based upon si!ilarity=for=us and therefore upon eApectations==precisely the kind of thing we wished to eAplain. This shows that there is an infinite regress in#ol#ed in 3u!e;s psychological theory.

3u!e / felt had ne#er accepted the full force of his own logical analysis. 3a#ing refuted the logical idea of induction he was faced with the following proble!< how do we actually obtain our

Page 1E of '(

knowledge as a !atter of psychological fact if induction is a procedure which is logically in#alid and rationally un"ustifiable? There are two possible answers< $1& %e obtain our knowledge by a non=inducti#e procedure. This answer would ha#e allowed 3u!e to retain a for! of rationalis!. $@& %e obtain our knowledge by repetition and induction and therefore by a logically in#alid and rationally un"ustifiable procedure so that all apparent knowledge is !erely a kind of belief==belief based on habit. This answer would i!ply that e#en scientific knowledge is irrational so that rationalis! is absurd and !ust be gi#en up. $/ shall not discuss here the age=old atte!pts now again fashionable to get out of the difficulty by asserting that though induction is of course logically in#alid if we !ean by ;logic; the sa!e as ;deducti#e logic; it is not irrational by its own standards as !ay be seen fro! the fact that e#ery reasonable !an applies it as a !atter of fact< it was 3u!e;s great achie#e!ent to break this uncritical identification of the 2uestion of factH2uid facti?==and the 2uestion of "ustification or #alidity==2uid "uris? $See below point $13& of the appendiA to the present chapter.&

/t see!s that 3u!e ne#er seriously considered the first alternati#e. 3a#ing cast out the logical theory of induction by repetition he struck a bargain with co!!on sense !eekly allowing the re= entry of induction by repetition in the guise of a psychological theory. / proposed to turn the tables upon this theory of 3u!e;s. /nstead of eAplaining our propensity to eApect regularities as the result of repetition / proposed to eAplain repetition=for=us as the result of our propensity to eApect regularities and to search for the!.

Thus / was led by purely logical considerations to replace the psychological theory of induction by the following #iew. %ithout waiting passi#ely for repetitions to i!press or i!pose regularities upon us we acti#ely try to i!pose regularities upon the world. %e try to disco#er

Page 19 of '(

si!ilarities in it and to interpret it in ter!s of laws in#ented by us. %ithout waiting for pre!ises we "u!p to conclusions. These !ay ha#e to be discarded later should obser#ation show that they are wrong.

This was a theory of trial and error==of con"ectures and refutations. /t !ade it possible to understand why our atte!pts to force interpretations upon the world were logically prior to the obser#ation of si!ilarities. Since there were logical reasons behind this procedure / thought that it would apply in the field of science also: that scientific theories were not the digest of obser#ations but that they were in#entions==con"ectures boldly put forward for trial to be eli!inated if they clashed with obser#ations: with obser#ations which were rarely accidental but as a rule undertaken with the definite intention of testing a theory by obtaining if possible a decisi#e refutation.

V
The belief that science proceeds fro! obser#ation to theory is still so widely and so fir!ly held that !y denial of it is often !et with incredulity. / ha#e e#en been suspected of being insincere==of denying what nobody in his senses can doubt.

-ut in fact the belief that we can start with pure obser#ations alone without anything in the nature of a theory is absurd: as !ay be illustrated by the story of the !an who dedicated his life to natural science wrote down e#erything he could obser#e and be2ueathed his priceless collection of obser#ations to the Royal Society to be used as inducti#e e#idence. This story should show us that though beetles !ay profitably be collected obser#ations !ay not.

Twenty=fi#e years ago / tried to bring ho!e the sa!e point to a group of physics students in Page @D of '(

Cienna by beginning a lecture with the following instructions< ;Take pencil and paper: carefully obser#e and write down what you ha#e obser#edI; They asked of course what / wanted the! to obser#e. 7learly the instruction ;4bser#eI; is absurd.13 $/t is not e#en idio!atic unless the ob"ect of the transiti#e #erb can be taken as understood.& 4bser#ation is always selecti#e. /t needs a chosen ob"ect a definite task an interest a point of #iew a proble!. +nd its description presupposes a descripti#e language with property words: it presupposes si!ilarity and classification which in its turn presupposes interests points of #iew and proble!s. ;+ hungry ani!al; writes Kat9 1' ;di#ides the en#iron!ent into edible and inedible things. +n ani!al in flight sees roads to escape and hiding places. . . . Fenerally speaking ob"ects change. . . according to the needs of the ani!al.; %e !ay add that ob"ects can be classified and can beco!e si!ilar or dissi!ilar only in this way==by being related to needs and interests. This rule applies not only to ani!als but also to scientists. Bor the ani!al a point of #iew is pro#ided by its needs the task of the !o!ent and its eApectations: for the scientist by his theoretical interests the special proble! under in#estigation his con"ectures and anticipations and the theories which he accepts as a kind of background< his fra!e of reference his ;hori9on of eApectations;.

The proble! ;%hich co!es first the hypothesis $3& or the obser#ation $4& ; is soluble: as is the proble! ;%hich co!es first the hen $3& or the egg $4&;. The reply to the latter is ;+n earlier kind of egg;: to the for!er ;+n earlier kind of hypothesis;. /t is 2uite true that any particular hypothesis we choose will ha#e been preceded by obser#ations==the obser#ations for eAa!ple which it is designed to eAplain. -ut these obser#ations in their turn presupposed the adoption of a fra!e of reference< a fra!e of eApectations< a fra!e of theories. /f they were significant if they created a need for eAplanation and thus ga#e rise to the in#ention of a hypothesis it was because they could not be eAplained within the old theoretical fra!ework the old hori9on of eApectations. There is no danger here of an infinite regress. Foing back to !ore

Page @1 of '(

and !ore pri!iti#e theories and !yths we shall in the end find unconscious inborn eApectations.

The theory of inborn ideas is absurd / think: but e#ery organis! has inborn reactions or responses: and a!ong the! responses adapted to i!pending e#ents. These responses we !ay describe as ;eApectations; without i!plying that these ;eApectations; are conscious. The new= born baby ;eApects; in this sense to be fed $and one could e#en argue to be protected and lo#ed&. /n #iew of the close relation between eApectation and knowledge we !ay e#en speak in 2uite a reasonable sense of ;inborn knowledge;. This ;knowledge; is not howe#er #alid a priori: an inborn eApectation no !atter how strong and specific !ay be !istaken. $The newborn child !ay be abandoned and star#e.&

Thus we are born with eApectations: with ;knowledge; which although not #alid a priori is psychologically or genetically a priori i.e. prior to all obser#ational eAperience. 4ne of the !ost i!portant of these eApectations is the eApectation of finding a regularity. /t is connected with an inborn propensity to look out for regularities or with a need to find regularities as we !ay see fro! the pleasure of the child who satisfies this need.

This ;instincti#e; eApectation of finding regularities which is psychologically a priori corresponds #ery closely to the ;law of causality; which Kant belie#ed to be part of our !ental outfit and to be a priori #alid. 4ne !ight thus be inclined to say that Kant failed to distinguish between psychologically a priori ways of thinking or responding and a priori #alid beliefs. -ut / do not think that his !istake was 2uite as crude as that. Bor the eApectation of finding regularities is not only psychologically a priori but also logically a priori< it is logically prior to all obser#ational eAperience for it is prior to any recognition of si!ilarities as we ha#e seen: and all obser#ation in#ol#es the recognition of si!ilarities $or dissi!ilarities&. -ut in spite of being logically a priori in this sense the eApectation is not

Page @@ of '(

#alid a priori. Bor it !ay fail< we can easily construct an en#iron!ent $it would be a lethal one& which co!pared with our ordinary en#iron!ent is so chaotic that we co!pletely fail to find regularities. $+ll natural laws could re!ain #alid< en#iron!ents of this kind ha#e been used in the ani!al eAperi!ents !entioned in the neAt section.&

Thus Kant;s reply to 3u!e ca!e near to being right: for the distinction between an a priori #alid eApectation and one which is both genetically and logically prior to obser#ation but not a priori #alid is really so!ewhat subtle. -ut Kant pro#ed too !uch. /n trying to show how knowledge is possible he proposed a theory which had the una#oidable conse2uence that our 2uest for knowledge !ust necessarily succeed which is clearly !istaken. %hen Kant said ;4ur intellect does not draw its laws fro! nature but i!poses its laws upon nature; he was right. -ut in thinking that these laws are necessarily true or that we necessarily succeed in i!posing the! upon nature he was wrong.l( 0ature #ery often resists 2uite successfully forcing us to discard our laws as refuted: but if we li#e we !ay try again.

To su! up this logical criticis! of 3u!e;s psychology of induction we !ay consider the idea of building an induction !achine. Placed in a si!plified ;world; $for eAa!ple one of se2uences of coloured counters& such a !achine !ay through repetition ;learn; or e#en ;for!ulate; laws of succession which hold in its ;world;. /f such a !achine can be constructed $and / ha#e no doubt that it can& then it !ight be argued !y theory !ust be wrong: for if a !achine is capable of perfor!ing inductions on the basis of repetition there can be no logical reasons pre#enting us fro! doing the sa!e.

The argu!ent sounds con#incing but it is !istaken. /n constructing an induction !achine we the architects of the !achine !ust decide a priori what constitutes its ;world;: what things are to be taken as si!ilar or e2ual: and what kind of ;laws; we wish the !achine to be able to

Page @3 of '(

;disco#er; in its ;world;. /n other words we !ust build into the !achine a fra!ework deter!ining what is rele#ant or interesting in its world< the !achine will ha#e its ;inborn; selection principles. The proble!s of si!ilarity will ha#e been sol#ed for it by its !akers who thus ha#e interpreted the ;world; for the !achine.

VI
4ur propensity to look out for regularities and to i!pose laws upon nature leads to the psychological pheno!enon of dog!atic thinking or !ore generally dog!atic beha#iour< we eApect regularities e#erywhere and atte!pt to find the! e#en where there are none: e#ents which do not yield to these atte!pts we are inclined to treat as a kind of ;background noise;: and we stick to our eApectations e#en when they are inade2uate and we ought to accept defeat. This dog!atis! is to so!e eAtent necessary. /t is de!anded by a situation which can only be dealt with by forcing our con"ectures upon the world. 1oreo#er this dog!atis! allows us to approach a good theory in stages by way of approAi!ations< if we accept defeat too easily we !ay pre#ent oursel#es fro! finding that we were #ery nearly right.

/t is clear that this dog!atic attitude which !akes us stick to our first i!pressions is indicati#e of a strong belief: while a critical attitude which is ready to !odify its tenets which ad!its doubt and de!ands tests is indicati#e of a weaker belief. 0ow according to 3u!e;s theory and to the popular theory the strength of a belief should be a product of repetition: thus it should always grow with eAperience and always be greater in less pri!iti#e persons. -ut dog!atic thinking an uncontrolled wish to i!pose regularities a !anifest pleasure in rites and in repetition as such arc characteristic of pri!iti#es and children: and increasing eAperience and !aturity so!eti!es create an attitude of caution and criticis! rather than of dog!atis!.

/ !ay perhaps !ention here a point of agree!ent with psycho=analysis. Psycho=analysts assert Page @' of '(

that neurotics and others interpret the world in accordance with a personal set pattern which is not easily gi#en up and which can often be traced back to early childhood. + pattern or sche!e which was adopted #ery early in life is !aintained throughout and e#ery new eAperience is interpreted in ter!s of it: #erifying it as it were and contributing to its rigidity. This is a description of what / ha#e called the dog!atic attitude as distinct fro! the critical attitude which shares with the dog!atic attitude the 2uick adoption of a sche!a of eApectations==a !yth perhaps or a con"ecture or hypothesis==but which is ready to !odify it to correct it and e#en to gi#e it up. / a! inclined to suggest that !ost neuroses !ay be due to a partially arrested de#elop!ent of the critical attitude: to an arrested rather than a natural dog!atis!: to resistance to de!ands for the !odification and ad"ust!ent of certain sche!atic interpretations and responses. This resistance in its turn !ay perhaps be eAplained in so!e cases as due to an in"ury or shock resulting in fear and in an increased need for assurance or certainty analogous to the way in which an in"ury to a li!b !akes us afraid to !o#e it so that it beco!es stiff. $/t !ight e#en be argued that the case of the li!b is not !erely analogous to the dog!atic response but an instance of it.& The eAplanation of any concrete case will ha#e to take into account the weight of the difficulties in#ol#ed in !aking the necessary ad"ust!ents== difficulties which !ay be considerable especially in a co!pleA and changing world< we know fro! eAperi!ents on ani!als that #arying degrees of neurotic beha#iour !ay be produced at will by correspondingly #arying difficulties. / found !any other links between the psychology of knowledge and psychological fields which are often considered re!ote fro! it==for eAa!ple the psychology of art and !usic: in fact !y ideas about induction originated in a con"ecture about the e#olution of %estern polyphony. -ut you will be spared this story.

VII
1y logical criticis! of 3u!e;s psychological theory and the considerations connected with it $!ost of which 1 elaborated in 19@6=> in a thesis entitled ;4n 3abit and -elief in ,aws;16& !ay Page @( of '(

see! a little re!o#ed fro! the field of the philosophy of science. -ut the distinction between dog!atic and critical thinking or the dog!atic and the critical attitude brings us right back to our central proble!. Bor the dog!atic attitude is clearly related to the tendency to #erify our laws and sche!ata by seeking to apply the! and to confir! the! e#en to the point of neglecting refutations whereas the critical attitude is one of readiness to change the!== to test the!: to refute the!: to falsify the! if possible. This suggests that we !ay identify the critical attitude with the scientific attitude and the dog!atic attitude with the one which we ha#e described as pseudo=scientific.

/t further suggests that genetically speaking the pseudo=scientific attitude is !ore pri!iti#e than and prior to the scientific attitude< that it is a pre=scientific attitude. +nd this pri!iti#ity or priority also has its logical aspect. Bor the critical attitude is not so !uch opposed to the dog!atic attitude as super=i!posed upon it< criticis! !ust be directed against eAisting and influential beliefs in need of critical re#ision==in other words dog!atic beliefs. + critical attitude needs for its raw !aterial as it were theories or beliefs which are held !ore or less dog!atically.

Thus science !ust begin with !yths and with the criticis! of !yths: neither with the collection of obser#ations nor with the in#ention of eAperi!ents but with the critical discussion of !yths and of !agical techni2ues and practices. The scientific tradition is distinguished fro! the pre=scientific tradition in ha#ing two layers. ,ike the latter it passes on its theories: but it also passes on a critical attitude towards the!. The theories are passed on not as dog!as but rather with the challenge to discuss the! and i!pro#e upon the!. This tradition is 3ellenic< it !ay be traced back to Thales founder of the first school $/ do not !ean ;of the first philosophical school; but si!ply ;of the first school;& which was not !ainly concerned with the preser#ation of a dog!a.1>

Page @6 of '(

The critical attitude the tradition of free discussion of theories with the ai! of disco#ering their weak spots so that they !ay be i!pro#ed upon is the altitude of reasonableness of rationality. /t !akes far=reaching use of both #erbal argu!ent and obser#ation==of obser#ation in the interest of argu!ent howe#er. The Freeks; disco#ery of the critical !ethod ga#e rise at first to the !istaken hope that it would lead to the solution of all the great old proble!s: that it would establish certainty: that it would help to pro#e our theories to "ustify the!. -ut this hope was a residue of the dog!atic way of thinking: in fact nothing can be "ustified or pro#ed $outside of !athe!atics and logic&. The de!and for rational proofs in science indicates a failure to keep distinct the broad real! of rationality and the narrow real! of rational certainty< it is an untenable an unreasonable de!and.

0e#ertheless the role of logical argu!ent of deducti#e logical reasoning re!ains all= i!portant for the critical approach: not because it allows us to pro#e our theories or to infer the! fro! obser#ation state!ents but because only by purely deducti#e reasoning is it possible for us to disco#er what our theories i!ply and thus to critici9e the! effecti#ely. 7riticis! 1 said is an atte!pt to find the weak spots in a theory and these as a rule can be found only in the !ore re!ote logical conse2uences which can be deri#ed fro! it. /t is here that purely logical reasoning plays an i!portant part in science.

3u!e was right in stressing that our theories cannot be #alidly inferred fro! what we can know to be true==neither fro! obser#ations nor fro! anything else. 3e concluded fro! this that our belief in the! was irrational. /f ;belief; !eans here our inability to doubt our natural laws and the constancy of natural regularities then 3u!e is again right< this kind of dog!atic belief has one !ight say a physiological rather than a rational basis. /f howe#er the ter! ;belief; is taken to co#er our critical acceptance of scientific theories ==a tentati#e

Page @> of '(

acceptance co!bined with an eagerness to re#ise the theory if we succeed in designing a test which it cannot pass==then 3u!e was wrong. /n such an acceptance of theories there is nothing irrational. There is not e#en anything irrational in relying for practical purposes upon well= tested theories for no !ore rational course of action is open to us.

+ssu!e that we ha#e deliberately !ade it our task to li#e in this unknown world of ours: to ad"ust oursel#es to it as well as we can: to take ad#antage of the opportunities we can find in it: and to eAplain it if possible $we need not assu!e that it is& and as far as possible with the help of laws and eAplanatory theories. /f we ha#e !ade this our task then there is no !ore rational procedure than the !ethod of trial and error==of con"ecture and refutation< of boldly proposing theories: of trying our best to show that these are erroneous: and of accepting the! tentati#ely if our critical efforts are unsuccessful. Bro! the point of #iew here de#eloped all laws all theories re!ain essentially tentati#e or con"ectural or hypothetical e#en when we feel unable to doubt the! any longer. -efore a theory has been refuted we can ne#er know in what way it !ay ha#e to be !odified. That the sun will always rise and set within twenty=four hours is still pro#erbial as a law ;established by induction beyond reasonable doubt;. /t is odd that this eAa!ple is still in use though it !ay ha#e ser#ed well enough in the days of +ristotle and Pytheas of 1assalia ==the great tra#eler who for centuries was called a liar because of his tales of Thule the land of the fro9en sea and the !idnight sun.

The !ethod of trial and error is not of course si!ply identical with the scientific or critical approach==with the !ethod of con"ecture and refutation. The !ethod of trial and error is applied not only by .instein but in a !ore dog!atic fashion by the a!oeba also. The difference lies not so !uch in the trials as in a critical and constructi#e attitude towards errors: errors which the scientist consciously and cautiously tries to unco#er in order to refute his theories with searching argu!ents including appeals to the !ost se#ere eAperi!ental

Page @E of '(

tests which his theories and his ingenuity per!it hi! to design.

The critical attitude !ay be described as the conscious atte!pt to !ake our theories our con"ectures suffer in our stead in the struggle for the sur#i#al of the fittest. /t gi#es us a chance to sur#i#e the eli!ination of an inade2uate hypothesis==when a !ore dog!atic attitude would eli!inate it by eli!inating us. $There is a touching story of an /ndian co!!unity which disappeared because of its belief in the holiness of life including that of tigers.& %e thus obtain the fittest theory within our reach by the eli!ination of those which are less fit. $-y ;fitness; / do not !ean !erely ;usefulness; but truth: see chapters 3 and 1D below.& / do not think that this procedure is irrational or in need of any further rational "ustification.

VIII
,et us now turn fro! our logical criticis! of the psychology of eAperience to our real proble!== the proble! of the logic of science. +lthough so!e of the things / ha#e said !ay help us here in so far as they !ay ha#e eli/l1inated certain psychological pre"udices in fa#our of induction !y treat!ent of the logical proble! of induction is co!pletely independent of this criticis! and of all psychological considerations. Pro#ided you do not dog!atically belie#e in the alleged psychological fact that we !ake inductions you !ay now forget !y whole story with the eAception of two logical points< !y logical re!arks on testability or falsifiability as the criterion of de!arcation and 3u!eJs logical criticis! of induction.

Bro! what / ha#e said it is ob#ious that there was a close link between the two proble!s which interested !e at that ti!e< de!arcation and induction or scientific !ethod. /t was easy to see that the !ethod of science is criticis! i.e. atte!pted falsifications. 5et it took !e a few years to notice that the two proble!s==of de!arcation and of induction===were in a sense one.

Page @9 of '(

%hy / asked do so !any scientists belie#e in induction? / found they did so because they belie#ed natural science to be characteri9ed by the inducti#e !ethod==by a !ethod starting fro! and relying upon long se2uences of obser#ations and eAperi!ents. They belie#ed that the difference between genuine science and !etaphysical or pseudo=scientific speculation depended solely upon whether or not the inducti#e !ethod was e!ployed. They belie#ed $to put it in !y own ter!inology& that only the inducti#e !ethod could pro#ide a satisfactory criterion of de!arcation.

/ recently ca!e across an interesting for!ulation of this belief in a re!arkable philosophical book by a great physicist==1aA -orn;s 0atural Philosophy of 7ause and 7hance.1E 3e writes< ;/nduction allows us to generali9e a nu!ber of obser#ations into a general rule< that night follows day and day follows night. . . -ut while e#eryday life has no definite criterion for the #alidity of an induction . . . science has worked out a code or rule of craft for its application.; -orn nowhere re#eals the contents of this inducti#e code $which as his wording shows contains a ;definite criterion for the #alidity of an induction;&: but he stresses that ;there is no logical argu!ent; for its acceptance< ;it is a 2uestion of faith;: and he is therefore ;willing to call induction a !etaphysical principle;. -ut why does he belie#e that such a code of #alid inducti#e rules !ust eAist? This beco!es clear when he speaks of the ;#ast co!!unities of people ignorant of or re"ecting the rule of science a!ong the! the !e!bers of anti=#accination societies and belie#ers in astrology. /t is useless to argue with the!: / cannot co!pel the! to accept the sa!e criteria of #alid induction in which / belie#e< the code of scientific rules.; This !akes it 2uite clear that ;#alid induction; was here !eant to ser#e as a criterion of de!arcation between science and pseudo=science.

-ut it is ob#ious that this rule or craft of ;#alid induction; is not e#en !etaphysical< it si!ply does not eAist. 0o rule can e#er guarantee that a generali9ation inferred fro! true

Page 3D of '(

obser#ations howe#er often repeated is true. $-orn hi!self does not belie#e in the truth of 0ewtonian physics in spite of its success although he belie#es that it is based on induction.& +nd the success of science is not based upon rules of induction but depends upon luck ingenuity and the purely deducti#e rules of critical argu!ent.

/ !ay su!!ari9e so!e of !y conclusions as follows<

$1& /nduction i.e. inference based on !any obser#ations is a !yth. /t is neither a psychological fact nor a fact of ordinary life nor one of scientific procedure.

$@& The actual procedure of science is to operate with con"ectures< to "u!p to conclusions== often after one single obser#ation $as noticed for eAa!ple by 3u!e and -orn&.

$3& Repeated obser#ations and eAperi!ents function in science as tests of our con"ectures or hypotheses i.e. as atte!pted refutations.

$'& The !istaken belief in induction is fortified by the need for a criterion of de!arcation which it is traditionally but wrongly belie#ed only the inducti#e !ethod can pro#ide.

$(& The conception of such an inducti#e !ethod like the criterion of #erifiability i!plies a faulty de!arcation.

$6& 0one of this is altered in the least if we say that induction !akes theories only probable rather than certain. $See especially chapter 1D below.&

Page 31 of '(

IX
/f as / ha#e suggested the proble! of induction is only an instance or facet of the proble! of de!arcation then the solution to the proble! of de!arcation !ust pro#ide us with a solution to the proble! of induction. This is indeed the case / belie#e although it is perhaps not i!!ediately ob#ious.

Bor a brief for!ulation of the proble! of induction we can turn again to -orn who writes< ;. . . no obser#ation or eAperi!ent howe#er eAtended can gi#e !ore than a finite nu!ber of repetitions;: therefore ;the state!ent of a law==- depends on +==always transcends eAperience. 5et this kind of state!ent is !ade e#erywhere and all the ti!e and so!eti!es fro! scanty !aterial.;19

/n other words the logical proble! of induction arises fro! $a& 3u!e;s disco#ery $so well eApressed by -orn& that it is i!possible to "ustify a law by obser#ation or eAperi!ent since it ;transcends eAperience;: $b& the fact that science proposes and uses laws ;e#erywhere and all the ti!e;. $,ike 3u!e -orn is struck by the ;scanty !aterial; i.e. the few obser#ed instances upon which the law !ay be based.& To this we ha#e to add $c& the principle of e!piricis! which asserts that in science only obser#ation and eAperi!ent !ay decide upon the acceptance or re"ection of scientific state!ents including laws and theories.

These three principles $a& $b& and $c& appear at first sight to clash: and this apparent clash constitutes the logical proble! of induction.

Baced with this clash -orn gi#es up $c& the principle of e!piricis! $as Kant and !any others including -ertrand Russell ha#e done before hi!& in fa#our of what he calls a ;!etaphysical principle;: a !etaphysical principle which he does not e#en atte!pt to for!ulate: which he

Page 3@ of '(

#aguely describes as a ;code or rule of craft;: and of which / ha#e ne#er seen any for!ulation which e#en looked pro!ising and was not clearly untenable.

-ut in fact the principles $a& to $c& do not clash. %e can see this the !o!ent we reali9e that the acceptance by science of a law or of a theory is tentati#e only: which is to say that all laws and theories are con"ectures or tentati#e hypotheses $a position which / ha#e so!eti!es called ;hypotheticis!;&: and that we !ay re"ect a law or theory on the basis of new e#idence without necessarily discarding the old e#idence which originally led us to accept it.@D

The principle of e!piricis! $c& can be fully preser#ed since the fate of a theory its acceptance or re"ection is decided by obser#ation and eAperi!ent ==by the result of tests. So long as a theory stands up to the se#erest tests we can design it is accepted: if it does not it is re"ected. -ut it is ne#er inferred in any sense fro! the e!pirical e#idence. There is neither a psychological nor a logical induction. 4nly the falsity of the theory can be inferred fro! e!pirical e#idence and this inference is a purely deducti#e one. 3u!e showed that it is not possible to infer a theory fro! obser#ation state!ents: but this does not affect the possibility of refuting a theory by obser#ation state!ents. The full appreciation of this possibility !akes the relation between theories and obser#ations perfectly clear.

This sol#es the proble! of the alleged clash between the principles $a& $b& and $c& and with it 3u!e;s proble! of induction.

X
Thus the proble! of induction is sol#ed. -ut nothing see!s less wanted than a si!ple solution to an age=old philosophical proble!. %ittgenstein and his school hold that genuine philosophical Page 33 of '(

proble!s do not eAist:@1 fro! which it clearly follows that they cannot be sol#ed. 4thers a!ong !y conte!poraries do belie#e that there are philosophical proble!s and respect the!: but they see! to respect the! too !uch: they see! to belie#e that they are insoluble if not taboo: and they are shocked and horrified by the clai! that there is a si!ple neat and lucid solution to any of the!. /f there is a solution it !ust be deep they feel or at least co!plicated.

3owe#er this !ay be / a! still waiting for a si!ple neat and lucid criticis! of the solution which / published first in 1933 in !y letter to the .ditor of .rkenntnis @@ and later in The ,ogic of Scientific 8isco#ery.

4f course one can in#ent new proble!s of induction different fro! the one / ha#e for!ulated and sol#ed. $/ts for!ulation was half its solution.& -ut / ha#e yet to see any refor!ulation of the proble! whose solution cannot be easily obtained fro! !y old solution. / a! now going to discuss so!e of these re=for!ulations.

4ne 2uestion which !ay be asked is this< how do we really "u!p fro! an obser#ation state!ent to a theory?

+lthough this 2uestion appears to be psychological rather than philosophical one can say so!ething positi#e about it without in#oking psychology. 4ne can say first that the "u!p is not fro! an obser#ation state!ent but fro! a proble!=situation and that the theory !ust allow us to eAplain the obser#ations which created the proble! $that is to deduce the! fro! the theory strengthened by other accepted theories and by other obser#ation state!ents the so=called initial conditions&. This lea#es of course an i!!ense nu!ber of possible theories good and bad: and it thus appears that our 2uestion has not been answered.

Page 3' of '(

-ut this !akes it fairly clear that when we asked our 2uestion we had !ore in !ind than ;3ow do we "u!p fro! an obser#ation state!ent to a theory?; The 2uestion we had in !ind was it now appears ;3ow do we "u!p fro! an obser#ation state!ent to a good theory?; -ut to this the answer is< by "u!ping first to any theory and then testing it to find whether it is good or not: i.e. by repeatedly applying the critical !ethod eli!inating !any bad theories and in#enting !any new ones. 0ot e#erybody is able to do this: but there is no other way.

4ther 2uestions ha#e so!eti!es been asked. The original proble! of induction it was said is the proble! of "ustifying induction i.e. of "ustifying inducti#e inference. /f you answer this proble! by saying that what is called an ;inducti#e inference; is always in#alid and therefore clearly not "ustifiable the following new proble! !ust arise< how do you "ustify your !ethod of trial and error? Reply< the !ethod of trial and error is a !ethod of eli!inating false theories by obser#ation state!ents: and the "ustification for this is the purely logical relationship of deducibility which allows us to assert the falsity of uni#ersal state!ents if we accept the truth of singular ones.

+nother 2uestion so!eti!es asked is this< why is it reasonable to prefer non=falsified state!ents to falsified ones? To this 2uestion so!e in#ol#ed answers ha#e been produced for eAa!ple prag!atic answers. -ut fro! a prag!atic point of #iew the 2uestion does not arise since false theories often ser#e well enough< !ost for!ulae used in engineering or na#igation are known to be false although they !ay be eAcellent approAi!ations and easy to handle: and they are used with confidence by people who know the! to be false.

The only correct answer is the straightforward one< because we search for truth $e#en though we can ne#er be sure we ha#e found it& and because the falsified theories are known or belie#ed to be false while the non=falsified theories !ay still be true. -esides %e do not prefer e#ery

Page 3( of '(

non=falsified theory ==only one which in the light of criticis! appears to be better than its co!petitors< which sol#es our proble!s which is well tested and of which we think or rather con"ecture or hope $considering other pro#isionally accepted theories& that it will stand up to further tests.

/t has also been said that the proble! of induction is ;%hy is it reasonable to belie#e that the future will be like the past?; and that a satisfactory answer to this 2uestion should !ake it plain that such a belief is in fact reasonable. 1y reply is that it is reasonable to belie#e that the future will be #ery different fro! the past in !any #itally i!portant respects. +d!ittedly it is perfectly reasonable to act on the assu!ption that it will in !any respects be like the past and that well=tested laws will continue to hold $since we can ha#e no better assu!ption to act upon&: but it is also reasonable to belie#e that such a course of action will lead us at ti!es into se#ere trouble since so!e of the laws upon which we now hea#ily rely !ay easily pro#e unreliable. $Re!e!ber the !idnight sunI& 4ne !ight e#en say that to "udge fro! past eAperience and fro! our general scientific knowledge the future will not be like the past in perhaps !ost of the ways which those ha#e in !ind who say that it will. %ater will so!eti!es not 2uench thirst and air will choke those who breathe it. +n apparent way out is to say that the future will be like the past in the sense that the laws of nature will not change but this is begging the 2uestion. %e speak of a ;law of nature; only if we think that we ha#e before us a regularity which does not change: and if we find that it changes then we shall not continue to call it a ;law of nature;. 4f course our search for natural laws indicates that we hope to find the! and that we belie#e that there are natural laws: but our belief in any particular natural law cannot ha#e a safer basis than our unsuccessful critical atte!pts to refute it.

/ think that those who put the proble! of induction in ter!s of the reasonableness of our beliefs are perfectly right if they are dissatisfied with a 3u!ean or post=3u!ean sceptical

Page 36 of '(

despair of reason. %e !ust indeed re"ect the #iew that a belief in science is as irrational as a belief in pri!iti#e !agical practices ==that both are a !atter of accepting a ;total ideology; a con#ention or a tradition based on faith. -ut we !ust be cautious if we for!ulate our proble! with 3u!e as one of the reasonableness of our beliefs. %e should split this proble! into three= =our old proble! of de!arcation or of how to distinguish between science and pri!iti#e !agic: the proble! of the rationality of the scientific or critical procedure and of the role of obser#ation within it: and lastly the proble! of the rationality of our acceptance of theories for scientific and for practical purposes. To all these three proble!s solutions ha#e been offered here.

4ne should also be careful not to confuse the proble! of the reasonableness of the scientific procedure and the $tentati#e& acceptance of the results of this procedure==i.e. the scientific theories==with the proble! of the rationality or otherwise of the belief that this procedure will succeed. /n practice in practical scientific research this belief is no doubt una#oidable and reasonable there being no better alternati#e. -ut the belief is certainly un"ustifiable in a theoretical sense as / ha#e argued $in section #&. 1oreo#er if we could show on general logical grounds that the scientific 2uest is likely to succeed one could not understand why anything like success has been so rare in the long history of hu!an endea#ours to know !ore about our world.

5et another way of putting the proble! of induction is in ter!s of probability. ,et t be the theory and e the e#idence< we can ask for P$t e& that is to say the probability of t gi#en e. The proble! of induction it is often belie#ed can then be put thus< construct a calculus of probability which shows us to work out for any theory t what its probability is relati#e to any gi#en e!pirical e#idence e: and show that P$t e& increases with the accu!ulation of supporting e#idence and reaches high #alues==at any rate #alues greater than one=half.

Page 3> of '(

/n The ,ogic of Scientific 8isco#ery / eAplained why / think that this approach to the proble! is funda!entally !istaken.@3 To !ake this clear / introduced there the distinction between probability and degree of corroboration or confir!ation. $The ter! ;confir!ation; has lately been so !uch used and !isused that / ha#e decided to surrender it to the #erificationists and to use for !y own purposes ;corroboration; only. The ter! ;probability; is best used in so!e of the !any senses which satisfy the well=known calculus of probability aAio!ati9ed for eAa!ple by Keynes Keffreys and !yself: but nothing of course depends on the choice of words as long as we do not assu!e uncritically that degree of corroboration !ust also be a probability ==that is to say that it !ust satisfy the calculus of probability.&

/ eAplained in !y book why we are interested in theories with a high degree of corroboration. +nd / eAplained why it is a !istake to conclude fro! this that we are interested in highly probable theories. / pointed out that the probability of a state!ent $or set of state!ents& is always the greater the less the state!ent says< it is in#erse to the content or the deducti#e power of the state!ent and thus to its eAplanatory power. +ccordingly e#ery interesting and powerful state!ent !ust ha#e a low probability: and #ice #ersa< a state!ent with a high probability will be scientifically uninteresting because it says little and has no eAplanatory power. +lthough we seek theories with a high degree of corroboration as scientists we do not seek highly probable theories but eAplanations: that is to say powerful and i!probable theories.@' The opposite #iew==that science ai!s at high probability==is a characteristic de#elop!ent of #erificationis!< if you find that you cannot #erify a theory or !ake it certain by induction you !ay turn to probability as a kind of ;.rsat9; for certainty in the hope that induction !ay yield at least that !uch.

Page 3E of '(

NOTES

1 This is a slight o#ersi!plification for about half of the .instein effect !ay be deri#ed fro! the classical theory pro#ided we assu!e a ballistic theory of light.

@ See for eAa!ple !y 4pen Society and /ts .ne!ies ch. 1( section iii and notes 13=1'. 3 3 ;7linical obser#ations; like all other obser#ations are interpretations in the light of theories $see below sections i# ff.&: and for this reason alone they are apt to see! to support those theories in the light of which they were interpreted. -ut real support can be obtained only fro! obser#ations undertaken as tests $by ;atte!pted refutations;&: and for this purpose criteria of refutation ha#e to be laid down beforehand< it !ust be agreed which obser#able situations if actually obser#ed !ean that the theory is refuted. -ut what kind of clinical responses would refute to the satisfaction of the analyst not !erely a particular analytic diagnosis but psycho=analysis itself? +nd ha#e such criteria e#er been discussed or agreed upon by analysts? /s there not on the contrary a whole fa!ily of analytic concepts such as ;a!bi#alence; $/ do not suggest that there is no such thing as a!bi#alence& which would !ake it difficult if not i!possible to agree upon such criteria? 1oreo#er how !uch headway has been !ade in in#estigating the 2uestion of the eAtent to which the $conscious or unconscious& eApectations and theories held by the analyst influence the ;clinical responses; of the patient? $To say nothing about the conscious atte!pts to influence the patient by proposing interpretations to hi! etc.& 5ears ago / introduced the ter! ;4edipus effect; to describe the influence of a theory or eApectation or prediction upon the e#ent which it predicts or describes< it will be re!e!bered that the causal chain leading to 4edipus; parricide was started by the oracle;s prediction of this e#ent. This is a characteristic and recurrent the!e of such !yths but one which see!s to ha#e failed to attract the interest of the analysts perhaps not

Page 39 of '(

accidentally. $The proble! of confir!atory drea!s suggested by the analyst is discussed by Breud for eAa!ple in Fesa!!elte Schriften /// 19@( where he says on p. 31'< ;/f anybody asserts that !ost of the drea!s which can be utili9ed in an analysis. . . owe their origin to Lthe analyst;sM suggestion then no ob"ection can be !ade fro! the point of #iew of analytic theory. 5et there is nothing in this fact; he surprisingly adds ;which would detract fro! the reliability of our results.;&

' The case of astrology nowadays a typical pseudo=science !ay illustrate this point. /t was attacked by +ristotelians and other rationalists down to 0ewton;s day for the wrong reason== for its now accepted assertion that the planets had an ;influence; upon terrestrial $;sublunar;& e#ents. /n fact 0ewton;s theory of gra#ity and especially the lunar theory of the tides was historically speaking an offspring of astrological lore. 0ewton it see!s was !ost reluctant to adopt a theory which ca!e fro! the sa!e stable as for eAa!ple the theory that ;influen9a; epide!ics are due to an astral ;influence;. +nd Falileo no doubt for the sa!e reason actually re"ected the lunar theory of the tides: and his !isgi#ings about Kepler !ay easily be eAplained by his !isgi#ings about astrology.

( 1y ,ogic of Scientific 8isco#ery $19(9 196D 1961& here usually referred to as ,.Sc.8. is the translation of ,ogik der Borschung $193'& with a nu!ber of additional notes and appendices including $on pp. 31@=1'& the letter to the .ditor of .rkenntnis !entioned here in the teAt which was first published in .rkenntnis 3 1933 pp. '@6 f. 7oncerning !y ne#er published book !entioned here in the teAt see R. 7arnap;s paper ;Neber Protokollstii9e; $4n Protocol= Sentences& .rkenntnis 3 193@ pp. @1(=@E where he gi#es an outline of !y theory on pp. @@3=E and accepts it. 3e calls !y theory ;procedure 8; and says $p. @@' top&< ;Starting fro! a point of #iew different fro! 0eurath;s; $who de#eloped what 7a!ap calls on p. @@3 ;procedure +;& ;Popper de#eloped procedure - as part of his syste!.; +nd after describing in detail !y theory of

Page 'D of '(

tests 7a!ap su!s up his #iews as follows $p. @@E&< ;+fter weighing the #arious argu!ents here discussed it appears to !e that the second language for! with procedure 8=that is in the for! here described is the !ost ade2uate a!ong the for!s of scientific language at present ad#ocated. . . in the . . . theory of knowledge.; This paper of 7arnap;s contained the first published report of !y theory of critical testing. $See also !y critical re!arks in ,.Sc.8. note 1 to section @9 p. 1D' where the date ;1933; should read ;193@;: and ch. 11 below teAt to note 39.&

6 %ittgenstein;s eAa!ple of a nonsensical pseudo=proposition is< ;Socrates is identical;. 4b#iously ;Socrates is not identical; !ust also be nonsense. Thus the negation of any nonsense will be nonsense and that of a !eaningful state!ent will be !eaningful. -ut the negation of a testable $or falsifiable& state!ent need not be testable as was pointed out first in !y ,.Sc.8. $e.g. pp. 3E f.& and later by !y critics. The confusion caused by taking testability as a criterion of !eaning rather than of de!arcation can easily be i!agined.

> The !ost recent eAa!ple of the way in which the history of this proble! is !isunderstood is +. R. %hite;s ;0ote on 1eaning and Cerification; 1ind 63 19(' pp. 66 ff. K. ,. .#ans;s article 1ind 6@ 19(3 pp. 1 ff. which 1r. %hite critici9es is eAcellent in !y opinion and unusually percepti#e. Nnderstandably enough neither of the authors can 2uite reconstruct the story. $So!e hints !ay be found in !y 4pen Society notes '6 (1 and (@ to ch. 11: and a fuller analysis in ch. 11 of the present #olu!e.& s /n ,.Sc.8. / discussed and replied to so!e likely ob"ections which afterwards were indeed raised without reference to !y replies. 4ne of the! is the contention that the falsification of a natural law is "ust as i!possible as its #erification. The answer is that this ob"ection !iAes two entirely different le#els of analysis $like the ob"ection that !athe!atical de!onstrations are i!possible since checking no !atter how often repeated can ne#er !ake it 2uite certain that we ha#e not o#erlooked a !istake&. 4n the first

Page '1 of '(

le#el there is a logical asy!!etry< one singular state!ent==say about the perihelion of 1ercury==can for!ally falsify Kepler;s laws: but these cannot be for!ally #erified by any nu!ber of singular state!ents. The atte!pt to !ini!i9e this asy!!etry can only lead to confusion. 4n another le#el we !ay hesitate to accept any state!ent e#en the si!plest obser#ation state!ent: and we !ay point out that e#ery state!ent in#ol#es interpretation in the light of theories and that it is therefore uncertain. This does not affect the funda!ental asy!!etry but it is i!portant< !ost dissectors of the heart before 3ar#ey obser#ed the wrong things==those which they eApected to see. There can ne#er be anything like a co!pletely safe obser#ation free fro! all dangers of interpretation. $This is one of the reasons why the theory of induction does not work.& The Oe!pirical basisJ consists largely of a !iAture of theories of lower degree uni#ersality $of reproducible effectsJ&. -ut the fact re!ains that relati#e to whate#er basis the in#estigator accepts $at his perilI& he can etst his theory only by trying to refute it.

9. 3u!e does not say OlogicalJ but Ode!onstrati#e J a ter!inology which / think is a little !isleading. The following two 2uotations are fro! the Treatise of 3u!an 0ature -ook / Part /// sections #i and Aii. $The italics are all 3u!eJs.&

1D This and the neAt 2uotation are fro! loc. 7it. section #i. See also 3u!eJs .n2uiry 7oncerning 3u!an Nnderstanding section /C Part // and his +bstract edited 193E by K. 1. Keynes and P. Sraffa p. 1( and 2uoted in ,.Sc.8. new appendiA GC// teAt to note 6.

11 Treatise section Aiii: section A# rule '.

1@ B. -age ;Pur .ntwicklung etc.; Peitschrift f. 3undeforschung 1933: cp 8. Kat9 +ni!als and 1en ch. C/ footnote.

Page '@ of '(

13 See section 3D of ,.Sc..8.

1' Kat9 loc. cit.

1( Kant belie#ed that 0ewton;s dyna!ics was a priori #alid. $See his 1etaphysical Boundations of 0atural Science published between the first and the second editions of the 7riti2ue of Pure Reason.& -ut if as he thought we can eAplain the #alidity of 0ewton;s theory by the fact that our intellect i!poses its laws upon nature it follows / think that our intellect !ust succeed in this: which !akes it hard to understand why a priori knowledge such as 0ewton;s should be so hard to co!e by. + so!ewhat fuller state!ent of this criticis! can be found in ch. @ especially section iA and chs. > and E of the present #olu!e. 16 + thesis sub!itted under the title ;Fewohnheit und Feset9erlebnis; to the /nstitute of .ducation of the 7ity of Cienna in 19@>. $Nnpublished.&

1> Burther co!!ents on these de#elop!ents !ay be found in chs. ' and ( below.

1E 1aA -orn 0atural Philosophy of 7ause and 7hance 4Aford 19'9 p. >.

19 0atural Philosophy of 7ause and 7hance p. 6.

@D / do not doubt that -orn and !any others would agree that theories are accepted only tentati#ely. -ut the widespread belief in induction shows that the far=reaching i!plications of this #iew are rarely seen.

@1 %ittgenstein still held this belief in 19'6: see note E to ch. @ below.

Page '3 of '(

@@ See note ( abo#e.

@3 ,.Sc.8. $see note ( abo#e& ch. A especially sections ED to E3 also section 3' ff. See also !y note ;+ Set of /ndependent +Aio!s for Probability; 1ind 0.S. '> 193E p. @>(. $This note has since been reprinted with corrections in the new appendiA Gii of ,.Sc.8. See also the neAt note but one to the present chapter.&

@' + definition in ter!s of probabilities $see the neAt note& of 7$t e& i.e. of the degree of corroboration $of a theory t relati#e to the e#idence e& satisfying the de!ands indicated in !y ,.Sc.8. sections E@ to E3 is the following<

7$t e& Q .$t e& $1 R P$t&P$t e&&

where .$t e&& Q $P$e t& = P;$e&&S$P$e t& R P$e&&& is a $non=additi#e& !easure of the eAplanatory power of t with respect to e. 0ote that 7$t e& is not a probability< it !ay ha#e #alues between =1 $refutation of t b e& and 7$t t& T R 1. State!ents t which are lawlike and thus non= #erifiable cannot e#en reach 7$t e& Q 7$t t& upon e!pirical e#idence e. 7$t t& is the degree of corroborability of t and is e2ual to the degree of testability of t or to the content of t. -ecause of the de!ands i!plied in point $6& at the end of section / abo#e / do not think howe#er that it is possible to gi#e a co!plete for!ali9ation of the idea of corroboration $or as / pre#iously used to say of confir!ation&. $+dded 19(( to the first proofs of this paper<& See also !y note ;8egree of 7onfir!ation; -ritish Kournal for the Philosophy of Science ( 19(' pp. 1'3 /f. $See also ( pp. 33'.& / ha#e since si!plified this definition as follows $-.K.P.S. 19(( ( p. 3(9<&

Page '' of '(

7$t e& Q $P$e t& = P$e&&S$P$e t& =P$et& R P$e&&

Bor a further i!pro#e!ent see -.K.P.S. 6 19(( p. (6. (E

Page '( of '(