Date post: | 03-Jun-2018 |
Category: |
Documents |
Upload: | archie-dei-magarao |
View: | 214 times |
Download: | 0 times |
of 25
8/13/2019 POPER
1/25
Cite this entry
Search the SEP
Advanced Search
Tools RSS FeedTable of Contents
What's New
Archives Projected Contents
Editorial nfor!ation Abo"t the SEP
Editorial #oard
$ow to Cite the SEP
S%ecial Characters
S"%%ort the SEP
Contact the SEP
&eta%hysics Research (ab)CS()Stanford *niversity
+%en access to the SEP is !ade %ossible by a world,wide f"ndin- initiative.Please Read $ow /o" Can $el% 0ee% the Encyclo%edia Free
Karl Popper
http://plato.stanford.edu/cgi-bin/encyclopedia/archinfo.cgihttp://plato.stanford.edu/search/searcher.pyhttp://plato.stanford.edu/search/searcher.pyhttp://plato.stanford.edu/tools/http://plato.stanford.edu/rss/sep.xmlhttp://plato.stanford.edu/contents.htmlhttp://plato.stanford.edu/new.htmlhttp://plato.stanford.edu/archives/http://plato.stanford.edu/projected-contents.htmlhttp://plato.stanford.edu/info.htmlhttp://plato.stanford.edu/about.htmlhttp://plato.stanford.edu/board.htmlhttp://plato.stanford.edu/cite.htmlhttp://plato.stanford.edu/special-characters.htmlhttp://plato.stanford.edu/fundraising/mailto:[email protected]://plato.stanford.edu/info.html#chttp://mally.stanford.edu/http://www-csli.stanford.edu/http://www-csli.stanford.edu/http://www.stanford.edu/http://www.stanford.edu/http://plato.stanford.edu/fundraising/http://plato.stanford.edu/fundraising/http://plato.stanford.edu/http://plato.stanford.edu/cgi-bin/encyclopedia/archinfo.cgihttp://plato.stanford.edu/search/searcher.pyhttp://plato.stanford.edu/search/searcher.pyhttp://plato.stanford.edu/tools/http://plato.stanford.edu/rss/sep.xmlhttp://plato.stanford.edu/contents.htmlhttp://plato.stanford.edu/new.htmlhttp://plato.stanford.edu/archives/http://plato.stanford.edu/projected-contents.htmlhttp://plato.stanford.edu/info.htmlhttp://plato.stanford.edu/about.htmlhttp://plato.stanford.edu/board.htmlhttp://plato.stanford.edu/cite.htmlhttp://plato.stanford.edu/special-characters.htmlhttp://plato.stanford.edu/fundraising/mailto:[email protected]://plato.stanford.edu/info.html#chttp://mally.stanford.edu/http://www-csli.stanford.edu/http://www.stanford.edu/http://plato.stanford.edu/fundraising/http://plato.stanford.edu/fundraising/8/13/2019 POPER
2/25
First published Thu Nov 13, 1997; substantive revision Mon Oct 9, 2006
0arl Po%%er is -enerally re-arded as one of the -reatest %hiloso%hers of science of the
12th cent"ry. $e was also a social and %olitical %hiloso%her of considerable stat"re) aself,%rofessed 3critical,rationalist4) a dedicated o%%onent of all for!s of sce%ticis!)
conventionalis!) and relativis! in science and in h"!an affairs -enerally) a co!!ittedadvocate and sta"nch defender of the 3+%en Society4) and an i!%lacable critic of
totalitarianis! in all of its for!s. +ne of the !any re!ar5able feat"res of Po%%er'stho"-ht is the sco%e of his intellect"al infl"ence. n the !odern technolo-ical and hi-hly,
s%ecialised world scientists are rarely aware of the wor5 of %hiloso%hers6 it is virt"ally
"n%recedented to find the! 7"e"in- "%) as they have done in Po%%er's case) to testify tothe enor!o"sly %ractical beneficial i!%act which that %hiloso%hical wor5 has had "%on
their own. #"t notwithstandin- the fact that he wrote on even the !ost technical !atters
with cons"!!ate clarity) the sco%e of Po%%er's wor5 is s"ch that it is co!!on%lace bynow to find that co!!entators tend to deal with the e%iste!olo-ical) scientific and social
ele!ents of his tho"-ht as if they were 7"ite dis%arate and "nconnected) and th"s the
f"nda!ental "nity of his %hiloso%hical vision and !ethod has to a lar-e de-ree beendissi%ated. $ere we will try to trace the threads which interconnect the vario"s ele!ents
of his %hiloso%hy) and which -ive it its f"nda!ental "nity.
Section $eadin-s8
9. (ife
1. #ac5dro% to his Tho"-ht
:. The Proble! of ;e!arcation
8/13/2019 POPER
3/25
indeed he initially chose the history of !"sic as a second s"bject for his Ph.;
ea!ination. S"bse7"ently) his love for !"sic beca!e one of the ins%irational forces in
the develo%!ent of his tho"-ht) and !anifested itself in his hi-hly ori-inal inter%retationof the relationshi% between do-!atic and critical thin5in-) in his acco"nt of the
distinction between objectivity and s"bjectivity) and) !ost i!%ortantly) in the -rowth of
his hostility towards all for!s of historicis!) incl"din- historicist ideas abo"t the nat"reof the 3%ro-ressive4 in !"sic. The yo"n- 0arl attended the localeal!"#nasiu#) where
he was "nha%%y with the standards of the teachin-) and) after an illness which 5e%t hi! at
ho!e for a n"!ber of !onths) he left to attend the *niversity of ?ienna in 9D9.$owever) he did not for!ally enrol at the *niversity by ta5in- the !atric"lation
ea!ination for another fo"r years. 9D9D was in !any res%ects the !ost i!%ortant
for!ative year of his intellect"al life. n that year he beca!e heavily involved in left,
win- %olitics) joined the Association of Socialist School St"dents) and beca!e for a ti!ea arist. $owever) he was 7"ic5ly disill"sioned with the doctrinaire character of the
latter) and soon abandoned it entirely. $e also discovered the %sychoanalytic theories of
Fre"d and Adler G"nder whose ae-is he en-a-ed briefly in social wor5 with de%rived
childrenH) and listened entranced to a lect"re which Einstein -ave in ?ienna on relativitytheory. The do!inance of the critical s%irit in Einstein) and its total absence in ar)
Fre"d and Adler) str"c5 Po%%er as bein- of f"nda!ental i!%ortance8 the latter) he ca!eto thin5) co"ched their theories in ter!s which !ade the! a!enable only to
confir!ation) while Einstein's theory) cr"cially) had testable i!%lications which) if false)
wo"ld have falsified the theory itself.
Po%%er obtained a %ri!ary school teachin- di%lo!a in 9D1>) too5 a Ph.;. in %hiloso%hyin 9D1) and 7"alified to teach !athe!atics and %hysics in secondary school in 9D1D. The
do!inant %hiloso%hical -ro"% in ?ienna at the ti!e was the $iener %reis) the circle of
3scientifically,!inded4 intellect"als foc"sed aro"nd oritI Schlic5) who had been
a%%ointed Professor of the %hiloso%hy of the ind"ctive sciences at ?ienna *niversity in9D11. This incl"ded R"dolf Carna%) +tto Ne"rath) ?i5tor 0raft) $ans $ahn and $erbert
Fei-l. The %rinci%al objective of the !e!bers of the Circle was to "nify the sciences)
which carried with it) in their view) the need to eli!inate !eta%hysics once and for all byshowin- that !eta%hysical %ro%ositions are !eanin-less a %roject which Schlic5 in
%artic"lar saw as derivin- fro! the acco"nt of the %ro%osition -iven in Witt-enstein's
Tractatus. Altho"-h he was friendly with so!e of the Circle's !e!bers and shared theirestee! for science) Po%%er's hostility towards Witt-enstein alienated Schlic5) and he was
never invited to beco!e a !e!ber of the -ro"%. For his %art) Po%%er beca!e increasin-ly
critical of the !ain tenets of lo-ical %ositivis!) es%ecially of what he considered to be its!is%laced foc"s on the theory of !eanin- in %hiloso%hy and "%on verification in
scientific !ethodolo-y) and reveled in the title 3the official o%%osition4 which was
bestowed "%on hi! by Ne"rath. $e artic"lated his own view of science) and hiscriticis!s of the %ositivists) in his first wor5) %"blished "nder the title&o!i' derForschun!in 9D:
8/13/2019 POPER
4/25
8/13/2019 POPER
5/25
entails that it is not) and co"ld not be) -en"inely %redictive. Psychoanalytic theories by
their nat"re are ins"fficiently %recise to have ne-ative i!%lications) and so are i!!"nised
fro! e%eriential falsification.
The arist acco"nt of history too) Po%%er held) is not scientific) altho"-h it differs in
certain cr"cial res%ects fro! %sychoanalysis. For aris!) Po%%er believed) had beeninitially scientific) in that ar had %ost"lated a theory which was -en"inely %redictive.
$owever) when these %redictions were not in fact borne o"t) the theory was saved fro!falsification by the addition of ad hochy%otheses which !ade it co!%atible with the
facts. #y this !eans) Po%%er asserted) a theory which was initially -en"inely scientific
de-enerated into %se"do,scientific do-!a.
These factors co!bined to !a5e Po%%er ta5e(alsi(iabilit"as his criterion for de!arcatin-
science fro! non,science8 if a theory is inco!%atible with %ossible e!%irical
observations it is scientific6 conversely) a theory which is co!%atible with all s"ch
observations) either beca"se) as in the case of aris!) it has been !odified solely to
acco!!odate s"ch observations) or beca"se) as in the case of %sychoanalytic theories) itis consistent with all %ossible observations) is "nscientific. For Po%%er) however) to assert
that a theory is "nscientific) is not necessarily to hold that it is "nenli-htenin-) still lessthat it is !eanin-less) for it so!eti!es ha%%ens that a theory which is "nscientific
Gbeca"se it is "nfalsifiableH at a -iven ti!e !ay beco!e falsifiable) and th"s scientific)
with the develo%!ent of technolo-y) or with the f"rther artic"lation and refine!ent of thetheory. F"rther) even %"rely !ytho-enic e%lanations have %erfor!ed a val"able f"nction
in the %ast in e%editin- o"r "nderstandin- of the nat"re of reality.
KRet"rn to Section $eadin-sL
3. The Problem of Demarcation
As Po%%er re%resents it) the central %roble! in the %hiloso%hy of science is that of
de!arcation) i.e.) of distin-"ishin- between science and what he ter!s 3non,science4)
"nder which headin- he ran5s) a!on-st others) lo-ic) !eta%hysics) %sychoanalysis) and
Adler's individ"al %sycholo-y. Po%%er is "n"s"al a!on-st conte!%orary %hiloso%hers inthat he acceptsthe validity of the $"!ean criti7"e of ind"ction) and indeed) -oes beyond
it in ar-"in- that ind"ction is never act"ally "sed by the scientist. $owever) he does not
concede that this entails the sce%ticis! which is associated with $"!e) and ar-"es thatthe #aconianMNewtonian insistence on the %ri!acy of 3%"re4 observation) as the initial
ste% in the for!ation of theories) is co!%letely !is-"ided8 all observation is selective and
theory,laden there are no %"re or theory,free observations. n this way he destabilises
the traditional view that science can be distin-"ished fro! non,science on the basis of itsind"ctive !ethodolo-y6 in contradistinction to this) Po%%er holds that there is no "ni7"e
!ethodolo-y s%ecific to science. Science) li5e virt"ally every other h"!an) and indeedor-anic) activity) Po%%er believes) consists lar-ely of %roble!,solvin-.
Po%%er) then) re%"diates ind"ction) and rejects the view that it is the characteristic !ethod
of scientific investi-ation and inference) and s"bstit"tes(alsi(iabilit"in its %lace. t is
easy) he ar-"es) to obtain evidence in favo"r of virt"ally any theory) and he conse7"ently
http://plato.stanford.edu/entries/popper/#Links%23Linkshttp://plato.stanford.edu/entries/popper/#Links%23Links8/13/2019 POPER
6/25
holds that s"ch 3corroboration4) as he ter!s it) sho"ld co"nt scientifically only if it is the
%ositive res"lt of a -en"inely 3ris5y4 %rediction) which !i-ht conceivably have been
false. For Po%%er) a theory is scientific only if it is ref"table by a conceivable event.Every -en"ine test of a scientific theory) then) is lo-ically an atte!%t to ref"te or to
falsify it) and one -en"ine co"nter,instance falsifies the whole theory. n a critical sense)
Po%%er's theory of de!arcation is based "%on his %erce%tion of the lo-ical asy!!etrywhich holds between verification and falsification8 it is lo-ically i!%ossible to
concl"sively verify a "niversal %ro%osition by reference to e%erience Gas $"!e saw
clearlyH) b"t a sin-le co"nter,instance concl"sively falsifies the corres%ondin- "niversallaw. n a word) an ece%tion) far fro! 3%rovin-4 a r"le) concl"sively ref"tes it.
Every -en"ine scientific theory then) in Po%%er's view) isprohibitive) in the sense that it
forbids) by i!%lication) %artic"lar events or occ"rrences. As s"ch it can be tested and
falsified) b"t never lo-ically verified. Th"s Po%%er stresses that it sho"ld not be inferredfro! the fact that a theory has withstood the !ost ri-oro"s testin-) for however lon- a
%eriod of ti!e) that it has been verified6 rather we sho"ld reco-nise that s"ch a theory has
received a hi-h !eas"re of corroboration. and !ay be %rovisionally retained as the bestavailable theory "ntil it is finally falsified Gif indeed it is ever falsifiedH) andMor is
s"%erseded by a better theory.
Po%%er has always drawn a clear distinction between the lo!icof falsifiability and its
applied #ethodolo!". The lo-ic of his theory is "tterly si!%le8 if a sin-le ferro"s !etal is"naffected by a !a-netic field it cannot be the case that all ferro"s !etals are affected by
!a-netic fields. (o-ically s%ea5in-) a scientific law is concl"sively falsifiable altho"-h it
is not concl"sively verifiable. ethodolo-ically) however) the sit"ation is !"ch !ore
co!%le8 no observation is free fro! the %ossibility of error conse7"ently we !ay7"estion whether o"r e%eri!ental res"lt was what it a%%eared to be.
Th"s) while advocatin- falsifiability as the criterion of de!arcation for science) Po%%er
e%licitly allows for the fact that in %ractice a sin-le conflictin- or co"nter,instance isnever s"fficient !ethodolo-ically to falsify a theory) and that scientific theories are often
retained even tho"-h !"ch of the available evidence conflicts with the!) or is ano!alo"s
with res%ect to the!. Scientific theories !ay) and do) arise -enetically in !any different
ways) and the !anner in which a %artic"lar scientist co!es to for!"late a %artic"lartheory !ay be of bio-ra%hical interest) b"t it is of no conse7"ence as far as the
%hiloso%hy of science is concerned. Po%%er stresses in %artic"lar that there is no "ni7"e
way) no sin-le !ethod s"ch as ind"ction) which f"nctions as the ro"te to scientifictheory) a view which Einstein %ersonally endorsed with his affir!ation that 3There is no
lo-ical %ath leadin- to Kthe hi-hly "niversal laws of scienceL. They can only be reached
by int"ition) based "%on so!ethin- li5e an intellect"al love of the objects of e%erience4.Science) in Po%%er's view) starts with %roble!s rather than with observations it is)
indeed) %recisely in the contet of -ra%%lin- with a %roble! that the scientist !a5es
observations in the first instance8 his observations are selectively desi-ned to test the
etent to which a -iven theory f"nctions as a satisfactory sol"tion to a -iven %roble!.
8/13/2019 POPER
7/25
+n this criterion of de!arcation %hysics) che!istry) and Gnon,intros%ectiveH %sycholo-y)
a!on-st others) are sciences) %sychoanalysis is a %re,science Gi.e.) it "ndo"btedly
contains "sef"l and infor!ative tr"ths) b"t "ntil s"ch ti!e as %sychoanalytical theoriescan be for!"lated in s"ch a !anner as to be falsifiable) they will not attain the stat"s of
scientific theoriesH) and astrolo-y and %hrenolo-y are %se"do,sciences. For!ally) then)
Po%%er's theory of de!arcation !ay be artic"lated as follows8 where a 3basic state!ent4is to be "nderstood as a %artic"lar observation,re%ort) then we !ay say that a theory is
scientific if and only if it divides the class of basic state!ents into the followin- two non,
e!%ty s"b,classes8 GaH the class of all those basic state!ents with which it is inconsistent)or which it %rohibits this is the class of itspotential (alsi(iersGi.e.) those state!ents
which) if tr"e) falsify the whole theoryH) and GbH the class of those basic state!ents with
which it is consistent) or which it %er!its Gi.e.) those state!ents which) if tr"e)
corroborate it) or bear it o"tH.KRet"rn to Section $eadin-sL
4. The ro!th of "uman Kno!ledge
For Po%%er accordin-ly) the -rowth of h"!an 5nowled-e %roceeds fro! o"r %roble!s
and fro! o"r atte!%ts to solve the!. These atte!%ts involve the for!"lation of theorieswhich) if they are to e%lain ano!alies which eist with res%ect to earlier theories) !"st
-o beyond eistin- 5nowled-e and therefore re7"ire a lea% of the i!a-ination. For this
reason) Po%%er %laces s%ecial e!%hasis on the role %layed by the inde%endent creativei!a-ination in the for!"lation of theory. The centrality and %riority ofproble#sin
Po%%er's acco"nt of science is %ara!o"nt) and it is this which leads hi! to characterise
scientists as 3%roble!,solvers4. F"rther) since the scientist be-ins with %roble!s ratherthan with observations or 3bare facts4) Po%%er ar-"es that the only lo-ical techni7"e
which is an inte-ral %art of scientific !ethod is that of the ded"ctive testin- of theories
which are not the!selves the %rod"ct of any lo-ical o%eration. n this ded"ctive%roced"re concl"sions are inferred fro! a tentative hy%othesis. These concl"sions are
then co!%ared with one another and with other relevant state!ents to deter!ine whether
they falsify or corroborate the hy%othesis. S"ch concl"sions are not directly co!%ared
with the facts) Po%%er stresses) si!%ly beca"se there are no 3%"re4 facts available6 allobservation,state!ents are theory,laden) and are as !"ch a f"nction of %"rely s"bjective
factors Ginterests) e%ectations) wishes) etc.H as they are a f"nction of what is objectively
real.
$ow then does the ded"ctive %roced"re wor5 Po%%er s%ecifies fo"r ste%s8
GaH The first is(or#al) a testin- of the internal consistency of the theoretical syste! to seeif it involves any contradictions.
GbH The second ste% isse#i(or#al,the aio!atisin- of the theory to distin-"ish betweenits e!%irical and its lo-ical ele!ents. n %erfor!in- this ste% the scientist !a5es the
lo-ical for! of the theory e%licit. Fail"re to do this can lead to cate-ory,!ista5es the
scientist ends "% as5in- the wron- 7"estions) and searches for e!%irical data where none
are available. ost scientific theories contain analytic Gi.e.) a prioriH and syntheticele!ents) and it is necessary to aio!atise the! in order to distin-"ish the two clearly.
http://plato.stanford.edu/entries/popper/#Links%23Linkshttp://plato.stanford.edu/entries/popper/#Links%23Links8/13/2019 POPER
8/25
GcH The third ste% is the co!%arin- of the new theory with eistin- ones to deter!ine
whether it constit"tes an advance "%on the!. f it does not constit"te s"ch an advance) it
will not be ado%ted. f) on the other hand) its e%lanatory s"ccess !atches that of theeistin- theories) and additionally) it e%lains so!e hitherto ano!alo"s %heno!enon) or
solves so!e hitherto "nsolvable %roble!s) it will be dee!ed to constit"te an advance
"%on the eistin- theories) and will be ado%ted. Th"s science involves theoretical%ro-ress. $owever) Po%%er stresses that we ascertain whether one theory is better than
another by ded"ctively testin- both theories) rather than by ind"ction. For this reason) he
ar-"es that a theory is dee!ed to be better than another if Gwhile "nfalsifiedH it has-reater e!%irical content) and therefore -reater %redictive %ower than its rival. The
classic ill"stration of this in %hysics was the re%lace!ent of Newton's theory of "niversal
-ravitation by Einstein's theory of relativity. This el"cidates the nat"re of science as
Po%%er sees it8 at any -iven ti!e there will be a n"!ber of conflictin- theories orconject"res) so!e of which will e%lain !ore than others. The latter will conse7"ently be
%rovisionally ado%ted. n short) for Po%%er any theoryis better than a 3rival4 theory 4if
has!reater e#pirical content) and hence!reater predictive po5er) than /.
GdH The fo"rth and final ste% is the testin- of a theory by the e!%irical a%%lication of theconcl"sions derived fro! it. f s"ch concl"sions are shown to be tr"e) the theory is
corroborated Gb"t never verifiedH. f the concl"sion is shown to be false) then this is ta5enas a si-nal that the theory cannot be co!%letely correct Glo-ically the theory is falsifiedH)
and the scientist be-ins his 7"est for a better theory. $e does not) however) abandonthe
%resent theory "ntil s"ch ti!e as he has a better one to s"bstit"te for it. ore %recisely)the !ethod of theory,testin- is as follows8 certain sin-"lar %ro%ositions are ded"ced fro!
the new theory these are %redictions) and of s%ecial interest are those %redictions
which are 3ris5y4 Gin the sense of bein- int"itively i!%la"sible or of bein- startlin-ly
novelH and e%eri!entally testable. Fro! a!on-st the latter the scientist net selectsthose which are not derivable fro! the c"rrent or eistin- theory of %artic"lar
i!%ortance are those which contradict the c"rrent or eistin- theory. $e then see5s a
decision as re-ards these and other derived state!ents by co!%arin- the! with theres"lts of %ractical a%%lications and e%eri!entation. f the new %redictions are borne o"t)
then the new theory is corroboratedGand the old one falsifiedH) and is ado%ted as a
wor5in- hy%othesis. f the %redictions are not borne o"t) then they falsify the theory fro!which they are derived. Th"s Po%%er retains an ele!ent of e!%iricis!8 for hi! scientific
!ethod does involve !a5in- an a%%eal to e%erience. #"t "nli5e traditional e!%iricists)
Po%%er holds that e%erience cannot deter#inetheory Gi.e.) we do not ar-"e or infer fro!
observation to theoryH) it rather deli#itsit8 it shows which theories are false) not whichtheories are tr"e. oreover) Po%%er also rejects the e!%iricist doctrine that e!%irical
observations are) or can be) infallible) in view of the fact that they are the!selves theory,
laden.
The -eneral %ict"re of Po%%er's %hiloso%hy of science) then is this8 $"!e's %hiloso%hy
de!onstrates that there is a contradiction i!%licit in traditional e!%iricis!) which holds
both that all 5nowled-e is derived fro! e%erience andthat "niversal %ro%ositions
Gincl"din- scientific lawsH are verifiable by reference to e%erience. The contradiction)which $"!e hi!self saw clearly) derives fro! the atte!%t to show that) notwithstandin-
the o%en,ended nat"re of e%erience) scientific laws !ay be constr"ed as e!%irical
8/13/2019 POPER
9/25
-eneralisations which are in so!e way finally confir!able by a 3%ositive4 e%erience.
Po%%er eli!inates the contradiction by rejectin- the first of these %rinci%les and re!ovin-
the de!and for e!%irical verification in favo"r of e!%irical falsification in the second.Scientific theories) for hi!) are not ind"ctively inferred fro! e%erience) nor is scientific
e%eri!entation carried o"t with a view to verifyin- or finally establishin- the tr"th of
theories6 rather) all 'no5led!e is provisional, conectural, h"pothetical we can neverfinally %rove o"r scientific theories) we can !erely G%rovisionallyH confir! or
Gconcl"sivelyH ref"te the!6 hence at any -iven ti!e we have to choose between the
%otentially infinite n"!ber of theories which will e%lain the set of %heno!ena "nderinvesti-ation. Faced with this choice) we can only eli!inate those theories which are
de!onstrably false) and rationally choose between the re!ainin-) "nfalsified theories.
$ence Po%%er's e!%hasis on the i!%ortance of the critical s%irit to science for hi!
critical thin5in- is the very essence of rationality. For it is only by critical tho"-ht that wecan eli!inate false theories) and deter!ine which of the re!ainin- theories is the best
available one) in the sense of %ossessin- the hi-hest level of e%lanatory force and
%redictive %ower. t is %recisely this 5ind of critical thin5in- which is cons%ic"o"s by its
absence in conte!%orary aris! and in %sychoanalysis.KRet"rn to Section $eadin-sL
#. Probabilit$% Kno!ledge and &erisimilitude
n the view of !any social scientists) the !ore %robable a theory is) the betterit is) and ifwe have to choose between two theories which are e7"ally stron- in ter!s of their
e%lanatory %ower) and differ only in that one is %robable and the other is i!%robable)
then we sho"ld choose the for!er. Po%%er rejects this. Science) or to be %recise) thewor5in- scientist) is interested) in Po%%er's view) in theories with a hi-h infor!ative
content) beca"se s"ch theories %ossess a hi-h %redictive %ower and are conse7"ently
hi-hly testable. #"t if this is tr"e) Po%%er ar-"es) then) %aradoical as it !ay so"nd) the!ore i#probablea theory is the better it is scientifically) beca"se the %robability and
infor!ative content of a theory vary inversely the hi-her the infor!ative content of a
theory the lower will be its %robability) for the !ore infor!ation a state!ent contains) the
-reater will be the n"!ber of ways in which it !ay t"rn o"t to be false. Th"s thestate!ents which are of s%ecial interest to the scientist are those with a hi-h infor!ative
content and Gconse7"entiallyH a low %robability) which nevertheless co!e close to the
tr"th. nfor!ative content) which is in inverse %ro%ortion to %robability) is in direct%ro%ortion to testability. Conse7"ently the severity of the test to which a theory can be
s"bjected) and by !eans of which it is falsified or corroborated) is all,i!%ortant.
For Po%%er) all scientific criticis! !"st be %iece!eal) i.e.) he holds that it is not %ossibleto 7"estion every as%ect of a theory at once. ore %recisely) while atte!%tin- to resolve a%artic"lar %roble! a scientist of necessity acce%ts all 5inds of thin-s as "n%roble!atic.
These thin-s constit"te what Po%%er ter!s the 3bac5-ro"nd 5nowled-e4. $owever) he
stresses that the bac5-ro"nd 5nowled-e is not5nowled-e in the sense of bein-concl"sively established6 it !ay be challen-ed at any ti!e) es%ecially if it is s"s%ected
that its "ncritical acce%tance !ay be res%onsible for diffic"lties which are s"bse7"ently
enco"ntered. Nevertheless) it is clearly not %ossible to 7"estion both the theory and the
http://plato.stanford.edu/entries/popper/#Links%23Linkshttp://plato.stanford.edu/entries/popper/#Links%23Links8/13/2019 POPER
10/25
bac5-ro"nd 5nowled-e at the sa!e ti!e Ge.-.) in cond"ctin- an e%eri!ent the scientist
of necessity ass"!es that the a%%arat"s "sed is in wor5in- orderH.
$ow then can one be certain that one is 7"estionin- the ri-ht thin- The Po%%eriananswer is that we cannot have absol"te certainty here) b"t re%eated tests "s"ally show
where the tro"ble lies. Even observation state!ents) Po%%er !aintains) are fallible) andscience in his view is not a 7"est for certain 5nowled-e) b"t an evol"tionary %rocess in
which hy%otheses or conject"res are i!a-inatively %ro%osed and tested in order toe%lain facts or to solve %roble!s. Po%%er e!%hasises both the i!%ortance of 7"estionin-
the bac5-ro"nd 5nowled-e when the need arises) and the si-nificance of the fact that
observation,state!ents are theory,laden) and hence fallible. For while falsifiability issi!%le as a lo-ical %rinci%le) in %ractice it is eceedin-ly co!%licated no sin-le
observation can ever be ta5en to falsify a theory) for there is always the %ossibility GaH
that the observation itself is !ista5en) or GbH that the ass"!ed bac5-ro"nd 5nowled-e isfa"lty or defective.
Po%%er was initially "neasy with the conce%t of tr"th) and in his earliest writin-s heavoided assertin- that a theory which is corroborated is tr"e for clearly if every theory
is an o%en,ended hy%othesis) as he !aintains) then ipso (actoit has to be at least%otentially false. For this reason Po%%er restricted hi!self to the contention that a theory
which is falsified is false and is 5nown to be s"ch) and that a theory which re%laces a
falsified theory Gbeca"se it has a hi-her e!%irical content than the latter) and e%lainswhat has falsified itH is a 3better theory4 than its %redecessor. $owever) he ca!e to acce%t
Tars5i's refor!"lation of the corres%ondence theory of tr"th) and in onectures and
e(utationsG9D@:H he inte-rated the conce%ts of tr"th and content to fra!e the
!etalo-ical conce%t of 3tr"thli5eness4 or 3verisi#ilitude4. A 3-ood4 scientific theory)Po%%er th"s ar-"ed) has a hi-her level of verisi!ilit"de than its rivals) and he e%licated
this conce%t by reference to the lo-ical conse7"ences of theories. A theory's content is thetotality of its lo-ical conse7"ences) which can be divided into two classes8 there is the3truthcontent4 of a theory) which is the class of tr"e %ro%ositions which !ay be derived
fro! it) on the one hand) and the 3(alsit"content4 of a theory) on the other hand) which is
the class of the theory's false conse7"ences Gthis latter class !ay of co"rse be e!%ty) andin the case of a theory which is tr"e is necessarily e!%tyH.
Po%%er offered two !ethods of co!%arin- theories in ter!s of verisi!ilit"de) the
7"alitative and 7"antitative definitions. +n the 7"alitative acco"nt) Po%%er asserted8
Ass"!in- that the tr"th,content and the falsity,content of two theories t9and t1are
co!%arable) we can say that t1is !ore closely si!ilar to the tr"th) or corres%onds betterto the facts) than t9) if and only if either8
GaH the tr"th,content b"t not the falsity,content of t1eceeds that of t9) or
GbH the falsity,content of t9) b"t not its tr"th,content) eceeds that of t1. Gonectures and
e(utations) 1::H.
8/13/2019 POPER
11/25
$ere) verisi!ilit"de is defined in ter!s of s"bclass relationshi%s8 t1has a hi-her level of
verisi!ilit"de than t9if and only if their tr"th, and falsity,contents are co!%arable
thro"-h s"bclass relationshi%s) and eitherGaH t1's tr"th,content incl"des t9's and t1's falsity,content) if it eists) is incl"ded in) or is the sa!e as) t9's) orGbH t1's tr"th,content incl"des
or is the sa!e as t9's and t1's falsity,content) if it eists) is incl"ded in t9's.
+n the 7"antitative acco"nt) verisi!ilit"de is defined by assi-nin- 7"antities to contents)
where the inde of the content of a -iven theory is its lo-ical i!%robability G-iven a-ainthat content and %robability vary inverselyH. For!ally) then) Po%%er defines the
7"antitative verisi!ilit"de which a state!ent 3a4 %ossesses by !eans of a for!"la8
8sGaH O tTGaH tFGaH)
where 8sGaH re%resents the verisi!ilit"de of a) tTGaH is a !eas"re of the tr"th,content ofa) and tFGaH is a !eas"re of its falsity,content.
The "tilisation of either !ethod of co!%"tin- verisi!ilit"de shows) Po%%er held) thateven if a theory t1with a hi-her content than a rival theory t9is s"bse7"ently falsified) it
can still le-iti!ately be re-arded as a better theory than t9) and 3better4 is here now"nderstood to !ean t1is closer to the truththan t9. Th"s scientific %ro-ress involves) on
this view) the abandon!ent of %artially tr"e) b"t falsified) theories) for theories with a
hi-her level of verisi!ilit"de) i.e.) which a%%roach !ore closely to the tr"th. n this way)verisi!ilit"de allowed Po%%er to !iti-ate what !any saw as the %essi!is! of an anti,
ind"ctivist %hiloso%hy of science which held that !ost) if not all scientific theories are
false) and that a tr"e theory) even if discovered) co"ld not be 'no5nto be s"ch. With theintrod"ction of the new conce%t) Po%%er was able to re%resent this as an essentially
o%ti!istic %osition in ter!s of which we can le-iti!ately be said to have reason to
believe that science !a5es %ro-ress towards the tr"th thro"-h the falsification andcorroboration of theories. Scientific %ro-ress) in other words) co"ld now be re%resentedas %ro-ress to5ardsthe tr"th) and e%eri!ental corroboration co"ld be seen an indicator
of verisi!ilit"de.
$owever) in the 9DB2's a series of %a%ers %"blished by researchers s"ch as iller) TichQ)
and =rnba"! in %artic"lar revealed f"nda!ental defects in Po%%er's for!al definitionsof verisi!ilit"de. The si-nificance of this wor5 was that verisi!ilit"de is lar-ely
i!%ortant in Po%%er's syste! beca"se of its a%%lication to theories which are 5nown to be
(alse. n this connection) Po%%er had written8
*lti!ately) the idea of verisi!ilit"de is !ost i!%ortant in cases where we 5now that wehave to wor5 with theories which are at besta%%roi!ationsthat is to say) theories of
which we 5now that they cannot be tr"e. GThis is often the case in the social sciencesH. n
these cases we can still s%ea5 of better or worse a%%roi!ations to the tr"th Gand wetherefore do not need to inter%ret these cases in an instr"!entalist senseH. Gonectures
and e(utations) 1:>H.
8/13/2019 POPER
12/25
For these reasons) the deficiencies discovered by the critics in Po%%er's for!al definitions
were seen by !any as devastatin-) %recisely beca"se the !ost si-nificant of these related
to the levels of verisi!ilit"de of(alsetheories. n 9DB
8/13/2019 POPER
13/25
nat"re of the social sciences which see5 to describe and e%licate the! syste!atically)
%artic"larly history. t is in this contet that he offers an acco"nt of the nat"re of scientific
%rediction) which in t"rn allows hi! a %oint of de%art"re for his attac5 "%ontotalitarianis! and all its intellect"al s"%%orts) es%ecially holis! and historicis!. n this
contet holis! is to be "nderstood as the view that h"!an social -ro"%in-s are -reater
than the s"! of their !e!bers) that s"ch -ro"%in-s are 3or-anic4 entities in their ownri-ht) that they act on their h"!an !e!bers and sha%e their destinies) and that they are
s"bject to their own inde%endent laws of develo%!ent. $istoricis!) which is closely
associated with holis!) is the belief that history develo%s ineorably and necessarilyaccordin- to certain %rinci%les or r"les towards a deter!inate end Gas for ea!%le in the
dialectic of $e-el) which was ado%ted and i!%le!ented by arH. The lin5 between
holis! and historicis! is that the holist believes that individ"als are essentially for!ed
by the social -ro"%in-s to which they belon-) while the historicist who is "s"ally alsoa holist holds that we can "nderstand s"ch a social -ro"%in- only in ter!s of the
internal %rinci%les which deter!ine its develo%!ent.
These beliefs lead to what Po%%er calls 3The $istoricist ;octrine of the Social Sciences4)the views GaH that the %rinci%al tas5 of the social sciences is to !a5e %redictions abo"t the
social and %olitical develo%!ent of !an) and GbH that the tas5 of %olitics) once the 5ey
%redictions have been !ade) is) in ar's words) to lessen the 3birth %an-s4 of f"t"re
social and %olitical develo%!ents. Po%%er thin5s that this view of the social sciences isboth theoretically !isconceived Gin the sense of bein- based "%on a view of nat"ral
science and its !ethodolo-y which is totally wron-H) and socially dan-ero"s) as it leads
inevitably to totalitarianis! and a"thoritarianis! to centralised -overn!ental controlof the individ"al and the atte!%ted i!%osition of lar-e,scale social %lannin-. A-ainst this
Po%%er stron-ly advances the view that any h"!an social -ro"%in- is no !ore Gor lessH
than the s"! of its individ"al !e!bers) that what ha%%ens in history is the Glar-ely
"n%lanned and "nforeseeableH res"lt of the actions of s"ch individ"als) and that lar-escale social %lannin- to an antecedently conceived bl"e%rint is inherently !isconceived
and inevitably disastro"s %recisely beca"se h"!an actions have conse7"ences
which cannot be foreseen. Po%%er) then) is an historical indeter#inist) insofar as he holdsthat history does not evolve in accordance with intrinsic laws or %rinci%les) that in the
absence of s"ch laws and %rinci%les "nconditional %rediction in the social sciences is an
i!%ossibility) and that there is no s"ch thin- as historical necessity.
The lin5 between Po%%er's theory of 5nowled-e and his social %hiloso%hy is his
fallibilis! j"st as we !a5e theoretical %ro-ress in science by deliberately s"bjectin-
o"r theories to critical scr"tiny) and abandonin- those which have been falsified) so too)
Po%%er holds) the critical s%irit can and sho"ld be s"stained at the social level. ores%ecifically) the o%en society can be bro"-ht abo"t only if it is %ossible for the individ"al
citiIen to eval"ate critically the conse7"ences of the i!%le!entation of -overn!ent
%olicies) which can then be abandoned or !odified in the li-ht of s"ch critical scr"tiny in s"ch a society) the ri-hts of the individ"al to criticise ad!inistrative %olicies will be
for!ally safe-"arded and "%held) "ndesirable %olicies will be eli!inated in a !anner
analo-o"s to the eli!ination of falsified scientific theories) and differences between%eo%le on social %olicy will be resolved by critical disc"ssion and ar-"!ent rather than
8/13/2019 POPER
14/25
by force. The o%en society as th"s conceived of by Po%%er !ay be defined as 3an
association of free individ"als res%ectin- each other's ri-hts within the fra!ewor5 of
!"t"al %rotection s"%%lied by the state) and achievin-) thro"-h the !a5in- ofres%onsible) rational decisions) a -rowin- !eas"re of h"!ane and enli-htened life4
G(evinson) R.#.n *e(ense o( .lato) 9BH. As s"ch) Po%%er holds) it is not a "to%ian ideal)
b"t an e!%irically realised for! of social or-anisation which) he ar-"es) is in everyres%ect s"%erior to its Greal or %otentialH totalitarian rivals. #"t he does not en-a-e in a
!oral defence of the ideolo-y of liberalis!6 rather his strate-y is the !"ch dee%er one of
showin- that totalitarianis! is ty%ically based "%on historicist and holist %res"%%ositions)and of de!onstratin- that these %res"%%ositions are f"nda!entally incoherent.
KRet"rn to Section $eadin-sL
,. (cientific Kno!ledge% "istor$% and Prediction
At a very -eneral level) Po%%er ar-"es that historicis! and holis! have their ori-ins inwhat he ter!s 3one of the oldest drea!s of !an5ind the drea! of %ro%hecy) the idea
that we can 5now what the f"t"re has in store for "s) and that we can %rofit fro! s"ch5nowled-e by adj"stin- o"r %olicy to it.4 Gonectures and e(utations) ::H. This drea!
was -iven f"rther i!%et"s) he s%ec"lates) by the e!er-ence of a -en"ine %redictiveca%ability re-ardin- s"ch events as solar and l"nar ecli%ses at an early sta-e in h"!an
civilisation) which has of co"rse beco!e increasin-ly refined with the develo%!ent of the
nat"ral sciences and their conco!itant technolo-ies. The 5ind of reasonin- which has!ade) and contin"es to !a5e) historicis! %la"sible !ay) on this acco"nt) be
reconstr"cted as follows8 if the a%%lication of the laws of the nat"ral sciences can lead to
the s"ccessf"l %rediction of s"ch f"t"re events as ecli%ses) then s"rely it is reasonable toinfer that 5nowled-e of the laws of history as yielded by a social science or sciences
Gass"!in- that s"ch laws eistH wo"ld lead to the s"ccessf"l %rediction of s"ch f"t"re
social %heno!ena as revol"tions Why sho"ld it be %ossible to %redict an ecli%se) b"t nota revol"tion Why can we not conceive of a social science which co"ld and wo"ld
f"nction as the theoretical nat"ral sciences f"nction) and yield %recise "nconditional
%redictions in the a%%ro%riate s%here of a%%lication These are a!on-st the 7"estions
which Po%%er see5s to answer) and in doin- so) to show that they are based "%on a seriesof !isconce%tions abo"t the nat"re of science) and abo"t the relationshi% between
scientific laws and scientific %rediction.
$is first ar-"!ent !ay be s"!!arised as follows8 in relation to the critically i!%ortant
conce%t of %rediction) Po%%er !a5es a distinction between what he ter!s 3conditionalscientific %redictions4) which have the for! 3fta5es %lace) then 4will ta5e %lace4) and
3"nconditional scientific %ro%hecies4) which have the for! 34will ta5e %lace4. Contraryto %o%"lar belief) it is the for!er rather than the latter which are ty%ical of the nat"ralsciences) which !eans that ty%ically %rediction in nat"ral science is conditional and
li!ited in sco%e it ta5es the for! of hy%othetical assertions statin- that certain
s%ecified chan-es will co!e abo"t if %artic"lar s%ecified events antecedently ta5e %lace.This is not to deny that 3"nconditional scientific %ro%hecies4) s"ch as the %rediction of
ecli%ses) for ea!%le) do ta5e %lace in science) and that the theoretical nat"ral sciences
!a5e the! %ossible. $owever) Po%%er ar-"es that GaH these "nconditional %ro%hecies are
http://plato.stanford.edu/entries/popper/#Links%23Linkshttp://plato.stanford.edu/entries/popper/#Links%23Links8/13/2019 POPER
15/25
not characteristicof the nat"ral sciences) and GbH that the !echanis! whereby they
occ"r) in the very li!ited way in which they do) is not "nderstood by the historicist.
What is the !echanis! which !a5es "nconditional scientific %ro%hecies %ossible Theanswer is that s"ch %ro%hecies can so!eti!es be derived fro! a co!bination of
conditional %redictions Gthe!selves derived fro! scientific lawsH andeistentialstate!ents s%ecifyin- that the conditions in relation to the syste! bein- investi-ated are
f"lfilled. Sche!atically) this can be re%resented as follows8
KC.P. U E.S.LO*.P.
where C.P. O Conditional Prediction6 E.S. O Eistential State!ent6 *.P. O *nconditional
Pro%hecy. The !ost co!!on ea!%les of "nconditional scientific %ro%hecies in science
relate to the %rediction of s"ch %heno!ena as l"nar and solar ecli%ses and co!ets.
=iven) then) that this is the !echanis! which -enerates "nconditional scientific
%ro%hecies) Po%%er !a5es two related clai!s abo"t historicis!8 GaH That the historicistdoes not in fact derive his "nconditional scientific %ro%hecies in this !anner fro!
conditional %redictions) and GbH the historicist cannotdo so beca"se lon-,ter!"nconditional scientific %ro%hecies can be derived fro! conditional %redictions only if
they a%%ly to syste!s which are well,isolated) stationary) and rec"rrent Gli5e o"r solar
syste!H. S"ch syste!s are 7"ite rare in nat"re) and h"!an society is !ost e!%haticallynot one of the!.
This) then) Po%%er ar-"es) is the reason why it is a f"nda!ental !ista5e for the historicist
to ta5e the "nconditional scientific %ro%hecies of ecli%ses as bein- ty%ical and
characteristic of the %redictions of nat"ral science in fact s"ch %redictions are %ossible
only beca"se o"r solar syste! is a stationary and re%etitive syste! which is isolated fro!other s"ch syste!s by i!!ense e%anses of e!%ty s%ace. The solar syste! aside) there
are very few s"ch syste!s aro"nd for scientific investi-ation !ost of the others areconfined to the field of biolo-y) where "nconditional %ro%hecies abo"t the life,cycles of
or-anis!s are !ade %ossible by the eistence of %recisely the sa!e factors. Th"s one of
the fallacies co!!itted by the historicist is to ta5e the Grelatively rareH instances of
"nconditional %ro%hecies in the nat"ral science as constit"tin- the essence of whatscientific %rediction is) to fail to see that s"ch %ro%hecies a%%ly only to syste!s which are
isolated) stationary) and re%etitive) and to see5 to a%%ly the !ethod of scientific %ro%hecy
to h"!an society and h"!an history. The latter) of co"rse) is notan isolated syste! Ginfact it's not a syste! at allH) it is constantly chan-in-) and it contin"ally "nder-oes ra%id)
non,re%etitive develo%!ent. n the !ost f"nda!ental sense %ossible) every event in
h"!an history is discrete) novel) 7"ite "ni7"e) and ontolo-ically distinct fro! every otherhistorical event. For this reason) it is i!%ossible in %rinci%le that "nconditional scientific
%ro%hecies co"ld be !ade in relation to h"!an history the idea that the s"ccessf"l
"nconditional %rediction of ecli%ses %rovides "s with reasonable -ro"nds for the ho%e ofs"ccessf"l "nconditional %rediction re-ardin- the evol"tion of h"!an history t"rns o"t to
be based "%on a -ross !isconce%tion) and is 7"ite false. As Po%%er hi!self concl"des)
The fact that we %redict ecli%ses does not) therefore) %rovide a valid reason for e%ectin-
8/13/2019 POPER
16/25
8/13/2019 POPER
17/25
stands in need of scientific e%lanation) and it cannot therefore f"nction as the fra!e of
reference in ter!s of which anythin- else can be scientifically e%lained or %redicted.
A %oint which connects with this has to do with the role which the evol"tion of h"!an5nowled-e has %layed in the historical develo%!ent of h"!an society. t is incontestable
that) as ar hi!self observed) there has been a ca"sal lin5 between the two) in the sensethat advances in scientific and technolo-ical 5nowled-e have -iven rise to wides%read
-lobal chan-es in %atterns of h"!an social or-anisation and social interaction) which int"rn have led to social str"ct"res Ge.-. ed"cational syste!sH which f"rther -rowth in
h"!an 5nowled-e. n short) the evol"tion of h"!an history has been stron-ly infl"enced
by the !ro5th o( hu#an 'no5led!e) and it is etre!ely li5ely that this will contin"e to bethe case all the e!%irical evidence s"--ests that the lin5 between the two is
%ro-ressively consolidatin-. $owever) this -ives rise to f"rther %roble!s for the
historicist. n the first %lace) the state!ent that 3if there is s"ch a thin- as -rowin- h"!an5nowled-e) then we cannot antici%ate today what we shall 5now only to!orrow4 is)
Po%%er holds) int"itively hi-hly %la"sible. oreover) he ar-"es) it is lo-ically
de!onstrable by a consideration of the i!%lications of the fact that no scientific%redictor) h"!an or otherwise) can %ossibly %redict) by scientific !ethods) its own f"t"re
res"lts. Fro! this it follows) he holds) that 3no society can %redict) scientifically) its own
f"t"re states of 5nowled-e4. GThe .overt" o( /istoricis#) viiH. Th"s) while the f"t"re
evol"tion of h"!an history is etre!ely li5ely to be infl"enced by new develo%!ents inh"!an 5nowled-e) as it always has in the %ast) we cannot now scientifically deter!ine
what s"ch 5nowled-e will be. Fro! this it follows that if the f"t"re holds any new
discoveries or any new develo%!ents in the -rowth of o"r 5nowled-e Gand -iven thefallible nat"re of the latter) it is inconceivable that it does notH) then it is i!%ossible for "s
to %redict the! now) and it is therefore i!%ossible for "s to %redict the f"t"re
develo%!ent of h"!an history now) -iven that the latter will) at least in %art) be
deter!ined by the f"t"re -rowth of o"r 5nowled-e. Th"s once a-ain historicis! colla%ses the drea! of a theoretical) %redictive science of history is "nrealisable) beca"se it is an
i!%ossible drea!.
Po%%er's ar-"!ents a-ainst holis!) and in %artic"lar his ar-"!ents a-ainst the %ro%rietyof lar-e,scale %lannin- of social str"ct"res) are interconnected with his de!onstration of
the lo-ical shortco!in-s of the %res"%%ositions of historicis!. S"ch %lannin- Gwhich
act"ally too5 %lace) of co"rse) in the *SSR) in China) and in Ca!bodia) for ea!%le)"nder totalitarian re-i!es which acce%ted for!s of historicis! and holis!H) Po%%er
%oints o"t) is necessarily str"ct"red in the li-ht of the %redictions which have been !ade
abo"t f"t"re history on the basis of the so,called 3laws4 which historicists s"ch as ar
and ao clai!ed to have discovered in relation to h"!an history. Accordin-ly)reco-nition that there are no s"ch laws) and that "nconditional %redictions abo"t f"t"re
history are based) at best) "%on nothin- !ore s"bstantial than the observation of
contin-ent trends) shows that) fro! a %"rely theoretical as well as a %ractical %oint ofview) lar-e,scale social %lannin- is indeed a reci%e for disaster. n s"!!ary)
"nconditional lar-e,scale %lannin- for the f"t"re is theoretically as well as %ractically
!is-"ided) beca"se) a-ain) %art of what we are %lannin- for is o"r f"t"re 5nowled-e) ando"r f"t"re 5nowled-e is not so!ethin- which we can in %rinci%le now %ossess we
8/13/2019 POPER
18/25
cannot ade7"ately %lan for "ne%ected advances in o"r f"t"re 5nowled-e) or for the
effects which s"ch advances will have "%on society as a whole. The acce%tance of
historical indeter!inis!) then) as the only %hiloso%hy of history which is co!!ens"ratewith a %ro%er "nderstandin- of the nat"re of scientific 5nowled-e) fatally "nder!ines
both historicis! and holis!.
Po%%er's criti7"e of both historicis! and holis! is balanced) on the %ositive side) by his
stron- defence of the o%en society) the view) a-ain) that a society is e7"ivalent to the s"!of its !e!bers) that the actions of the !e!bers of society serve to fashion and to sha%e
it) not conversely) and that the social conse7"ences of intentional actions are very often)
and very lar-ely) "nintentional. This is why Po%%er hi!self advocates what he Grather"nfort"natelyH ter!s 3%iece!eal social en-ineerin-4 as the central !echanis! for social
%lannin- for in "tilisin- this !echanis! intentional actions are directed to the
achieve!ent of one s%ecific -oal at a ti!e) which !a5es it %ossible to !onitor thesit"ation to deter!ine whether adverse "nintended effects of intentional actions occ"r) in
order to correct and readj"st when this %roves necessary. This) of co"rse) %arallels
%recisely the critical testin- of theories in scientific investi-ation. This a%%roach to social%lannin- Gwhich is e%licitly based "%on the %re!ise that we do not) beca"se we cannot)
5now what the f"t"re will be li5eH enco"ra-es atte!%ts to %"t ri-ht what is %roble!atic in
society -enerally,ac5nowled-ed social ills rather than atte!%ts to i!%ose so!e
%reconceived idea of the 3-ood4 "%on society as a whole. For this reason) in a -en"inelyo%en society %iece!eal social en-ineerin- -oes hand,in,hand for Po%%er with ne!ative
"tilitarianis! Gthe atte!%t to !ini!ise the a!o"nt of !isery) rather than) as with %ositive
"tilitarianis!) the atte!%t to !ai!ise the a!o"nt of ha%%inessH. The state) he holds)sho"ld concern itself with the tas5 of %ro-ressively for!"latin- and i!%le!entin-
%olicies desi-ned to deal with the social %roble!s which act"ally confront it) with the
-oal of eli!inatin- h"!an !isery and s"fferin- to the hi-hest %ossible de-ree. The
%ositive tas5 of increasin- social and %ersonal ha%%iness) by contrast) can and sho"ld besho"ld be left to individ"al citiIens Gwho !ay) of co"rse) act collectively to this endH)
who) "nli5e the state) have at least a chance of achievin- this -oal) b"t who in a free
society are rarely in a %osition to syste!atically s"bvert the ri-hts of others in the %"rs"itof idealised objectives. Th"s in the final analysis for Po%%er the activity of %roble!,
solvin- is as definitive of o"r h"!anity at the level of social and %olitical or-anisation as
it is at the level of science) and it is this 5ey insi-ht which "nifies and inte-rates the broads%ectr"! of his tho"-ht.
KRet"rn to Section $eadin-sL
/. *ritical 0aluation
While it cannot be said that Po%%er was a !odest !an) he too5 criticis! of his theoriesvery serio"sly) and s%ent !"ch of his ti!e in his later years endeavo"rin- to show that
s"ch criticis!s were either based "%on !is"nderstandin-s) or that his theories co"ld)
witho"t loss of inte-rity) be !ade co!%atible with new and i!%ortant insi-hts. Thefollowin- is a s"!!ary of so!e of the !ain criticis!s which he has had to address. GFor
Po%%er's res%onses to critical co!!entary) see his 3Re%lies to y Critics4) in P.A.
http://plato.stanford.edu/entries/popper/#Links%23Linkshttp://plato.stanford.edu/entries/popper/#Links%23Links8/13/2019 POPER
19/25
Schil%% Ged.H) The .hilosoph" o( %arl .opper) ?ol"!e 1) and hisealis# and the -i# o(
)cience) edited by W.W. #artley .H
9. Po%%er %rofesses to be anti,conventionalist) and his co!!it!ent to the corres%ondencetheory of tr"th %laces hi! fir!ly within the realist's ca!%. /et) followin- 0ant) he
stron-ly re%"diates the %ositivistMe!%iricist view that basic state!ents Gi.e.) %resent,tenseobservation state!ents abo"t sense,dataH are infallible) and ar-"es convincin-ly that s"ch
basic state!ents are not !ere 3re%orts4 of %assively re-istered sensations. Rather they aredescri%tions of what is observed as inter%reted by the observer with reference to a
deter!inate theoretical fra!ewor5. This is why Po%%er re%eatedly e!%hasises that basic
state!ents are not infallible) and it indicates what he !eans when he says that they are3theory laden4 %erce%tion itself is an active %rocess) in which the !ind assi!ilates data
by reference to an ass"!ed theoretical bac5dro%. $e accordin-ly asserts that basic
state!ents the!selves are o%en,ended hy%otheses8 they have a certain ca"sal relationshi%with e%erience) b"t they are not deter#inedby e%erience) and they cannot be verified
or confir!ed by e%erience. $owever) this %oses a diffic"lty re-ardin- the consistency of
Po%%er's theory8 if a theoryis to be -en"inely testable Gand so scientificH it !"st be%ossible to deter!ine whether or not the basic %ro%ositions which wo"ld) if tr"e) falsify
it) are actuall"tr"e or false Gi.e.) whether its %otential falsifiers are act"al falsifiersH. #"t
how can this be 5nown) if s"ch basic state!ents cannot be verified by e%erience
Po%%er's answer is that 3basic state!ents are not j"stifiable by o"r i!!ediatee%eriences) b"t are acce%ted by an act) a free decision4. G&o!ic o( )cienti(ic
*iscover") 92DH. $owever) and notwithstandin- Po%%er's clai!s to the contrary) this itself
see!s to be a refined for! of conventionalis! it i!%lies that it is al!ost entirely anarbitrary !atter whether it is acce%ted that a %otential falsifier is an act"al one) and
conse7"ently that the falsification of a theory is itself the f"nction of a 3free4 and arbitrary
act. t also see!s very diffic"lt to reconcile this with Po%%er's view that science
%ro-ressively !oves closer to the tr"th) conceived of in ter!s of the corres%ondencetheory) for this 5ind of conventionalis! is ini!ical to this GclassicalH conce%tion of tr"th.
1. As (a5atos has %ointed o"t) Po%%er's theory of de!arcation hin-es 7"ite
f"nda!entally on the ass"!%tion that there are s"ch thin-s as critical tests) which eitherfalsify a theory) or -ive it a stron- !eas"re of corroboration. Po%%er hi!self is fond of
citin-) as an ea!%le of s"ch a critical test) the resol"tion) by Ada!s and (everrier) of the
%roble! which the ano!alo"s orbit of *ran"s %osed for nineteenth cent"ry astrono!ers.#oth !en inde%endently ca!e to the concl"sion that) ass"!in- Newtonian !echanics to
be %recisely correct) the observed diver-ence in the elli%tical orbit of *ran"s co"ld be
e%lained if the eistence of a seventh) as yet "nobserved o"ter %lanet was %osited.
F"rther) they were able) a-ain within the fra!ewor5 of Newtonian !echanics) tocalc"late the %recise %osition of the 3new4 %lanet. Th"s when s"bse7"ent research by
=alle at the #erlin observatory revealed that s"ch a %lanet GNe%t"neH did in fact eist) and
was sit"ated %recisely where Ada!s and (everrier had calc"lated) this was hailed as byall and s"ndry as a !a-nificent tri"!%h for Newtonian %hysics8 in Po%%erian ter!s)
Newton's theory had been s"bjected to a critical test) and had %assed with flyin- colo"rs.
Po%%er hi!self refers to this stron- corroboration of Newtonian %hysics as 3the !oststartlin- and convincin- s"ccess of any h"!an intellect"al achieve!ent4. /et (a5atos
8/13/2019 POPER
20/25
flatly denies that there are critical tests) in the Po%%erian sense) in science) and ar-"es the
%oint convincin-ly by t"rnin- the above ea!%le of an alle-ed critical test on its head.
What) he as5s) wo"ld have ha%%ened if =alle had notfo"nd the %lanet Ne%t"ne Wo"ldNewtonian %hysics have been abandoned) or wo"ld Newton's theory have been falsified
The answer is clearly not) for =alle's fail"re co"ld have been attrib"ted to any n"!ber of
ca"ses other than the falsity of Newtonian %hysics Ge.-.) the interference of the earth'sat!os%here with the telesco%e) the eistence of an asteroid belt which hides the new
%lanet fro! the earth) etcH. The %oint here is that the 3falsificationMcorroboration4
disj"nction offered by Po%%er is far too lo-ically neat8 non,corroboration is notnecessaril"falsification) and falsification of a hi-h,level scientific theory is never
bro"-ht abo"t by an isolated observation or set of observations. S"ch theories are) it is
now -enerally acce%ted) hi-hly resistant to falsification. They are falsified) if at all)
(a5atos ar-"es) not by Po%%erian critical tests) b"t rather within the elaborate contet ofthe research %ro-ra!!es associated with the! -rad"ally -rindin- to a halt) with the
res"lt that an ever,widenin- -a% o%ens "% between the facts to be e%lained) and the
research %ro-ra!!es the!selves. G(a5atos) . The Methodolo!" o( )cienti(ic esearch
.ro!ra##es) %assi!H. Po%%er's distinction between the lo-ic of falsifiability and itsa%%lied !ethodolo-y does not in the end do f"ll j"stice to the fact that all hi-h,level
theories -row and live des%ite the eistence of ano!alies Gi.e.) eventsM%heno!ena whichare inco!%atible with the theoriesH. The eistence of s"ch ano!alies is not "s"ally ta5en
by the wor5in- scientist as an indication that the theory in 7"estion is false6 on the
contrary) he will "s"ally) and necessarily) ass"!e that the a"iliary hy%otheses which areassociated with the theory can be !odified to incor%orate) and e%lain) eistin-
ano!alies.
:. Scientific laws are e%ressed by "niversal state!ents Gi.e.) they ta5e the lo-ical for!
3All-s are4) or so!e e7"ivalentH which are therefore concealed conditionals they
have to be "nderstood as hy%othetical state!ents assertin- what wo"ld be the case "ndercertain ideal conditions. n the!selves they are not eistentialin nat"re. Th"s 3All-s are
4 !eans 3f anythin- is an-) then it is4. Since scientific laws are non,eistential in
nat"re) they lo-ically cannot i!%ly any basic state!ents) since the latter are e%licitlyeistential. The 7"estion arises) then) as to how any basic state!ent can falsify a
scientific law) -iven that basic state!ents are not ded"cible fro! scientific laws in
the!selves Po%%er answers that scientific laws are always ta5en in conunction 5ithstate!ents o"tlinin- the 3initial conditions4 of the syste! "nder investi-ation6 these latter)
which are sin-"lar eistential state!ents) do) when co!bined with the scientific law)
yield hard and fast i!%lications. Th"s) the law 3All-s are4) to-ether with the initialcondition state!ent 3There is an-at 44) yields the i!%lication 3The-at 4is4) which) if
false) falsifies the ori-inal law.
This re%ly is ade7"ate only if it is tr"e) as Po%%er ass"!es) thatsin!ulareistential
state!ents will always do the wor5 of brid-in- the -a% between a "niversal theory and a%rediction. $ilary P"tna! in %artic"lar has ar-"ed that this ass"!%tion is false) in that in
so!e cases at least the state!ents re7"ired to brid-e this -a% Gwhich he calls 3a"iliary
hy%otheses4H are -eneral rather than %artic"lar) and conse7"ently that when the %redictiont"rns o"t to be false we have no way of 5nowin- whether this is d"e to the falsity of the
8/13/2019 POPER
21/25
scientific law orthe falsity of the a"iliary hy%otheses. The wor5in- scientist) P"tna!
ar-"es) always initially ass"!es that it is the latter) which shows not only that scientific
laws are) contraPo%%er) hi-hly resistant to falsification) b"t also why they are so hi-hlyresistant to falsification.
Po%%er's final %osition is that he ac5nowled-es that it is i!%ossible to discri!inatescience fro! non,science on the basis of the falsifiability of the scientific state!ents
alone6 he reco-niIes that scientific theories are %redictive) and conse7"ently %rohibitive)onl"when ta5en in conj"nction with a"iliary hy%otheses) and he also reco-niIes that
readj"st!ent or !odification of the latter is an inte-ral %art of scientific %ractice. $ence
his final concern is to o"tline conditions which indicate when s"ch !odification is-en"inely scientific) and when it is !erely ad hoc. This is itself clearly a !ajor alteration
in his %osition) and ar-"ably re%resents a s"bstantial retraction on his %art8 aris! can
no lon-er be dis!issed as 3"nscientific4 si!%ly beca"se its advocates %reserved thetheory fro! falsification by !odifyin- it Gfor in -eneral ter!s) s"ch a %roced"re) it now
trans%ires) is %erfectly res%ectable scientific %racticeH. t is now conde!ned as
"nscientific by Po%%er beca"se the only rationalefor the !odifications which were !adeto the ori-inal theory was to ens"re that it evaded falsification) and so s"ch !odifications
were ad hoc) rather than scientific. This contention tho"-h not at all i!%la"sible
has) to hostile eyes) a so!ewhat contrived air abo"t it) and is "nli5ely to worry the
convinced arist. +n the other hand) the shift in Po%%er's own basic %osition is ta5en byso!e critics as an indicator that falsificationis!) for all its a%%arent !erits) fares no better
in the final analysis than verificationis!.
KRet"rn to Section $eadin-sL
Bibliograph$
orks B$ Popper
&o!i' der Forschun!. "li"s S%rin-er ?erla-) ?ienna) 9D:>.
The Open )ociet" and ts ne#ies. G1 ?olsH. Ro"tled-e) (ondon) 9D.
The &o!ic o( )cienti(ic *iscover". Gtranslation of&o!i' der Forschun!H.
$"tchinson) (ondon) 9D>D.
onectures and e(utations: The ro5th o( )cienti(ic %no5led!e. Ro"tled-e)
(ondon) 9D@:.
The .overt" o( /istoricis#G1nd. edH. Ro"tled-e) (ondon) 9D@9.
Obective %no5led!e: -n volutionar" -pproach. Clarendon Press) +ford)
9DB1.
+nended uest; -n ntellectual -utobio!raph". Fontana) (ondon) 9DB@. 3A Note on ?erisi!ilit"de4) The
8/13/2019 POPER
22/25
ealis# and the -i# o( )cience. Ged. W.W. #artley H. (ondon) $"tchinson)
9D:.
The M"th o( the Fra#e5or': n *e(ence o( )cience and ationalit" . Ro"tled-e)
(ondon) 9DD::,>ac!illan) (ondon) 9D@1.
Cornforth) . The Open .hilosoph" and the Open )ociet": - epl" to *r?
.opper@s e(utations o( Maris#. (awrence V Wishart) (ondon) 9D@. Corvi) R.-n ntroduction to the Thou!ht o( %arl .opper. Gtrans. P. Ca!illerH.
Ro"tled-e) (ondon V New /or5) 9DDB.
C"rrie) =. V "s-rave) A. GedsH..opper and the /u#an )ciences. Nijhoff)
;ordrecht) 9D>.
Ed!onds) ;. and Eidinow) . Witt!enstein@s .o'er: The )tor" o( a TenMinute
-r!u#ent .
=rnba"!) A. 3s the ethod of #old Conject"res and Atte!%ted Ref"tations
"stifiably the ethod of Science4)
8/13/2019 POPER
23/25
ohannson) .- ritiAue o( %arl .opper@s Methodolo!". Scandinavian *niversity
#oo5s) Stoc5hol!) 9DB>.
0e5es) . 3Po%%er in Pers%ective4)Metaphilosoph" G9DBBH) %%. :@,@9.
0e"th) $. 3?erisi!ilit"de or the A%%roach to the Whole Tr"th4).hilosoph" o(
)cience9DB@) :99,::@.
0"i%ers) T. A. F. 3A%%roachin- ;escri%tive and Theoretical Tr"th4)r'enntnis1-) 9D1) :
8/13/2019 POPER
24/25
Shear!"r) ..olitical Thou!ht o( %arl .opper. (ondon V New /or58 Ro"tled-e)
9DD@.
Si!5in) C..opper@s 8ie5s on Natural and )ocial )cience. #rill) (eiden) 9DD:.
Sto5es) =..opper: .hilosoph", .olitics and )cienti(ic Method. Polity Press) 9DD.
Stove) ;..opper and -(ter: Four Modern rrationalists. Per-a!on Press)
+ford) 9D1. Schil%%) P.A. GedH The .hilosoph" o( %arl .opper. G1 ?olsH. +%en Co"rt Press) (a
Salle) 9DB
8/13/2019 POPER
25/25