,

Draft: No'l; for Quotation

THE STRUCTURE OF SCIENTIFIC REVOWl'IONS

Thomas S.

Kuhn

Contents
I.

Introduction.

... . . . . . . . ..

II. The Route to Normal Science

II

III.

The

Nature of Normal Science.

IV. Normal Science as Rule-Determined

..

..

23

36

V. Anomaly and the Emergence of Scientific Discoveries

45

Crisis as the Prelude to Scientific Theories.

61

VII. The Response to Crisis.

73

88

VI.

VIII.

The

Nature and Necessity of Scientific Revolutions.

IX. Revolutions as Changes of World View.

X. The Invisibility of Revolutions

XI. The Resolution of Revolutions

XII.

Progress through Revolutions.

llO

138

149

166

I.

Introduction

The study of h1storYll.as not been's. u:iw source for the West's can-

ceptionof

Viewed&s a reposi-

tory for more. than anecdote orchronoloe;y. history cOUld proiluce a decisive
transformation in the image of science by wbich ve are
image has previously been

possessed.

That

I'lrawn. even by scientists themselves, mainly :?rom

the study of finished scientific achievements as these ere recOl'ded in the

claee1es and. more recently. in the textbooks from which each new scieu. "Inevitsbly.....

. .

a1m ot such books is perS1.ll1sive and pedagogic; a conception

than an image of a national. culture

guage teJct;.

"

them and in

,,"

""

ways;

,a.tourist brochure or a lan-

This monograph attempt a to Bhow

conception of eciencethat
searcll

..... the

'

""

ha.vebeen micled by

Its airuis.s eketch of the quite different

',"

can emerge .. :?r0l)l thl!

llistoriClll record of the re-

. ,'.

Even trOll! history> hOlfever, tbatnew C()!lCePt:l.on

will not

be

ing if' h1storicalda.ta continue to bescught nnd scrutinized mainly "to

t'.!l-

.lIver questionn pOGed by the .unhiatoriceJ. .Iltcrectype <b.'awn from flcience

teAts.

Tholle texta. lw.ve . for .

. often seemed to imply that, the con-

t(lll-(; of Gcience ie uniquely exemplified by the observatioIla, la-..

/

J.

au:'!.

together with the logical opers:tiolle cll!ploJ'-eU "hen rolE.tine; those data to
the textbook's theoretical generalizations.
tion of science vith profound developmental

The result Me

e. con;:ep-

If science ie the constellation of facte, theories, and methods

lected in current

then scientists are the men uuo, successfully or

not, have striven to

constellation.

one or another element to that particular

Scientific development becomes the piecemeal process by

which these items have been aclded, sin,gJ.y and in combina:l;ion,

'';0

t.he ever-

growing stockpile that constitutes scientific teclm:t.que o.:nd lmwlee.ge.

,W.a

history of science becomes the discipline uhich chronicles both these succees:l.ve increments end the obstacles that have inhibited their acc1J.lliUlation.
Concerned. vith scientific devel.opr.lent, the historian then appears to have
tl1C

main taslts.

On the one hs:ad, he muct de1;e..'"l/l1!1e bY whet man and at tlhat

point in time each contemporary ecientii'ic fact, law. and theory

covered or invented.

dis-

liaS

On the other, he must describe and explain the ccn-

gel'ies of error, myth, and superstition which !leve 1.uh1bited the more rapid
of the constitucnt of the modern science

Much recearch

Im!l been directed to these ends, and 1:J0lt-e still ill.

In recent yss.:t:a, h01r<lVer.

II

:few

of science h",ve been

lug it more a.,d more difficult to fulfill the funetions

scientific develepment-by-accunmlatiun essigns to them.
au incremental process they

COl1.t::eived

the concept of
lw cUronicl(;T:' oJ.'

that additionnl research makes it

harecsr, net easier, 'co sumTer queatiOnfl like:

truo

When .>as O::ygen diecovsiO'ed?

of emrgy ccnser'li"ation7 . Incr;;aaingly> n few of them

cultiesin diatil)gUishill5.the "scientific". comr.onent ofpaet

end belief i'rom 'tihll.ttheir predecessors had readily labelled Herror" and.

that those once current v1awse!

as a whole, neither leao

scient1:t'ic nor more the product of human idiosyncrasy than ou:r.O\I!l.

llone

of these notf disce..1:'dedbodies of theory &ldobservation ce-.!l,even in retl:Ospect,.bed:l.v.l.de.d 1ntotwoparts,.Qne.belong1ngto 't.!le. scionti:f'ic stocl;pile

because it. is preserved .in eurrent.texts,.theother to mytholoBY because

torian had thought he.stu.d1cd .must include bodiel3of belief' incoIil.PGL'i;ibJ.e

lIith those current today..or.1tmllst exclude views. of. nature v.!.thout 'Il'hich
olJr '(mn ll'ou1.d not have come

'being.

The llame hiztor1cal research th&t

displays the difi'icult:l.esin:!.eoJ.e.ting indiv:l.dmt.l.

E1.tlci. diSCOVel'-

iee gives g:r.ctind forprof'oUllddou'bta about . the cumulative .procees thr0\1Olh

wien these

eont:r:!.but1olUl to science 'Were. thoue,ht

'1;0

he,Ye been

revolution int.'ie study of eciEmce; th01.1gb one thetis still in its c,"-,.'J.y

they insist upon studying the OD:lniolls of t1;&t g;:ot-'l? fJ,i:W. ethel:' ailll:Uc.r

ones from the vie-dpoint--ucuslly very

scienc:e--which g;ives

froul thst of

opiniolls the mm:im"um :lntert'.!ll cohe::euce /too.

the closest possible fit to nature.

through the

'Ii'Ol'i!:C tllat l'OSUU,

varks perhaps bellt exell1J?lii'ied in the llritings of Alm:andre Koyrii, science

does not seem altogether the sume

as the one discussed by writ

ers :1.n the older historiographic tradition.

By implication. ct leest,

these historical studies suggest the possibility of a new

This

of science.

aims to delineate that imase and to make it flLl1y explicit.

What aspects of science

emerge to prcm!nence from the

toward science here to be eleborated and

Fixst. at least iu.Ol'

der t,f' prellentat1on. is the in(lU:fi'iciencyof methodological

by

themeelves. to dictate

D.

eciellt1i'ic queot1ona.

Inst..-ucted to examine electrical or cilemical phenom-

ena, the

1&'.0

,ho ift

scientific lEay
clusions.

unique eubstallt:l.\re conclusion to

of these tields out '\rho

l!lI?.!ly

lma.Hl l!ht

reach any one 01: a =ber 01:

sorts of

it ie to be

con-

Jlmong those leGitill!<,te po03s1011:1.t1es, the :;:.articul2.r ccnclEsions

he docs arrive at are lU'obnbly determined by his prio:;;, e::q:orieo.ce in other

fields. by tbo accidents of his investigation,

l!!8.li:ol,lJl.

Hb2.t beliei'D a1Jout.

chemistry or

Ga:j.

by his O"Jil iniiviQtcal

the starn does be brillg to the nttte,y

"nieh of the many conceive.ble

vu.nt to ""he r,:;,'.f fiola. dees he elect to pel'form f:l.rat?

Ani

relee.ll'pects of

l:l.ltetheoeare oft.,nc'sriditial

.
terminants of scientificde'Velo71ment ...
tion II tbll.t the

..

note .

ofmoet aciences have been characterized by

cont1nua.1. competitionbetlleen a'number"of'd.1Btlnct-'V1:evs .of- nature, each

partially derivedi'rom.. sndaU roughly, compatible,v1th,the,.d:l.ctll.tes of

scient1fic observation and method.

8choo16

wal!!

not

one

What differentiated 'I;hese variO'..tS

all "scien-

ti:l'icH--but whatve shs.Ucome to CIlJ.l the:l.r1ncamnenaurablev.ays of seeinzthe w.rldand Of doing science in it.

Observation ana' eXJ,>llxience can

drnatically restrict the range Of .admissible SCientific belief. else there

. ll\1ttl:leY.C:SJlllot .

..

...bqdy of

suchbel:l.et.An apparently arbitraryelement, compounded of'persontll'an:l

is a.lWllyt3&, tormative ingredient of the beliefs es

:poused by a giVen scient1f1ccOmmlmity at a given time.

That element otarb1tr&'1neSB does

scientit'.c group could in'acticelts trooe at

ceived beliefs.

that any

an

witliout SOllie set of ra-

Nor does 1t mekelees consequential the jport:l.c1.uar con-

stellation towl:lch the group, at agireu time, is in fact cOl:r.ll1tt.od.

Ef-

fecti ve reeearch ecarcely beg1ne-bd'ore a ec1entif1c c = 1t.y thinke :i. t

has acquired :firm answcra toqv.est:l.one like:

W'aat arc the t'lmClament;cl en-

tit.ies en: "hieh the uniJerae ie cO!llposed? nOlI do '.;heee in'tercGt with Cll.;:h

other end with the senses? What questions may lezitimately bo csked about
such entities snd "What techniques employed in seeking Elolutiono'l At. leust
:tn the tr.c.ture GCienCCB, anr;wcrll (or full substitutes for
t1,('lns like 'these ('-ol:"e firr::ly'

in "the

to que:;,

:tr-.;it1.ation that

eo.uce.ticn 13 booth rigO!'ous end rigid,

hold cn the scientific n1ind..

aUfil:re:,."s c:omG' to

u.eel.:)

That they can do so dool} It''olch to t;cccun:i:,

both tor the peculiar efficiency of "l;!-..2 no!1'".At ree.cc,zoch ac'.;:\. \';ct.y ar.d.

tohe direction in wich it p.-oceeda at any given t1.me.

n = l science in Section III,

search e.s a atrenuoU!3 ana.

"''hen e;{e!"..:.tu:!.ng

fiho.ll want finally to t',eaeribe the;t re-

lie

attelllpt to force nature into

tual. boxes svp}?lied by professional education.

concep-

S1.!i1UltrulCO'-'OJ.y; li'O oll",ll

wonder whether research could proceed without such boxes,

e.1elll0nt of arbitrariness in their
their eubsequent

fCle'

the

end.. occaz;:l.o:ll1.l1y, in

Yet that element of arb!trarinc!ls ie prel'lent, i!.I'.!l it too l:ias

1m

im-

portant ef:fect on z;cieutitic dcvelop!!len'G. one '1M.ell ..'ill be e:mrnined in

detail in Sact5.ons V and VI.

r,o=l

scientists inevltebly epend aJr.oet

the a.ctiv:!.ty in uhieb.

t.heix time, is predicated on the

assumption that the llc5.entUic collliIl1.lIdty km:n;ra l',hat the w,')l'ld 18 15.1;:.",.
Much of the en'l;erp-.t'"ioe 9 s success

to defend thet.

tt!HI'llll!;ptl.on,

fl'"O!J.

"t.t111il'igl:lens

it neceseary at consideruble corrh.

aCillC!, i'c:,r ey,.s.urple, au:ppreoscs

nevelt.ies becaur.re

nCC:'DE'-ar5.1y cubvercive of its beeic co-.auitmcnts.

aG those ccrotri

the

No=l

flO

l-:'):-:.g

:retcin an clem\?ut of the m"bi trn.."7, the v'ery no:t.UX'0

of nOrlnnl re8earcb. encU!:>"',;c tha;'c novelty shall not .be E"r.1ppress&:l for.

So.nretltnos e norlml p::'o'!Jlem; one ';;'ha't cugh:'\i to be solvable by

ruleg and

resistu the reiter:z.t8a onslnv,ght of thc cblGl;rfi mern-

7
'Which

e:l):ec-

tatioo..

!D. , these. and otller ;roys

astray.

And when it dooll--:vhen, tha'G is, the pro:fcs,,1cn can )10 101lscr

doe" a'1d. Wlct go

eva.de811omal1esthats'.1bvertthe e..':istiIl$tre.ditiOll- of (:cient:Lf'ic prs:::tice-then" 'begin "th!l, .

to

I!I.

new set c cOIllrilitments and

sio!!.?,tIMt

..;;;yo! l'I't1.Ct:tci!'.g science.

T'ill1 extl'u-

ordinary episodes in vhich that shift ofprofess:ton&icommitments occurs

the ones

in this monograph as acientific revolutions.

cOlllplements to

T'ney are the

a.ctivity of normal

science.
The llIostobVious

of scien'.ific revolutions exe those famoua

episodes in Bcient1f1cdevelopment tbb.t, have

'been labeled revolutions

before. There:f'ore.:tn Sections VII an.clVIII, where the ne.ture 01' scientific
revolutions i6 fir!!t directly scrutinized,

'Ire

shall deal repe"tedly 1d,th the

episodes in the history 01' at lellat the physicnl science these display lrnat
all

revolutions areaboui; ... Each of

theIll

necessi'uateu. the

nity's rejection of ella time-honored lilcientifictheoryin favor or

illcompatible 'With. it.

Each produced.

1:.

C0'U'':::U''

COllGGguent shift in the go.'l.la of

in thelll'oble."Js avo.llnble for scientific ,scrutinyJ and. :tn the

ste.'I'1dards by 1>-m.ch tr.e pro1'es::;:!.on 6.eternined what shol.',ld count o.S e.ll o.amis-

;mys tM;t

1f!'!

shell uJ.t1ma.tely need to describe as a tranefol"'lll.'li;ion of the

acteriet.:i.cs of cc:i.eutii'ic

8 .
These characteristics emerge with particular
say, the Newtonisn or the

frau a stu-iy of,

It is, h01lever,

Revolution.

lJ.

tal thesis of this monogra.ph that they can aleo be retrieved from the study
of meny other episodes which were not so obviously revolutiontlry.
far smaller professional group effected by them

For the

s equations were as

revolutionary as Einstein 's, and they Wl';'re resisted accordingly.

The in.

vention of other new theories regularly, and appropriately, evokes the same
response from some 01' the specialists on
they impinge.

are of special compatence

For these men the ne-:-l' theory implies a chsIJge in the rules

governing the prior practice of normal science.

therefore, it

reflects upon much scientific work they ha7e already

That is why a new theory, however special its range of application, is seldom or never just an increment to what is already k!lmm.

I'GS

sOBiml1at:!.on

requires the reconstruction of prior theorY and the reevaluation ot prior,

fact, an intrinsically revolutionary precess that is completed seldom by a
single man and naver overnight.

No wonder historians have had difficulty

in precisely dating this extended

that their vocabulary impels them

to view as an isole.tsd event.

Nor are nov inventions of theory the only scientific events which have

impact

the

in

they occur.

The

cOl!l1!litments 'Which govern ncl'1llc.l science speci;('y net only what sort6 (Yi: entities the univorse does cOllt5.in, but alBO, by implication, those thnt it
does not.

that n

It follows, tbOU8h the point will

ext.ended diccu5s;l.on,

thst of, say: 07.ygen or X-Reys docs not

or:.e more it-2m to tbe PCIfJlat:i.on of the scientist; S ,(-Tcrle..

ndG

tJl"'o-tJj'ctely it

newark at theory thrOUGh lIh1ch it deala 'ldth the world.

and theory are not

Sc:!.ent:t:!'ic :f'-sct

perhaps''lflthln a sillgle

<

tradition atnormaJ.,..scient1fic. practice;

The.t.is. lilly

<iis'

covery is not simply factual in ita import and why the scientist's world
is qualitatively tranaformed as well asqwmtitativel.y enriched. by :f\mdamentalnovelt1es 01' e1ther fact or theory.

This extended conception at the nature at scient;1flcrevolutiollS is

the one to be delineated in the pages \Thich tollow.';.dmittedly the extension stra.tnscu8tomary uea,se,];lerhapsbeyond the l1cenacallowed even to

metaphor.

Newrtheless. I shllll. contiIl1!e to speak o;t'<e'l."en diecoveries.e.a

revoidtionary, because it is .1usttlieposslbility of relatii'l!!: their_Btruc

" ' , __ M""_"'_''''

""

,_''', , _ ' ' ' ' , , ' __ ' ' ' ' __

"'"

'''_ " " '"

_"

'"'''''' ''''''

tu..-e to tba:t &f . say. the Ca,pernican Revolution that makes the extended

conception seem to me 60 important.

Theprecedlng dicc1.!.ssioniniicates... how

the com,plementarY not.:!.onsaf normal ecienceWll.tofscientiticrevolut:to!1!l

will be developed in the eight eectiOllS immedia.tely t.o follm: The oncs

that :1'01l0-W" them attempt to dispose of tbrea lema.1ning

quoetion:::.

Section X, by discussing the textbook tradition. coneidcr6'fby 6c:1.l'utific

revolutionahaveprevlcusly been so difficult to see .. Section XI o.escribea

the revolutionary cO"'.P<'ltit:i_on bet...-een the proponents of the 01(, nCrJ"..al5cienMflc tradition and the adherents of the new one.

It thus conoidal'S

the process l!ilich lllUllt eomehC".;' replace- the coni'1rmat$.on or falsification

procedm'ea made femililll' 41' cur uaual in:3ge of acienco.

COlIrpGtit:i_on b$tween

10

progress.

For thnt question, hO'.-ever, thitl mouosrapll 'Will p:ro'l5.,1e no mOTe

than the main li-nes 01' answer, cne which depends 1-!pon

of

the scientific c0lil!llUll1ty that require llI'J.ch arld1tior..al o:qJlo;:ation em!;.

study.
IT"ne f1naJ. vereicn Yill x'equire one or
this point.

They ,.111

1:,,10

additional ps.rag;rapha at

footnote and bibliography polic;",

the relation of this :fon! ot the flCnogrli.ph to its i'ull<:!l:' v!)l'siou, JuotH'y

t181 acknowledgmente.]

or more past ec:l.entif'ic achievements tha.t have CO'.:le to be accepted. as established

Paraa.:@i1sbY the

tif':!.c qommunity.

ov'erimelfuingmaj6r1ty

of

thepr.ofessionalscien-

Todaytheee parsdigmaareto be

elementary and. advanced.

SUch texts Bll.-Pound the body of accepted theorYt

illustrate many or ell of its lluccessfulapp1:tca.tione.andcompare these 6.pplications wIth eltempllll'Y obsen'atians and el>.-periments.

Betoreoooks111re

thelle became popuJ.arearly in the nineteenth century (a."!d. until even. mere
recently 1nthe more recently matured sciences),meny6f the
of ecienCBserved.a similar !Unction.

classics

Aristotle's !hysics.,Ptolemys

llewton's Princ1l?ia and 9gt1c!tl!. Franklin's Electricity, Isvoisier's

Chemill:tr;y......and ..

time

and m::,uy other ;rorks served tor a

implic1tly to define the legitimate problema end methods afa research field

tor succeeding generations ot

They

able to do so be-

cause they shared two essential cheracteristica. Theil' achievement

ficiently unprecedented to attract an enduring group of adherents

open-ended to leave all sorts of problems for the

suf

practitioners to

from

These character2stics of pnradisws reguire!urtller elaboration a.id

shortly recei v,," it. 1'1.-1; fir st the I'ole of paradigms must be cle.rified 'DY
noting a sense in \1hich i;hc.:e can be such a i:htn$ as scientific practice

11.

12
physical optics.

of \laves and.

Tcday's physicc ta:ttboo1::s tell

of particles.

Btv.zlent th,at liid;. 5.s

prccee(ls accoz'dillgly,

01'

according to the rom-e elabors;te end. Lw,themstic.,l cnaracteri"s.tion from "ilieh

this

verbalization is derived.

ever. scarcely half a century old.

That chr.racteri:!;a;l;:!.on of light is, h<:;!,Before it

'I.'aS

dc-"eloped by P1e.nck, E:l.n-

stein, and. others early in this century, 11hysice texts tm:ght that light

'il8S

t.e.nsveree wave motion, a paradigm derived uJ.tiDately frCl'!l! the opticC'.l writiDgB of YOtlllg and. Fresllel in the early nineteenth century clld. cmbG.o.ied 111

most physics teaching frClll about 1830 cu.

Nor .'as ths l1'<lve theo!'Y the first

to be embraced by almost all pre.ct1.t:l.ollerz of optical adell.ce.

DV.Z'ing the

eighteen'vn centu.."'Y the psradigm for this field '!>."as provided by

Jl.t that time pbya-

which taught that light 'Ima lll9.terieJ. cOl"puaclcs.

ic1.ats Bought, as the early '!lave thsol'j.ots did not, for ev:!.e.ence of the pressure exerted. by light particle a impingiDE on solid bodies.

rmrolutione, and the success:!. ve trl1.!lcit5_OD from one

be exrunining :I.t

Il,S

such in the next three sectiO'.ls.

peTli.d.i(!%

But

to Imothsr via

cC!r:.cerns uo

hel."e is only the contrast bet'htl9n that pattc..-m .C'wd the one chnrac-terizing

'.

13

ss

..:i;l8rilcUJ;ru.- clustercf OXrG

"---------,,---.-.,,

which its own theory could domoet to ",::;plain.

dealt with by

!l'! !.!elaborations,

or theyrellainc.(t as

problGfas

for further research.

At varioust'JUQ.s. e.ll toone schools

significant contributions to

phan=1ll, WJd. techlligueo tra.n. 'lrilich liewton <Ire;, the

--'-______

first nearly uniformly accepted paradigm for physical optics.

Any defini-

tion of the scientist that excludes.st least. the more creative members of
these various schools i;illexclu.de their mod.e:rn IltlCCesaore as well.
men were scientists.

Yetanycne

Those

a survey of physical optics be

fore Newton may well conclude that, though the field's practitioners were
scientists, the mit .resuJ,t

or

their activ.ltY'llllssomething leaothan ec1ence.

Baing able to take no cC!l!lllon body of belief tor grsnted,each -writer en phys-

ical optiCS Zelt forced to bUild his field a!lew from ita foundations.
doing 80 his choice of l!UPPorting

,."

"',,

'

In

and e;-q>erimcnt ;'7ltl relatively

free, for there .Wlsno standard llet.oi'methodeOl' of phenomena that every

ax>tical writer

forced to employ and e:r.p1.s.in.

d.rcumstances, thed:!.eJ.ogv.e o:f'thereaultingbooks

Inevitably, Ullder these

often directed as

ll"o'.,ch to the members of other flcho()ls,as it.Ull.s to.. nature.

That pattern is

Lot unfamiliaz' :i.n a number of creative fields today,nor is it

,rith sign:l.fictmt discovel'Y and in'.rent:!.on.

It 16 not, ho-ever. the l?at.tern

of develC1'ment toot phyeicsJ. optics acquired after Newton o.nd that other
natural sciences lila-!re f'smilim: today.
The histCl'""l.,! of elec-crice.l :cescnrcn in the firr!# ha 1f 0'1 the eighteentb

All their numeroue concepts of electricity had sO:llcthlng

c=oll--they

were partially de....ived. fran one or another version of the mechanico-corpl.lSculnr philosophy tlw.t guided a.1.1 scientific reseru'ch of the day.

In addi-

t1oo, all were components of real scientific theories, 0-1 theories, that
is, which had been draw 1n part :'rem experiment and observation and tthich

partially deter'Jl'n"a. their choice and in"OOl'llretat10n of

undertaken in research.
though most of the

no lllore then

Il.

Yet thov.gi:l all the experilll'ants

problems
electrical end

l"ee.rl each other's works, their theories had.

family resemblance.

One early grcnip of theories

se'li'Emteenth-century pre.ct:!.ce, re

gar.ding attraction and frict:i.onaJ. generation as the tunda!!:0!:tll.l electrical

phencmena.

This group tendec'!. to traat repulsion ao a fleccndm'y ei"l'ect; due

to scme sort of mec!J.s.::licaJ. rebounding v.no. also to pOfJtpalle fer 8S long e.s
pOl'lllible bo'ch discussion e.n.d syatemat:l.c :research or!

electrical conduction.

took attraction and

(Jl;her

to be equelly

s ne'lrly discovered

tt

the

t01Ul

io their crl11l,

of

electricity end modified their theor:f.e!: end x'osearch accordingly.

this group is remarkably

fa::- the lr.utual rCIlvJ.zioll of

(Actw..lly,

Franl:1in's 'cheory never gutte ecaounted

n<:gs.t:tvely C;l."'Xged bcdies.)

But they h,;(l as

15

initwtu.."U,

______ had diff'iculty:.'econ.::1ling its tb.eca'Y .."ith a number ot' {;"tt;:'act:tvc w::d rc-

pulaive efi"ecta.

Only through li'rantlin's wca'kd:l.datheo:ry srise which

couldaccov.nt with aomethitlg.like equal facUityforverynearly all theee

co\lldanddi4proviAeamws?;quent generation
of "electricians" withe. CCl!D!ilon :pa:t'adigm for ita. reaeexch.

Excluding thosefields,likemathelll8.t1cll ar,d. astronomy, "'hose t'irst

firm.paradigms date from.prehistca'y and: also those. like

which arosc by division .and. reccmb:l.zJation of apecia.J.tics. ab'ev.d.y matured.

the situations outline.d abovo are historically t:l'p1cah
my

Though itinvclveo

to em,ploy the tmfortunates1!llplif1cation 'Ghat tags an. extendeCl

hiator1ca1 episode with a single and somewhat.a:ib1tlll.:t'iJ.y chosell r.ame (Ne'b-ton

or Franklin, above), I

t!ll:'G

awnSf'.

tundamentaldiaagracments chsr-

acterize.d, far example, the study 01' motion before Aristotle and. of statics
before .llrch1medes,the .study. .

of

and.Boerhaave, and of' historical geology before Hutton.

the study of heredity, for example, the first

bai'ara Boyle
Inl>ar.te of b:l.ol().z;f;
received ps:radigms

6."e still more recent, and it relll!1.ins an open question 'Whutpsrts of EC-:::!.aJ.
science M'ile ;yet 8cqu1reil. I!Uch paradi@llllatall.

dic.atc for parE'..digm ell of the. facts

Hi3tory. Busgellts thQ.'t; the

cc.Qu possibly PCl"t!:'.in to the de-

1.6

the:t lie ready-to-hand. The resultill3 pool of f!lc'ca cont..

t,hoae ac.:!s:s-

sible to casua.l obaez"Ifl1:tion a.,d e.."qw.riMnt 'l;ozethe:r ,lith !lome ot the Illore

Illaltillg,

and metel.J.urg;}'".

It is psrtly because the craf'ts are t;he oue :::-000-

ily accessible source of facta wIu.ch

not have ooan cflsuulJ.y discovered.

that technology has so often ple.:>""d a v:I.tB1 :role in the emergence of new
!lciencos.

But though this BOllt of fact-collec'c:i.ng has been essential. to the

origin of many significant SCiences, anyone who

say. Pliny's ency-

clop.-.ed:l.c writings or the :flaccID.an nn:tural histories of the seventeenth

,,'ill diseovGr that it :pro:;.ucea a morass.

CI1.l1 the 1iterature that zel'mJ.tl:l eeie'ltific.

One sO!llehmr hesii;ates to

The Baconiml "histories" of

U<Slllt, color, wind., mining, and co on are filled "tI1th intOl:'lllaUCol:l, SOOle of

it reccmiUte.

But they .1uxtepo::e facts (so.y, heating by m2Xtm:e; ths:t "\Jill

17

statics, dynamics, ll.!ld

Grlfcics,do:l:e.cts collected'l>r:tthsoUttle

be done in no other 'its:y.

This is thes1tua.tion1ibic:'b. creE,tea the schoolscharacte;:>i:;:tic of the
eerly stages of a science's de-.;elO.',}ment.
Ilreted in the absence 01: at lee.at 80me

UonaturaJ. h:!.s'(;oI"",f can. be inter-

t body of intertv!ned thco-

;retica.l andmethodolog1cal belief that perm:l.ts selection, evru.uat:LCll, and

critii::iam.

If,thntbQi;y'ofbeUefiEl notalrcl.l.dy implicit in the collection

of :f'ac:tB--in 'Which CCl.semore than

facta" 'are

hlmd--it must be ex-

bta.current metaphYsic, by another scienc:e, or

by personal and !iistoricsJ. accident.

st6,ges

No lronder, tlwn, that in the early

of any sciellce different men. ccnfi-ont:Lng the seme

:;.leo un:l.que :i-a its degree to the fields we c!lll sCience, is 'Ghat eu,,!;. .....1_

tid o.ivergenceilcrh(l"U:W. ever lareely diSF.J?.it'S!Il'.

Fox they do d.isa.ppeer to e. very considerable e;{tent ana th:m. :;.p:;;e.rcn"c,J.y

tr:tl;u:rgh of one of the pre . . :;r..lXadigm schools. 't1hicn, because ot" its Qt.11 C!1C:::-SC-

18

fruit 01' their ci'torts ;ras

wytl.;)n j!'.l',

been discovered by a man el:ploring

ck"rice

carrua.l.ly

llE.tlU'e

lras in :fa.ct it:uiepandently dev(':lClled by

O!.'

never have

at rsndom but which

least t\;O ill.'.reati5Gi;ora in t.hc

Ii;!;

particularly concerned. to e:;,.-plain that strange

ticularly revealing piece of

though one which

'iias f)'cill

electrical repulsion.

tmeble

'&0

!llld.,

in the event, p..'U'-

Eis success in doing so

provided the most effective of the

ter than ita

,inion

Almost i'rom the lltm7t at: his electrical ree",a::'ches Fl"a:.1Uill

early 1740' s.
'liaS

Go

that ma.de hie theory a paredigm,

a.cccunt for quite all t.ile !mmm. case a of

To be accep'Geo. a.s a perai!.igm a theory'

l1IUOt

seem bet-

but it need not, and in tact neve:;:, does, explain

all the facts that ita proponeate feel it should.

What the fluid theory of electricity

rected to secondary er to overly call1l1ex

li'reed from the concern

for the

which held

of eJ.ectl."icity,

1..oul5.

and all el:ctr1ce.l phenomena, -the UJ"11ted g;t'ouy

19

"'!'rUth emerge a more readily from eloror tllanf'rCll'i collflauion."

based. resea:rchin . the ne:..":I; .. l:Iec'cion.. Oll1;fllUSi.; fil.'si.; note briefly ho;; the emergenee of .a}laradigm a.:ffects . the structure of the profession.

Wilen, in the

development ofa natural science "anindj.vic1ua.lOl'" groupctiratproiuces

Il.

synthesis able toattrsctl!lost of; the ne::tgenerat:!.on' s praci;itioncro, the

by the:l.l' members' convei'aion to the new paradigm.

me.n

B1.ltthel:'e are all1a.;y-s soma

cling to one Ol'811other of the older views, alldthey e.l:'e. simply read

out Of the profession. vhj;chthereattel' :l.gnO::GS the:l.l' work;

implies a new Slld more r1giddefinition of the field.
unable to accOIll!!Iodate t!l.e:l.l'

new :pm-a<li&,'Ill

Those um::l.ll:l.ng or

.itmus'prcceed
in isolation or attach
,

themaelves to some other group . Historically, they hIl."eoften simply stayed

in the depar'-uments Of philosophy by "hich Bomany of" the spec:!.f'-lsciences
have been

This h:l.nte--wat could at lcast't'.e e.l:'gued.--tliat it io

just their reception of e ]}In'ad:l.gm that trallSfo:rins into S 1,lrofcosio!l a group

that had previously just been intoreflted. in theotuay of }w.tlll'e,
sciences (though'ilot in f:!.elds 1:1.100 medicine. technology, a..'1d. la17,
:pri.ncilW-l

;lh050

'tltre is an external social need.), the fOl"Ill<Ltion of

cia1 journf'-ls, the

a special

L'l the

in the

of'

apscialis'.;s'. societies; an.o. the claim for

have

been

with a

20

first principles

can be left to the

justifying the use of each

of textbooks.

scientist can beg"-Jl his research

Given a DUradiBP, th0 creative

it lea.ves off

can thUD conceu-

trate exclua1vely upon the llubtJ.Got rulll. most eaotel'i,c aspects ot the natural
phenomena that concern bis group.
communiques begin to chmlge in ,rays

As h.e tlOO!l this. hmrever I h:!.s research

evolution hall been too little

studied but yhose modern eoo-prlxh,.cts are obVious to all e.l1d. oppressive to

to anyone vho might be 1ntel'ected in tIle subject matter of the tield.

Instead they m.ll usually eppesr ao brief er'tic!es addresseel only to proi'ea-

and

prove to be

only Cn0.8 able to read the payors aadre8sed to

them.
Todey in the sciences books are al""aya either texts Qr

reflectio!l5 tlpOl1. O!le c.*ct ox' anothe:- Cot: tb.e ecient1..1':i.c life _ Tk.l.e sc:tentiot wh.o writes one is .more 11.kely "co fine. bis profeaaioual reput..

velopmcnt of the verioun ncicnccs did

to profes2iollaJ.

im-

book o:rdiniirily J}osseo3 the name

that it still retains in other creative

.... -, "'".....

/:...

21

educated a.udience.

In dy-aamics research become 3imilarly esote::c-ic in the

later Middle .d.ges, and it .recaptured general intelligibility only briefly

Electrical research began to require

which had guided medieval research.

translation-for the laymanbei'orethe end of the eighteenth century) and

most other fields of physical science ceased to be generally accessible in

Duril'.g .the emne tvro. centuries silll:t1ar. trana5.tiol1s can be

the nineteenth.

ieol.ate.d in the various parts of. the biologi,cal sciences.

social. sciences they may 'I>"'ellbe occurring today,

j.n

purts of the

Althov.gh it hac become

custoruary,end. issu.,:,ely prOJ:ler, to deplore the 'irideninsgLllf that ee1=tes

.the professional scientist.from his.

in othe.r fields, t.oo lit'Ue

. attention is paid to the essential relationsh1p bet-J'een that Gulf and 'ehe
mechanisms .int1;'i061c to ecientific.ad"snce.
Ever since prehistoric antiquity one field of tltudyafter another ha.s
crossed the divide bet"!reen lfhat the historitm might can its pre-history as
a ec::i.ence and it.shis'vory proper,
been

GO

sudden

have implied..

0.'

so. unequ:l.vocalas mynecssflar:l.ly schellllatic dis.cussioll may

llu.t neither oove they been histOrically f7!!.dutl.l, coe:rteno:!:'!c,

that is to say. w.!.tu the

occu.rl"ed.

These transitions tQlllatur1ty have aeld.om

develo1ll1lsnt of the fields 'rlthin

ft:r:tters on electricity duril1g the r.irst four decac1.es of the. eigh-

century

&!lore infoxl""t-S.tion

elcctr1.cel

Cot.l.1omb I o.nli Volta in the lust ttird ef the

from those of GraY', Dll Fay: and. even

the first time ene.bled to take the

....'1:;."Y

'chan

these eru."ly eighteenth-century elecJ:;rical

sixteenth century.

ZV..l'thel' l"2ilnO"';;cd.

the

t:tn...;z of

trots those of the

17'!.() aud 1780 electricians 'irere tor

of their field for granted.

From that point they pushed on to mcre concrete and

problems, and

increasingly they then l'eported their results in articles addressed to

other electricians rather than in books addressed to the learned world at
large.

As a group they achieved 'ilhc.t astronOlilers h::lil gained 1.11 alltiquity,

students of motion in the DI.1ddle Agee, of phYGical optice in the 1n.te seventeenth century, and of historical geology in tho e:>J."ly nineteenth.
'chat is, achieved
aea"t."ch.

Co

They had,

that proved. able to gui.de the lmole group' [) re-

Except with the advantage of hindSight, it 1.s lo..ard to i'im], another

cr:!. ter:!.on that so clearly Pl"Oclaims

Il.

field a eCience.

then is eo

--------

and esoteric research that it,p reception by the Groul? llrmits?

moment the second. of these questions

anm.rer

to

the

t'irat will permit

us

:1.0 the

mOl'e

aid
Bye.

tcpi'oCried to it ;

For the
a BchelllUtic

paradigm I mean

a :fundamentsi sCientific achievement that!J.as been a.cce:pted as such by a

group concerned to explo!.t it further.

Fund.!J.Dlentalheremee.ns tlmt the

paradigm bll.ildll i'l'om first principles to

some t:l.methereatter, return.

group practice neea no'!;, 101"

Sc1entificmeans ,thoughthis is no defini-

tion Otthat vexing ter:n,that the paradigm inclUdes :l.l!lP1icitor explicit

.. la.mland pr:!.nciPleiabouttheelementsoftfuichnat\1remaybe conceived to

consist and about the
.

-.'

in which these
.

..

affect both observers

.

. ... ...- .. -and-each,othel".--.A-pm'adigm.,._...there:f'oro....._is_..at. .

more loosely,. a conceptual Ilche!;1c.

:Bu.,;

or, .

and this ,,1.111ate1' prove oriti-

cally important, a. paradigm lmlst alr;o be SCl:1retbing more.

It must. that is,

from its first enunc:!.a.tion be exhibited as givinG order to come concrete

rar.ge of natural phenomena.

These phe1lomena are the theory's knmm Iil.:ppli

cations, ani, v1 thin the p:rxadigm. they are Inver 'luite eep..'U's,ble from the

cepted as such.

Hore :lJ:n:portaut; appl:tce.tiolllO azoc need.ed to s\.ll?ply at least

part of the P<!!'ad'.e.:Iil'E! llleaning.

Stue.entG of Iiemon's 1'r1nci;pi8, like those

of any other scientific classic or tarot, learned the lll!lanings of concs!".s

like force, mass. rrpace, und time less f.rom the

23

auccess:ful than their competitor5 iu solvir,g a fe;, problEma that the

ot practitioners

to recoenize ns acute.

To be

grOt\)?

is

not, hO'lfever, to be either completely successful ,;'itt a single problem or

particularly successful 'il1th any lro:ge n'.lmber.

The Bl,ccesa of

Il.

parae.ism--

'lfnether Aristotle's, Ptolemy's, Neilton's, ui'loiBier's, or Mex;;ell's--is

initially aJ:waya a potential success

complete examples.

potentiality, au
i'actsthat th;;:

Normal science cooaiGtB in the actualization of that

achieved by

pOints need to ba una.slstccd.

science.

the knowledge of those

dieplaya as l=ticula,ly l'evealing, by incl'eaoil1g

the e::rtent of the r.:atch

Slut still 1n-

in selected

thoEle facts and the llaJ:uligm' s predi.ct1ons>

M(j!'ping-ull

Closely examined.$

Il.re wlmt eng::5e moot

or

contenJlSc-

25
.

..............

these _ere def'e-cts, cuI; they have

virtues.

by

focusing a.ttention upon e smi:.ll rll.llgc of relatively eccteric problems the

. parad:!.gmf'orces ... Bcientista to illvestigate sOlile part ofrw.tUZ'ein a detail
.. and depth thatwouJ,g,otherw,tseba un1maZ:1!reblj:!, . The. arees i:wescigatoil.,

normal. science With ita problems, are, of' course,

rainisculejthe enterprise.now underdiscuGsion has drastically restricted

vision .. But that restriction tu...""nB out to dono harm"at least nct to the
development of science.

Normal science. as we shall begin to see in the

next section. has abuilt,..iu mechanism 1.'hich.ensures ttsabandom.nent wen

its guiding paradigm ceases to function f'ruitfully,

At that point scien-

tists begin to behave d1fferently. end the nature of their research proble1l16 changes <Satin the interim, during the period when the paradiglil is
euccesaful, the profession "!dUlle-ve solved. problems .they could scarcely
have iIP.agined and vouldnever .. have ."Mer"taken withcut. corlllll1tment to the
para.?igm.And at leas'l;pa..""taf that achievement:. always proves to be perma-

nent.
To displaj" rllOl.'e clearly. ,,'ha.t is meant by no.."'mal 0"' paradigm-baaed rcIl'illll'ch. let me nowat'csl1lpt to classify snd illustrate the problems of which

normal science principslly COnB1sts.

For convenience I. postpone theoretical

activity p.nd begin with fact gathering, that 19, with the experiments and

observations described in the technical journals throU3h Which SCientists

:!.nform their professional colleag'..lesof the results of' their con-tinuing research.

choice to

Cll ,;hia:t e.er.ects of nattl,l.'e do SCientists ol'di=ily rc;port'l

'1hat

V(.'7st:lgation, end they are neither elw--ays ncr

is that cle.sG of i'e.cts tlhich the )Jaradigm has sho;m to be

vee.1ing of the na',ure of things.

di.Et:.hJ.ct..

.e-

:By eEPloying thei;! in salving problems the

paradigm has made them l10rth o.etel'lll1nii:lg both 'doth lilore precision an.'I. in a

larger variety of situations.

At one

tual determinations bF.ve included:

or another tness significant fac-

in astroncmy--ete1.lar position and

nitude, the periods of eclipsing binaries and of planets; in physicB--the

specific gravities and compressib1litiee of materials,

lengths and

spectral intensities, electrical conductivities and contact potentials; end

points

ill chemietry--compoaition and cO!!lbining Wights, boiling

of solutions, etructursl fOl>mlle.e

activities.

SIlo.

Il.ci6.ity

SIlo.

Attempto to in-

crease the accuracy and scope vith which facts like these are known
a signii'ico.nt fraction of the literature of
science.

tnd. Observational

l\ga.in aLd again cotiplex Ilpzcial ap..=atus has been

for

such purposes, and. the invention, const.--uction, o.nel deplo;yment 01' '"hat ap

paratuB ha.s demenO.ed first-rate talent I

cial backing.

time I and cCllsi<lerable tiILO'.u-

Synchrotons end radio-telescopes are only tho moet recent

e:;:.aropleo of the lengths to lmich acientiata lI"ill go 13: a p:n,oo.igm aSGurea

thei!! that the facts they seek e.re ll.'llOl'to.nt.

From Tycho B;:slle '(;0 E. O.

renee a nunber of them have acquired great reputations not from

novelty

of their diacov"riea but frcr.:l the preci:::!.o!l, z'eliability, mil. ",cope of the

of fact.

...f.O
:;'ec t en.

-t

1:otl\)...:-;<;.:

r ... of ...
::..<........

. . . ;.;.,

'>--' C.;:.?,.:-.r.
'- ..,

.' .,.. .,..,'.r.

-..'.,'. -,..

..... ---

..

."

27
"'ce.n-be cOJJl!l1ll'ed:directlYl>"itnJ.?l'edct.:!:Olla from the':9$"adigm'.;!leory.

shall see ,when I shortly tl.lXn,i'rom'Ghe c;;""l,Wri'Jl$z(.;ruto

leme ot: normal science, there are Ilclda.n muny a.reas in

AS1re

theoreticalprob-

a scientific

theory,-pa.rticularly a mathemat:1cal theory, can be directly

.In.th

theory 01' relativity. li'u:rthel:'li1w-e, eVGn in those aroos uhGre 8J?j}lication is

possible, it often demilnde theoreUcal and 1.1strumental allproxima.tlo11!J ',;hat

severeJ.y.1.1mit -t;he agreement toea expected.

findiDgnewareas in

the..t

or

agreement can bo demonstrated at ell presents n

constant challe!'..ge to the skill end :1JusginaMon of the experiment.alillt rmd

observer.

Specialtelellcopell todell!lonstrate the Coparnicsn prediction of

annualperal.lax; Atwood' s machine. first invented. almost

century aftei' the"

Pr1nc:!.1?ia, to g:l.vei the first uneqUivoetlldemonatrat:i.on of Newton's

l".&l.w; Foueault' s appare.tuB to show the speed. 01" light greater :I.n air than in

lreter; or. the gigen'ticbubble .cilambel' designed to, del!l?l1strate theexis'.:nce

oftheneutr:!.no;tl:tesc p:l.ecee of apeciaiallparatulil s.ndJilaIl..y others like th\)m

illustrate

:imuensc etf'o..rt and ingenu:l.ty that have been required to ori.ne;

implicated directly in the design of

i-Jithout the?rin:::ip,j&; for.0l'.!\l!lple,

lr1pm'attl3

able to solve the problem.

meaS'.l,;:'Cll!entG

made ,lith the Atvrc"d. ma-

28
undertaken to ar'i;iculate the paradigm tileo-.L'Y, t.h"i.t ie, to

some of

its'residusl ambiguities thus permitting the tlolution of problems to

the paradigm h;:l.1l previously only drawn at'tenUon.
the most important of all, and its description

Thill

proves to be

its subdivision.

In the Dore mathemticlll sciences some of the experiments aimed at articu-

Jation are directed to the determination of physical constants.

for er-=1?le. indir.e.ted that the force

distance would be the same far all

Newton's

unit masses at unit

of matter at all positions in the

But his own problema could be solved without even eatimatix:g the
,
size of this attraction, the univerBal gravitational cOl1stan't, and for a

\miveree.

century attar the Principia appeared no one else vas able to devise appal'stUB

a.ble to determine it.

the 1790's the last.

N01' 1m9

Cavendish's famolls determiIIBtioll in

Because of its central position in physical theory,

',ll1pl'oved values of the sravitational coostant have been the object of repeated efforts ever since by a number of outstanding experimentalists.
Other examples of 'the seme sart of continuing work 'Would includ,e determinatj.olla of the astrollomical unit, JI,-;ogadro's nlSber, Joule's coefficient, the
electronic charge, and so on.

Few of these elaborate efforto uould have

been conceived an,l none wculd have been cru:ried out v.ithout a paradigm
theory to defiuG the problem and to guarantee the existence of a stuble sol.utton.
Efforts to

fI,

paredizm ere net, hmmvel', l-eetri,cted to the

29
isnotapparentth8.t aparadigmisprerequisite1;o-thediScovcl'ycf la'i,s
____

- We

hear that_theY_Il.:>:'e found by examiniU$ It-easuremen-ta un

dertakeu_i'or theirotm sake Wld 'llithout theoretical commitment.

of'ferstlo'supporti'or-so excessively Be-coman a method.

But history

Boyle I s experiments

---wcrenotconceivable -. (and- if. conceivecL'lould_'ha.ve__

tat:l.on or none at allFuntil air lms recognized as' all elastic fluid _to "1hich
all the elaborate concepts of hydrostatics could be applied.

Coulomb's

L'lUC-

cess depended upon his constructing specialapparlltus to measure the force

oe'cl1e6n pOint charges.

(!rhoae ,'mo had previously measured electrical:f'orces

using ordinary pen balemces, etc., had found no_ consistent or simple regularButtl:ul.tdesign.

ity at all.)

depe!lilcl'lll!<?l1 Franluin's previOUS

recognition that every particle of electric fluide.cts upon every other at a

__

the force betwensuch particles--the ouly for<:e v.hich

might safelY be assumed a simple function of dlstancethat Coulomb ,-ms look-

ing. -

couldalsobe\lSedt;)illustre.te

Q.uantitative

JAWS emergethrClll&li parildigmarticulation. ' In fact; so general cnd close is

the relation

qualitative paradigm and Q.uantitative

t$t, since

Galileo, such la..Tshave very otten been correctly and repeatedly guessed by
'choseii"lio held the paradigm for some ;real'S before apparatus could be designed
f.'or their experimental determino.t10n.
Finally, there is a thil'd sort of ezpel'iJne!1t -which aims to articulate a

pal.'ad1gm, one which is particularly prevalen', in those periods. end. sc:l.ence3

"hi,eh deal more "l-1ith the qualitative than'l.1.th the q"..tantits.'cive aspects of

nat.ure g s regule=:lty.

Often a pnred.1e;m de<:.relCj?ed.. far

O!'le

set

of phenomena

is

30
to the

ot intelest. For example, 'the paradigm a:U:i?lications

area

lll'l,1

of: the

caloric theory of beat \lere to heating sn.d. cooling by mixtures and by change
BUG heat could be released or absorbed in

of state.

e.g., by chemical combination, by friction.

ana.

lJ'.aly

other

beSides,

by comprcsoion or absorption

of a gG.s, and to each of these other phenOflCn!l. the theory could be applied

in several lrays.

If '"he vacuum lw.d a hea'G cacpac1ty, for exe.mple, heating by

compression could be explained as the result of mLxiug gas

might be due to a change in the specific heator gases

sure.

void.

or

it

changing pres-

lind there ,,>ere ueveral other explanations besides.

MElny experiments

,rere undertaken to elaborate these various eJ.lllanstions and to distir.zu;!.sh

them; all these

arise from the caloric theory sa paradigm,

and all exploit it in design of experiments

sults.

Once the phenomenon

all fuJ.cher

or

heating by

ana

in

of re

had been established,

in the area were paradigm-dependent in this vray.

Given the phenomenon hO'\l else could an expel."iil1cnt to elucida.te it have been

elloaen?

Raving discussed the empirical problems of normal scientific reseerch

at 'Ghis length

terparts.

permit an abridged examination of

The theoretical problems of normal f.'cicnce fe.ll into Ve;,'Y

'the sane classes as the e):perimental

'(;ooo1'e'(;1co.l ,{ork, but

ana

Observational.

this case a vcry smeJ.l part,

tsx'iGt:1cs,

of ast:::,"onr:mj
0=

the

the

r:.CB.,;:J.y

A llurt of no.';)""l

use of ex:i.oti!'..g themy to pred:i.ct factual ini'o::.'mation of

Ths

COU11-

simply in the.

vl:.lue.

of radiO :propsgation cm"vcz: are

of. lens ch,&'acof uo:'k

31

tain

/l.

greatwmy theerstic:?.!

ot: problelC8 !l'hieD.,

the nCll-

___.._._. ____.... _. . .

tr:!.ns:tcally T&1ua.ble bu;t becauGc they can be confx'onted directly vith

-------'pe-r...J:!!ne""
..

1eta clieplay ('. nell

ClC-

of

to incres,ee the precision or an c.pplicatlon tile.t llalLall'eady been sade.

T"ne need for trol'k of this sort a.-r1ses from

'l1l1!;:t

I previously caned the

illllllQnee diffiCl.'.ltiee oftenencot'.ntered in developing points of contact be-

'tween atheol'Y aIld. ns.tllre. The.se difficulties can be b:defly illustrated by

an examination

w: the hiotory ot: dynamics after

.teerith centu.--y those scientists '\/ho Mceptedthe

By the early eigh-

lie-Jton.

as

took

the generaUty oi' its conclusions 1:1:' granted, and. they had every reason to
do eo.

No other 'Work lm0\'1l1 to the hietcry of science has permit"(;ed so large

a eimultansous increase. in

on
....

"l'IA
..... A.. '1i.".,WI ...,

--

tiOll.l end ad hoc

and an ill1porte.n:!;

... .1.

fOl'lrr'.lle.

science at the time,

....

scope m!.7. );ll'ecie1011 of reeearch.

......
,,,J.

"

4,.t.lJ.l.4

"""'

r'
t";.:1r..
oloU.;.g.

he had elEo been

lJ-T+.h
1 .... _

of these

__

..........

0'2
r. ....1 .;.1._
.'. ;.:c.v..
.....

to lie:d-,c Eeyle' s rr."

fo:r the opesd of c01.llld in air.

r.lUCCCIlS

F:ll' th'l!

Given the f;w.te of

'!;as il;.tencdy im-

quired in order to pI'o\'ide the epeci&l dI1.',,;a that the conCl'etc

of Newton' B paradigm dell'.anded.

Similar di:fi'icultie!1 in obtaining agreement

existed on the side of theory.

To apply h:1.a

to pendulumCl J fGr

Neilton ..-as forced to treat the bob as a. masfl-point in order to prolride a

unique dei'initiOD of pendulum length; most of his theorems J the few eJ>cep-

tiona being hypothetical and preliminary. aleo ignored the effect of air
resistance.

These vere sound j?hysical. approximations.

as

apprOX:i.matlon they restricted the agreement to be eJ:pected bet,reen Newton's

The !1ame dUi'icultles a:pr,ear even

p:edictiona GIld. actual experiments.

clearly in the application of rJelf!;on's theory to the heavens.

titative

SillWle guan-

observations indicate that the plcnets do not qUite cbey

lillpler's la'fS, and N5mtou's theory indicate a tw.t they ohoulc1110'I;,

To de-

l'ive thoBe laws i{e-.n;on had been forced to neglect all gl',witational attrac'\;ion except thai; betv.een indiv'.dual p1al2ets alld the sun.

also

Since the plw..eta

each other, only apprmtimate agl.'cemsnt bet,,-een the Cl)lllied

end telescopic obee.-vation cOttld be eA1?ected,

All in the case of;' pendulums the

zactory to those "ilo cb'Gained it.

of those

,T20

aw.eCilWllt

Obta.:!.l'lcd was more tlmn satis-

No other theory could do nem-ly

questioned the validity at He.lton' Il ,.,ork dia

GO

50

,.ell.

bec(">Uee

33

Jilutue.J.lyai:i:ract:tngbcdies.
eupied many

ot

'Il

These ;pz'oblcms and.lllSllyothers like themcc-

best ms.thema;i;:l.c:io.ns au."t'ingthe e:i.ghteenth andesrly

nineteenth centuries The Bornoullisj,EuJ.er,LP.grenge. lE.plece,

ana. Gause,

aU.. d.ilisome of .the1rmos,!;bri,111antvork !')nproblemlll!.).lllCd to :tm.prove the

match betlreenl.'ie1f,1;on's parad1gm.lJludnature.

Manyoftlieso same fl.gures

s1!1ru1taneO'J.slyworked to develop the mathemll.ticareqUired' for <lPlllicatiorul

that Nemon had

not

even attempted, producing,

en

ii1lll!eDSe

literature and sOliWvery l'cwer1'1llrnathell'at1caJ. techniques for

and for the problem of vibrating strings.

These proDlems of e;pplicat:lon a.c-

count 'for whatisprobablythemo.ltbl:tlliant end conauming scientific wOrk

of the eighteenth century.

And they could rl/la.diJ.y be d\tplicated by 1m

examlbat:l.ono:f' thepoBt":paredigm periottinthe developmento!" thetmod.yne.mics.

the_Vi! theory of ligiit,eJ.ectromsgnetic theory,or any other bra:::lch of

science whose fundtillnental le.i'S EIl'e. tully quanti tati ve.

At least in the more

mathemstical" sciences moe1;theoret:l.cal,vork isof.thio sort .

ll'<lt it is not all oi'this scrt.E,,-cninthe

sCiencee

are alaotlieorErf,:!.cal'problemo'oi'pa7.'a.d.:lgm articuiation,a.uci In the

there

periods of mm'o q:ml:l.tative sc:tcntif.'ict'.e.el"l>l"ilnt thelle. pl'oblelEs

p::'oilJ.ems, in both ',he more quantitative Emd mor-s

cc.ienceUt e.1!11 61nrjilly t.t cll3rific<1.tion by reforl!lUlation.

'ra.e E:2:.n.ciJliq. fo"!!

exa.::lpJ.e. did n,jt al"lmys !!l'ove en eesy \;c.r!( to a.pply pa..""'tly bec:ml3e it l'etfdnee. some of. the C1V.!llBines!.I inevitable in a first v(mture s.nd because S'.)
1'[;t.lch

1:'(#9 metil'd.ng waG only :i.mplieit in

brilliant mathematical

et:d.cnvored t.o l:*c:i:'crmu.lc:te

leflBons of. the Prin.cip:ta in a log1c&11y more coherent vel"s:i.o!l., Dill) 'Gha.t

lIculd 'he less eqtrl.,'ocal in its applications to the nC;,ily els'borate problems

Similar reforlllulat:!.oll3 of. a pm:>ail.igm have occu...-re.d. repeatedly in all

of: the sCiences, but most of' them have lll:'oduced wore :mlnrt3lltial cb.e.ngas in

the p.'l.l'ai.'!.igm than the reformulatiollS of the Principia oited above.

Such

changes rellUl.t frOll! the empirical .'ork prev.1ously (lescribd as ailted a',; para.digm articulAtion.

arbitrary.

Indeed, to cll:>seify that

of work a9 <;l!(iliricaJ. we.a

than any other sort of normel research the problems of

paradigm artic1.1.1e.t1on are simultlmeously theoretical a.."ld. exr."crimentaJ.; the

e,mmples given previously

mll serve equally ...-ell here.

construct h:1.e egui:p:nent and make measurements

electrical theory in determining how his

of hit; meaouren:ents

lI'aS

he could

i'C Coulomb had to c!!I:p1oy

should be bUilt,

a refinement in that theory.

O:r

the

['.sam,

is of this sort.
{", ... 4-

_,,;' ... V;

35
theUterat1iriFof-norlt:aracience'-15otliempiricar:::n<rtlleCi'etical;-. -
not. of cauree. quite exhaust the entire
also extraordinary problems, and1t may

ofm:ience.

lTell

do

There aJ:e

00 their resolution that "'"kes

the scientific enterprise -as eo whole lIopa..--t:1c:u1e.r ly W()rth while.

trll.ord1nm:y.problems.are not to be.h&dfor the aI3K1:ng.

ex-

Theyemeree only on-

special occasions preparlld by tlw advance of normal l"eaearch.

Inevitably,

therefore, theovennelmine; majority of the problems under.takenbYeyeu the

very best scientists usuallY falls into one of the three categories outlined above.

Work under the pa:t'eil.igm can be conducted in no other way. and

to desert the paradigm iato cease practicing the sc1ence it defines.

shall shortly discover that such desertions do occur.

about vhichscient1fic revolutions turn;

We

They are the pivots

But betore beginning the study of

such revoiut:!.ono,letuiihave one lasta.ti!l morepanore.rnic looket the normal

scientifj.c purSUits l>'hich prepare the ...-ay.

have Just considered is hov little they aim to produce

eithel' conceptua.l or pnenomenal.

have repeatedly

The elaborate projects that scientists

the

to

or theo-

retical prediction of, say, v.svelengths or planetary motions increase the

and precision of the

cauce directed so

But be-

available to science.

to phenomena already known, they are partic

ul8xly unlikely to disclose

new order in nature or to add a new sort

of fact to the liet with which science deals.

Every&hing but the mont eso-

ter:lc detail o the result (e.g., the lest t.ro decimal. ;places to the right
in a mm "e.vel.ength meo,surellWllt) is knC1<m it! ad'lranee.

An el'.p-erimeut

theoretical manipulation whose outcome does not closely coinciQe

0):

ex

I19c'o;at1on tells nothil'.g about nature but s:!.lllply rei'lec';;a upon the zkill and
aometirees even upon the

of the man

it.

Advance e7.p6ctations are not, of COUTee, often so :preciSE as in the

or v.svelength detel'miua.tio-A.

Particularly in work v.hoee aim is

there is usually genuine latitU:;1e of e:::pectution, and.

e Bort of factual llovd'..,- can then emerge.

urementa need

perea. for

S.l1y

perhaps; heve :I:'itted just an invel'sc egnare lal;; the men

one of sevez-el different results.

anticip&.ted, (E'.rul

,.
un.;.":!

-l-'

The Z"cllUltf: of Coulor:;1J' fJ meas-

' .... ,. \'.t

..

aei!imiJ.eble)

,',;.

' .. )...; ..

.... ; ...:.v\";I..-.

'.-. ('.'

Nevertheless, the ral1fJe of

is al1!aYO smal.l,

-:-:.r:-,:-:."
.'.-.,."r. . -:-- - " ' - '
_._,'

..,.'_,

the PTCjcct.

. ..,,:>.:.:_'o;':.:.
,., .:.;-.

>

which they derived:

--,

Ther-sfo:re they rer.15.ined mere facta I unrelo.ted &'110.. u.n..

relatable to the continuing progress of electrical resecrch.

Only in retl'o-

. about, whatcharacteristlcs Of electrical phenolll('::ila they display.

C01l1cmiJ

and his contemporaries" of course, also posl3essed this later paradigm or one

that,

appUed to the problem of attraction, yieldB the same el.-:pecta.

tiona.

is 'Why Coulcmb _e able.todG-sign

assimilable by paradigm articulation,

---

'l;hat gave a result

But it is also ,illy that result sur-

prlsed no one and lItIy several of Coulomb.' II cOl1temporaries had been able to

____

predict it 1ri ad"-rulce.

..__

Even the project whose goo.l is p3radigm-al't:l.culaUon

does not a5.m at the unex.pected novelty.

.............c ....... _ .......,:.............................................................................

But i:( the a.im

of:

.....................

n.o:t'I!lal science is not major I!1lbets.\'ltive nmrelties--i,

/

failure to come near theant:l.c:!.pat.ed result is Usually failure a.s a scientist--then why are these problen1s ul1i!.ertali:eu. at a.ll'l
a1rel1.Qy

Part of the answer has

developed . To scientists. &t least, the results gail1ed in no!'-

ms1 research are significant because '.:hey e.adto theacope end precis:1.on

vith liniel:! the

be ap,pl:!.ed.

Thatanm;er. hom:lVer, CImll.crG accouut

fez' the enthusiasill and devotion th"..-\;. scientists (,d.splay fer the ;probler;s of

nO'.cml?l reaearch.

No

OIl.e

<l.evotee yee:rs to. flay, the a.e,,<:'J.opment of s.

spectrometer or the production of an ilJl.prcveil solution to the !ll:oblem 0;:

38

so lII1lch of it is repetition of

that have been ce.rdcd otyl; bci'cro.

That, I think, is the key of fascina'Gion of the norms.1 research problco.

Though its outcome can be anticipated, often in detail so
mains to be known is in itself
remains very much in doubt.

that whE.t re

the way to achieve 'i;hat

:Bringing a norme.l research }?::'oblem to a conclu-

sion is achieving the antiCipated in a neUl,>ay, ana. it requires the solution

of all sorts -of complex instrumental,

and mathematical puzzles.

The roan liho succeeds proves hill!seJ.:f an expert pu:::zle-solver.

The problems ot

research are puzzles, and their fascination is

largely that of the puzzle.

The psradigm from which they derive defines

both the natUl'e of the solution

that solU'l;ion nmst be unl'.ertaken.

the rules under which the search for

problem (or puzzle) at all.

Wi'thout the paraa.igm there would be no

Eq1:al1y important, it is the scientist's

ment to the ptJ.re.digw which assures him that the problems it poses

have a

Without that felt guarantee the problems of normal resesrch,

eolu-(;ion.

Hoother conceived. or not, would ecarcely be v.nclertr;:ken.

COl1ll:li'l;-

Perhaps all this

normal research eeem uninteresting, but it ought not do so.

Any

chess

player w-lll testify that to knOl{ the rules and the sl:istence of a solution
dec3 not deprive a chess prOblem of challenge end fascination.

On the con-

without that knO'.>'leo.ge there ll01.1ld be neither puzzle to solve nor game

to play.

This reference to chen:: can, ho,:O'i'$1', introduce no more th<:" a me'Gs.llhor.

Ae a rr.etaphor it ca',chee, ! beHove, the

::.nii fc.scination of normal

science" vhcu,5h probably not the iU'l;ensiJIi:r of th;;. addiction g.sna:,;-c:'::;ea. by the
J

..
'C::.0

.......
(;,
......;:0\...:;. ... ;...-; ...
. 1

... , ... .:_ ....

... -'

....

::".:;... ,.,..;'...:,:,....',. . .".

--

- -

....

C".,..r
-

.f-."

<:.. : ....

--

..

.... _..

.._"-

39
is .conducted.

Chess ba.a

same can be played by machine.

that can bew.:.a.ee:r;plicit: thl1.t is "'hy the

lire there equivalent rulea for tl:.e

to do normal science?
let me say at once that I am not altogether sure of the anlTh"r to

these questions.

For some j<!are I took

it tor

tliat

a.

seto! rulea

eui'ficient to detemine normal 8c1enti:f'1c practice was implicit in the textbooks and the training procedures that

the student to the profes-

During that time my discu6s1ous.of norl!lal sCience made no reference

sion.

to pIlIlldigme.

Inatead they referred to the speci:t'ic

mitments--methodolog:lcal, theoretical, and

ular problem"solViilgtr8.ditionpol3S1ble.

of

COln-

n:eJl:e a pm"tic-

on that view periods of normal

science w"re periods of consensua, during which the entire scientific commun1tyagreedabouttherulesofthegame.Al!dscienti.fierevolutions were
then the episodes through l,hich the rules of the game were changed.

That is

the position implied, for the salce of preliminary

the in-

troduction to this monogra.ph, and I st111 tind it extraordinarily tempting.

But it is very probably wrong, and we can learn something more ebout both
normal science and. revo1utiol16 by exploring first ',he position's temptati.ona

and then the reasons for reSisting them.

Anyone
111tely

bas tB.lked much with contemporary SOcial SCientists is

have observed their frequent and deep concern

&'001.11;

the <lefini

tions of their various fields and I).bout the problesa, methods; end stel1'J,rc'.s
of solution legitlme;(;{; 1'01' their specialties.
this sort

Among many of' them

of

repeatedly discussed, aud the discussiousproduce little or no

40
They were, for example. particularly ilUport&nt in tha

cf

teenth-century chemistry and of late Gigh-te:enth- ,r..nd. ee.l"ly

geology.

Furthermore, they recur

GeV(l!l-

duriug periods of sc:l,entific

revolutions. the tiMes t;hen para-diems Chllnge.

transi tio!'! f.'om Ne,,-'.;onlv.n

to relativistic and quantum mechanics evoked mauy deba'"es about

and standards of physics, some of wich still continu,e.

'ene

nature

There are people

alive today lTho can remember the simila.r argumento engendered by l<a:,mell's
electromagnetic theory and by otatistical mechanics.

PJii earlier still the

assimilation of Newtonianiem gave rise to a particularly famous series of debates vith Aristotelians, Cartesians. and I.eibnitzia.ne about the standards
legitimate to science.

Contro-.ereiee of this sort, though not

on this

scale. are a consistent feature of scientific revolutions and a v,sual :feature

of pre-paradigm scientific development.

But they seldom occur dvring periods

of normal science, and it is therefore very temptil".g to suppose that normal

science depends upon a consensus among practitioners about the nature of
their field, i'l;s problems; aml its methods.

Appal'ently, if scientists do not

d1.sagree about the nature of their field, that must be bacause they s;;ree
about :Lt.
Yet tha't cOnclueioll is Ilurely not neCCfllml"Y and is very ;pxobo.bly lllis'taken.

Certainly there are a great many th:!.:ngs about ,mich all scientis'Gs

do agree, and
In the first

probably could be explica.ted and tranetormed to l'u1eE.

there are at least a fe\Ol endurir,g cOll!lllitmenta withou'c

Which no man is a scientist.

The sc:tentist must, for eXaILple, be concerned

to underctand the 1;orld end to demonstrate previously unno'l;;l.ce(1. aspects of

41
of

uei,i

his observational

or to

. ''of
..

There must be other l'Ules like these that hold for a.llscientiots.

In ad-

axid.thatthusdist1l'1guishit frOlJl its predecessors axidS'.lccessQJ;s can be

made explicit.

NS1;ton'sI.awswere l'Ulesfor the Nemonislls .. but not for

the .Aristotelians or Einsteinians.

Mass was a fundamental ontolog1cal oat-

egory forprerelativiBticphysiCsji though it is no .longer that today.

Or

i:tgaill,thoUghChemietsbelieved1nelemertts b6thbefore and after I.avois1er,

the Chemical Revolut10n:traJisformed the chemist's convictions about what an

" ........... ,.......,...

..
purele.boratbry suDstIlrices.

Before that timetheyhSdbeen conceived as sub-

stances which lnprinCiple could not be isolated for laboratory .investigation . TheaeeXrimples illustrate the sorts of rules that ce.n be made e:i:plicit
and that 'help todistingu1shone trad1tion of normal reEearch practice from

Undoubtedly other such l'Ules would be discloeedby the study of pv.rticnlar normal-scientific traditions.

We shall, in fact, be

many rnor.e

'Of thamin tlie'pageathat follow. . Nevertheless,thenumber'O:f: rules ths.t

call be educed. by suchatudy never seems su:f'ficient to.define .the puzzles the',
sc:1.entists normally undertake or to rcstrictscient11'ic .attention to their
pursuit. Furthsl'more, it is by no means cleG,!' that scientists l10uld or even
could agree aboute.ny set of r'.:!lee su:f'fic:!.entlylnrgep..ud prec:i.se to have
that effec'i;.

Asl, ellY rand-om group of physicists;

aotrollomcrs, or

get almost as = y different

there

ImlTh-el:'S Coil

Ul"e

scientists in

-YG\J_'7

-.

r.;,'.!';n2c_

and. -none of theil: &llSwcrs is likely to fit prec1.3cly til", iull range vi: ccntinuing professional reaearch.

Or Oxam:tUe the texts tllrO"<lZll ,lhich at-:.zdents

learn to practice normal science and the ClaGB1cs through which an older
generation was

to its professioilll.l :?ield.

l,gain you "ill :rind

many explicit rules, laws, and principles, but again they will collectively
fail to dei':tne the traditional problems and procedures of the profession.
If' a BUi'i'icieni; body of roes is to be found at all, it

simply from a textbook's discussion of

not be retrieved

dci'in:l.tiona, and theories.

disC"<lseiion ".oilJ. y ....oYiu.e same rules for noml science, but

That

must be

sought implicit in the semple problems that bulk so large both in science

texts end laboratories

in the

part of every scientific cla.ssic.

app11ce.tionsthat are en integral

That is

wy,

at the start ol this aectiCll,

I insisted so strongly that its applications are a

ot every parsdigru.

But if sampl.e problems and applicaticns m'e pext of uhat =);;es possible
e. tradition of

UOl'r:lel

reaec::ch, then that trs.dition need. not be entirely de-

'Germined by discoverable rules.

Applications end problems need

im!?ly

rules in order to deterwine nOl'1l1$.l science.' Rather than learn r.ulea the
scientist can, and. in some part clearly does, learn by pr<:;ctic1ng on paradigm
problema.
through

In contcmpore.:ry scientific education thin process of educ:;rc:!.on

bezins in high achool or in the first

con.tinUetl s"i:;et'..d:i.ly to" or more often

Subsequant prof'ee3ioMl l'e3earcn. is no

of college

the c.O'..;to::.:'al dissertation.

I'!hsrp and r,il!dd0n depal'ture.

Its

43
-

---.

and less than a set. of rules .tor .the co!lduct. of .the... scientific life.

It is

because they learn in this vay thatscientista can so regularly agree in

'.'theirevaluationsofpart.1cular-problems.and.particular.solutions without
manifesting anysimil.eragre"ment
to underlie their Judgments.

thefu1laet of rules that apl?3ar

One can model lfork upon.aparlidigm or recog-

nize work mOdeledoncne 'ifithoutbe1ng entirely able to say what it ie tOO.t

gives the model its status.'
This view. of the ...y in which parad:tgms determine. normal science seems

to me ;particularly plausible because of' 'Ghe assistance it gives nth

__ .....

_ .. ".,,-

problems that we shall ellCounter below .... Hot all the debates that are evoked

__

__

,, __ ,,_.,."_,,'_." _ _ . _ " " " _.... _ _ . _____ ...

_____

_ _

___

__ ,_ ... ""' ___ ,,

revolutionarycbange in 8c1enti:f1ctheory are defini-

bye. proposal for

8.

the profession' saEsim11at1on

by

__ "." __ ..:.....""" ___

of the new paradigm.

Some

are, of course, ;particularly those' that revol"/e around the. choice between
tlro cOl!l1}eting setsoi' Bc1ent1:f'1c lawliorsometbing .else of the sort.

But

other debates" part1cu18rlysome .of thoflewh1ch:attempt to define .the proper

,

suoject ma.tter of a particular science, seem rather to die out

Clergl'01l!ld 'Without any consensus .having. been reached.

sc:timce

irl't<:l1'

Ol'

to go

Though .relewnt to a

thefpllrOOigmhasbeen chocell... Problems can be selected

solutions evaluated m.thout sareement about

00-

14'10.

During p!:)l'iods of ncr-.ral

science' they are therel'O"...e silnplynot often nor seriouaJ.y diccucsed,

Ins'Gcad.

they renain as lo.tent srsaa of d:l.sagreamer;.t until b;,'ought to light by a

philosopher or by a new flcien'Gif:i.c revolution.

The absence of disagreernent

A second and more important problem about flcielltii'ic rcYol1.\ticns is

even more sierJificantly affected by recognizing that a nor-wal

trac.ition may be deteTlllined by moo.el.:i.ng sa well as by Tolles.

E,rery physi-

cal scientist today learns the 1&'Irs of:. say, quantum IlWchar;icB and mos'l;

emv1.oy them at

point in their reeeerch or tee-ciling.

But they do net

all J.e&'n the same applications of theee lalrs, m:ld. they ere not therefore
aU af:fected in the

S8!o.'C

ways by changes in Quantl!lll-mecillm1caJ. practice.

On the I'cad to professional specialization

a fe'\l'

physical scientists en-

counter only the basic principles c:t.e q1.1lmtUl!l mechanics.

o-thers st,my in

detail the paradigm applicatione of these principles to chemistry, still

others to the physics of' the solid state, and

80 on.

Wast gw::.ntUl!l meche.nics

means to each of them depends upon what cour13es he has had., what te:dis he:

has read, and which journals he studies.

It follows tbat, though a change

in qm1.ntum mechaD.1cal. law will be revolutl=y fur all of these groups, a

change that reflecto only on one or another of the pm:adigm s)?plice.tions of

quantUl!l lrechanics need be revolutionsry orJ.y for the me:nbere of a perticular

professional Bub-specielt7.

For the rest of the profession and for those

who prac';;ice other physical sciences that cheuge need

at all.

DO'!;

be revolu.Uo:anry

In short, though quantum ll'.echanic:s (or }Jewl;onian dynClllic:n, or e lec-

theory) is a paradigm for many SCientists, itl> role as pe.radigm

is not the same for all of them.

It can, there1'ore, siur<l.l:..al1.eously

several traditiolW of normal science

ovsl"lap without beive; coextensi 1{<:).

Ae a resu.lt, a revolution proo.uced ,ritl"dn one of these trruli',;ions 1nll not

l1ecessE'xily e::;:teno. to the ot.h.era e.s

In the sectioZl that follol:a im1!le-

a highly cumulativeenterpr:!.ae, e1ll1.!'..entlystlccessful 1:0. its a1m, t.he steady

In all these

reapectsit fitswithgree,i; prec:i.ilionthemost usuelima,ge oi"acientific

hesitate to equate this scirt of science wlth the

vhole, for one standard productotthe scientific enterprise isstillm1.s

science s'*Iksno noveltiea of factortheol'Y, and"

cessfUl,f:l.nds none.

New Wld'unsuepected.phenClllena are, however: repea.tedly

uncovered'bt sc:!.elltif':1.c rellearch,

nelf theories h:J.ve ega1.nand

_. ...... __.......... ._._l'.gain.beelLin;rented.'tiy_sc:l.elltistl...

. . tlle ....

tiftc enterpriSe has Cl.evelopedauniquely

llUl"pl'isks ofth1,ssort.

mth

pooerful

If this

ofedence is to be recon-

..e.yof inducing lla.rad:Lgmchalige.

funilaInentalno"relt:l.es of fact e.ndtheory do.

game playedund,eI' one

techniquei'or produc1.1'.g

has already been said,then research under a paradigm must

bea particularly

tion of

IlUC-

set.

The.tiswhat

PrOduced inadvertently by a

set of ruies, theirassilllilatiou requires the elaboraAfter theybave beCCllleparts of science, tbeenter-

pl":!.se, at least of'thoscspecialtsts:l.n'lihoee perticula.r field the novelties

lie, is never quite i;lie same again.

We !:Just now ask am; cha,\1gea of thio sort can come b-,"oout:

t:irst discoveries, 01' novelties of tact, e.ml then inventions, ox'nov0:ltieo

1;5

fincl that \lJ.scover:l.eo are not ieo1ated' evento but ex'tenC'.oc. e;giso:l.es Hi',;h
a l'egu.larly recurrent structure.

Discovery conmencea

the

miu!'eUGSS

of an0ll!3.1y, i.e., with the l'ecognition that nature han somehow v5.oJ.atcd

the paradigm-induced expectationo that govern normal

It then

.71th a more err less extended eJ..'Plare:tion of 'Ghe area of ancrualy.

And it closes only"ilen the paradigm theory bas been adJueted so that the

anroalous has become the expected.

AElsim1lat1ng a

netf

sort at fact d.emands

a merre-thlln-additive adjUBtmellt of theory, and until that adjustment is completed--until the scientist has learned to see nature in a differeflt way-the new fact 1a not quite a scientific fact at all.

That conception of

dil1co-;rery as a process :I.Ilitiated by ano:aaly and completed. by a simultaneous

adJ1.,stment r:tf: fact and theory is the one to be discussed in this Election.
In those that foL1ow we shall explore the rather similar'structure of the
episoo.es that eventuate in the prcduct10n of JD3.jor theories.

There too 'We

shall note the VIa;{ in which tact and theory are cntengled lfithin the expec-

tations induced by a. paradigm.

In particular, .re shall then .;ant to say

that one important function of a new

With a

theory scientists can see

viously sCen in another 'my.

is the cre&tion of new facta.

phenomena tPJat they had pre-

Only in the textbooks written after recogni-

'cion haa occurred are fact and theory once again firmly separated.
'1'0 see hOI. closely factual end theoretical novelty are intert.rined in
scientific discovery examine a pru.-ticulerly :tamous erJample 1 the disco"ery
of

At

three different reen have a J.eg:l:timate claim to it, and

Ilcvel'al other che::!liat.s

in the early 1870' s, have had enriched air in

..toPl'e)fiU'e .C1r.rok,tively purG

. of'. the
ever, ignorahis

;me the .Swedish 5,pothcc:;,ry, C. tl, Scheele.

We !mY,

it .mSllot. published until ozygen' c di,scovery

had,repeatedly...been,announcedelsevhere...an,d .;thus .. had . no....eft.ect . . ullon .the

:. historical. pattern that most concGI'nl3 us her.e...The second. in time to ea.. tablisha. claim '!raG. the British Bcient1.st.. d. .di-vine, Joseph

wo colleted the gas released. by hes.ted.reO;.oxide o:f'!llercury e.s ena item

in aprolonged>nOrmaJ.invest:!.ge.tion' of. the "e.irs" evolved by a la:r.ge. number

of solid subatanees. In lTI4 he identified theg&s .thus. llrcdueed as nitrous oxide and in 17'75. led by f'urthertests, as camnonair w.!.th less than
i'1;6

usue.l quantity of phlogiston. . The chiI'd Cla.:i.ma.nt, . Lavoisier , started

theworkWich led him to'OXygen after Priestley's experiments of 17'74 and

pcssibly"asthe'result of 'e.hintfrom: Priestley.

Early in lTI5 wvoiflier

reported that the gas obtainedbY'heatingtheredoxide of mercury ws "ail'

itself euttio ,lii;houtaJ.terai;ion [e:l':cept the;t] it ccmes out more pure>

l!!orereap1rable."

By lTI7,probably ;i1iththee.saiatlmce

'fr()l!l l'r1es-t;ley, Lavoisier had concluded", thai;

one

of the

ot a second hint

,,;as a d:i.stincts]fscies;

.a.tmcrlphere .. a ccnclurJion that

Prieatley. ..:was . .I!.eVel .able.i;o . . e.ccept"

Thisr.attern ot:d:!.scoveryraiaes a,.queBtion .'hich, at

in its

seconiiand more fund.ame.:ltal:f'orm. can be asked about every no....el phenomenon

that has e ....er entered the consciousness, of scientists.
Lavois:!.er. ifeii;her.

dlo

first'ctiscoveredoxygcn'l

Was it. Priestley or

In either

C3IlS--a

form of

i;he q'<lsationthat could. be asked.. even if only the man proclaimed discovGrcr

0.0 SO

precisely becau.el there ie no

2.!!S'4'C:':-

oi'

th0

tind

io sought.

been contested since the 1780' a--is e. eYlllptcm of lloL:ethlng aske1'; in the inlllge

of science that
ezample.

cl,iecovery

GO

fun&l.mente.l a role.

Priestley's claim to the discovery of

Look once moroat our

is bused

01'1ty in iaolating a gas that 1ms later recognized as a

his prispecies.

But Priestley's sample ,raa not pure. and. > i:t' hcl<Uug impure

in

000' s

hands is to d..tscover it, that b(ltd been done by cw:ryone who ever lJo'GUer1 a"t-

moephel'ic air.

Bes:!.des) if ?.cieat1ey was the discoverer, when was the dis-

co"ery i!'.de'i'

In 1774 he thought he bad obtained nitrous oxide, a species he

eJ.ready

in 1775 he eaw the gas as (lephlogisticated n:!x, ..."hich ie stHl

not Ol:Jlgen nor even,

fOT

phlogistic chemists I a quite ulle:.-pacted sort of go.s.

e claim !!lay 00 strQ<l-{cr, but it prec.:zuts the

ref1..l.ae the paJ.l11 to PricctleyJ l-re C2Jlnot

p7:cbler:,s. I f

S9.!ilO

,Ie

i:ti to L3vo:'i.sier fer the

of

IT{5 lfh:!.ch led him to iif:!ntify the gas as the uair itself elltiro." P!.-(,)!l1.\!iKlbly
uo waU

fOT

the work of. 1if6 and 1Tn "Hhich led IE-'IO:!. sicr to

i.n 1777 e.nd to the end of his life Is.Yoicier iU;:!ioted

Hr..::inciple"

m:dtGd.. ir'lth calor:tc, t.he

of reat.

S: ;(!'Yi;

merel y

O:'Jrgen ,:aG

Shall

'tIe

therefore

like the discovery of oxygen.

Though ,llldoubtedlycorrectthe

>

"OXygen was disco"ered, II mislenda by 5U(Ogest:l.ng that discove:d.!lg SCJ!lletl.d.ug

isa"singleosimpleoactoasslml1a.bletooouroutrual... (and alllo . . questionable)

Il0

readily assume. that disco\>-ering,

like fleeing or touching; should be unequivocally at""..ributiible to an i!i.di-

vidualandtoa moment

'.n time . :But the latter attribution is aJ:IIS.Ys imIgnoring Scheele. we .can sai'ely

possible and the fOl'lllel' often is as ..well.

say that oxygen .had not been discovered.before lT74 and we would probably
also say that it. had been d1scoveredby l7T7or shortly thereai'ter.
wUhin

But

limits or others lilte. them sny attem;pt to date the. discovln'Y

mustinevit.ably be arbitrary. Furthermore . itl!lust be arbitrary Just because diacoverillS "a newsOl'tof phenO!lJenonis necessarily a com,pJ.ex event

.-

which involves recogniz1ng. both that someth:l.ngis.and what it is.

Note,

. for example ,that:tfoxygenwereac'!;W:illy,dePhlogistlca'csd. air 1 we shou.ld

hsd discovered it thoughw-e

. still notknawquitewhen.
fllct Wldassimllation

to

w9u1d "

BItt i f bothobserva'cion and conceptualization,

theory, are inseparably linked in theproceao of

conceptual ca.'.;egoriesarepreparedinll.dvance, in lihich cane the phenomenon

"would not be ofa new sort, can discovering that and discovering voot occur
eff()rtlesSly, together, and in an instant.
BItt, granting that discovery involves an ehtended, thouel not neccssar-

ily long, procesa of conceptual assimilation, can

say that it involves a

his

paJ?""l'S

froo lIT{ on *,s

lJ.Qt 0

much the discovery of

as the

OlWi},Ci:Z!.

g,en theory of canbuation" and. th.a.t theory 'V;-as the keystone for: a
Uon or chemistry so vast

it is usually called the Chemical Revolu'Gion.

Indeed, if' the discovery ot rurygen llsd

been

= intim:,.te part of the

emergence of e new paradigm for chemifJtry, the question of priOl1ty from

which

began would never have s"omeil. so iPl)?ol't:m:t.

In this ct'.se

fiG

in

o'chera the value :placed u);lon a new pilenomenon and th'(l.s upon its a.iocovoloer

varies 'IIith our estimate of the extent to which the phcmcmenon yicla't\:1d
p.'U'sdigm,:tndtlced &nt1cipations.

Hoticl;l, however, since it will be

J.a.tel", that the d1acov0::-y of oJ::yj!;en

not by itself the ce:llse of the change

Long before he ;Played any ps;rt in the discovery of the

in chemical theory.

new gas, Lavoisier

*'3

liaS

convinced both that s01ll!;lthing .ms 'If.t'ong 'With the

theory and 'w.s:t burning bodies l1.baorbed some

phere.

thing

of th<l etmoll

That much he had recorded in a eeF.led note deposited v1'c.h the Secre-

tary of the French Academy in lT72.

much

What the ;Tcrk on oxygen did

*'!)

to give

form aud structure to Lavoisier's earlier sense that some'1m!)

amiss.

It told bim s. thing he

WE

alrelJ.dy

to diacever--

the nature of the ,substance that combustion removes f:-cm the at.!Uo!f,k'ihe!Oe. ,
ThEt advance l:'l.iiSl'eness of <liff:tculties in the

cO!1:lbust:J.on Iu"Uct be

$.

to sse \fhat

phlOZis'i;cll

of

significa.nt purt of "t-That enabled I.e.:VOiDl.C:C "'';0 sr;:a :tn

m-a.ch like Priestley 4 a n gas t:'hs.t

sce there himself.

hf!..d

be04

ul1c.1.blc "co

Conversely1 the fact that a. maja;; psr(migm rCYiaion ,*,,""aO

Sa'l:T

r:!"uct be

pr:tnc::':psl

r;;;iUSo-n

yhy

51

coverieocan come abou.t,these Elxa",plea

to

be

di.:I.'ferent

The:1'irst,X-rays,:L1i! a.
clAss,10"case., of

, a type '\,hiob occurs

quentlytha.n the impersonal

easily to realize.

mCl'e

fre-

of eCien:t:!.f'icrcporting e.1l0l! us

story opens on the dey that the physicist Roentgen

interrupted a nor:malinvestigat:!.onof cathodersys because he had. not:1.ced.

that e. 'barium platinocyanide screenst some distanee from hiezhielded F-.p-

paratus glOtTed when thedisclliirge WEi in proeMS.

ll'urthel:' inverrtigatious--

theyretjuiTedseven hectic;reeksdu.Tillg which Roentgen rarelyle:ft the

laboretory--in::1icated that the cause ot the glO1f came in ctraight lines from
the cathoo.e ray tube,that the radiation 'cast shadows, c(mld !lot be defleeted by a msgnet, and l!lilch elSe besides.

"Roentgenbndconv:l,nced himBel:f''thai;

BsforellllZlcuncing his discovery

",""Ss not due to cathode l7aye

but to an'agentmthat least some sill!ilar'ltyto light.

EVen so brief' an epitome reveals striking resemblances to thed:l.r;covel"

the pll.J;og:l.SCOll parad:i.gm:

Roentgen '0 d1.scovery c,omr.enced 'ldt.h the l'ecogni-

1;1011 that his screan gl(f"ed "hen it' should not..!n both cases the pcrc'=pUon of anomaly , that:!.!'.: > of a :phenOmenon for "'hich t.he parad5.r,m lw.d nat;

prepared. the investigator: plv..yed. an essential role in prepa."t'ing the l-re::

52
aSllimilation.

ought

'Ire

At 'I.-hat. point in Roentgen's invostigl'!:tiollJ for e::liLlplo,

say that

lllld actl121.1y been discovered.?

>

cony c;;:,ss 1

at the first instant, '\!hen all t.hat hl:.d been noted was a clewinG nCl'cen.

At least one othel" investigator had. Been that 6101/ sud, to hie
chagrin, disco;rered nothing at ell.

llor, it is almost es clero:, can the

moment of discovery be pushed back to

vestigation; by which time Roentgen

Il.

'!.'nll

point during the last lo-cek of inexploring the properties of the

:new radiation he had alread,v discovered.

emerged in

We

between November 8th

CI.Ul

say only that X-ra.ys

Decembc1' 28th of 1895.

In a third nrea, hOl.'ever 1 the existence of cignificant r.a:rallels be-'

tween the discoveries of oxygen

ana

of X-rays is far less apparent.

Unlike

the discovery of oxygen, that of X-rays n'tl8 not, at least fer a decade
after the event, implicated in any obvious upheaval in scientific theory.

In wat sense then can the assimilation of that discovery be auid to have

necessitated. :paradigm
strong.

The cae

1'01'

denyir.g euch a cr..c..nge is very

To be sure the paradigms subecribed to by Roentgen e.Jli his coa-

temporaries could not have been used to predict X-rays.

(!<!a.lrnell'" elGc-

troosgnetic theory had not yet been generally accep',;ed, and. the electron

theory of cathode ra:ys hed net even

!l.Il.-1ounced..)

But neither did those

paradigms, at J.east in fIily ob"liC'.ls senee. prohibit the exis'cence of X-rays

as the phlogiston theory bd prohibited laVOisier's interpretation of
Priestley' e gso.

On the contrary, :!.n 1895 acccpt,ed

practice admi'.;'.;ed e.

theory P,J::d

of 1'O::-dl9 of rr..dis.tioll--visible, infrared, end

53
New elements- ;!erestillbeiDgfitted
____ day.

Theil'.

Roent,,"Cll' B

had become a stWldard pro.ioct for normal scienc:e "__________ _

Ult us first be clear that X-rays '\:lere not

BO

received.

Though

Roentgen had previoulllY'eetablished a fine reputation an an "expcriJlw,ntalist.

_his_ .announcement:lnDece1llber, 1895.

W.II.

those most concerned, w1ththe incredulity and.skepticism 'Which we shall

henceforth repeatedly encounter as Signs: of resistance to paradigm char>l)e.
IDrd Kelvin, to c1teonly an especially prominent example, immediately pronounced X-rays the product oi'an elaborate hoax,

Even without benefit of

Freud, one m1ghtguess from theee circumstances that to some scientistll the
ass1milation of X-rays meant the surrender of sanethiDg else.

Furthermore,

only one more step is needed to recognize that what. had to be. given up, or
at least strenuously reexamined. was the current
of

numbel" of paradigm laboratory procedures.

BUd design

If Roentgen's apparatus had

produced X-rays, then a .number . ofother European. experimentaliatS . .lllUst. for

sCIlletime have been prodUCing those rays in-their ow laboratories but 1.-11thout knowing it.

Perhaps those rays,'Illrl.ch me-ht well have. other unackuowl-

edged sourcea too,wre implicated :In behavior previouoly explained 'lfi'Ghout

reference to them.

At thevl'ryle:::.Gt, ecveral.sorts oi'loDg faml.lw o.:ppa-

ratuB often deployed on experiments quite unrelated to X-rays 'lfould in the

future have to be shielded with lead.

Some previously completed scientific

work on normnl paradi&''lll projects might now have to be done over becauoe
earlier scientists bad tailed to recognize and control a relevant variable.
X-rsyu,

clC'".k"lain of

is true, opened up n
sciel:,ce.

B"'..1.t

neY

field end

'GtU!l

adc1.ed to the potential

also, end. th:ts is i"lm! the more important

previously parCldigmatic appUcationc of theor.y tbeu' :t'ight to th&t tit.!.",.

In ShOl-t, if" a pro:adigm incluiies e'...pplicattonll, aEi ! hs.v," alre::Cl.y in-

sisted. that it IllUst, then mere failu.r-e to tuke an

flo,,'t of phe-

nomenon into account may often imply that that sort of phenC:ll<::Doll cannot
exist.

Even when it violates no explicit theoretical preo.ict1on, the

emergence of a nefT sort of phenomenon can necessitate pe.,'s(liem chsnge.

There is even, for example, some evidence that so readily assimilable a
discovery as that of Neptune evoked resistance that may

have been

due to its destructive impact upon previouelyparadigmatic astronomical

tools.

Neptune's existence end approximate position and motion lTere first

predicted from Newtonian theory to account for observed d:!.screpanc1.eB between the predicted and observed motions of Uranus.
of observed anomaly.)

(Note, again, the role

A felT astronomers uere skeptical even about Neptune's

existence, and all but a very fev displayed a remarknble reluctance to

search for the neii' planet even 1/'herr told just linere to look.

Some of tha.t

resistance was very probably due to their recognition, corwci=

O!'

ur.con-

scious, that previously paradigmatic tnbles for the motien of Satu::'n t,nd
perhaps aleo of Mars 'ITouId have t.o be recomputed. if the unp:clcedented :flTediction of Adams end. Le Verrter produced a planet.

Only if.

11',"

U!1licrata.'1il

haw a par::il.igm can, partly by ,1.irect statement a.'1d :partly tlll'oue;h :l.ts applications, restrict the field of phenomena acce!!sible to Bcienti.i'ic invcl'!t:1,gat10n, shall 'We realize hOt-T the discovery of X-rays, for exauwle, couJ.d open
a strange nelr 'World to

60

many scientists.

And if X-rays had not had that

effect, they could sco'::.:"cely he.v: :paxt1cipated

ef:i"ect1 vely in generating

crisis th=:\.t leO. to the tv;er;.tleth-cent.u::y rcYolution :tn ph.yaic;s.

55

..

longs to a class that may be dCBcril;led .all theory-induced.

Initially, the

term may seem paradoxical. because everything said eo flU' muatindicate that
discoverieslihose outcome
of'

...

'Ill"e.perts

. inno

I ll,ve,f'or example,

preViouslyref'erredto the discoveries...of' new chemigal elements during the

second half of the ninet.eenthcentury as proceedingfrOlll,normal sCience in
this way.

But, as we shall see. in the next section, not all theories. are

paradigm theories... Both during the. pre-p&radig1ll periods descr5.bed in Section U,and during the erise8that lead to. lerge-Bcale chaIlges of paradigm,
sCientist.susually develop many quite flpeculative and unartigulated theories ,
which call themselves point. the way to discovery.

Usually. however. that dis-

covery-is not quite the one anticipated by the speculative and tentative
hypothesis.

(Yukaw' s meson theory of nuclear forces provides one

.It. helped scientists to.lloe meson ..

examined

But that

was not found until later.)

. inphot()8l'aphs.theyhad repaated.ly
not the one .. predictedby Yuka'1a.

It

Only as both experiment and tentative theory

are articulated to II. match. does the discovery emerge and the '.;heory .become

a . parad1gm.
Thediecovery ot: the Leyden jar diSi?lays all these features as well as
the others we have observed before.
tor electrical research.

When 1t begen, there w"as no paracligrn

L'lotcad, a number of theories, aU derived 'rom

crelatively.accessible phenomena, were .in competition, and. none of'these

theories seemed to handle the whole variety of electrical phenomena very
;lcll.

Thllt failure is the first of several.e.no:.aliea that provlo.e bacJ;erounc1.

f{1r ""he

of the 1.e:ruell jc;r.

One of

conwsting schools of

electr:\.c1,ne took electricity to be a fluid, aud that cOIlcuption led. s.

n<llDber of men to attempt bottling the flu1d by holding a ,mtor-filled
glass v1al in their handa end touching the wter to a cond,1.'.ct.or (mspended.
from an active electroetat:\.c generator.

chine and to"o1Ching the water (or

iii

On removing tho jar from the tna-

conductor connected. with it.) ,.'1th his

free hand: each 01' these investigators experienced

a severe

shock.

ThODe

first experiments did not, however. provide electriCians with the Leyden
That device emerged more slowly, and it is agll.i.n :l.mpocoible to say

jar.

.just when its discovery vas completed.

The initial. nttempte to store

electrical. fluid worked only because investigators held the vial in their

handa while standing upon the ground.

Electricians had still to learn

that the jar required an outer as veIl as an inner conducting coating and
\

that the fluid is not really stored in the jar at all.

SOUle"llnere in the

course of the investigatlons that showed them this and that introduced them
to several other anomu.lous et'fectll, the device which we call the Leyden jar
emerged.

Furthermore, the e:cper1ments which led to i ts

of them perfoI'med by Franklin,

many

elso the ones which necessitated the

drastic revision of the fluid theory and ?hich thus

the

digm for electricity.

To e. greater or lesser extent (cc=esponding to the ccnt;tnuUlll
the shocking to the tmt1dP'ltod reault) the characteristics

COEmOll

frOD

to the

three el;amplcll ebove s.re charac'i;eristle of all diBcoveriea frO!!! ullich

sorts of llhenomsDlJ. elnerge.

1le"""

Thoee ohsractel'.atico are: the prior awareness

of anO'J:cly, the sradu.u ano. Ilizml teneous Elll!ergence of both observational

and

l"ecosn:l.ticn, s.nd. the ccnnt;;Cr:..;ent change of perao.igl!l categorien

57

process itself' In .
kuO'"rIn

outside

experiment

identify; on short

deseryea to be

_____ .__

end Postman asked experimental

ana.

to

controlled exposure a series of'"play1ngcards.

Hoot

of' thecardsyere ZlOrma],il>utafC\i' 'iTereme,de anOJilalo.\ls ,e.g. , .. 1. red .. fi va

of clubs and. a black seven of dismonda.

Each experimental run ../sa consti-

tuted by the. display. of a single. card to a single' subject in a . series of

gradually. increased exposures.

After. each exposure the subject ws aslted

what ,he had seen,< 8lld therun1l83 terminated on correct ictentii'ication.

On the shortest exposures, of

incorrect, were produced.
varying from subJect

no. identifications, correct or

But; ""ithe. small increase of'expoaure (the amount

subject) identifications vere regularly produced for

all cardS".Por thenormaJ. cerds these identifications wre usually correct,

but the ano:nalousclll'da were un1formly1dent1fied.. lv"ithoutapparent hSllitation. or puz1.lement, as. normal... The red ...five 01'. clubs. might,
identii'iedas a five ofclubs ... of hstl..'"'tsror ofdiamonG.a.

fOl"

cXfi.ll!J?le, be

WithO"llt awareness

of trOUble it .ra.a1mmed1ately fitted to one of the available cOZlce:ptual.

ca.tegoriea for plnying cards.

One vould. not even like to say tOOt the} sub-

jectll had seeu something different from l."ht<.i:. they identified.

With a fur-

ther :!.ncreaae of exposu..--e to the 8nooaloUll cards subjects did begin to heeitate and. to diaplo.y !l'lr-areness of aIlOlllaly.
five of clubs aame
thing

1o/l'Ong

Exposed, for e:l'.ample to the reel.

.say: tlw.t's the five of hearts, but there's some-

with it"-thered seems alli'ully purple.

Further increase of: e}:-

posure resulted in stUl more hesitation and. confufJio!l until finally, and.

58
anomalous cards they would have no fur-cher difficult.iea uith the ot;h"rs.
A few subjects, however, were never able to make this tronsition.

their

awareness of anomaly and. of their inability to reflolve it increas:;d, they

became so confused that they felt forced to v.lthdralf :!.'rOll! the eA'jy.:riment
elltirely.

In the next section we slmll occasionally see scientists behav-

ing this way too.

Eitherae a metaphor or because it reflects the nature 01' the mind,
tlmt psychological experiment provides a wonderfully simple and cogent
schema for the process of scientific discovery.

In SCience, as in the play-

ing card experiment, novelty emerges onlyvith difficulty manifested by resistance-against a background provided by expectation.

Initially, only the

anticipated and. usual is experienced even under circ1llllstances where anomaly

:l:s later to be oboerved.

Further acquaintance, however I does result in

n=eness of something wrong or does relate the effect to somthing that has
gone wrong before.

That awarenees of anomaly opens a pariod in ....hich con-

cep-tual categories are adJusted and the

At that po:!.nt t.he discovery has been completed.

is seen for '\/bat it is.

I have already urged that

that proceas or one very web like it is involved in the eUlergcnce of all
fundamental scientific novelties.
the proceoe,

'Ire

Let me now point out that, recognizing

can at last begin to see 'Why normal SCience, a purouit not

directed to no,clties and tending at first to suppress them, should neverthelesa be quite
In the

130

effect:!. ve ill cauling them to lU'ise.

of any science the first received paradigm is uau-

ally felt t.o account quj.te succes:::fully tor most of the oba0r'J'J.t;ions and

59
equipmentithedevelopment-ofcan-esoteric,-voc:e.bUililry-eru:iskiJoJ.s,-alld-a;cefi.nement. of concepts. tbat-1ncreasinslylessens their resemblance to. their
usual commonsense

Conce1vablythatsort of elaboration could

be accomp11shedby-the.relat:l.'Ielyrandom.fact"collect1ngand.theor1z1ng of

even

euc:hal'levelOImlent 1Gun:Likely,

hao a use that lYill be

In any case,

more tully in t.he

60
surrendered, reaistance. e;um-an"teee that Ilcientill'ts will not be lightly
distracted and that tae anomalies that lead to

will peu-

trate existing knowledge to the core. The very fact that a significant
scientific novelty so often emerges simultl).Ucouely fron several labora-

tories is an index both to the strongly traditional nature of normal

science and to the

with 'fhich that traditional purllUit pre-

pares the way for its !Tan change.

All !;he discovex lee c:onetdm"ed 11. S<:lc: tlonV1rere-cau8e3\li'-or-corr"---.....:u-'tributors to

psra.d1gJJl

change.

Furthermore, the

in 'Which thelle

discoveries were implicated were all destruc:tive all veil ae constructive.

Atter'lille .disc6,ferylladbJen aririiiDlrated,

for atrl.der

reed

l..:ereable

range oti'laturai phenomena erto account :for.

those previously known with greater prec1sion.

to

some of

But that gain vas achieved

only bi declsrlDg scme previous paradigm beliefs or procedures no lOI18er

paradigmatic and, simUltaneously, by replac1ng those components of the previous paradigm vith o1;hers.

Shiits of thiB sort are, I have argued, asso-

ciated vithall dis.coveries achieved throushnOl'lllal science excepting only

the tmsurprising ones that had been antici:pated in all but their details.
Bu'tcdiscoveriesarenottheonlysources.of . theaedestructive-eollst-"'Uctive
paradigm changes.
lar but usually

In this section we shal1besin to consider those simi-

larger shifts that result from the invent10n of new

theories.
Having argued already that in the sciences fllct and theory, a,ieeo,e:::r
and invention, are not ca'tcegorically and pennanently dist.inct, <re

ticipate overlap between this section and the last.

an-

(The impossible SUZ-

gelltion that Prientley first discovered oxygen and. w-voisier then invented
i t haa its attractions.

Oxygen naB encountel'ed as discovery, above; ve

shall meet it again as theory, below.)

theories

In taking up the emergence of new

shall inevitably extend our understanding of discovery as well.

Still, ave:;.-In]? is not id:ttity.

The aOl'to of d:i.l':covcl'ies considered i.,'). the

61

aa the Cqpernican, Nevtonian. Chemical, and

Revolutions.

\/ere they responsible for the flome"ilat smaller,

Ner

more excluoively

professional, changes in paradigm prod.uced by the ",-ave theory 01' light,

the dynamical theory of heat. or

electromagnetic theory.

can theories like these arise from normal science. an activity

Hew

even less directe to their pursuit than to that of discoveries?

If allsreness of anomaly plays a role in the emergence of all new

sarts of phenomena, it should surprise no one that a similar but more profound awareness is prerequisite to all changes in theory.
historical evidence is, I think, entirely unequivocal.

On this point

The state of

Ptolemaic astronomy "'as a scandal before Copernicus' announcement.

Gali-

leo's contributions to the study of motion depended closely upon difficul-

ties discovered in Ariatotle's theory by scholastic critics.

new

theory of ligbt and color originated in the discovery that none of the
ej:isting pre-paradigm theories would account for the length of the spectru.m, and the wave theory that replaced Nevtoll' s ",-as announced in the ;:lidst
of gt'OIdng concern about anomalies in the relation of diffraction and :polarization effects to

theory.

Thermodyn!u:!1cs was born from the col-

lision of two existing nineteenth-centuryphya1cal theories; and quantum

mechanics from a

of difficulties surrounding black-boay radiation.

specific heats, and the photoelectric effect. Furthermore, in all these

cases except that of' Nel.<ton the aimreneGS of anomaly had lasi;ed so long
and

so deep that one can apl?l"oprilltely describe the :f'ielda ai'by it as :I.n a a'.;o.te of gI'm:ing cri,r.:is.

Gcc:l0

dcctruct:tn::l

mr.jo:c

Be::auoe it demands lerge-

in the

________

1nnecurity is generated by the Fer3iotenttailure

nOl'llll!l science. to come out as they should.

thepuzzlca of

Failure of exist:i.::lg rules is

the prelude to. a. search for new ones.

emergence of Copernice.nl1stronomy.

When its

Ptolemaic

system, was tirBtde\'Clopeddur:Lng the .lasttwo centuries before Christ

I1nd. the first two

it lms ad!n:l.rably succesef'ul inpredict1ng the

changing posit10naof' both stars and planets . !lo other.S!l..cient eyste:m

had performed. so .well; for the stare Ptolemaic astl'onomy i8. still widely
. .:uned,today ..

dictions were as good aoCopex'nicus'.

...

But to be admirably successful is

never, for a scientifictheory,to be completely successful.

"

"""

,,_n<_'M

, _ , . . . . . . . " ' " ' ' ' '

__ ,

n,,,_,,, ''''',.,'''' ________ ,, __ ''' __ ""' __ ''" .,,"'_""'" .,,'"'''_''''' ____ ._""'''''' ",_ .. '''''' "" ____ '"

.pre-

_____ "" """

,_ .,. "" "."_,,

__ "'" _" .,,""

Withrespect

both to lllaneta:ry position and to precession of the eqm,noxes predictions

Ptolemy's 'system never qu1teconformed with the beat .a;r&i1able

observations.

Further reduction otthose minor discrepano::iesconctituteci

many oftbe principal problems of normal astronomical reeearch for

of

Ptolemy's succesoors JUDt as a similaJ:" attempttobrinS celestial observe."

t1oD.andNewtonian theorytogetherpro.vided normal reeearch problema
Newton's e1ghteen'.;h-century successors.

fOl

For some time astronomers had

every reason to. Bup:p0lle that 1;hese attempts WOUld. be as successful as those
which had led to Ptolemy's system.

Given a particular diecrepancy: aatron-

omara wereinv-n;riubly able to eliminate it by me.ltillg come particular adjustmant 1nPtolemy' s system of comp;:rv.nded circl()ll.

B-..:t as time mmt on

0. mru1

then its accuracy and tha.t a diGcl"elle.ncy ca:!"Z'ccted. in

to

plo.ce

lil':cly

up in another.

Because the astronomical tradition

lIUC

intel'r1T.),lted from

outside and becauoe, in the absence of pI'illting, cOlllllllmication between

tronomers ws restricted, these difi'icultiea were ollly alcnrly rec052lized.

But awareness did came.

By the thirteenth century Alfonso X could prcdaim

tllet, i f God had consulted him when cl'eating the universe, He would hllve l"e-,
ceived good advice.

as

In the siXteenth celltury Copernicus' C0'I1Orker, Domenico

Novara, held that no system so cumbersome and inaccurate ao the

had become could possibly ba tnm of uature.

in his Preface to the

And. Copernicus h:l.mself ,;.rote

that the astronomical

inherited had finally created only a. monster.

he

By the early sixteenth century

en increa.sing number of Europc' fl best astrono;ners

recogn:tzir.g tns.t the

aat!'onomical paradigm tras failing in application to ita m:n i;rllditionai. problems,

That l'ecognition was prerequ.1.e:l.t.e to Copernicus' rejection ot the

Ptolemaic paradigm and Me search for a new one,

His famous preface still

one of the classic dcacr1ptions of a crisis

only ingl'edient of the astronomical crisis that faced Cope?nicu8.

tended treatment would also

dil3CUflS

An

the social pressure for calendn.r ref:czm,

." pressure which l1l1'..a.e the puzzle of precession particulE.rJ.y u:q:;ent..

clition,

D.

In ad-

fuller e.cco:4"1t ,,"(Jul!l COIloider medieval c!'iticism of /;risi;otlc, the

rise of Henaias.nce Ncopl&tonl.mn, e.v.d other s:tell1i'icant hillto!'l.cnl

beside(:,

Blxt technical brcwiOlm '\f:luJ.d still remain th" core of the

s:t'ea -in .mich, because e;:l.ven pm:ticul!i.r

curs.

thebl'e&l;.io!1n first oc-

Though 1mmeneely important, issues of that sort are out-of-bounds

for this monograph. -. Particularly,tlleyare out-or-bounds- in, this section

whOseob,1ect ill Only to show. that ,a recQ'gtdzable techni.cal breakdown :1nthe

practice

or

normal science precedes the emergence ot a nell theory.

If' that, much is cltlar in the case

or

the Copernican Revolution, let us

tlirnfrom it toa seCOm and rather different eJci.mple', the crisis which preceded the emergence of IavoisieX" s oxygen theory of combustion.

In the

18'70' s many factors canbined to generate acriais inchemiatry, and hi.Btor-

18.llB are, not altogether agreed about either the:l.rnature or their relative
imPortance.

But two ot them are generally accepted as of first-rate signif-

:!.ce.nce:the rise ofpIieUmiitic chemistry ana'" the ques1;iOl1 of weightrele.tions.

"

.,

., __ ",.' _____ '"

'"

__ "". ____ '''''_,

""'"

-,,,-,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,-,,,,,,,,,,,,,,,,,,,-,,,,,,.,,1,,,,,,,,,,,,,,,,,,

The history of the first beginG in the seventeenth century with development

of the air pump and 1ts deployment inchemice.1 e:'lle:r1lnentetion.

During the

follOl.'1ng century, us:J.nt that putnpalld 8: nUll!ber of other pneutatil:: devicee,

chemists came increasingly to real:tzethat air must be an active ingredient
in cilel!>icl).]. reactions.

But,

'With

a few excepiionsBo equivoCEll the.t they

mo,y not be e):ceptions a.t 1).].1, chsmiflts continued to believe that a1.1' ,!U.s the

only sort of gas.

until 1756, when Joeeph Ble.ckshcwed tont tuea sir (C02)

>Jas conSistently distingu:l.eb.able from normoJ. air ,. t'.o I3elllplee of ear. m.;re
thought to be distinct only in their impurit.ies.
lli'ter Ble.cl;: B work the investiga.tion of eaaBe )?l'oceeded r&pldly, most
notab:ty in the hende of Os-vents.ish, Fr5.:?:stley, and. S-:ueele, \o,.Tho tcgether de-

66
the phlogiston theory li'm often el!Jployed. 11; ill their design a.!ld iril:el']?l"etation of

Scheele actually til"at produced. oJ,"Ygen 'by an elabo-

rate chain of experiments deigned to dephlogbticate h,"at"

Yet the net

result of their exper1m':=nts lias a ....ariety of. gas samples a.nd gas properties
so elaborate that the phlogiaton theory proved increasingly little able to
cope with laboratory experience.

Though none of these chemists suggested

that the theory should be replaced, they were unable to apply it consistently.

By the time Lavoisier began his experimento on airs in the early

1770's, there were almost as many versions of the phlogiston theory as

there were pneumatic chemists.

Thnt proliferation of veraione of a theory

is a very usual symptom of crisis.

In his Preface Copernicus complained of

it as well.
The increasing

and decreasing utility of the phlogiston

theory for pneUlUatic chemistry ""ere not, hm.-ever. the only source of the
criGin that confronted Lavoisier.

He i<"Ba also much cOllcerne,d to explain

the gain in '!>'eight ths.t moat bodies experience when burned or roasted, and
that e.gain 1e a problem tlith e. long prehiBtory.

At leaat a few Islamic

chemis'.. s had lmOlm that some met&ls gain wight when ro"llted.
teenth century several investigators had concluded. :from this

a roasteo. metal takes up some

:from the ntJnosphere.

In the seven-

fact that

f)1l.mc

But

seventeenth century that concluSion seer::-ed UIllleCGOSory to mO:3t cue:nir:r'cs.

It chemical l'cacticns could alt.er the volume, color, and. toxture of the :tngredients, why sl:lO'.11d. th(,y not E'.1ter Height as we1l1 He:l.;ht
talten to be the"

of q'J,antity of

VT"-S

no'c ahT.Ys

Eceidea,. -YJ'2:1.g:l:rt-gain on

BpOlll:lee. to the J.1!:9blem or weight-ge.in beceme inCl'easingly dii'f'icul t

tain.

main-

Partly because the bals71ce _e increasingly ueed as a stalld.ord chemi-

cal tool and partly because the developmcntofpneums:ticchemiatxymade it

.to retain. the. gaseous pt:od.Ucts ot reactions. chern-

...

lets discovered more and more cases in which Wight-gaj.n accOJn1laIlied roactlng.S:!.multaneously,tb.e gradual aso1mllatloIi of Newton's gravitational
theory led chemists tOinsiat tilnte;ain in wa1ght must mean gain in quantity
of matter. Those conclusions did not reeultinreJecticn of thephlcg:f.ston
theory, for that theory cOUld be adjusted in many ways.

l-'ert.aptl. phlogiston

imdnegativewes.ght,or llerhalls fire particles or something else entered the .

roosted body as phlosiston left it.

There ,rere other

beeio.ee.

did IiOt lead to rejection, it did lead

to an increaa:!.ng numooroi'special etudieain-wilich thiS problem bulked
large.

One of them, "On phlogiston conSidered as lLsubatance uith weight

9.1ld.[SrialYzed)in termsoi' theveightcllE'.ngcs it produces in bodies with

which it uIiites," was read to the French Academy early in 1772, the year at
whcaeclose Lavoisier deposited his famous eealednotewith the Academy's
. Secretary.

Before that note

lr.;l,9

written a problem that had been a:tthe edge

of the chemist's conecioUsnessfor many years had become an outstanding lxnsolved puzzle.

Many different versions of the phlogiston theory "ere being

els.borated to meet it.

theory "'tre-n.

Like the problemlJ of pneumatic chemist!".! tho ..e of

'I'hough still

Cll0. truoted. a.s

Co

vOXk-t"lg

e. j,)CtJ:'ud.igra

66
"the competing schools of the

period,

of crisis.

Consider now, as a third and final

the late

century crisis in physics that :preparea the way for the emerGence of
tiv1ty theory.

One root of that

teenth century when a number of natural

can be traced to the late seven

moat notably Leibnitz,

criticized Neaten's retention of an up-dated version of the classic conception of absolute space.

They wre very nearly, 'though never quite, able to

show that absolute positions and absolute motions were vlthout any function
at all in Nctltcm' s system, and they did succeed in at least hinting at the
considerable aesthetic appeal which a fUlly relativisitic conception of space and motion would later come to display.- But their critique ;;as purely
logical.

Like the early Copernicaos who criticized Aristctle's proofs of

the earth's stability, they did not dream that tranition to a relativisitic
system could be directly justified by observation.

At no :point did they

relate their ...'ie-.rs to any problems that arose when applying Neutonitm theory
to nature.

As a result, their views died'TUh them during tbe early decades

of the eighteenth century to be resurrected

nineteenth "hen they had

11

in the last decades of the

very different relation to the :practice of l?bo1cs.

The technical problems to .bieh a relativistic ph:i.lozo1?hy of space waD

ultlmately to be related began to enter normal science

the

of the ;re,ve theory of light ef-i;er about 1815 though they evoked no crisis
until the 1890's.

I t light 10 ..-ave !notion prop&Sated in a

aether

69

.accuracy 'tiO';pl'cw1iiti-relevant informaUon.'andthec'detectionofsether-dxifi;

_ _ _ _ _

..__._..........

research.

Much special equipment was built to reaolve it. That equipment.

however ,detected noobservabledrl1't, and the problem wBtheretore trans. f'erred from the elcper1mental1stst;ndobseryers to thetheoret1c1ans who had
no great difficulty with It. .During the central decades at the century
Fresnel, Stokes, and otbersdevised n\linerousertlculi1.tlons ot theaether
theory deaigned to expla1l1thefalluretoobllerve drift.

Each of' these ar-

ticulatiOns assumed that amov11l8 body draga somefract10n of theaether

mthlt, and each "as sufficiently successful 'to explain not only the negatiVCl..re.su1ts of celest1a,lobserw.tion but also those of terrestrial experimentation includ1l1g the fl1lllousexper1ment of Michelllon fUld Morley. There

waS

still no confl1ct exceptinSthlitbet'ifeenthe various srticulauons. and,

in the absence of relevant experimental techniques, that conflict never be-

cMleacute.
The situation clui.ilged asain only l."iththe g:.oadualacceptsnce of' Maxwell's electrOllll!gnetic theory of light lnthe last t'ifOdecades of the nine. teenth century.

Though Maxwell hillleelfbel1eved in the aether, he made no

use of its mechan1cal properties in hiefiiIeJ. theory.

In particular. hie

discussion of theelectranagnetic behavior ofbod1es 1n motion made no

reference to any sether's l>e1!l8 dragged with them..

BIlt, in the absence of:

aether drag, a 'Iofuole series of earlier obser\'8.',ione to detect motion through

the e.ether beCm!1e anomalous.
long aeries of brilliant

The years after 1890 therefore ""itDessed a

ar,Cl theoretical attempts both to de-

70
results eguivccaJ..

The latter proO.uced a nWdler 01' prO!:liSing cta:cts, par-

those of Lorenz and Fitzgerald,

they also disclosed still

other puzzles and finally resulted in Juct that proliferation of competing

theories 1Ih1ch we bave previously found to be the conc01ll1te.nt of crisis.
It is against tbat historical settina tbat Einstein's special theory of
relativity emerged in

1905.

These three examples are almost entirely t;ypical.

theory emerged only after

pronounced fs.llure in the normal problem-solving

activity. Furthermore, except in the case of

ternal to science played

In each case a novel

where factors ex-

particularly large role, that breakdol!ll and the

proliferation of theories Which is its sign occurred within a decade or at

most two of the new theory's enunciation. The novel theory seems a direct
response to crisis.

Rote also though th1s may not be quite so typicaJ., that

the problems w-lth respect to 1Ihich breakdown occurred wre all of a type
tbat bad long been recognized.

Previous practice of normal science hQd

given every reason to consider them solved or all but solved, which helps to
explain why the sense of failure, when it came, could be quite so acute.
Failure with a new sort of problem is often disappointing but never surpris

ing.

these

problems nor puzzles often yield to the first attack.

Finally,

share another characteristic which may help to make the. case

for the role of crisis impressive.

The solution to each of them had been at

least partially antiCipated during a period when there vas no crisis in the
correoponeling science. and, 5.n the absence of criois, thODe ll.!lticipationfl roel
been ignored.
The only cOnT.Qlete cntiicip&tion iD c,lso the nort feJIJot!.!J) tha.t of Co-pern:t-

71

troIlOmy IIl1ght have bego.lnits' developillent' eighteen centuries earlier than

it did.

But that is to ignore all. histOrical conte:&t.

'rlhenAristurchus'

suggestion'l/asmade ,., the..,yastlYcmore.. reasonable,geocentric ,system, had. no

needs that a hel,i()Centr1c system might evenc:ollceivably have :L'ul.filled.
The whole development of. Ptolemaic astronomy, bOth its triumphs and ita

breakd6wn,'falls in thecentur1es'after'Aristarchus'proposal.
there

were no obvious

BeSides,

reasons for taking'.l\ristarchus seriously.

E'ren Co-

pernicus' moreelal;lorate proposal was neither Simpler nor more accurate

than PtoJ.elllyssystem.Ave1lable observational tests,jas we shall see
more clearlybelow,providedno',basis for a cho1ee.betw-een them. . Under
those c1l'cUDlstances onsot the factors' the.t'led astronomers to Copernicus
(and

one

that ciSUldnct"have"ledthemto'Ar1stllrchus) was the recognized

crisisthat'lll'idbeen'respousiblefor innovation in the tirs't place.

Ptole-

maic astronomy he.d failed. tosolve.itsproblemlJ;.:tne .time had come to give

.acomp:e't.itor;a chance.
anticipations.

our

other tlro examples provide no similarly full

nut eilrely one reason, why the theories of combustion-by-

Il.tmosplleric-e.bsorpt:1ondevefopeo.ln theseVeIlteenth century by Rey , Hooke,

and Mayow fe:tled to get

a suf'f1c1ent

heerl.ngllaS that they m!<,cle no cont!l.ct

recbgnizedtroubJ.espotin normal scientific

neglect by

and nineteenth-century scientists 01'

And the long

rcla-

tivisticcriticli must largely have been due, aSlzaS arll;ued above, to a sitlilar :failure in confrontat1on .
Philosophers of science bave repeatedly delllonctrated. that more than
one theoreticnlconstruct:!.on canalHayabe pl!lccd upon a Biven collection
ci' <le-ta..

History of science indicates tha:t, p:E:t'tiCW.Ul"ly in. t.he early

72
develoJ,)lllental stages of a new perad.igm, i t ie not eyen supremely ... ifficult
to invent BUch alternates.

that invention of

is just

scientists do not and probably ought not undert!!ke except (i.uz'ins the prc

stage of their science's development and at very npecial occasions

during its subsequent evolution.

While the tools a paradigm supplies con-

tinue to prove capable of solving the problems it defines, acience moves

fastest and penetrates most deeply through coni'ident employment of thoHe
tools.

As in manufacture so in science. retooling is

reserved far

occao1on that demands i't.

Illl

to be

The significance of crises is

the indication they provide that au occasion for retooling has arrived.

emergence of I),ovel theories and ask mmt hew scientists respond to thei;:
existence.

Part ot the auswer l as obvious us it. is importent, can be rcnot:l.ngfirst

what

scient:1.sta ne ...erdoHhen collfronted. by even

eevereand prolongedanon:-e.l1es.

Though they may begin toloee faith und

,
then to consider alternatives, they do not pronounce
that
has led them into crisis invalid.

In part that generalization is simply

a statement from historic fact, baaed upon

and, more extens:!.Yely, below.

like those given above

These hint ,rhat our la.ter examination 01'

:pa.radigm;;;rejectionvilld1sc!osemore:f'Ully:onceithas achieved thestatus of paradiem, a scien'cific theory is declared invalid only if an ulteI'-

by the historiccd study of scientific development

all resembles the

methodolog5.cal Dtereotype of falsification by direct compeJ.'ison u:tth natuxe.

That remark does not mean that scientists do not reject scientific

theories, nor that experience and e:<periment are not essential to t.he process in which they do so.

But it docs mean--what iYill

bE! a

tX'al point--that the act of judgment \'Thich lead.s scientists to reject a

paradizm '.;heory is alw-ays baaed upon more than a cCh."Iparison of that theory
"ith the world.

The decision

'1;0

reject one pnradigro if! ullmyo nirr.l),ltan:;-

ously the deci.sion to accept ano'.;her J and the judgment leading to that; decis2.0n

the

of'

parL4dignw lrl th

other ..

73

nile!.

c<?,ch

developing it my argument ,rill itself begin occauionally to

this monog-!'aph' s main theses.
purely factual.; they

The reasons :ror do;;.bt sketch"", aboye

that is, mere cQunterina'.;al,ces to

epistemological theory.

D.

prevllle)"!t

As such; if my present pOint is correct; ';;hey

can at best help to creat a crisi.s or, mOl"e accurately, to reinfoZ'ce one
that is already very much in existence.

By themselves they cannot 1:U10-

.Till not falsify that theory, for its defenders Will do> what He have al-

ready seen scientists doing llhen confronted. With criSiS.

vise numerous articulations and

They u:I.J.l de.,

of their theory in

order to eliminate any apparent coni'l1ct.

In fact, many of the relevant

modifications and qualifications are alread.y in the literature.

J,

fore, my counter',nstances, are to constitute more than a minOI' '.rritant,

that nill be because they help to :permit the emer8ence of a ne'.; theory
within which they nre no longer a sourco of trouble.

Furthermore, if a

typical pattern uhich ue shall later observe in scientific revolutions is

applicable here, they ,,111 then no longer seem to be oimply facts.

Prom

..'ithin the theOl"y they may instead. seem to be something very much like
tautologies, representable by statements '!;hich could not have been othCl'-

It haa often been observed, for example, that 1'Io"ton's Second 1m.; of
motion, though it took centuries of difficult factual and theoretical rcsearch to achieve, behaves for those committed to l!cW".;on B thecr;r 'lory
ouch like a purely logical stllter.;(mt that no amount of obzel'vation could

In

o;

,,'

0"'"
;...",

... -

IX vIe shall

.. .- ., .
q

.....

r :::::::
_.

__

._

:1,,',,-,'
_

:,.' '_'

_._

that the chemic;;..l lB.1-' of

.-_

__

,',

1"">'"'-'"'.''' ", '.",',

','" -,'.,.',',,:":.
-

--

pro-

v :.'
..

75

gestthat :f'rom the vie'iPoint to.;arCls science beiI1-8 developed in this monograph, scientists tail to reject paradigms l;hen :faced. H:l.th coimterinstances
simplyllece.use they could not do so and stili remain scientists.
Though history 10

of eX!l.!llples, some men have un-

doubte:ily 'beendxiven to desert science because of their inability toaupport crisis.

Like art:l.sts, creative scientists must occasionally be able

to live in a world 'out ofjo1nt'.

I have elselrheredescribed '.;hat !"..ecessity

as "the essential tension" implicit in 'much scientific research, and a psy'chologistwhoinvestigates artistic creativity has ,since""taken, Q"fJ'cr the
term.

But that rejection of' science m,taver of another,occtlpation is, I

by them-

Delves ca.'l lead. 'Once e. firstperadigm through uhichto view natu=e has been
1ouna.:;there iEi'I'li!riPlynoSUchthirtgaareseO.Tchuithout a para.dl.gm.

To re-

ject one paradigm Hithout simultaneously substituting another is therefore

to reject science itself.

lna.'1..

NrJ.O. that act reflects not on the paradigm

en.

Irtevi.tablyhem.ll be seen by hia colleagues as, "the carpenter ,':10

blames his tools."

The same :point csn be tlade at least equally effectively in revel'se:
there is no auchthing as research

counterinataI'..ces,.
.

For what is

it that di:f'ferent:l.a.tcs nonta1 science !rom science in a cr:i.sio state?

sv.rely, that the former confronts no' couuterinstanceEl.

On the

Hot,

to do that, e .3., geometric optics, have ehorUy ceased trJ yield research
problems at all and have instead become toolo for engineering.

F.:Xce:;::>t:i.ng

those that are exclusively instrmnental, every problem that normal science
sees as a )?Uzzle can be seen, from another viewpOint, as a counterinstance
and thus as a source of crisis.

Copernicul; sa1f as counter instances ,That

mest of Ptolemy's other successors had seen as puzzles in the

tween observation and theory.

Iavoisier sa" as a

be

what

Priestley had seen as a successfully solved puzzle in the articulation of

the phlogiston theory.

And Einstein ea,! as countcrinotances .mat Lorentz,

Fitzgerald, and others had -seen as puzzles in the articulation of Newton's

and MaJ(well' s theories.

Furthel-nlOre, eYen the_existence 0:>: crisis does

not by itself -transform a puzzle into a counterinstance.

sharp dividing line.

There is no ouch

Instead, by proliferating versions of the paradigm,

crisis loosens the rules of normal puzzle-solving in ways that ultimately

permit a ne,; paradigm to emerge.

There are, I think, only t"10 v.lterna.tiv-es:

,
either no ccientif'ic theory ever confronts a counterinstance, or t.U ouch
theories cor.front counterinsta.nces at all times.
How can i-G ever l>.e.ve seemed otheruise2
lead to thp

That quention tr-.lct

and critical elucidation of philocophy, ano. those:

topics are here barred.

But 'ie can at lea.:;t note t,l'O reaGons 1;hy science

haG seemed to provide so apt an illustration 0:>: the eeneralization that

truth and falsity are uniquely a.'l.d unequivocally determined by the eonfron'.;ut:1.on of

"lith :fact.

Normal science docs and must contimmlly

.. i vc to b:::-ing theory (4'1(1 fact into clOSer und closer agrec:aent, end. that

e.ct:tv:U::y curl c7,;::ily be

77
a puzzle for v.hose very existence the vaU<l1tyof thoparadiQU ,".lot be a:;

e to achieve a
the theory.

ec:'.cnt;J,s-G u:1I1

Here even more than above the proverb applieo:

carpenter 'who blames his tools. II

"It is a 1'00:-

In addition. them=cr

Ucicilce

pedaScgyente!rigleifctiscusslon ofa"theory wltlircmrtrlts ont ts"exemple.ry applications bas hel:petltorew.forcee. cOnfii'mation-'thcorydra""Ilpred.ominantly

from other

SMCCO;;

Given the slightest reason for dciing so. the man ,rho

rea.d.s a sCience text can easily take the applications to be the evidence
1'01'

the theory,thereasonswhy ito1.lghtto be believed.

But science stu-

dents accept theories on the authorit.yof teacher and te:f:t I not because of

Whatalt.ernattves havethey,orlThat COlllJ.letence'l- Theaplllica.

tiona given in to)..-1;s are not there as evIdence butbecaullc learning them is
llartof-rearningthe::paredi.gmat.the.. base . . ofcurl'ent...:practice......!f..applies..,
tions .rereeducedasevidencEl, then the very failure of te)(ts to suggest alternative interpretations or todisc;u.Ss problEims:;f'orlrhich sc:ientists have
t'ailed to produce plIl'f.ul.:!.gmsolutions would convictthe:trauthors ot' eX'';l'eme
bias.

There is not the slightest reason for such an :tniictr,;e:lt.

Row then: to ret1.U'tl to the in! tial question, do scientis'ts rcsp::m.o. to

the a,m:rcnescofan anomaly in the fit between theory ll..'ldncture?

has

just been sa.:i,d indicates that even a discrepa.ilcy u,lflccountably larger t.':lan
that experienced in other ap:pl:tcatiol1s of the theory need not dra..? any very
llrofouild reopcnse.

There are always some discrepanc'.es.

stubborn ones usually rert}?ond at laet to normal 'Il1'D.ctice.

Even the most

Vel'y often seien-

....

78
of that observed.

As Europe's best mathell'.atica.l physicists continued to

westle unsuccessfully ...'1th tho \Joll-knoul2 discrepancy, there "Wero occasional proposals tor a modification of Nouton's inverse squaxo laW'.
no one took these proposals

VOl'Y

But

seriously ano. l in practice, this patience

major anomaly proved justified.

Cls.:traut in 1750 \.'l11l r;ble to ShOll

. that only the mathematiCS of the application had been

tonian theory could stand as before.

and that New-

Even in eases 1{hore no mere lllistcl{e

seems quite possible (perhaps because the mathematics involved is simpler

or of a familiar and elsewhere succesaful sort). persistent and recognized.
anomaly does not always induce crisis.

No one seriously questioned Ne\T-

tonian theory because ot the long-recognized discrepanCies between prediCtions from that theory and both the speed of sOl.'m and the motion of Mercury.

The first discrepancy _'l1S ultimately and quite une:l'pectedly resolved

by experiments on heat undertaken for a very different purposej the second

vanished with the general theory of relativity after a crisis which it had
had. no role in creating.

Apparently neither had seemed eufficienUy funda-

meu;;a.l to evoke the malaise that goes ,lith crisis.

They could be recog-

nized. as counterinstancea and stin be set aside for later ;rork.

It fo11o'l1'B that if .an allOm!lly is to evoke crisis i t
more 'Ghan just an anomaly.

\lsu<;lly be

There are aluayo difficulties aor.:e;{hel'e in the

paradigm-nature fit; most of them are set right soonel or la.ter, often by
processes that could not have been foreseenj the scientist ,mo pauses to
e:t:"llline every anomaly he notes \/ill seldom get any s1g11if'.cant 17ori, done.
He have to as}: ,;hat it io that mokec an anomaly seem llorth concerteD. scrut:i.ny, an. to tho;:;

thc:r;; is proba.bly no fully genc:::-cl

G.!lB1f0r.

The

79

alizat10ne 01' the paradigm as thepl"oblem of

..

diet for th.ose

lfltO

accepted Maxvrell' Btheory . Or. aI3 in the Copernj,can Rovolu'.;ioll, an allO:mly

.. vi

apparent . . f'undamental.,importmay.evoke .crisis, .ifthe.applications

. 1nhibits ha'V'e

for

calendar des1!l'1-. and astrology.

development of .. normal

or,. as

in eiit1lteollth.. century chemistry1 the

may transform an anomaly tiliat hrulprev.l.cuflly

been only a .vexation into a source 01' crisis: .the. problem of weight relntions had avery different status a:f'ter the evolution of pneumo:l;ic chemical
techniques . Presumably there. are still other circUID.stances that can make an
anomalyparticularly . pressing. and ordinarily .sevel'a! of 1;hese w111 combine.
We have alz:eady noted, for example. that one source of the. crisis that confronted Copern1cuswas the.Dlere length of time during ,,'bich astronomere had
wrestled llnsuccesstully.uith the reduction(jf the residual diGcrepouciea in
Ptolemy'. s syst",m.
Hhen; .forthese. reasons or.others lil;:ethem,. an anomaly comes to seem
more than jus'.; another puzzle
of normal sCience, the . transition t.o crisis
.
,

-'

and to e;ctraore.d1nary science .ha.s begun.

Thq anoml.lyitself

11I0regeIlerallyrecognizedas GUchbythe PI'o:r",e.sion.

DOlt COines

More... and more

is devoted to it by more end more of. the :field I S moat eainent men.

to be

If it

still continues to resiEt, as .it usually does not, Illany of them may cOr:!e to
view its l-esolution as -.....
the subject metter of their discipline.
.
fj.eld

ferent

no longer look qu1tethe !lame as it hadeal'1ier.

Ii'or then: tho

Part of its dif-

results simply fror.l the new fixation point of ncient:ific

80
tnade availa.ble.

The ea.rly a:'taclto upon the reo:i.c.t2.nt pre-blen ,,:";'11 have

tollol-red the paradigm rtllea quite clonely.

with contint:ircg rocista.l1ce

more .and more of the attacks upon it unl have inyolved lome minor or not
DO minor articulation of the para:iigm. no tlTO of them quit\! alike, each
partially BuccesB1'ul. but none sufficiently so to be accepted as paradiGP.l
by the group.

ThrQugh this prolii'cI'ution of divergent srticulat;ions (they

'1,'111 come to be iocree.eingly dcscrib",d aa

adJustments). the rules

of normal science become increasingly bl1.1l"I'cd.

Though '.:here still ia u

pal'e.dlBID. fev practitioners prove to be en'.::h'ely agreed abO'.1t

Even formerly paradigm

it is,

of solved probler:lo are calleel in quest:ton.

Copernicus complained that in his day e.stronomers vere so "j,ncollsicterl't; in

these [astronomical] investigations "that theY' cannot eYen e';pls.in
observe ';;he COllstant length of the seaconal year. rr

"Kith them," he

01'

tinued, "it is all though an artist lIere to gcther the hs.nO.s, feet, heau. and
other members for his images from divol'se mcx1.els, each part cxcelJ.cn>.;l,;
eh-e:IlIl, but not related to a SiiJ.i!;lc body, uno. since they in no ,my [Latch
each other, the result "Would be monstel' rather than man."

1'0-

by current uoage to less florid language, l!rote only, "It

co

if the ground had been pulled fl'om wder one, with no firm fOU11O..s.ticn to
be seen any'(;uere on \7hich one could have built."
Such explicit recognit1.ol1a of
effec,.l.;.o of crisis do uot en-"irely

'\f'1ut call ,\;'I"G say tha.t these efi'cctG c,l:e?

m'e

upon ito conscious

rc.:::c" bUG the

....

Only ttro of them Gccm to 'bG un:t-

81
.

and'lllol'e'cleaHyde;;"

fined.incrisis.

And all crises close ,.;itt the emergence of a roSY'clll'ldi- . ___._._

date for paradigm and with the subsequent battle over ito acceptance.
Thesearesubjectstobe considered in later sections. but

"tTe

1I!Uatantici-

pate.. a .. bit,of...:wbat ..l1111 .there...h.e. Baid.in:.order toccnnplete .theacrems.rl;:s .

about the evolution and anatomy of the crisis state.
The transition from a paradigmllhich has evoke"- crisis to a ne,' one
from which

ti.

nel; tradition of normal' science can emerge iafar from a

cumulative process. one achieved by an articulation or extension of the

old paradigm.

Rather it is a reconstruction of the field from new t'unda-

men:ta.ls. a recolllltruction which cho.nges some of the field I a moat elementary

theoretical generalizatiollll as well as many of its paradigm methods and applications.

.lill bea laree but never

complete everlapbetween the problems l1hicncanbe solved by the old and

by the mmparadigm.BIl:1; there

mod.es of solution.

all10be a decisive dii'ferencein. the

When the traneition 1scomplete, the profession ,rill

have changed its view of the field,its methods,and i'.;s goals.

One l)cr-

ceptiveh:!.storian, vievin\ a classic case of a science's reorientation by

l?1J,radigmchange, recently described it as "picking up the

end

s'.;ick, tf a process which involves "he.mUng theaame bundle of

0:('

as be-

fore, but placing them in a nl1m system of relations lr.i:th one another by
giving them a dil'i'erent frer,ie.forl<;."

Others who have noted this aspect of

scientific advance have et:11'hasized ito similarity to

g'"stnl t.

tl

change in

Tho mw,ka on llSper which were first aeen as e. 1;'ird arc

seen

82

\le have already examined some of the prchlatls

Priestley saw oxygen

dephlogist:l.cated a:tr.

by nc.ying

In

does not preserve the gestalt subject IS f'reedo:. to

between ways of seeing.

the sCientist

bacl: and forth

Nevertheless, the IlWitch of 6(ostalt; pru:ticl.llarly

because it is today so familiar, is a useful

occurs in full-scale paradigm shift.

Until. Section

for what

rt we can go no

1'u:r-

ther.
The precei'.ing antiCipation rne.y help us see crisis as I':'.n appro]?!':l.ste
prelude to the emergence of new theories, particu.l.arly since we have already eXlllllined a smaller scale version of the Ilame process iu dj.l3cullsing
the emergence of discoveries.

Just becaulle the emel'e'mce of a new theory

breaks with one tradition of scientific pl'o,ctice and introduces a new oue
conducted under different rules and within a different universe of

it is only: likely to occur ,then the first tradition is felt to haye gone

badly astray.

That remark is, hem'ever, no

Incre

than a prelua.e

vestigation of the cl.'inis-state, and, \.!Ilfor',;une.te1.y,

',;0

the in-

questions to ,,111c11

it leaiJ.B demand the co.mpetence of the psychologist even mOY'e than that of
the historian.

lm;-like?

"mat is

research like!

IIOVi is e.norco.ly made

How do Gcientiste proceed "hen anare only that soneth:'.ng has

gO'"e t'undJricntally wrong at a level '.1 til vhien their trc:in:i.:ng he,s not
I

equi!>,Pcd them to deal?

Those qucaticns need fer more

it. ought not. all be historical.

und. J.esn complete than

\,;'hat

l ..hnt

c.l1d

11il1 ncc(-;sso.ril;r be norc ten-

112.5 gone befot'c ..

83

full extent of the crisis 1n pneuma.ticchemistry . Or again, Thomas

Young's .f1rstaccountsof thetlave theory of light appeared at a very

...that ..would .be almost...

unnoticeable. except that, w1th no assistance from Young, it had grOlm

to an international scientific scanaAl within a decade of the time he
first wrote.

In cases like these one censay.only that a m1nor breFlk

dOlmof thepal:adigm and the very:t'trstblurrillgofits. rule 11 for normal

science 'tlere sufficient to induce a new way of looking at the field.
Uhat . intervened between the firet Bense of trouble and the recognition
of an available alternate must.havebeen largely unconscious.

Hith the

techniques apprapriatetothiB monogrephwe cannot even suggest t1hat .may

have transpired.
In other cases, hmrever,for example those of Copernicue,Einstein,
and contemporary nuclear theory, considerable time elapses betlreenthe

i'irstconsciousnessofbreakdown and the emergence of a ne-..r paradigm.

When that occurs, the historian may capture at least a fe., hints of yThat
extl'aordinaryscience is like.

Faced f:ithan admittedly funiiamen'"al

anomaly in theory, the scientist's first effort w1lll often be to isolate

it more precisely and to give it structure.

On the. one hand, though nmi

a:m,..x'e that they emmot be quite right, he will push the rules of normal
science harder than ever to see, in tho area ofdif'ficulty, jUGt whel"e and
ho../ far they can be mcUe to lTork.

On the other hand,

the ",:::ec of

had been ,rilsn dis:played in experiments whoae outco:ne

kno,m' in auva!Jce.

m.lS

to be

And in 'che letter effort, ma;.e than iT.! 14<;( ot)ler pm't

of the post-paradigm development of science, he will J.ook aJ.n:ollt like our

most prevalent image of the scientist.
I

He vill, in the first place, often

seem a man searching at random, trying experimnta ;lust to [lee wtJ.at trill
happen, looking for an effect whose nature he cannot qu1.te gueas.

S1mul-

taneausly, since no experiment can be conceived without &ome sort ot

thet'l'Y, the scientist in crisis '1."111 constantly t):y to genel'ate specule.-

ti'1e theories (e.g., Yuka;.'S's meson) \,h1ch lll!lY, i f successful, 6.:1.oc10se

the road to a new paradigm and which, i f unsuccessful, can. be surrenil.ered

with relative ease.

The noteboolts lm1ch reCord Newton's optical researches CJr. the rema;,?2s
of Dalton's notes on his lTork with Dlixed sases provide classic e:;;am1?len of

ttc

more l.'G.m1.o111 sort of research proclucea by the awareness of e.nomv.ly.

But; probably the best illuetra.tions ofaJ.l come from contcr.rvore.ry research

in field theor:r and on fundamental particles.

In the absence of a crisis

"h:l.ch has !!!!W.e it nececae.ry to see JUDt hOll far

rules of n.ormal science

carl. stretch, would the immenae effor'G rcqu:f.red. to detect the netltrino

leemed ,1U:Jtj.:!.'ied'l

Or, if the rules had not

h!'.ve bc,,:m either llUU..,gost,oj. or

br.oken dmm nt sa"::

l.:md:l.scloGeo. 1')o1nt, Hould the radical hypothesis

h.l-;oVe

or.

parity

Th:ts sort of cx"t;raox'dina1:Y l'eseEXch is cftec t thOl'.gil by no I:l2nnD

'1:0 the extent that normaJ.research l10rk can be conrl.ucted by using the

paradigm aa a model, rules and 8.sIlUlllpticna, need not be mae.e e;,:.plicit.

Section .III we notedtha.t the full eet of rules

In

by philosophical

analysis ' need not even exist .. But.tha.t1s notto .. say,that"the search:for
aesumptions (even for nonexistent ODes) can not be., an effective w.y.to
weaken the grip of a trad1tion upon the mind and to suggest the basis :for
a new ODe.

It is no. accident that the emergence of Newtonian physics in

the seventeenth century and. of' relativity and quantum mechanics in. the
twentieth should have been both preceded and accompanied by fundamental

philosophical analyses of .the contemporary :research

Nor is it

an accident that in both the.se periods the so-called "though'l;

eo critical a .role in the prClgl"ese 0:1' resecn,:ll.

,_ ""'" ,. , , '" ,_'"''

"'''".,,'' ..

_"_""" ,

""'''''''''''''''''''' _"""

, .. m,.

,C

._'_,.''_"" '"

As I

.' __ ,," __ " _.'_." ,

have sho-.m elsewhere, the aualyticalthought ex.per1l!!entat:l.on that bulks so

larGe in the

of' Galileo,Einstein,Bobr, and others a.,e psrfectly

ce.lculated to expose the old paradigm to existing knowledge in

isolate the root of crisis

that

clarity unattainable in the laboratory.

With the deployment, singly or together. of . these extraordinary pI'Ocedures, one otherthill3 may occur.

By concentrating scientific

unon
- a narrow area of trouble end by p.reparing the llcient1fic mind to rccog.

nize e;:pel'imcnUl-l anomalies :for lihat they are, crisis

often
prOliferates
.
.
new discoveries.

We have

ting1l1ahes !.?voisier' s

noted haw. the al?areneasof crisis dlson

from PrIestley's I and.

viaS

not

86
like polarization by reflection, wel'e
centl'ated work in an ares of trO'.lble
discovery,

a l'O(l1lJ:c

makeD

of tho

likely.

that con

(!.1:lJ.us>

the

lfho reeia

just starting work for the Academy's prize oesey on double

was

refraction, a subJect widely

knOhll

to be in 'an unsatisfectory state).

Others, like the light spot at the centor of the shadow of a circular disc,
were predictions from the new hypothesis and ones whose success helped. to
transform it to a paradigm for later work. And still others, like the
colors of scratches and of thick plates, were effects that had often been
seen end occasionally remarked before but which, like Priestley's oxygen,
had been assimilated to well-known effects in ways that prevented their
being Eeen for lrhat they uere.

A similar account could be given of the

multiple discoveries which, from about 18515, "IIere a constant concooi'l;ant

of the emergence of quantum mechanics.

Extraordinary research must have still other

end cf

fecte, but in this area ue have scarcely begun to discover the queotlons
that n"ed
The

'.;0

be asked.

Perhaps, hmrever, no more are needed at this pain',;.

preceding remarks should suffice to chow how crisio simultaneously

lcoaens the stereotypes &nd provides the incremental data necessary for a
fundamenta). paradigm shift.

Sometimes the shape of the nell :;>aradj.gm is

foreshadO'Jed in the structure which elttrao:rdinary resc2.rch ha:3 given

anomaly.

the

Eins'.;ein ;11'0te that before he had any oupstitute for classical

mocilenicG he

see the interrelation bet\reen the knmm anO"JiaHcs of

black-body radiation, the photoelectric effect,

often no such structuz'c

.
u,::.

',;0

c-.1",'..:
. . ,.. .. J (.;-.;....__: .... .

specific heats.

consciously seen in atlvo.nce..

'.'.,
l-'-"
h,

-,-l .

.. --

r. .....
."'.',', -v,,..' .,.,,.w
,..........'"'
,
. , , ' . ,
-..... -'.'

'".l

tte

("
- ..

-t"- .

More
ne-VT

,;..-.. .:..-,:,.:.,-

--WbattlieIiature

_____

ind::t;;;

invents (or find-she lias invented) a nelr liay of givin,,'1 o:rderto data
now all aesembled;.;..musthere remain inscrutable and may be perillilnently
Let 'UsherenoteOIllyone' th1ng about
.tbese--t'Undamental-inventionoof

tl,:

00.

menuho aChieve

new; parad.:!.gm .havs_been eithel' ..veryyoung

(unaer28) or very new to the field whose pe.:radigm they change.

haps that point need not have beenma.deexpl1cit.

And. per-

For. obviously these are

the men who, being little committed by prior practice to the traditional

rules of normal science, are particularly l1kelytossethat those ruies no

longer define a playable game and to conceive another set that can replace
them .'

VIII.

The Nature and Necessity of Scientific Revolutions

These remarks permit us at last to turn to the problems 1fhich provide

this monograph with its title.

WhAt are scientific revolutions, and what

is their function in scientific development?

Much of the an61fSr to these

questions has been anticipated in earlier sections.

In

the

preceding discussion bas indicated that sCientific revolutions are here

taken to be those noncumulative developmental episodes in which an older

parad18Jl1 is replaced in whole or in part by an incompatible new one.

Only

through such epiSodes, I have suggested, do :t'Unds!llental novelties usually

enter the sciences.

There is more to be said,

and an essential

part of it can be introduced by asking one further question.

metaphor, should a change of paradi8Jll be called a revolution?
of the vast and essential differences betMeen political

Why,

even in

In the face

and SCientific de-

velopment, can the metaphor much finds revolutions in both disclose ar.ry
significant parallelism?
I believe that it can and that one part of the parallelism must already be apparent.

Political revolutions are inaugurated by a grmnng

sense, often restricted to a segment of the political community, that existing institutions have ceased adequately to meet the problema posed by
an environment that they have in part created.,

In much 'ehe same

scien-

tific revolutiOns are inaugurated by a growing sense, again often restricted

to a narrow professional subdivision of the scientific COmmunity, that an
existing parad18Jll has ceased to function adequately in tile exploration of
an asooct of nature to llhich that "ps.r&digm itself had pl'evi01.1s2y led the

way.

In booth political and scientific develcllmeut the sense oi: mo.lf"1nction

llhich can lead to crisis ie prerequisite to revolution.

88

Furthermore,

89
though it aamittedly_stra,1ns the metaphor, that

holds not only

far the majar paradigm .. changes, like. those due to COp1lrnicutl e!!d lavoio1er.
but also .far the far smaller ones associated. with the assimilation of

sort ofllhenolilenon.

--------.,noteci at the

end of

X,.rays.

Ii.

new

ScientifiC reVOlutions. as we

Deed seem revollltio-aery only to those lihofle_ __

paradigms are a1.'f'ectea, by them.

Toou'\;siders they .may.like the Balkan

revolutions of the

opmental-process.

Astronomers, for example, could accept X-rays as a mere

twentieth century. seem

addition to knowledge, far their paradigms

parts of the devel-

unat'fected by the existence

of the newradiation . But for men like Kelvin, Crookes, and Roentgen,

. vlioseresearchdealt wlthradiat10n theory or with cathode ray tubes, the

emergence of X-raylinecessar1lY vi()lated one

as i t created another.

!Ilbat 18 whY .

sowitlilng's first

going wrong With normal research.

This genetic aspect of the parallelism between political and scientific development has already been developed quite fully.

Though its 81g-

questionable, its .existence should no longer be open

to. doubt.

The parallelism has, however, a second and more profound e.spect,

. and thes1gn1ficance of the first depends upon it.

folitical revolutions

aim to. change political institutions. in l>"ays that those institutions themselves

Their success. therefore n.ecessitates the partial relin-

quishment of one set Of institutions in favor of another, and in the interim

society is. not fully governec. by i.nstitutions at all.

Initially it is cri-

sis alone that attenuates the role of political institutions as we have already eeenit attenuate.therole at llI3Xadigms.

In incl.'easi1lg llumbe,,'s indi-

viduals become . 1ncreas:!.ngly estra.'lged from political life 8.3ld behave more

90
and more eccentrically within it.

Then, as the crisis deepens, many of

these individuals commit themselves to some concrete proposal for the reconstruction of society in a new institutional framework.

At that pOint

the society is divided into competing camps or parties, one seeking to

defend the old institutional constellation, the others (usually only one)
seeking to institute some new one.

And, once that polarization has oc-

curred, political recourse fails.

Because they dii'fer about the institu-

tional matrix within which political change is to be achieved and evaluated, becauae they acknowledge no supra-institutional framework for the
adJudication of revolutionary difference, the parties to a revolutionary
conflict must finally resort to the techniques of mass persuaSion, often
including force.

Though revolutions have had a vital role in the evolu-

tion of political institutions, that role depends upon their being partially extra-political or extra-institutional events.
The remainder of this monograph aims to demonstrate that the historical study of paradi@D change reveals very similar characteristics in the
evolution of the sciences.

Like the choice between competing political

institutions, that between competing paradigms proves to be a chOice between incompetible modes of communtty life. Because it has that character,
the choice is not and can not be determined merely by the evaluative procedures characteristic of normal science, for these depend in part upon
particular paradigm, and that paradi@D is at issue.

II

When parediSDIs enter,

as they must,into a debate about the choice of a paradigm, their role is

necessarily circular.
paradigm's defense.

Each group uses its mm pa.."'8digm to argue in that

91
The.. result1ng circularity does not',ot course, .render the arguments
,

. ineftectuaJ, or evenlYTong.

The man who premises a paradigm uhen arguing

in its detense csn,nonethelessprovide a clea.r exhibit of what scientific

. practice willbe.l1ketor those.who . adopt the new view of'nature, and that
exhibit can.be :lmmensely persU8stue,oftencompelltngl;y eo.
its.torce, the.statusot the circular 8l'gument, is, only

Yet, whateve'''z---

It cannot, as we shall . . discover in. the pages ,to tollaw, be made logically
rorevenprobabilistically compelling tor those who refuse to step into the
circle.

The prem1ses and values shared.bythetwo parties to a debate

over paradigms are not sutticientlyextensive to make logic the sale arbiter of choice.

As iii polltical revolutions ,so iii paradigm choice--there

is no s'Mn!!erd .. higher.. than ..the ... assent .. ot. the .. relevant,.cOlllllJUllity.

cover haw scientifiC'revolutions are. eftected,..
examine not only the impact of ,nature and

lie..

To dis-

shall therefore have to

ot logic but also the techniques

ot persuasive 8l'gIl1I1entation eUec:t1ve w1,1;hinthequite'spccial groups that

conetitute the community .of scientists .
'To,discover'Wy this;issuc of'parad1gmchoicecsn naverOO settled by
logic alone, we must.shortlY.exaIIIlnethe natureof the differences that
separate the proponents of a. traditi.onal paradigm from their revolutionary
successore.
the next.

That examination iethe

object otthie section and

We have, however,already noted numerous. examples of: such dU-

ferenc:es, and no one .will doubt that history can SUpply many others.

is more likely to be doubted than. their eXistence .. and llhat must therefore
be .cons1dered fir.st., is. tblit . Buch exwrrDle.s pl'O'.rideessential ini'ormation
about the natu-"6 of science.

Granting that parad1gmreject5.on has been an

historic fact, does it illuminate more than human credulity and confusion?

Are there intrinsic reasons why the assimilation of either e new sort of

phenomenon or a new sCientific theory must demand the rejection of an

older paradigJD.?
let us first notice that, i f there are such reasons, they cannot be
merely logical.

In principle, a new phenomenon might emerge tdthout re-

:rJ.ect1Xlg destructively upon any part of past scientific practice.

discover1Dg life on the moon would today be. destructive of existing para_
digms (these tell us things about the moon that seem incompatible with
life's existence there), discover1Dg life in some less well known part of
the galaxy would not.

By the same token, a new theory does not

conflict with any of its predecessors.

to

It might deal exclusively vith

phenomena not previously known, as the quantum theory deals (but, significantly, not exclusively) with subatomic phenomena unknown before the
twentieth century.

Or, again, the new theory might be simply a higher

level theory than those known before, one which linked together a whole
group of lower level theories without substantially chaDging any.

Today,

the theory of energy conservation provides just such links between dynamiCs, chemiStry, electricity,

thermal theory, and so on.

Still

other com.patible relationships betlleen old and new theories can be conceived.

any and all of them might be exemplified by the historic

process through which science has developed.

development would be genuinely cumuJ.ative.

And, if they

scientific

New sorts of phenomena would

simply disclose order in an aspect of nature where none had been seen before.

In the evolution of science new knowledge lrould replace ignorance

rather than replacing knowledge of another and

sort.

93
Ofcoursecscience(Orsome"otherenterprise; perhaps less eff'ect.ive)
.' might have develaped in that. fully cIlmulative manner. !/.any people have
believed. that it did. so, and 1II0st still seem to BUPi,lPSlfthatcuimuation
....

ideal which historical development would display i f only

it had not so(otten' been distoi Wil by ho.man :td1osytlcrasy. There ace imlIortimt rell.sonsi'Or thatbel1ef.... In Sec.tion'IXwe shall discover how
closely,the View. of. science..;as-'cUmulativeis entaDgledwitha dominant
ep:l.stemologythat.takes;kJiowleage'lio beacenatriictionplaced.directly
upon raw sense' data; by

And in Section X we shall examine the

... "sti-ongsUppol't provided to the sSllle histOriographicsclmaby the tech-

......................c.

n:l.ques OfetfectiVe sciencepede.gogy.

Nevertheless , despite the immense

plausibil:ttY'of 'thIlt'[ideal1ma.ge, there" is increasiDg reason ,to wonder

Whethi!ritC:an})CIssibly bean 1IDage

! science .' After

the. pre-paradigm

period theass1:mulationOf all new-theOries and9f. almost all nev sorts

ofphenomerie.;'hBs ti1rnedoutto:deman(Lparad1gm. destruction and. a. consequentcOrifJ.:l.etb'Ertwee:d'competiDgi.schools..of, scientific .thought.

Cumula-

t:1ve;a.cqiilis:l.'Ii:l.cm of'tfrlBb.t:f.cipated'hoveltiesproves to be an.almost nonex. istentexceptiOh'tCf the rule of.. sc1ent:L:f'icdevelC)pl!lent. The man who takes
. historici'act se:riOuslylnust suspect that science. does not tend tOlrord the
:l.dell.l Whicib.6ur1.Ill8Se, of its cUIinllativeness has. suggested Perhaps:l.t is

atlotlierI'lOrtofehter:p:rise.
ti',however'.ii,esistant facts

us that. far, then a second

lOOk. a.tthlil!?;i-OUDd we have' alreadycoVereil may. suggest that cumulative

acquisition'ot noVeltyis not only. rare in fa.c'\; but impro'bsole in pI'i.ncipie.Normal bas:l.c research,which !!!.cUIinIla.tive. owes its succpas to
the ability of scientists regularly to select problems that can be solved'

with conceptual and instrumental techniques close to "!;hose already in existence.

(That is why an excessive concern 'tdth useful problems, regard-

less of their relation to existing knovledge and technique, can so easily

inhibit scientifiC development.) The

is striving to solve a prob-

lem defined by existing knowledge and technique is not, hOlrever, just

looking around.

He

knows lIhat he wants to achieve, and he designs his in-

stuments and directs th1s thoughts accordingly.

Unanticipated novelty,

the new discovery, can emerge only to the extent that his antiCipations

about nature and his instruments prove wrong.

Often the ill\po,ltance of the

resulting discovery will itself be propOZ'tional to the extent and stubbornness of the anomaly which foreshadOlred it.

ObViOUSly, then, there must be

a conflict between the paradigm that disclosed anomaly and the one that
later renders the anOlllaly lavlike. And, since the older paradigm had prev10usly been constitutive of some part of SCience, the cOlllll!1.Ulity concerned
resists the new one while it can.

Paradigm destruction is, to a greater

or lesser extent, characteristic of each of the discoveries examined in

Section V, and I nO'l1 suggest that it had to be.

Those examples did not

confront un with mere h1storical accident. There is no other effective

way in wbich discoverles might be genel.'ated.

The same argument applies even more clearly to the invention of new
theories.

There are, in principle, only three types of phenomena about

which a new theory might be developed. The first consists of phenomem

already well explained by existing parad:l.gms, and thase seldom provide
either motive or point of departure for theory construction.

When they do,

as with the three fsmoun anticipatiOns discuilsed at the end of Sec'cion VI,
the theories that result are not accepted because nature provides no

95

for dis.c:r:l.mine.tion.A.secondclass of phenomena. con81$1;s of those

nature is indicated by existing paradigms but whose details can be understood only through furthertlleory articulation. '!I!Ilese iUoe the phenomena
. upon which scientists do do researchifltichoftlletime,llut that research

aims at the
new

OlIeS.

iri1Iention of

Only when these attempts at articulation fail do scientists

enc:.ounterthe. third type of phenomena, the rec08l1izedenomal1eswose

characteristicfeature is
isting paradigms.

This type alone gives rise to new

to ex

As the

next sectionw111 disclose more fully, paradigms provide all phenomena

except. anomalies mth a theory;;.Cietierillinedplaceu"theaclent1st"s field
of nslon.

. ...

But ifI.. newtheories are called'forth to resolve anomalies in the relation of an existing theory to nature, then thellucceesful new theory
must somewhere permit predictions that aredi'ferent from those derived
from. its predecessor.

That difference could riot occur i f the two wre

logically caupatiblei 'IntheprocessofbeingasaimiJ..il.tedthesecond must

displace the first . Even a theorylikeeneret:! conservation, wich today
seems a logical superstructure that rel.8.tes to 'nature only through iIldependently established theories,did not develop historicallY't>"ithout
digm destruction.

:p!U'D.-

Instead, it emerged frOlll a c:risisone of wose essen-

tial ingredients w.stheincom;patlbllitybetween Newtonian

some recently formulated consequences of the caloric theory of heat.

and

Only

after the caloric theor:rhsd been rejected could energy conservation become part of science. And only after it l!adbeen:part of science fer some
time could it come to seem a theory of a logically higher type, one not in

confl1ct with its predecessors.

It is bard to see how new theories could

arise without these destructive challges in beliefs about ns.t14"'e.

Though

reductionism remains a log1cally permissible view of. the relation between

successive sc1entific theories, 'it is en historical

A century ago it would, I think, have been possible 'to let the ease
tor the necessity of revolutions rest at this pOint.

But today, unfortu-

.nately, that cannot be done because the view of the subject developed
above cannot be maintained :I.i' the most prevalent contemporary interpreta.
tion of the nature and function of scientific theory 1s accepted.

That

interpretation, closely associated with logical pOSitivism, would restrict

the range and meaning of an accepted theory so that it could not possibly
conflict with any later theOl"Y that made predictions about Bome of the
same natural phenomena.

The best known and the strongest case for this

restricted conception of a scientific theory emerges in discussions of the

relation between contemporary Einsteinian dynamics and the older dynamical
equations that descend from Newton' s Principia. From the viewpoint of
this monograph these two theoriee are tundBmentally incom,patible in the
sense illustrated by the relation of Copernican to Ptolemaic aGtronomy;
Einstein's theory can be accepted only
ws wrong.

the recogn1tion that Newton's

TOday this remains a minor1ty view.

We must therefore e;:....,.,'!ne

the most prevalent objections to it.

The sist ot these objections can be developed as follows.
tic dynamics C8Jl1lot have shown

Relativis-

dynamics to be wrong, for Nelf-

tonian dynam1cs is sti:ll used '!.i. th great success by most engineers and, in
selected applications, by me.ny phySicists. Fu..thermore, the propriety of
this use of the older theory can be proved :f'rO!ll the very theo.."'Y that has,

in'otherappl1cations,'replaced it. Einswin'stheory can be used to

shaw that pred1ctions1'l'OlIf Newton'seq1lations will be as good as our meas-

in. au' a;p;pl1cationB'jjat1sfy1ntf'iFsmallnumberof re

strl.ct1veconditions .' For.ex.il.m.Ple.,1fNMoniantheory is to provide a

gooa:

So1.ut1On, '. tberelattve veloCities of tbe "ood:1es consid-

eredlllU8tbe smalli' com,pared l11ththevelocityof'.light.SUbject to this

cO!ld1tionand aifew otherB,Newtonian theory seems.tobe derivable from
,l!lI.Ds1;eintanOf' Whicb.it is therefareaspei::iU'case; .'. (Actually;no one

bas yet produced Bucb.a derivation, atleasttor.E1nstein t e general theory.

'1'heassUiDI>ti'on tha.tone exists seems to be;anact oftaithbased upon the
.veryphilosophice.l. position tha.t .. thederivation is used to support.
exietenceof

The

affectm;yp6int. and I

shall therel'ore assume tha.tone Vill Ult1mately.:beprovided.)

. But,theob'jectioncontiiiuesjno the017 can possibly confl1ct with

oDe of its Special:c:cases., If: Einsteinian science: seems 'to lIBkeNewtonian

w:t'CIll!:,tbatisonlY because" some: Newtonians \rere so "incautious

'affto'cJJHm:thit'Newt(ifilMtheory'ytelded'elltirely precise'resUlts or that

it Was. WJ.id at wry high relative velocities. ,. Since they coUld not' have
. ha4:aut eVidence tor sueb. clB.1ms,theybetrayed the standards of science

. when they mBae them.

Insotar'aB Newtontantheory .was ever a trUly scien-

t1fic theory supported by valid. evid.ence,1tstlll is.

Only

e;ttravagant

elaims 1'ilrthe theory-.i.elaims that were never. properly parts of science--

canhe.ve beenshOil'liby. E1nstein to be. 'i/1'Cli.'Ig. Purged

extravaSances

01'. these merely h1l!llall

.. been challenged and. cannot be.

Some ve;r-fiili'tof this argUment 1squiteSufficient "to make any theory

ever ulled by a significant group 01' eom.petent scientists :!.mmul.1eto attack:.

98
!rhe much-maligned phlogiston theory, for example. gave order to a large

number of physic:aJ. and chemical phenomena.

burned--they were rich in phlogiston--and

It explained why bodies

why

the metals had so

many

more

properties in comon than did their ores. The metals were all compounded
from d1i'f'erent elementary earths combined with phlogiston, and the latter,

common to all metals, produced comon pro:perties.

addition, the phlo-

giston theory accounted for a number of: reactiOns in which acids were
formed by the combustion of aubstances like carbon and sulphur. Also, it
explained the decrease Q:f' volume when combustion occurs in a confined volume of air--the phlogiston released by combustion "spoils" the elasticity
of the air that absorbed it Just as fire "spoils" the elasticity of: a

steel spring.

If these were the only phenomena that the phlOgiston theo-

rists had claimed for their theoryI that theory could never have been
challenged. A similar argument will suffice for any tlleory that has ever
been successfully applied to any l'8lIge of phenomena at all.
But to save theories in this way their range of application must be
restricted to those phenomena and to that precision of observation with
wich the experimental evidence in hand already deals.

Carried just a

step further (and the step can scarcely be avoided once the first is taken)
such a limitation prohibits the scientist from claiming to speak "scientifically" about any phenomenon not already observed.

Even in its present

form the restriction forbids the scientist to rely upon a theory in his
own research whenever that research enters an area or seeks a degree of

precision for which past practice with the theory offers no precedent.
Logically these prOhibitions are unaxceptionable.

But the result of ac-

cepting them would be the end of the research through which SCience may
develo:p further.

99
By now that point too is virtually a tautology.

Without commitment
..-

to a paradigmtherecou.1dbe. no normal science. Furthermore, that cOllllllitment muet -extend to'areaeand to degreellotprecision:l;orwhich . there is
no full precedent.

If it did not, theparadigmcouldprovide.no puzzles

thithad;.notalread;y been solved.

Besides,

that depends upon commitmenttoaparsdigm.

it;

is not only XlO1'IlI!d science

If. eJl:ietingtheory.binds the

scientist ollly with respect. to existing applications, then there can be no

8u1.'Pffses,anomal.1es, or<d:'isea.But these. are just the signposts that
point thewaytoextraord1naryscience.

If positivistic res'l;rictions on

. the raDgeota theory's legit:1mate al'Plicabil1ty.are . taken literally, the

meC::bairl.emtba.tte'ils the sc1ent1f'ic cOlll!lllin1ty wl:la.1: problems may lead to

. fimdamentalchange mustcea.se4otunction.And when that occurs, the comIIIUIlit'y Willinev1tablyreturh. to. something very like its pre-paradigm
state,s. collditiou'in which "all practitioners do science b.ut in which
their gross product "sc::arcelyresembles. science at. all.
woMer that
whii:liZ'fme

Is .it really a:ny

the price of eign1tics.nt scienUficadvance isa commitment

importantj"there is a revealing logical

lacuna in the positiv-

ist's argUment and one that Will reintroduce us immediately. to the nature
of revolut:l.onaryc:ballgEl; . Can Newtonian. dynamics really be derived from

relativistic clynamics'l What would such a derivation look like? Imagine

"a set cif stateinen.ts,
relativity theory.
resenting

spat:l.al

.' !:ar'which'together embody the lawa of

These statements contain:va.riables. andpll.rameters repposition, time, rest mas, etc.:, and from

together

With the a;pparatus 'of logicano.mathemattcs, .is deducible a ;;ihole set of

further sta:tements including some that can be checked by obaer-rotion.

To

.....

100
prove the adequacy of Newtonian

as a special case, we must add

to the El 's additional statements, like (V/C)2

of the parameters and variables.

1, restricting the range

This enlarged set of statements is then

manipulated to yield a new set, NlI N2 ,

lim,

which is identical in

form 'With Newton's laws ot Motion, the law of gravity, etc. Apparently
Newtonian dynamics has been derived from Einsteinian subject to a few limiting conditions.
Yet the derivation is spurious.

Though the Ni's are a special case

of the laws of relativistic mechaniCS, they are not Newton's laws.

Or at

least they are not unless those laws are reinterpreted in a way that would
have been impossible until after Einstein's work.

The variables and param-

eters which in the Einsteinian Et's represented spatial position, time,

mass, ete., still occur in the Ni's, and they there still represent Einsteinian space, time, and mass.

But the physical referents of these Ein-

steinian concepts are by no means identical with those of the Newtonian

concepts that bear the same name.

(Newtonian mass is conserved,; Einstein-

ian is convertible 'With energy,; only at lOW' relative velocities may the
two be measured in the same way, and even then they IlIUst not be conceived
to be the same.)

unless we change the definitions of the variables in the

Ni's, the statements we have derived are not NewtOnian.

Ii' we do change

them, we cannot properly be said to have derived Newton's laws, at least

not in any sense of "derive" nOW' generally recognized.

Is not this need

to change the meaning of established and familiar concepts central to the

revolutionary impact of Einstein's theory1 Though subtler than the changes
from geocentrism to heliocentrism, from phlogiston to

or from cor-

puscles to waves, this transformation of established concepts is no less

101
decislvelydestrllctive of a prevlouely esteblished paradigm.
.. ' cometoeeelt. aSa

Wernay even

revolutionary reorientations in the

sciences .. Juet.becauselt. did ..not invo.1vethe.introduction of new objects

or.concepts,. the transition frOlll Newtonian to Einsteinian mechanics illustrates With part1CUlai' Cia1'ltyt1leeCieD.t1flc revoJ:ut1Oh as a displacemen""t

o.f theconceptuaJ:.'network throughwhich.csc1ent1ste view: the world.

These remarks should suffice to. show wat might, in another philosoph-

:l.c:8l cJ.:l.mate,bI'I.VeheeJita1tenforgranted.. At least for scientists most

between a d1scarded.scientific theory and its

iiucc:essorerereal.:!rhoughanout-o.f-date theo.ry can always be viewed as

'a special case of lts,;up-tc-date.iiuccessor,it must be transformed. for the
.

one.tbat.canbe undertaken" only with

the advantages ofliindelght,the explicit guidance of the more recent theory.Furthermore, evenlfthat.transformatlonvere a legitimate device to.
employ'in;interPretilig the cIder theory, "the result o:f'1tsappl1cat1on
would be a theo.ry so restricted that;lt could onlyree1;ate lIhat was already

mow;

'Becauseof.itseconOlilythat zoestatement would.have utility, but it

could. not Suftlced'or,the.. gul4,ance.Clfresearch.

; let us, therefore,now takeltforgre.ntedthat,the differences betweeneu.ccees1Ve;Parlidigms;ereboth necessary and irreconcilable.
then

say moreaxpl1c1t1ywhattsort.,of,d1fferencesthese

pII1"enttY,ile has ,already been.1llustrated

are?

can

"Ire

The mest ap-

SUccessive paradigms

tell'us differentthil18sabcut. thepopulat1onof. the universe and abeut

that poplllatlen' sbehavior. ;!rhey di:f':f'er" tbat is, about .such questions sa
tb.e'ex:l.stellce of subatomic particles,the.lll!l.:teriaJ.ity ef light, and the
. coIlservation>of heat 'or ct energy.

These are the Dubstantive differences

102

between successive paradigllls, and they require no further illustration.

But paradigms differ in more than substance, for they are directed not
only to nature but also back upon the science that produced them.

They

are the source of the methods, problem-field, and IltandardS of solution

accepted by any mature scientific community at any given

As a re-

sult the reception of a new paradiglll often nece.ssitates a redefinition of

the corresponding science.

Some old problems may be relegs.ted to another

science or declared entirely "unscientific." Others that were previously

nonexistent or trivial maYi with a new paradiglll, become the very archetY,Pes of significant scientific achievement.

And. as the problems change.

so. often. does the standard tbs.t distinguishes a

scientific solu-

tion fram a mere metaphysical speculation, word game. or mathematical

play.

normal scientific tradition that emerges from a scientific rev

olution is not only incompatible but often actually incommensurable with

that which has gone before.
The

of Newton's work upon the normal seventeenth-century tra-

dition of scientifiC practice provides a striking example of these subtler

effects of paradiglll shift.

Before Newton was born the "new science" of

the century had at last succeeded in rejecting Aristotelian end scholastic

explanations conducted

terms of the essences of material bodies.

To

say that a stone, fell because its "nature" drove it toward the center of the
universe had been made to look, what it had not previously been. a mere
tautological word-play.

Henceforth the entire flux of sensorya.ppearances.

including color. taste. and even weiglIl;. were to be explained in terma of

the size. shape. pOSition, -and motion of the elementary corpuscles of base
matter.

The attribution of other qualities to the elementary atoms 'Was a

103

resort to the occult aDd. theref'ore out-of'-bounds f'orscience.

caught the new spirit precisely when he ridiculed the doctor who. explained
opium's ef'f'icacy as a soporif'icby attributiDgtoit a dormativepotency.
During the lasthalf' of' the seventeenth century I!laIIYscientistspref'erred

1'ollnel shapeo:rthe opillmparticles enabledthelll--toO--.lSKOIOO'l;;thD.----

the nerves about which they moved.

In

an

earl1erperiod explanations in terms of occult qualities had

been an integral part of' productive scientif'ic work . Nevertheless, the

seventeenth century' s new commitment to mechanico-corpuscular explaDation
proved immensely frUitful f'or a.number of' sCiences,riddiDg them'of' problelliS that had 'def'ied
replace them.

others to

In dynamics, f'or example, Newton's three Laws of' Motion are

iEu';saprOduct of' novefexper1meiits ih!l.Dofthe attempt to r!id.iiterPret

well-known observations in terms of' the motions e.nd interactions of' primary neutreJ. corpuscles.

Consider just one concrete illustration.

Since

nelltreJ. corpuscles cOUld act on each other only. by contact, the mechanicocorpuscularviev of'ne.turedirected scientific . attention to a.brand new
subject of' study, the alterationof'particulate motiOns by collisions.
Descartes announced the problem aDd provided its f'irst

solution.

Huyghens, Wren, e.ndwaUis carried it still further, partly by experimentingwith colliding pendulum bobs but mosUy by applying previOUSly wellknown" characteristics of motion to the new problem.

And

Newton embedded

. their results in his Laws of Motion. The equal "action" aDd "reaction"
of' the Third Law are the changes in quantity of' motion experienced by the

two parties to a collision., The same c!1ange-of."motion.ll1lPplies the definition of' dynamical force implicit in the Second law.

In this case as in

,.

104
many others during the seventeenth century, the corpuscular paradigm bred

both a new problem and a large part of that problem's solution.

Yet, though much of Newton' s work ;ms directed to problems and embodied standards derived from the mechanico-corpuacular vorld view, the
effect of the paradigm that resulted. from his work was a further and partially destructive change in the problems and standarda legitimate for
science.

Gravity, interpreted as an innate attraction between every pair

of particles of matter, was an occult qua.lity in the same sense as the

scholastic's tendancy-to-1'all had been.

Therefore,wh1le the standards

of cor;puscularism remained in effect, the search for a mecba.1:lical explanation of gravity was one of the most challensing problems for those who accepted the Principia as paradigm.

Newton devoted much attention to it and

so did many 01' his eighteenth-century successors. The only apparent option
was to reject Newton's theory for its failure to explain gravity, and that
alterlUl.tive, too, vas widely adopted.
mately triumphed.

Yet neither of these views ulti-

Unable either to practice science without the PrinCipia

or to make that work conform to the corpuscular standards of the seventeellth century, SCientists gradually accepted the view that gravity was indeed innate. By the mid-eighteenth celltury that interpretation had been
'almost universally accepted, and the result was a genuine reversion (which
is not the same as a retrogression) to a scholastic standard.

Innate at-

tractions and repulsions joined size, shape, pOSition, and motion a.s physically irreducible primary properties of matter.
The resulting change in the standards and problem-field of physical
science was once again consequential.

By the 1740's, for

elec-

tricians could speak of the attractive "virtue" of the electric fluid

105

a century before.

Astheyd1d so. electrical phenomena increasii!gly dis-

played' anorder'd11'ferent"frOlll the 'one''the:lha:diiliawwhenVie'iTed as the

ef1'ectsOf amechail1caJ.effluviumtllatcOuld act onlYbycontact.
'ti1Ctilar.'whetieleetdeal aetion';at-a-cl:tstanee

bEicame

In par_

a subject for stuay

uitSOlrlt r1ght.thephenoinenonwe nO'lfcallcllarg1l:!g by induction could

'be'recOgnized as one Ofi'ts effects.

PreviouSlY, when seen at all, they

lI.!iifbeen attriliUted'to'theleak8gea iIieVItable"'1ifeny electiii:allaboratOiy.Thatnew

vietf ofinlluctive effeets l.'as,u tUrn, the key to

Frank-

lin'saUalysis Of the Leyden .1arantl thus 'to the emergence Of a new and
NMOntan parMigm for electricity.

Nor

were

dyilamics and. Eilectricity

the onlY'scient:tficf1e1.ds affected by the legit1mizat1onofthe search

for forces inhate to matter.

Tilelal'ge bOdy of eighteenth";century liter-

atureon cme'micaJ. affinities and. replacementsEiriesalso der:l.ves from this

supra:"mec:haiiical agpect OtNewt0n1anism.Chemists . who .bel1evedin these
'. duferential attractions between the various chemicaJ.species setup previouilly'jm'imaginelfeX,per1ments and searched'for'nevsorts"of reactions.

oata and the, Chemical.concepts develaped in that process the

later 'work OtIavoisierand"

moreparticularlYiOfIlaltonwould be incom-

prehensible.Changesin the standards

permissible problems,

concepts, and.ex;planations can transform a science.

I

In the next section

sentle in which' they transform the world.

Other examples of these nonsubstantivedifi'erences betn-een successive

parad1gmscan'be retr:l.eved from the history of. ,anyscience in almoet s:tJ.Y
periodofiistlevelOpliient.

other and far briefer illustrations.

let us. be content with just two

Before the Chemical Revolu'.;ion one

106
of the acknowledged taske of chemistry lres to account for the qualities
of chemical substances and for the changes these underwent during chemical
reactioils. With the aid of a small number of elementary "principles"--of
which phlogiston

'liaS

one--the chemist

'liaS

to explain why some substances

are acidic, others metall1ne, combustible, and so forth.

Some success in

this direction had been achieved: we have already noted that phlogiston
explained why the metals were so much alike. and we could have developed
a simila:r argument for the acids.

lavoisier's ref'orm, however, ultimately

did a'WlI nth chemical "principles," and thus ended by deprivilfS chemistry
. of' some actual and much potential explanatory power.

To com.pensate for

this loss the search for an explanation of qualities and. their changes
widely declared unscientific.

During

'!laS

much of the nineteenth century scien-

tists took such explanatiOns to be no part of' their function.

Or again,

Clerk Maxwell shared with other nineteenth-century propo-

nents of the wave theory of light the conviction that light waves must be
propagated through a material aether.
support such waves
raries.

W8

Designing a mechanical medium to

a standard problem for lDBIlY of hia ablest contempo-

H:l.s own theory, however, the electromagnetic theory of light,

gave no account at all of' a medium able to support light waves, and it
clearly made such an account harder to provide than it had seemed before.
Initially Maxwell's theory was widely rejected for those reaSOllB.

But,

like Newton's theory, Maxwell's proved difficult to dispense with, and as

it achieved the status of a pare.d1@!l in the last decades of the nineteenth
century, the community's attitude tovard it changed. li!.axwell's insistence
upon the existence of a mechanical sether looked more and more like lipservice, which it emphatically had.not been, and the attemptn to de6ign

107
suchan aetherial medium were aba1ldoned.Sc1entistsnolouger thought it
uIlscient11'icto speak of an electrical "displacement" mthoutspecifyiug
What was beingd18p18ced. 'Theresult,' again. was a
8JId

"Of

new' sdt

of problems

lnthe event, had much to de with the emergence

Thesecbaracteristic shifts intbe scient:l.fic cOlllllllUlity' s concept:l.on

of itsleg1timate problellisand litandards wouJ.dbeve lessslgnificance to
'thilimonograph' sthesi:S '11'

'one' COUld

frOm some'methodologiCally lower to BOme hiSher type.

.

.'

--

.'

In thatcaee their

effects, too,Would seem cumulative. No wOXiderthat some histor:l.allS have

" "

'

arPa'tbat tbe Mstory of'science 'reco:rds 'Ii 'continuing' increase in the

matUrity and refinement of man'sc:onceptioJl. of ,the ,nature of science Yet
the case for CUIII1ll8.t1ve development of science's problems and. standards :l.s

even hardertomakethaJI. the case for cumuIation,of theories.

The attempt

to eXplain gravity, though:fruitfUlly abandoned bymoste:l.ghteenth-century

sCient:l.sts, w,snet directed to all. intr:l.ns:l.cally,:l.llegit1mateproblem; the
obJec:t1OJ1.sto:l.nnateforces wereneither :iJ:lherently'unsc:l.entific nor mstapllYSiCe.J.<:I.Ji some pejoratiVe sense.
permit a

There are rio extern8.l standards to

i::itthat sort. what occurred wsneither a decline nor a

rais1ilgof standards but s:Lmply e. change demanded by the adoption of a

paradism.

Furthermore,that change has since been reversed

aga:I.n. In

tl1etwent:i.eth centuryEinsteinbesBUcceeded in explaining

and.

could be

gravitational attractions, and that eXplanation has returned science to a

set of canons andprobleIl!B that are;inthispsrticular respect, more like
those of Newton's lli'edecessors than of hie succesS'ora. -Oraga:l.n. the development of quantum mechanics has reversed the methodolcg1calprohibiticn

108

that originated in the Chemical Revolution.

Chemists nO',/" do a.ttempt, and

with great success, to explain the color, state of azgregat10n, and other
qualities of the substances used and produced in their laboratories.
similar reversal may even be underlro.Y in electromagnetic theory.

Space,

in contemporary physics, is not the inert and homogeneous substratum em

ployed in both Newton's and Maxwell's theories,; some of its

properties

are not unlike those once attributed to the aether; we may someday come to
know what an electric displacement is.

B.1 shifting

from the cognitive tQ the normative functions of

paradigms, the preceding examples enlarge our understanding of the ways in

which paradigms give form to the scientific 11fe. Previously lre had principally examined the paradigm's role as a vehicle for scientific theory.
In that role it functions by telling the scientist about the entities lihich

nature does and does not contain and about the 'lIBya in 'ihich those entities
behave.

That information provides the map wose details mat'.1re scientific

research elucidates. And, since nature is too complex and varied to be explored at random, that map is as essential as observation and experiment
to science's continuing development.

Through the theories theyembody,

paradigms prove to be constitutive of the research activity.

They are also,

however, constitutive of science in other respects, and that is nOW' t.lJ.e

pOint.

In particular, our most recent examples show that p&'a.digms provide

scientists not only with a map but also with some of the directions essential for map-making.

In learning a paradigm the scientist acquires theory,

methods, and standa,:us together, usually in en inextricable miXture.

fore, wen paradigms change, there

a."'E!

usually aignific6.J-:;t shifts in the

criteria determining the legitimacy both of problems and of

tions.

There-

solu-

109

.. That observation returns us tothe:,pointfrom.wh1chthis section began,'foritprovidesourf:!.rst explicit indication of why the choice between
competingpatiI!.dlgms regularlyra:l.sesql1estionsthat cannot be .resolved by

the cr:!.teria O:tnormaJ. science . To the: extent,as

..

:!.t is in-

i:1that"tw()scfent1:f'1c schools disagxeea'bout-wha;tis eo problellHlIld---. .

what a solution; they w:i.llinev1tably talltthrough. each other when debating

thereJ.ativemeritsaf tbeir respective paradigms.

In the circular argu-

ments that regularly result eaclrparadigmwillbe'shown: to:eat:l.si'y the cr:!.tEiriathat it d:l.ctatesfor:l.tselfand t()fallshort . of a' fev of those dictated by its opponent.

There are other reasons, tool.for the failure of

logiCe.l contact that coIiSlsttmtlycli8.racitEidZes Paz'8d:l.gmd.ebates.

For ex-

8ni.Ple;:since rio paradigmeversolvesallcthe. problems.:!.t.def:l.nesand since

notw()'parad1gms leave all thesameproblemsunsolved, ..paradigm debates al-

ways :l.rivolve thequeBtion:which problems is it more significant to have

'solved?

L:l.kEitheissueot competing standai'ds, that question can be an-

sweredol1ly in terms ()tcl'iter:l.athat lie outside ofnorma.1scie!lce altogether, arid there a:reothersuch:!.saueS'lles:l.des. Ultimately,in Section XI,
they w.l.ll driva us to conclude that there is no cri tel'ionsorelevant to
p8.rsdiSJII selection

as

persuas:!.venessin the competition for the . allegie.nce

of thEl ecientific community.

sh1ftsas revolutions.

l-esPectin which

That is amain reason for describing parad:l.gm

But before.taltingthatstepwe must examine one more

paradigms Eil'e. necessarily self-justifying.

To this point

I have argued only that paradigms arecons1;itutive of science.

Now I wish

to display a sense in llhichthey a:reconstitutive of nature as well.

IX.

Revolutions as Changes of I10rld View

Examining the record of past rel3earch from the '11'!mte.ge of contemporary

historiography, the historian of science may be tempted to exclaim that

.

'When paradigms change the world itself changes with them.

led by a new

scientists adopt new instruments and look in new places.

Even

more important, during revolutions scientists see new and different things
'When looking with familiar instruments in places they have looked before.
It is rather as if the professional COmmunity had been suddenly transported
to another planet where familiar objects are seen in a different light and
'Where they are joined by many unfamiliar ones as well.

I need not say that

nothing of quite that sort does occur: there is no geographical transplantation; outside the laboratory everyday affairs usually continue as before.
Nevertheless,

changes do cause scientists to see the vorld of

their research-engagement differently.

Insofar as their only recourse to

that 'World is through what they see, we may want to say that, after a revolution, scientists live in a different world.
It is as elementary prototypes for these transformations of the scientist's world that the familiar demonstrations of a .switch in visual gestalt
prove so suggestive.

What were ducks in the scientist's world before the

revolutiou are rabbits afterwards.

The man wo firl3t Baw the exterior of

the box from above later sees its interior from below.

Transformations

like these, though usually more gradual and almost always iI'".!'eversible, are

common concomitants of scientific training.

Looking at a contour map 'che

student sees lines on paper, the cartographer a picture of a terrain.

ing at a bubble-chamber

Look-

the student sees confV,sed and broken

lines, the physicist a record of familiar subnuclei'll' events.

no

Only after a

111
number of such transformations of vision. does . the stUd.entbecome an inhabitant of the. scientist's world, seeing.what the scientist seGS and
responding

as

the scientist. does.

The world which the student then enters

is not , hCJWever I fixed once and for. all by the nature of the environment,
on the one bend, endot sCience,on-the-Gthel'.Rathel'.. i t is determine.tid----.1ointlyby thE! en:v1rolllllent.and the. particular normal scientific .tradition
which the stUdenthes been trained to.pursue.

Therefore, at times of revo-

lution,when the .normal SCientific..tradit10ncbanges,.thescientist' s perception of his envirOlllllent mustbereeducated--in acme familiar situations
.he must learn to see a new gestalt.

ADd. afterhe.he.s.c done so..the world of

-his research Will seem, here and there,incammensurablewiththeone he had

inhabited before..

That iSaDother.reasonwhy schools guidedbyd1fferent

paradigms are alwayssl1gb.tly at

In their most usual form, of course, gestalt experiments illustrate

"

only
the. nature of perceptual transformations.
<,
."

.,

" , , ,

the role of paradigms or of

_."

,_

"..

',0

prev1.
..
OUB.

" " "

They tell us nothing about

"

ly assimilated experience in. the proc-

""

"

--"

'

,,--

ess of .. perception. But onthatpoint.thereis.a.richbcdy of psychological

literature,.much.of it stemming from thepioneer1ng work of the Dartmouth
Eye Institute.

An experimental sub.1ect who puts on goggles fitted .:ith in-

verting prisms initially sees the. entire world upside down.

At .the sta...-t

his percElptual apparatus functions Bsit had been trained to function in

the absence. of the. prisms, and the result. is extreme disorientation, an

acute personal crisis. .But ai'terthesub.1ect has begun to learn .to deal
with his new world,1 his..

visual field "flips over, It sometimes after

an intervening pericdin vhicbvis:!'oD is simply obscure.. Thereafter cb.1ects are again seen right side up as they had .been before the prislII3 were

112
put on.

The assimilation of previously anomalous visual data. has reacted

and changed those data themselves.

In the most literal sense imagin-

able the man accustomed to inverting prisms has undergone a revolutionary

transformation of vision.

The sub3ects of the anomalous playiDg card eJ.."per1n:-ent. discussed in

Section V experienced a quite similar transformation.

Until taught by pro-

longed exposure that the universe contained anomalous cards, they saw only
the types of cards for which previous experience had equipped them.

Yet

once ,experience had provided the requisite additional categories, they were
able to see all anomalous cards on the first inspection long enough to per, mit any identification at all.

still other exper1lllents demonstrate that

the perceived size, color, and so on of experimentally displayed objects

also varies with the subject's previous training and

ing

SUrvey-

the rich exper1lllental literature from which these eT.amples are drawn

makes one suspect that something like a paradigm is prerequisite to perception itself.

What a man sees depends both upon what he looks at and

also upon what his previous visual-conceptual experience has taught him to
see.

In the absence of such training there can only be. in William JameD'

phrese, "a bloomin' buzzin', coDfusion."

In recent years several of those concerned with the history of science

have found the sorts of experiments descl'1bed above immensely suggesti\'e.

N. R. Hanson, in particular, has made brilliant use of gestalt demonstrations to elaborate some of the same aspects of scientifiC theories that
concern me here.

other colleagues bave repeatedly noted that his'Gory of

science would make better and more coherent sense if one

suppose that

scientists occasionally experienced shifts of perception like those described

113
. above.YetFthoughpsychological expei'iments ere suggestive" they C9.lll1ot,
1nthe D8tureofthe case, be more than that. Theydod1splay character:f.stfcs'ofpereeption'thatcould .becentral to. scientific .development, but
they do not demonstrate that the careful and controlleaobservation exer-

cisedb1 _research sc1entfstatallpartake" of those

Furthermore the very D8ture of these exper1ments makes anyd1rect demonstration of that po1nt1mposlJ1ble.

It histOri<:aJ. example is to make these

.psychOlOg1CeJ.exper1Dliffits. seelll relevant'Wemustffrst notice the sorts of

evidence that We

may

and may not expect . history to provide.

Thesub.1ectofa gestalt,delllonstrationlmows that his perception has

shiftedbecal1se.hecan make.:!.t

i3hifi bacIt

and

forth repeatedly while he

holds the'SSIIIC''bOokOr'piece, of'paper1n his"hands.

Aware that nothing 1n

,....... ... ............................

. his'envfroni1lenthas 'cllarlgedmth . his ". perception, he' directs. his attention

1Jlcrea.singly nottothiffigure(duck or rabbit) but to the lines on the
:PaPer he is

at.

Ultimately, hEf may even learn. to see .those lines

w1'thoutseeing either of the figUres;, and he may. then say (what he could
nOtlegitiirilitelY'haweai"d"ea.rlierthat>itts'these'lines which he reallY
sees but that

he

sees them alternately!! as duck and !! a rabbit.

By the

same token the sub.1ect of the anomalous card experiment knows (or, more
accurateiy,c:anbepeisuaded) that his perception must have shifted. because

a. h1gherau'thor:LtYi

theexper1menter,'assures .him that,rega.rdless of what

he saw, be. was looking ata black five of. hearts all the. time.

In both

these cases, as 1n all similar psychological experiments, .the effectiveness

ofthedemonstrat1ondependBupon .. i ts . . being. analyzable in. th:l.S way.

UnJ.eaa

therewere'anexrernalBtandAl'd with reapect1;o vrhich a Il'.ntch of vision

coUld be demonstrated, no conclusion about alternate perceptual possibilities could be drawn.

114
With scientific observation, however, the situation is exactly reversed.

The scientist can have no recourse above or beyond 1tant he sees

with his eyes and instruments.

If there were some higher authority by

recourse to which his vision might be show to have shifted, then that
authority would itsel1' become the source of his data, and the behavior of
his vision would become a source of problems as that of the el>.-perimental
subJect has for the psychologist.

The

same sorts of problems would arise

if the scientist could switch back and forth like the subJect
talt experiments.

The

period during which light

'WaS

o:f

the ges-

"sometimes a wave

and sometimes a particle" was a period of crisis--a period when something

was wrong--and it ended only with the development of wave mechanics and
the realization that light was a self-consistent entity different from
both waves and particles.

In the SCiences, therefore. if perceptual

switches accompany paradigm challges, ve may not

test to these challges directly.

scientists to at-

Looking at the moon, the convert to

Copernica.nism does not say, "I used to see a planet but now see a satellite." That locution would imply a sense in which the PtoleU!Aic system
had once been correct.

Instead a convert to the new astronomy says, "I

once took the moon to be (or, saw the moon as) a planet, but I was mistaken. " That sort of statement does recur in the aftermath of sc:!.entific
revolutions.

If, as I suspect, it ordinarily disguises a shift of scien-

tific vision or some other mental transformation with the same effect, we
may not expect direct testimony to that effect.

Rather ve must look for

indirect and behavioral evidence that the scientist with a new:paradigm

sees differently from the way he had seen before.

115
Let us then return to history e.nd.s,sk whai;sorts oftrensformat1ons
in the scientist's world the man who believes there may be !r.lch things can
. retrievefrom.it.Si1-William

discovery. of Uranusprpvides a

. first example and one. that closely parallels the anomalous card experiment.
On

at J east seventeen d1 f'fCrt!nt occasions between 1690 Ang 1781. a number

of astronomers,

of Europe's most eminent. Observers, had

a star in positions that. we nmr s1lpIlose must have been occupied at

the time by Uranus..

0lIe. of

observers-in .this.. group had actually

seen the star on four successive. nights in 1769 without noting the motion
which would have. suggested another identification.

Herschel, when he first

observed the sameob.1ec1;tweive>years later,Ma. so W11;iia IIl1ich-impro'led

.

te1escopeot his awn manufacture.

/'

,'"

","

"--".,,,

As a result hews able to notice an ap

,-,

parentdiiic:':'iiizetbiitwsatleastunusU8.l.for

'"

aWry,

and he therefore postponed identification pending further scrutiny.

That

scrutiny disclosed Uranus' motion amoDg the stars, and Herschel therefore
announced that he had seen a new comet!

Only several months later; after

.fru1tless.attempts . to fit ..the.observed motiC!Jltoa

cometEiry

Lexell suggest that. the orbit . was probably planetary.

or1l1.t. did

When. that suggestion

was accepted there were s.evera! fever stars and one more planet in the
world of the professional astronomer.

Acelest1al body that had been ob-

served .ofi'-end-on for almost. a century WB seen d11':f'erenUy after 1781 because, like ananomalouB playing card. it could. no longer be fitted to the
perceptU8.l. categories (star.or comet) provided by the pa-"'adigm that had
previously prevailed.
The shift of vision that enabled . 8stronomers to see Uranus, the plaret,
.;

does not, .however, seem to have affected only the perception of that

..

u6
previously observed object.

Its consequences

more far-reaching.

Probably, though the evidence is equivocal, the minor paradigm cbenge

forced

Herschel helped to prepare astronomers for the rapid discovery,

after 1801, of the nlllllerous minor planets or asteroids.

Because of their

BJDall siZe, these did not display the anomalous magnification that had
alerted Herschel.

Nevertheless, astronomers

to fini additional

planets were able, with standard instruments, to identify thirteen of

them in the first fifty years of the nineteenth century

The history of

astronOlllY provides IIlaIlY other examples of J.lIU"8digm-:l.nduced changes in

scientific perception,

SOllIe

of them even less equivocal.

Can it, :for ex-

ample, conceivably be an accident that Western astronomers first saw

change in the previously immutable heavens during the half-century after
Copernicus' new pa.'"8.digm was first proposed?

The Chinese, 'Whose

COSIllO-

logical beliefs did not preclude celestial change, had recorded the appearance of new stars in the heavens at a much earlier date.

Also, even

without the aid ot a telescope, the Chinese had occasionally reported the
appearance of sunspots centuries before these were seen by Galileo.

Nor

were sunspots and a nelf star the only exmnples of celestial chsnge to
emerge in the heavens of Western astronomy immediately

Cor:ernicus.

Using traditional instruments, some as simple as a piece of thread, late

sixteenth century astronomers repeatedly discovered that comets wandered
at will through the space previously reserved for the immutable planets
and stars.

The very ease and rapidity with lIhich astronomers

BaW

new

things lIhen looking at old objects with old instruments may make us ,,"ish
to say that, after Copernicus, astronomers lived in a different lmrld.

any case, their research responded as though that were the case.

In

ll7
The.preced1Dg eX8Jll.Ples are selected froD! astronomy beca.use reports
of celest1aJ. observation are frequently delivered in a vocabulery con-

s1Btingof relatively pure observation terms. "Orilyfromsuch reports can

we hope to retrieve anyth1ng like a f'ull parallelism between the observationsof' se1entiatiHiM those Of thep!l1ellologist' sexper:i:loontBl subJects.
But we need not 1nsiston so :fUll aparallel1sDi, and we have much: to gain
byrelax1Dg

om. standard.

If we can be content with the everyday (as

. againstthephllosophieal).useof the verb "toseei"ws'ma.y quickly recognize that wei1S.ve already encountered . many other' examples of these shifts
mseientific perceptionwh1ch accompany paradigm cbsnge. T'nat extended

Use 'Of""perceptiOll"

a!ld'of'''iieeing'' Will' shortlyi'eqU1re expliCit defense.

but let me first illustrate its usein.practice.

:Wok Bga1n for a moment at two of our. previous examples .tram this
history

the seventeenth century, when their re-

search was guided byene or another effluvium theory, electricians repeatedly saw chaff perticles' rebound from, or fall off of, the,electrified
'bodies that had attracted them. Atlesstthat'iswhatseventeenth-century
observers' said, they saw, and we 'have no.' more reason to doubt their reports
Of percept:!.en than our own.

Placed before the.same apparatull a modern ob-

server would see electrostatic repulsion (rather than

or gravi-

tational rebounding), but historically, with one universally ignored exception; electrostatic

not seen '. at all until Hauksbee' s

J..arge"sealeapparatUShadgreatlymagnified 1tseffects.

Once he had

learn.ed to see the repulsion produced by contact electrification, hm,rever,

. Hauksbee"was rapidly able to recognize ,a number of other repulsive effects
as well.

TheelectricaJ. phenomena visible in,!;he early eighteenth century.

118
vere both subtler and more varied than those seen by observers in the seventeenth century had been.

Or again, after the assimilation of Fre.nklints

paradigm, the electrician looking at a Layden jar saw something different

from what he had seen before.

which neither the

The device had become a condenser. for

nor glass was required.

Instead, the two con-

ducting coatings--one of which had been no part of the original device-emerged to prominence.

As both written discussions end pictorial represen-

tations gradually attest, two metal plates with a nonconductor betwEen

them had become the protot:me for the class.

Simultaneously, other induc-

tive effects received new descriptions end still others were noted for the
first time.
Shifts of this sort are not restricted to astronomy and electricity.
We have already remarked some of the similar transformations of vision
that can be drawn from the history of chemistry.

lavoisier, we said, saw

oxygen 'Where Priestley had seen phlogisticated air and where others had
seen nothing at all.

In learning to see oxygen, however, lavoisier also

had to chsnge his viev of many other more femiliar substances.

He had,

for example. to see a cOlllJ?ound ore where Priestley end his contemporaries
had seen an elementar,r earth, end there were other such changes besides.

At the very least, as a result of discoveriIlg oxygen, lavoisier saw nature

differently.

And. in the absence of some recourse to .thath;i'Pothetical

fiXed nature which he "saw differently, n the principle of' economy will

urge us to say that, after discovering oxygen, laVOisier worked in a different world.
I shall inquire in a moment about the possibility of avoiding this
strange locution, but we first require an additional example of its use,

119
this,. one. deriving. from one of thebeBt

of the 'W"Ork of Galileo.

Since remote. EIlltiqu1ty most people have seen one or another heavy body
Mng1ngbackand .fo..-thon a string or chain UIltil.it: fiD8.lly comes to

rest.'lo the Aristotelians, who believed that a heavy body is moved by

its

OlIn JIat'1l!e

from a higher position to s ststeot natural restst a

lower one, the swinging bOdy wassiDIply falling with difficulty.

Con-

strained by the chain. it could achieve rest at its lowpoint only after
.. a tortuous motion and a considerable t1me. Ge.J.ileoron the other hand,
look:i.Jlg at the sv1 n g1ng bOdy. lava pendulum, a body which almost succeeded
in repeating the same motion over.EIlld over again

!!!!:

infinitum. And, hav-

ing seen that much. Ge.J.ileoobserved other properties of the pendulum as

well and constructed
new

many

around them.

of .the most significant end original parts of his

Froathe properties of the pendulum. for exWnple.

Galileo derived his only. full and sound arguments for the independence of
we:l.glltand rate-of-fall as well as for the relationship between vertical
height and termiilal velocity of motions dovn incl1nedplanes.

All these

natural phenomenahe.8aw" dU'f'erently.from. the. way. they. had been. seen before.
Why did that shift of. vision occur?fhrough Galileo's individual

genius, of course. But note. that genius does not here manifeBt itself in
more accurate or ob.1e.ct:l.ve observation of the mringing body.

Descrip-

tively the P.ristotel1an perceptionis.1ustas accurate. .When Galileo reported that the pendulum'sper:l.odwas1ndependentof amplitude for amplitudes as great as 90, his view of .the pendulum led hiDI to see far more
regularity than we can now discover there.

Bather, what. seems to heaG

been involved was the exploitBtionby.geniuB of perceptual possibilities

made ava1lableby a medieval paradigm

.1.'aS

no-" raised quite

120

an Aristotelian.
terms of the

On the contrary, he

'llS.S

trained to analyze motions in

theory, a late medieval paradigm which held that the

cont1nu1ng motion of a heavy body is due to an internal pOm!r implanted

in it by the projector which initiated its motion.

fourteenth-century scholastic who brought the

Jean Buridan, the

theory to perhaps

its most perfect formulation, is the first man known to have seen in a
mrillg1ng body any part of what GeJ.ileo saw there.

He described that mo-

tion as cme in which increasing 1DJ.petus is 1DrGlanted in the bob by its

gravity during the downswing; that impetus is next consumed in raising
the body against gravity on the upsving; then gravity again carries the
body down, 1DrGlanting increasing impetus until the low point of the mng
is reached; end so on in a process "'hich repeats itself symmetrically again
and again.

Clearly that is very close to the view of the pendulUlli from

which Galileo started. At least in Buride.n's case, and almost certainly

in Galileo's as well, it was a view made possible by the transition from

the original Aristotelian to the scholastic impetus paradigm for motion.

Unt:l.l that scholastic paradigm ws invented, there were no pendulUllis, but
only S'I11nging stones, for the SCientist to see.

Pendulums were brought

into existence by something very like a paradigm-induced gestalt switch.

Do we, however, really need to describe what separates Galileo from

Aristotle, or lavoiSier from Priestley, as a transformation of vision?

Did these men really
of ob3ects?

things when looking at the same sorts

Is there any legitimate sense in which we can say that

pursued their research in different worlds? Those questions can no longer

be postponed, for there is obviously another and far more usual 'loIay to

describe all of the historical examples outlined above.

SUrely most

121
readers will want to say tlle.tl>1hatchailges 1dth a paradigmirtorily the
scientist's interpretation of observations which themselves are fiXed
once:imd for 8.l1 by 'the nature :oftlieeiiv:l;rOxnnent'and'of'the:perceptual
ap;pa.ratlis.On this VieW, Priestley-'aildIavo1sier b6tlisawoxygen. but

they:lDterpretfldiiheu ooseHaiiCnff 4ifferefttlyj1lr:!:stotle Slid 68111eo.----both aawpe;du1uma, but they differed in their interpretations of what
thElybothhad seen .
Iatmeaayat once'thatthis very usUaJ. viewofwhe.t occursl1hen
'scientists change . their minds about 1'undamental ,matterscanbe neither

all wrong nora mere mistake.

Rather it is an essential. part ofa philo-

and
with DesCSrtesin the role played. for dynamics by Newton.

philosophy well.

Thatparad.:l.gm

lteeJqlloitation, like that

of dyJ'lamicsitseJ.f. l:ui.s been fruitful of a fundamental understanding that

perhaps could not .havebeen achieved in another way. But. as: the example
of NeWton18n a.yD.amics aJ.so:l.nd1.cates. even the most striking past success
prov:l.desno guare,uteetliatcris1s canbe indefinitely postponed. and to-

day-research tn partsofphilosopny. psycholosy, linguistics. and even

afth:l.story. e.nconverge to sUggest that the traditional paradigm is
somehow askew.

That failure to fit :l.s Iilso made increasingly apparent by

'. the . historical. stud;\' of science to which I necessarily direct most of my

atteritionhere. :'
Nonear these crisis-promoting subjects has yet produced a Viable
al.ternate tothEitraditional epistemologiCal. paradigm, but they do beg:ln
to sUggest what some of that parad:l.gm'livcharacteristicswi:t1be.

I am,

for example,scutelyawre of the difficulties inherent in saying that,

122

when Aristotle and Galileo looked at swinging stones, the first sail con-

--

strained fall, the second a pendulum.

The same d1i'i'icuJ.i;iea sre pre-

sented in an even more fundamental form by the opening sentences of this

section: though the world does not change with a change of paradigm. the
scientist afterwards ttOrks in a different world.

Nevertheless, I am con-

vinced that we must learn to make sense of statements that at least resemble these.

What occurs during a scientific revolution is not fully

reducible to a reinterpretation of indJ.Vidual and stable data.

In the

first place, the data are' not unequivocally stable. A pendulum is not a
fa]Jing stODe, nor is oxygen dephlogisticated air.

Consequently, the

data which scientists collect from these diverse objects are, as

shortly see, themselves d1ff'erent.

shall

Furthermore, and perhaps more impor-

tant, the revolutionary process by which either the individual or the community makes the transition from constrained fall to the pendulum or from
dephlogistice.ted air to oxygen is not one that at all resembles interpretation.

How could it do so in the absence of fiXed data for the scientist

to interpret? Rather than being

an interpreter,

the scientisl; who em-

braces a new paradigm is like the m&n1itl!U'ing inverting pr:l.sms.

SUddenly

he confronts a world which, though indubitably the same in il;s totality,

has been transformed through-and-through :l.n its details.
None of these remarks is intended to indicate that scientists do not
characterist:l.cally interpret observat:l.ons and data.

On the contrary,

Galileo interpreted observa"tions on the pendulum, Aristo-cle observatiolls

on falling stones, Musschenbroek obBervat:l.ons on a charge-filled bottle,
and Franklin observations on a condenser.

t:l.ons presupposed a pa:t'.adigm.

-.

But each of these int.erpreta-

They were parts of normal SCience, an

123

. enterprise which, .as.we . have already. seen,a:l.ms to 'refine ,cextend, and

articulate a paradigm that is ,already in existence.

Section III pro-

,videdmany elr8IIIples in which interpretation played a central role, and

tht!seexample.s., typify, theoverwhelmiDgmajorityofresearch.
v1l'tueef,.an ae<lepted pa.ra.digmjllmeu

In each of

,met

was, .1dlat...instruments miSht be used to,retrieve it, .andwhat concepts were

,relevant ..to its inter,pretation.

Given a paradigm, interpretation of d.e.ta

',is central to the enter,prise.that' explores. it.

BUt that interpretive enterprise--endthisvasthe burden of the
paragraph before l.a.st--can only.articulate a paradigm, not correct it.
ParidigJIIs arenotcorrig:tble by'normal scien-deat.all.. Instead, as we
, have already. seen, . nm:maJ; .. science...ultimatelyleads. ,.onlY.tothe.. recognition
of, anomalies and to . crises.

And the.se

are terminated .. not by de.liberation

and interpretation, but by a relatively sudden and una.tructUl"ed. event like

the gestalt switch.

Scientists then of1;en speak of the "scales falling

.fran the eyes" or of the ,"l1Shtning flash" that "inundates" a previously

obscurepuszlerenabling ,its components to be seen in ane\i';1Bythat for
the first t1mepermits

luminationcomes'in sleep.

On other occasions. the relevant il-

No ordinary sense of "the. terminterpreta'i;ion

fits these fl.a.shes of intutti.on through which a new paradigm is born.

Thougil such intuitions depend.upon,the experience, both anomalous and congruent,gained with the' old paradigm.. they are not logically or, piecemeal
linked to particular items of that experience as en.interpretation would
be.

Instead, they gather up large porUons of that experience 'cogether

'endtranstorm, it to the rather. different bundle of ex:periencethat .,-ill

thereafter link piecemeal to the new paradigm but not to the old.

124
To learn more about what these di:t':f'erences in eXllerien(!e can be, re-

turn for a moment to Aristotle, Galileo, and the pendulum. 1<lhat data did
the interaction of their different paradigms' and their common environment
make accessible to each of them?

Seeing constra.1ned i'all, the P.risto-

telian would measure (or at least discuss--the Aristotelian seldom measured) the weight of the stone, the vertical height to llhicn it had been
raised, and the time required for it to achieve rest.

Together With the

resistance of the medium, these were the conceptual categories deployed

by Aristotelian science when dealing With a falling body, and normal research guided by them could not possibly have produced the 1..io7ll that C.aUleo discovered.

It could anly--and by another route it did--lead to the

series of crises from which Galileo's view of the swinging stone emerged.

AS,a result of those crises and of other intellectual changes beSides,

Galileo saw the sWinging stone quite differently.

Archimedes'

w"Ol'k

on

floating bodies made the medium nonessential. the impetus theory ;rendered
the motion symmetrical and enduring. and Neoplatonism directed Galileo' s
attention to the motion's circular form.

He therefore measured only

weight, radius, angular displacement, and time per swing, which were precisely the data that could be interpreted to yield Galileo's laws for the
pendulum.

In the event, interpretation proved almost unnecessary.

Galileo's paradisms, pendulum-like regularities were very

ble to inspection.

Given

accessi-

How elee are we to account for Galileo's discovery

that the bob's period is entirely independent of amplitude, a discovery

Yhich the normal science that stems fron Galileo had to eradicate and which
we are quite unable to document today.

Reb'Ularities that could not have

existed for an Aristotelian (and that are, in fact, nOiIhere precisely

125
exemplified by uature}.were _consequences of 1mmediate experience for the
. man who saw theswing1ng stone as Galileo diii.
<Perhaps that example istoofanciful since the Aristotelians recorded no discussions of swinging stones.

On their paradigm it lms an

the simpler case,. stones.falling without. uncommon constraints, and the

same differences of visionshoir there.

Contemplating stalling stone

J\ristotle,Baw a change of state rather thana process.

For him the rele,.

vant parameters.ofa motion.. were. therefore total distance. covered and

.total elapsedt1mewh1ch yield what we should now call not speed .but avere.gespeed.

the stone was.:l.mpelledby its nature to

reachitsfinal. resting pOint, Arist.otlesaw .the relevant.,distance parameter at any instant. during the. motion as the. distal'l.ce to'tllef:iriii.J. end
point rather than as

origin of motion.

parameters.underlie aDd.

sense to most of his

Those conceptual

"laws of mo-

.. tion." .Pal:tly through the. impetus paradigm, however ,and pertly through
a doctrine known as the latitude:.of forms, scholastic cr:l:ticiam cha.."lged
this way ofView1ng motion.

A stone moved by impetus, gained more and

... more of it while receding from its starting point; distance

than

therefore became the relevant parameter.

rather

In addition,

Aristotle's notion of speed was bifUrcated by the scholasti.CB into concepts that soon after Galileo. become our average speed

and.

instantaneous

speed. But when seenthroush_the paradigm of whichthafJeccnceptioIls

were a part, the falling stone, like the pendulum, exhibiteo, its govern"'iog laws almost' ortinspection. . Ga1ileowas not. ons.ot.the first men to

suggest uniform acceleration. Furthermore, he had developed that theorem

126

together With many of its consequences before he experimented

clined plane.

an in-

That theorem 'WaS another one of the network of new regular-

ities accessible to genius in the world determined jointly by nature and

by the paradigms upon which Galileo and his contemporaries had been rained.
Living in that world Galileo could still, when he chose, eJ..1}lain why Aris-

totle had seen what he did.

leo's

Nevertheless, the immediate content of Gali-

falling stones was not what Aristotle's had been.

It is, of course, by no means clear that

"immediate

lie

need be so concerned With

that ie, the perceptual features which a

paradigm so highlights that they surrender their regularities almost upon

inspection.

Those features must obviously change With the scientist's com-

mitments to paradigms, but they are far from llbat we ordinarily have in
mind when

lre

speak of the raw data or the brute experience f::oom which

scientific research is reputed to Illoceed.

should be set aside as fluid, and

lre

Perhaps immediate experience

should discuss instead. the concrete

qperations and measurements that the scientist performs in his laboratory.

Or perhaps the analysis should be carried further still from the immedi-

ately given.

It might, for example, be conducted in terms of aome neutral

sense-datum language designed to conform to the retinsl imprints which

mediate What the scientist sees.

only in one of these 'Ways can 'We hope

to retrieve a realm in which experience is again stable, once and for all-in which the pendulum and constrained fall are not different perceptions
but rather di:i.'f'erent interpretations of the unequivocal data proVided by
observation of a ew1ng:l.ng stone.
But is sensory experience fixed and neutral? Are theories

man-

made interpretations of given data? The epistemological paradigm that has

dom1nated WesternphUosopby for three centurlesdlctateslm 1mmed1ate and

In the absence ofa develapedalternatlvei I find it

I th1Dk ltno lODger func

.. tloils . effectlvely,and thee atteroPtsto makelt.do so through the introducttonOfan opGlat1cmal or of a sense-aatumlanguage now seem t;ome entire"'lny..---hopeless.
!he operatlonsaDdmeasurementstbi1.t

e.

sclentlstUndertakes ln the

J.81'iorator;yarenot"thegiveii" of expenence btit rather "the collected-withdifflculty." !hey are not ybatthe IIcientll1tsees,;;.;;at leallt not before hls
researChlsYell ad:vanced aDd hlsattentlon focused.

Rather they are con-

elementazoyperceptlOUS, aDd as suCh

. they are selected,forthe close scrutlny ofnormsl research only because

. they

of an accepted para-

digm.Far m6reclearly than the immediate experlence from wD.1ch they in

. part derive , . aperatlonsand.D!euurements. are .ps.r8dlgm-determ1ned.

SClence

doss not deal in allposslblelaboratory.msn1pulations . Ins'tead lt selects

those relevalittothe.juxtapositlonofaps.r8dlgm with the1llImediate expedence tbat tbatps.r8digm bas partially determined.

As a result, scientlsts

with .. differentparailieps engage indifferent concrete laboratory manipulations. Themeasuiements to be performed on a pendulum are not the ones
relevant to a . case of .constrained .fall..Nor. are ..the operatious relevant
for the elucidatlon of oxygen' s properties . uniformly the same as those requiredYheliinvestigating thecbi1.racterlstlcs of dephlogistlcated n1r.
As for a sense-datum language, ...perhaps .. one will yej;pe d.evlsed.

But

three centuries after Descartes our hope for .suchan eventuality still depsnds excluslvelyupon.a theory of perception and of the mind.

And modern

128

psychological

is rapidly proliferating phenomena

which

that theory can scarcely deal. The duck-rabbit shows that two men 'ldth the

same retinal impressions can see different things; the

prisms

show that two men with different retinal impressions can see the same thing.
Psychology supplies a great deal of other evidence to the seJne effect and
the doubts that derive from it are readily reinforced by the history of at

to exhibit an actual sense-datum languase. No current attempt to

achieve that end has yet come close to a generally applicable language of
pure percepts. Alld those attempts which come closest share one characteristic that is likelY to render them functionless for the acquisition of
knowledge.

They presuppose a paradigm, taken either trom a current scien-

tific theory or from some traction ot everyday discourse, and they then try
to eliminate
from it all nonlogical and nonperceptual terms.
,

'l'he.t effort

my, of course, ultimately be success1'ul. In a few realms of diacO'.l%'se

Carnap and Goodman have already carried it very far.

But

there is no rea-

son to suppose that a language derived by the purification of anyone paradigm will prove transportable to another.

It ever developed and deployed,

a sense-datum language derived in that way will probably tend to the Game
r1gidification of experience as the paradigm through whose
was derived.

it

It it is to be serviceable for research, a sen3e-datum lan-

guage must be capable of describing experiences which violate current theory as wall as those which conform to it. Philosophical

has

to date failed to provide even a hint of what such a language would be like.
Under those circumstances we may at least suspect

ocientists are

right in principle as well as in practice when they treat oxygen and pendulums (and perhaps also atoms and electrons) as the fundamental

129
of their i1lllDediate experience.

As "11. result of: the paradigm-embodied expe-

dence 01' the race, tile culture, and, tinally, the profession, the world

populated

mthplanets 'andpendll1ums,

and othersuchbocHes besides. Compared With

these ob3ects otper.eept:l:ou, both meter st1ck teadlngs and retinal

':

con-

pr1nts are elaborate constructs to:which.experiElUcehas direct access only

when the sc1entist,torthe: special purposesof his research, 'arranges
tbAt one

or: the

other should do so. . Th"8.t is not to suggest that pendulums ,

say,. are the only th1ngs.ascientist couldposs1bly see .

at a

sw1ng1Dgstone. We have already noted that members of another scientific

cOllllllimity coUld see constrained fall.

But 11; is t() suggesttbAtthe

is in

pr1nciplemoreelementarythan see1ng.a.peildulum.'The alternative is not

some lIypothetical"t1xed" vision but vision through another paradigm, one
whichmakesthesw1.ng1ng stone something .else .
All of 'this may seem .more reasonable it we now reemphasize a characteristicot:perception.thathas most recently been noted only in passing.
Neither scientists rior la;ymen learn to seethe . world piecemeal or item-by

when all the. conceptual and manipulative categories are pre-

pared in advance".e ;g. ,for the discovery of an additional trans-Uranic

element orfortlle sight of anell'.house--both.scientists and laymen sort
out whole areas from the flux 01' experience together.

The chUd who trans-

fers the word "Mama" from all humans, to all females, and then to his
motller is not Just learning what

or .who his mother is.

te.ne()usly he is.learning some of the differences between males

Simul-

and females

as well as something about the ways in which all but one feme.ls ,Till behave

130
toward him.

His reactions, expectations, and beliefs--indeed much of his

perceived world--changes accordingly.

:By the same token, the Copernicana

who denied its traditional title "planet" to the sun were not only learning What "planet" meant or what the sun was.

Instead they were changing

the meaning of "planet" so that it could continue to meke useful. distinctions in-a world where all celestial bodies, not Just the s;m, were seen
differently from the way they had been seen before.
be made about any of our earlier examples.

The

same point could

To see oxygen instead of de-

phlogist1cated air, the condenser instead of the Leyden jar, or the pendulum instead of constrained fall was only one part of' an integrated shift
in the scientis.t's Vision of a great many related chemical, electrical, or
dynamical phenomena. Paradigms determine large areas of experienoe together.
It is, however, only after experience has been thus determined that
the search for an operational definition or a sense-datum language can begin.

The

scientist or philosopher who asIts what measurements or retinal

imprints make the pendulum What it is must already be able to

pendulum when he sees one.

I f he saw constrained fall instead, his ques-

tion could not even be asked.

And i f he saw a pendulum, but saw it in the

same way he saw a tuning fork or an oscillating balance, his question

could not be answered.

Or at least it could not be anm.-ered in the srune

way, because it would not be the same question.

Therefore, though they

are always legitimate and are occasionally extraordinarily

tions about retinal imprints or about the consequences of

queslabo-

ratory manipulations presuppose a lTorld already llerceptuelly and conceptually subdivided in a certain lTay.

In a senee such questions are parts

131
of normal sCience, for they depend upon the eilstence of a paradigm and.
they receive dii'1'erent answers as a result of paradigm change
.To c()%lc1Ude .this seCtiOll let UB henceforth neSlect retin8.J. impl'essl()D8 and agaiJI.., restrlct attentlon' to the laboratory' operations that

prOVide the sc1ent;1st w1th concrete thOugbf:ragmentary1nalces

to what

he has already seen.. One way in which such laboratory operations change
With parad1gms has. already been observed repeatedly "After a sCientific

Decome "1rrelErvant

revolutlQ11 ..

and

are replaced .by others instead.OIIedoesnOtapply all the same teste

.

. to oxygen as .todephlogisticated air.

tot!1.,l..

But c:hSllges of this sort are never

lIhateverhe may then see, thsscient1st 81'tera revolution is

stilliooldngat . the same -world. Furthermore,though he

have

ay previously

them differently, much of his language and most of his lab-

. oratory instruments are still the samess they vere before.

As a result,

pt:!st1'evolutionary sclence invariably includes .'JiIaDy of the same manipulations, performed With the same instruments and. described
\

in the .same

terms,

as Its prerevolutionary predecessor .I1'thoseenduringmanipulations have

.been c:haDged at" all, the Change mustl1ee1ther in their relation to the
paradlgmorin their concrete results.

I now eusgest, by the introduction

"01' one. lastnevexam,ple ,that both these . sorts 01' c::hanges

to occur.

CSZl

be observed

Examining. the work of Dalton and hisccntelllpOra.ries we shall

discover that one and the same operation, when it attaches to nature
through a different paradigm, can become an index to a quite different as-

.pect.of nature's regularity.

In addition,

we

shall see that occasionally

the old manipulation ln its D.ewrole will yield di1'1'erentconcl'ete results.

132
Throughout much of the eighteenth century and into
European chemists almost universally believed that the
of which all chemical

mutual affinity.

nineteentb
atoms

consisted were held together by forces of

ThUB a lump of silver cohered because of the forces of

affinity between silver corpuscles (until after lavoisier these corpuscles were themsleves thought of as compounded from still more elementary
particles).

On the same theory silver dissolved in acid (or salt in

. water) because the particles of acid attracted those of Silver (or the
particles of water, attracted those of salt) more strongly than particles
of these solutes attracted each other.

Or again, copper would dissolve

in the silver solution and precipitate silver, because the copper-acid

affinity was greater than the affinity of acid for silver.

other phenomena were explained in the same way.

A great many

In the eighteenth cen-

tury the theory of elective affinity was an admirable chemical paradigm,

and

deployed in the design and analysis of chemical ex-

perimentation.
Affinity theory, however. drew the line separating

mixtures

from chemical compounds in a way.that has become unfamiliar aince the assimilation of Dalton's work.
two sorts of processes.

Eighteenth-century chemists did recognize

When mix.ins produced heat, light, effervescence

or something else of the sort, chemiCal union was seen to hEtve taken place.
If, on the other hand, the particles in the mixture could be distinguished

by eye or mechanically separated, there \laS only physical mb;ture.

the very large number of intermediate cases--salt in

But in

alloys, glags,

oxygen in the atmosphere, and so on--theee crude criteria were of little

use.

Guided by their paradigm, most chemists viey)"ed this entire'

133
'intermed1aterange'aschemicaJ.';becausethe processes 01' which it conslsted1iereaUgovernedby 1'orceso1'thesSme sort.
,oxygen
'"

Salt in

or

.were.1ustasmuchexampleso:f''chem1cal combination as

. was;thecombinatlonproducedby. OJCidi;1ngcopper. ;ThearguJ!ents 1'or

VifiW1ng .solutions as compounds were very strong. . A1'1'ini ty theory i tsel1'

was 'welrattested., Besld:es, the' 1'ormatiolio1'e.com.PoUnd accOUnted fora

solutton'sobserVea.homagenelty.

If/'rbrexamplei'oxygen and riitrogen

were ol'llymiXed and. notcolllbined' ill theaas;pHe:t;&, .

heavier gas,

oxygen, should settle to the bottom. . Dalton,who took the atmosphere to

be a mixture, liaS never aatis1'actorily able to ' expJ.a1l?i oxygen' sfailure

to do so.

Theass1milatlon 01' hlsatomic'theoryult:l.mately created an

'anomaiy; where.'there'had beennone'beforeC;;"

one

is temPted 'to sayth8t thechemistri who viewed solutions as com-

pOuxlds d11'1'eredtrom1;heir successors

oniy over e. matter of definition.

In one sense that may 'have been the case.

But that sense is not the one

which makes definitions mere conventional conveniences.

In the eighteenth

centtiryin:lxtureswerencitfli1J.y cl1stinS\11shed:1'romcom.Pounds by operattonal tests., and;perhaps they could not have been.

Even if' chemists had

looked for such tests,they wouldhaVesifught Criteria thatlDade the solution a compound.

Themixture-colllp0und distinction vas part of their

p8ra.dism--partot the way they v1ewedtheir wliole field 01' research--and

as such it'waspr:l.ortoanyparticular.laboratory test, though not to the
acculllUlated experience 01' cheinistryasa whole.
But while Chemistry
pUfied laws

liaS

viewed in this way , chemical phenomena exam-

from those which emerged with the assimilation of

Dalton's newpe.radiSin; III particular{while solutions remained compounds,

amount of chemical experimentation could by itself have produced the

law of fixed proportions.

At the end of the eighteenth century it was

widely known that !!.2!!!! compounds ordinarily contained fixed proportions

by weight of their constituents.

For some categories of reactions the

German chemist Richter. bad even noted the further regularities now embraced by the law of chemical equivalents.

But no chemist made use of

these regularities except in recipes, and no one until almost the end of
the century thought of generalizing them.

Given the obvious counter-

instances, like glass or like salt in water. no generalization was possible without an abandonment of affinity theory and a reconceptualization
of the boundaries of the chemist I s domain.

That consequence became

ex-

plicit at the very em of the century in a famous debate between the

French chemists Proust and Berthollet.

The first claimed that all chemi-

cal reactions occurred in fixed proportion, the latter that they did not.

and each collected impressive experimental evidence for his View.

Never-

theless. the two men necessarily talked through each other, and their debate was entirely inconclusive.

Where Berthollet saw a compound that

could vary in proportion. Proust saw only a physical mixture.

To that

issue neither experiment nor a change of definitional convention could be

. relevant.

The two were as fundamentally at cross-purposes as Galileo and

Aristotle bad been.

This was the situation during the years when John Dalton undertook
the investigations that led finally to his famous chemical atomic theory.

But until the very last stages of those investigations, Dalton

a chemist nor interested. in chemistry.

neither

Instead he was a meteorologist in-

vestigating the, for him, physical problems of the absorption of gases by

, 135
vater andot.vaterbythe-atmosphere.

Partlybecauseh1str8.irliDg was in

.a d1t1'erent.
.J:1e...

'!fOr1tin that specialty,

porary
...
,'.,,". -- chelll1sts
;, .:

,,'

th ...a.parad1gm . . different ti-om .iha.i ot COlltem-",','

--

niixture
oisa-ses or the ab-

one iii wbiCh forces

sor;pl;lou of a gas in water as a p1IySicaI .process,

ai't1nity

ot

him, theretore,theobserved homoseneity ot

.solutious. vas a.problem,but. one which he.thOll8ht he coUld solve if' he

cOUld determine the :l;"elative sizes and. weights
cles 1n.hisex,per1mentsl mixtures.

or

the variOUII atomic parti-

It was todeterm1lle theEe sizes and

tirlally . turnedtochemistrY, BUP.PoS1ns tromthe start

that,.1n ..the .restricted range ..otreactious which he took to be chemical,

. atomscoulcl OIllycombirleone-to-OIle. or.in some other "simple'lfhole number

ratio.

That physically. natural ass1llllPtlO1l didenablebimto detel'!ll1ne the

SizeS. and veights of elementary

stan'!;. prpportiOll.a.t!;Lutology.

but it also mdethelaw of con-

ForDalton,anyreactionin'which the ingre-

dien'!;s didllOtenter 1nt1Xed. ;proportion W2sipso

cal proceSS.

not a purely chemi-

A lawWhlchEilq?er1ment cOUld not have establiSlied before Dal-

ton's .. wC)l,"k .l:leCli!llle, . .

that work vas accepted" a cOllstitutive principle

whichnCl sinsle set of. chemical measurements could have upset.

As a result

. of wbe.t is perhaps our fullest example of asclenttticrevolution, the same

chemical manipulations assumed a relationship to. chemical generalization
vel"Y d11'1'erent.. from the one they had had betore.
Needless to say I DaltOll' fl cOllclusiClns were widely attacked when first
8III10unced.

Berthollet, in particular,was never convinced.

Considering

the nature of' .. theissue, he need not. have been,B'.1t tor most chemists
DaltOll'Snew parad1gmproved conv1ncirlgwhereProust'.shad not been, for
..;:..'

"

.
. ,.,

136
that paradigm bad 1IqpJ.:l.cations far m.der and more important than a new criterion for distiDguiahing a mixture from a compound.

Ii', for

atoms,cOUld combine chemically only in simple whole number ratios, then a

reexamination of existing chemical data should disclose examples of multiple as well as of fixed proportions.

Chemistll stopped writing that the

two oxides of, say, carbon contained 5fi1, and 72$ of oxygen by weight; instead they wrote that one we1ght of carbon wOUld combine either wit/ll.3
or with 2,6 weights of oxygen.

When the results of old manipulations were

recorded in this way, a 2:1 ratio leaped to the eye, and this occurred in
the analysis of many weU-known reactions and of new ones besides.
dition, Dalton's paradigm made it possible to assimilate Richter' a

In ad-

and to see its full generality.

WO!'k

Also, it suggested new experiments, par-

ticularly those of Gay :Wssac on combining volumes, and these yielded

still other regularities, ones that chemists bad not previously dreamed of.
What chemists took from Dalton was not new ex,per1mental laws but a new way
of practicing chemistry (he himself called it the "new system of chemical
philosophy"), and this proved so rapidly fruitful that only a few of the
older chemists in France and Britain were able to resist it.

As a result,

chemists came to live in a world Where reactions behaved quite differently

fram the way they had before.
And as all this went on, one other tn,>:l.cal and very 1mpor"'-..ant challge

occurred.
shitt.

Here and there the very :nUlller ical data of chemiatry began to

When Dalton first searched the chemical literature for data to

support his PhYsical theory, he found some records of reactions that fitted,
but he can scarcely have avoided finding others that did not.
measurements on the

oxides of copper yielded, for

Proust's
an

137
,weight-ratio of 1.47:1 rather th8n the 2:1 demanded by the atomic theory,

and

-is just the man vilo m1ght have beenex.pected to achieve the

Daltonianra,tiO. He w.s,. 'tliatis, a f1neexper1iilentaJ.:J.st, and his view

relation betweenlrdXtu:res aDd caDpouD.cle was ve;:rclose to Dalton's.

BUt 1t1Bhai'a tb

III!Ike nat'i.l:i'e fit

a. Paradigm.

That 1s 'Wl'I.Y the puzzles of

science are sochtl.lleng1nge.nd aJ.so why

v1thoutaparadigmso seldCimlelld

undertaken

to arri conclusions at all . Cheinists

on the .evidence,

'.
,

'-

"

-',"

-,','-,

.muchofthB.tvas still negative.

-'

for

Instead, they had to beat nature into

.. line, aprocessWhich,in theevell.t;,took 1Wn0st another generation to

canplete.W1ienit wasdoll.e, even the percentage CanPoSition of well-known
compOundswasd1tterent. b t is the lalltofthe senlles in which we may
'illmt.te>

say that a1'tera revolutioxi sc:ientilltllli've in a: d:l.fferent world.

----

x.

The Inv1s1bilityof Revolutions

We have by no means finished our discussion of scientific revolutions.

In part1cular, if science does resemble the process sketched to this point

in this

there are two pressing questions that we must still ask

about them. First, how are scientific revolutions

What argu-

ments can persuade a scientific cOll!lDUllity to abandon one time-honored way

of pract1cing science in favor of another? Row is the conversion to a new
world-view tranBlllitted from the individual who invents it to his group?
fbat is the problem ord1nar1ly described as the nature of proof or of verificat10n in the SCiences, and we shall consider it in the section that fol10IIII this one. As we do so, a second problem will come to seem very urgent
indeed. We have already seen that, given -a scientific revolution, the preand post-revolutionary generations must be slightly at cross-purposes.

The

worlds in which they live and practice science are not everywhere quite the
same, nor are the rules identical which govern the normal practice of the
two

cOll!lDUllities.

It follOllll that, though either community may hope to per-

suade the other, neither can hope to prove its case.

Proof can al'WaYs be

evaded by insisting that the world is not that way (the sample, is e. compound not a miXture) or that some of the procedures involved in the proof
are themselves not scientific.

If, however, scientific revolutions are re-

solved by persuasion rather than by proof, we shall have to ask how persuasion in the sciences can function as it does.

Why,

if there is no such

thing as scientific proof, should scientific development seem quite so different from that of, say, art or religion?

Can the image of science de-

veloped to this point in the monograph be reconciled with what is generally

called the fact of scientific progress?

138

139
I think it can and shall. therefore shortlY. begin to

path

alcmgwhich. that. reconclliation.is to be effected. But 1'U'8t one last at

. .

the .existence and nature 01' revolutions ie

caUedfor.
txatl
.

display revolutions simply by iUus-

on,lI.pdHtllE!!lei:U;ustrijjlive .examples could be multiplied

clearly

=na::u:::s::;:e:::BIII::: .

these examples , which were deliberately selected for

customarily

CIIIIIUl.ativ'eadditi6lls tosc1ent11'i" knowledge.

not.as revolution,- out as

That

could equally

1IeU be <taken of any additionalillustrationa, anr'.. these would probably be

ineffective.
",

'."

"',,

"

I suggest. that there. is,exceUe:tt reason for the usual view

',0

,,"'

"-,,

"

.that.lllllkes revolutions. so J1e!ll'ly invisible.

:,oth scientists and laymen

te.kemuch of.their1mageofcreat1vescientif;,c activity from

tative source

that systematically

an: authorifunc-

. Only when, the nature of.1;hatauthorit.y1s rer.:ogn1zed and analyzed, can one
hOReto.make,h1,storica+ examplej,'ullya.f'fective. Furthermore, though the
point. can be j,'ully developed only ir. my concluding section, the analysis
llCM,reqll4'ecin:P. begin to. indicat.e.oneofthe.aspects. of SCientific york
which most .clearly distinsuishes .itfromevery other creative pursuit except perhaps theology .
As ..the.llource of authority.. I. have in mind pr1ncipally textbooks of

science. together with. both thepnpularizations and. the philosophical works

modelled on them.

All three of these categories"';.until recently nc other

significant sources of information about science have been available except

:through .the pract:l.ce of research--have one thing in common.

They address

themselves to an already articulated body of problems, data, and theol".I',

140
most often to the particular set of paradigms to which the scientific community is committed at the time they are vritten.

themselves

aim to communicate the vocabulary and syntax of a contemporary scientific

language.

Popularizations attempt to describe these same applications in

a language closer to that of everyday life.

And philosophy of SCience,

particularly that of the English-speaking vorld, analyzes the logical

structure of the same completed body of scientific knowledge.

'.rhough

fuller treatment would necessarily deal vith the very real distinctions

between these three geD1'es, it is their similarities which most concern

us here.

All three record the stable outcome of past revolutions and thus

display the bases of the current normal scientifiC tradition.

To fulfill

their function they need not provide authentic ini'Qrmation about the way
in which those bases were first recognized and then secured to the profession.

In the case of textbooks, at least, there are even good reasons

why, in these matters, they should be systematically misleading.

We noted in Section II that an increasiDg reliance on textbooks or
their equivalent was an invariable concomitant of the emergence of a first

paradigm in any field of science.

Tbe concluding section of this mono-

graph vill argue that the dom1nation of a mature science by such texts significantly differentiates its developmental pattern from that of other
fields.

For the moment let us simply take it for granted that, to an sx-

tent unprecedented in other fields, both the layman's and the pl'actitioner's knowledge of science is based on textbooks and a few other types
of literature derivative from them.

Textbooks, hO".rever, being pedagogic

vehicles for the perpetuation of n01"_1 science, bave to be

in

whole or in part whenever the language, prOblem-structure J or standards of

141

normal science ,change.

Inc short, they have to be rewritten in the si'ter-

math of each scientific revolution, and, once rewr1tten,they ine\"itably

disguise not'only the role but .the .very'.existenceoftherevolution that
. ,producedthem .UDless he.has persoIial.ly experienced a revolution in his
own.l1fet1me,thehistor:I:Cal sense of. the. working scientist and of

reader of textbook.l1teratureextexids<only to the outcome of the most recent revolution 'in the field. ,.
'l'extbooksthus beginbytrunca.t1ng the scientist's senseot his discipline' shistOl7', and they thenproceed.to supply a substitute for What
theyhave.el1m1.nated.

Characteristically, textbooks. of Ilcience . do contain

JUst a bit Of h1storr,eitherinanintroductOrychapter or, more often,

in.scattered.referencesto.the.:great heroes. ofan earlier age. From such
referencesboth stUdents and.prof'essiOna1s.do come to ,feel like participants :lna.long-standing historical tradition.

Yet the textbook-derived

tradition in which scientists come ..to sense.their.participation is one

that, in 1'act;neverexisted.F<irreasons.that. are.both obvious and highly
tunctional,sciancetextbooks (axidtoomany of the older histories of'
science) refer only to that part of the work of past scientists that can
be made to seem contrlbutions to the statement and solution of.. the text's

paradismproblems.Partly by'selection and partly by distortion, the

scientists otearl1er ages are... implicitly represented. as having worked upon
tbesame set of fixed problems and in accordance with.thesame set of fixed
canons' that the most recant revolution in .scientific theory and method has
made to seem sciel'!tific .No vonder-tAAt jtjJCtPooks an,dthe historical tro.ditiontheyimplyhave .to 'b.e rewrittensi'ter eachscienUi'1c revolution.
And

no.'Wonderthat, 8stheyarerewrit'cen,

entirely cumulative.

once again comes to seem

Scientists are not, of course, the only group that tends to see its
discipline's past developing linearly tOliard its present vantage.

The

temptation to write history backwards is both omnipresent and perennial.

But scientists are more affected by the temptation to rewrite history,
partly because the results of scientific rese8.lch show no obvious dependence upon the historical context of the inquiry, and partly because,
at any given time, the scientist's vantage seems so secure. More historical detail, whether of science's present or of its past, or more responsibility to the historical details that are presented, could only
give artificial status to human idiosyncrasy, error, and confusion. Why
dignify what science's best and most persistent efforts have made it possible to discard!

depreciation of historical fact is deeply, and I

think functionally, ingrained in the ideology of the scientific profession,

the same profession that places the highest of all values upon factual details of other sorts. Whitehead caught the unhistorical spirit of the
scientific community when he wrote, itA science that hesitates to forget
its founders is lost." Yet he was not quite right, for the SCiences, like
other professional enterprises, do need their heroes and do preserve their
names.

Fortunately, instead of forgetting these heroes, scientists have

been able to forget or revise their works.

The result is a persistent tendency to make the history of science

look linear or cumulative, a tendency that even affects scientists looking

back at their own research.

For example, all three of Dalton's incompat-

ible accounts of the development of his chemical atomism make it appear

that he was interested from an early date in just thOse chemical problems
of combining proportiOns that he

WEI

later famous for having solved.

Actually thoseproplemsseemonly.to .. ha.ve. occurred to him with their solutions and then notuntll his awn creative vork'ifas very nearly complete.
What all of Dalton's accounts omit 8l"E!the r:eyolutionsry effects of applying to chemistry a set of questions
and concepts
to
..
. . previously.restr1cted

.",

"

-,

"".

physics and meteorology . That is what Dalten-tid. and. the result

1M!

reorientationtowarcithefield. a reorientation that taught chem1sts to

ask new questions about and to draw !leW, C:()IIclusions frODl old data .
Or

asain.. Newton wrote that Gallleohaddiscovered

that. the constant
..
.

force of gravity produces uniform acceleration. .In fact. Galileo' s kinemat1c theorem does take that form.when embed\ied in the ma1;rix of Newton's
.
dyzIam1cal cOllCepts.ButGalileo.se.1d nClthingofthe sClZ't.
.

awn

falling bodies

l:T1s diSCUB-

al,lnde.sto forces".lDUchll'lfJstoa uniform

grartta.tiOiiai forcethBtcauses

bodies .to, fall .. By. crediting .to Galileo

the. answer to a quest10n that Galileo' s paradigms did not permit to be

asked. Newton' s account hides the, effect ofa small .but rev()lut10nary re,

.,',',

'M",,"""''''_

,_"

,,,,-. __ . "

,,,

,-""

formulat1on in the quest10ns that sc1entists. asked . about. motion as well as

in the .ansvers .they felt able to accept.

But1tis just th1s sort of

change. in the. formulation

of quest10ns and answers
that accounts. far lIIore
:
..
-

._,"

'

thannovel.emp1rical discover1es. f()r.

Galilean and fran Galilean to

"

from Aristotelian to
By disguising such

change!!. the textb.ook's tendengy to,lIIakethe development of science linear

h1des a process that lies at the, .heart of the.lllost sigm.ticant ep1sodes of
scientific development.

examples display, eachwii;llin the context of a single

revolution, the beginnings ota

completed by poatrevolutioD8l'Y science texts.

that is regularly
But in. toot completion more

ie involved than a l!IUl.tiplication of the histol:'ical

lustrated above.

il-

Those misconetrwctiona render revolution: invisible;

,

the arrangement of the still visible material in s.ciencs tEl:t:l ilnplies <.
process which, i f it existed, would deny revolution!! a function.

Because

they aim quickly to acquaint the student w.!.th 'What the contempol'al'Y Gcienn
tific community thiDks it know, textbooks treat '"he various e:x;r>eriments,
concepts, laws, and theoriee of the current normal science just as separately and as nearly seriatim as possible. As pede.gogy this technique of
presentation is unexceptionable.

But when combined w.!.th the gene:t-eJ.ly un-

histOrical air of SCience

and w.!.th the occasional systematic mis

constructiOns discussed above, one strong 1Jllpression is ovel"imelm1ngly

likely to follow: science has developed to its present point by a series
of individual discoveries and inventions which, ,taen gatherod together
constitute the modern body of technical knowledge.
the scientific enterprise. a textbook presentation

Frro the begiiming of

acientiets

have striven for the particular objectives that are the

ments today.

paJ.'<ldiS1l!

achieve-

One by one, in a process often compared to the addition of

bricks to a building, scientists have added one more fact, concept,

or theory to the body of information supplied in the

1&.;;,

science

text.
llerbaps I need no longer insist that this is not the way a sciene;,

does develop. Many of the puzzles 01'

normal science did not

exist until after the most recent scientific revolution.

Very tew of

them can be traced back to the historic beginning of the sc:l.cuce w'lthin
which they now occur.

Earlier generatiolls :pursued thei:" OH'""

.nth their awn instruments and their

0'\IIl

canons of !!olut:Lon.

n.:>:!:, is i1;

just the problems tb2.thavechanged.-Ratherthewllole

of fact

and theory which the textbook puradigm fits to natm'a has shifted.

Viewed.tbrOUghanolder paradigm llIlt)lfe' itsel1'was notqu1 te <the same

aDd could not be studied1D. gUtte tilflssme

example; amen

chemical

;;,my.

Is the cOnstancy of

fact ofei:PeriencetJlat chelll-

lstscould have discovered by exper1ment v.l.thin anyone of' th(:l'worlds

withinwh:l.ch chemistshavepracticed'lOi':f.s it rather one element--and

. an iiidubitabliFone; atthai:-'-'fiia iiewfii.ll:r:f.cO'tassoc:latedfact

and 'che-

orythat Dalton flt1;ed .to the earllerexper:f.ence of . chemists as a "mole,

chaDgingthatexperlencein the'prcicess'lOi', by the same token. is the
constani: accelerai:lon prOduced byacollstant force a mere f'actthat students of dynam1cs have always sollgll'br

'rather theamrwer to a

questlonthatfirstaroseonlYi,iUhfii Net1'tcm:l.ari.theory aruf\1hicn that

theory could anSwerirom the body of infoX'lll!ltion availablebefOl'e the
question was asked?

These questions are here askedaoout

the piecemeal-discovered facts of atenbook presentation.

appear as
But obviously

they have iDi;pl1catloD8 aswllforwhatthetextpreaents astheol'ies.

Those theories, of coUrse , do "fit the facts,," but only by 1;J;e.us:f'orming
prevlouslyaccessible informa.t:f.onintofactsthat.for the preceding
paradigm, had not exis1;ed at all.

And

that means, as we have by now al-

ready seen, that thebriestoo do not evolve piecemeal to t'it facta that

were

there

all the

time.

Rather they emerge togethervlth the facts they

fit fram a revolutionary reformulation of the preceding scientific tradition, s. tradition within wllich the lmcniledge-mediatedrela.ticnship bet\1eer:.
the sCientist and n!l.ture wes not qti!te the same.

146
One

last example may clarify this account of the imT-act ot textbcok

presentation upon our 1mage of scientific development.

Every elementl:J:'Y

chemistry text must discuss the concept of acllemlcal clement. Almost

always, when that notion is introduced, its origin is attributed to the
seventeenth-century chemist, Robert Boyle. in whose Sceptical Cllyeist the
attentive reader will find a definition of "element" quite close to that
in use today.

Reference to Boyle's contribution helps to make "l;he neo-

phyte aware that chemistry did not begin with the sulpha drugs;

addi-

in

tion, it tells him that one of the scientist's traditional tasks is to

invent concepts of this sort. As a part of the pedagog:l.c arsenal that
makes a man a SCientist. the attribution is immeneely successful.

Never-

theless. it illustrates once more the pattern of historical mistakes that

misleads both students and laymen about the nature of the scientific enterpriee.
According to Boyle, who was quite right, his "definition" of an element was no more tll&n a paraphrase of a traditional chemical concept;
Boyle offered it only in order to argue that no such things as chemical
elements existj as history," the textbook version of Boyle's contribu"!;ion
is quite mistaken. 'l'hat mistake, of course, is trivial, thoue-I} no

so, I inSist, than any other misrepresentation of data. Wnat is not trivial, however, is the impression of SCience fostarcd unen this s.).'t of militake is first compounded and then built into the technical atruc'GUre of
the text.

Like "time," "energy," "force," or "particle," tha concept of

an element is not the sort of textbook ingredient that is often "illVen"l;ed."

or "diacovered" at all.

Boyle's det'inition, in particuk"r,

back at least to Aristotle and

Lavoisier into modern texts,

Yet .thatls" not .to .. E!ay. that E!cience. has .. possessed. the modeI'Il concept of
an element since antiquity.

Verbal definitionB'likelloyle' s have little

scientlflccontentwhen.considered by themselves.'TheY's->-e not full

logict+l spec:l,1'1catlons otmeanirjg(if there are such}, but more nearly

point gain .full

Jlhen related. within a text or other syStematic presen-

tat10n, to .other.sclent1fic concepts, to IIIEIl11pulat:!.veprocedures, and. to

paradigmappl1cat1onE!.

It tollow!lthat concepts l:l.ke that o;t"sn element

can scarcely be invented 1ndep$ndent '01' context. 'Fu.."rthermore, given the.

context ,they,1"8l'ely requirelnventlon because they are already at hand.
Bo.thBoyle and lavoisler c:lUIIfged the.'cheri11cal signii'icanceof "element"
ln :1mportantways." But theydld.not.invant the notion or even change
thevel'bal:formula that serves as. lts "definitlon. n Nor, as we have
seen, d1d Elnateln have to invent or even expllc:!.tly redet1heepace and
time ln order to give them.new mean1ng within the

of his work.

What then wsBoyle' s hlstoric.al functlon in that part 01' his work
that 1ncludesthe1'amous "definition"? Hews the leader ota sclen. tl1'lc revolut1on that,. by cha:og:lngthe.relatlon 01' "element," to chemical IlI!l!l1pulatlon and chemical theory, transtormed the notion into a tool
quite different trom what it had been before.and tranatormad both chemistry and. the chemist's world in the process.

other revolutions , in-

cluding the one that centers aro1.1I!.d lavoisier, were required to give the
concept lts modern ':t'orm and function.

But Boyle provldes a typical exam-

ple both of the.process lnvolved.ateacho:t' these stages and.of what happens to that process when existing knowledge is embodied :!.na textbook.

148
More than any other sitlgle aspect of science 1 that :pedagogic form has, I
think, determined our image of the nature of science and of the role of
discovery and invention in its advance.

If the description

ot sCielltif1cdevelopment; so far achariced has been

at all persuasive I 'it

f6rmost'readersla:i'Uiid8lDental problem.

What is the process bywh1cha new candidate tor paradigm replaces its
predecessor? TO this po1lit we !lave seen iIOIf a paradigtD preViously

suc(fellll""----

tul in gll1 d1 ng pro1'essional research encounters first anomaly and then l i t

the anomaly endures l crisis.

In add1tionl we have examined at least a tew

occasional characteristics of the extraordinary' research 'I!b1ch is engen-

dered by the recoSDition ot crisis and which precedes the rapid emergence
at a new candidate tor paradigm.

Fin&llYI we have compared the new para-

d1gm with the oldl noting part1cularly the incommensurability ot the new
and old l an.'1ncommensurab111tyman1test in changed rules

tor scientific be-

havior and, in the new entities and regularities which scientists can see in
the world.

But clearly in tollowing this path we have mieced a step.

The

new candidate tor paradigm that emerges :from a crisis, emergeD tirst in the
mind of a single 1IId1v1dual.

It 1s he who tirst learns to see SCience and

the world d11'f'erentlyl and usually his ability to make the transition is
tacilitated by two circumstances Which. ere not common to most other members at his protession.

Invariably his attention has been intensely con-

centrated upon the crisis-provoking problems; usually, in add1t1on, he is

a man so young or so new to the crisis-ridden tield that practice has committed h1m less deeply than most of his contem,poraries to the world-view
and rules determined by the old paradigm.

to convert the entire profession or

ROlf is he a,ble, 'What must he do,

relevant professional sub-group to

his way ot seeing science and the world? What persuades the group to abandon one tradition of normal research in favor of another?

150

In the present state of the literature on scientific method one point

requires emphasis at the start:

the preceding questions represent the

only formulation which the historian can

for the philosopher's in-

quiry about the testing, verification, or falsification of established

scientific theories.

Insofar as he is engaged in norrual SCience, the re-

search worker is a solver of puzzles, not a tester of paradigms.

Thougn

he may, during the search for a particular puzzle's solution, try out a
number of alternative approaches to the solution, rejecting those that
fail to yield the desired result, he is not testing the paradigm when he
does so.

Instead he is like the chess player who, with a problem stated

and the board l1terally or mentally before him, tries out various alterna-

tive moves in the search for a solution. These trial attempts, whether by
the chess player or by the SCientist, are trials only of themselves, not
of the rules of the game.

The pa...'ad1gm itself is never at issue: it is

not being tested, end, in the puzzle-solving Situation, could not be.
Puzzle-solving can continue only so long as the paradigm itself is taken
for granted.

Therefore, as we have already seen, paradigm-testing occurs

only after perSistent failure to solve a noteworthy puzzle has given rise
to crisis.

And even then it occurs only after the sense of crisis has

evoked an alternate candidate for paradigm.

In the sciences the testing

situation never consists, as puzzle-solving does,

son of a single paradigm with Datura.

in the

Instead testing occurs as part of

the competition between two rival paradigms for the allegiance of the
scientific community.
If closely examined, this formulation displays tUlexpected and, I

think, significant parallels to the

most popular contemporary

151
"phUosophicaltheories about ,verification. ,Few: philosophers of sCience
stills.eek

criteria

the verification of sCientific theories.

..t"JlQ ,\;llE!or'l.!=!1J!

.. to, allposBible relevant

... teats,
',.

,-

. they., ask not:wbether:the theQry bas been verified .but rather about :Lts
prObab:tltty hi thel:l;ght of theevideneethat-aetually exists .

that quest:Lonons'1mportant : sChool is driven, as .1 am llere, to compare

, theabil1tyof a number oi'.different theor:Lestoexpla;Ln the evidence at
:hand. Tl:i8.tiris:LstenceonCompa1'ing theQr1es points one. of .the directions

inwh1Ch;t belie-v:e .future discussiQ1l8 of verification must go.

In the1r

fOl"lllS,:however,.probapilisticverification theorles demand the

cOlII,(iBrisonof all theories,thatcouJ,d possibly be. constructed for. the ex-

::plaDation"of a given body

And. the very concept of

suCh a construction.be8iris by raising all the problems. discussed previously. in connection With

and then gives rise to

others besides. , S1nc:e,J:may.

culties,letme only urge that the

spaqe to eJ!al!l1ne those diffi

by sc1en-

tists concernedWithver1fication are al'llllYB actual proposals of whiCh

there,areseldommore,tban tW!l . 1 can discover neither a need for nor a
route to theconstructiono:t

Verification is like

natural selection: 'it picks out the most viable emong a number of' actual
alternat1ve9in a

Whether thet choice is

the best that could have been made. ifst1l1.other alternatives bad been

., available is e. question that can not usefully be

Karl Popper, 'Io.l1ose news are obviously far c:lose1' to mine than are
those of the probabilists, appears to.av01d,the problems of all-poBsibletheories by denying the existence of any

procedures at all.

152
Instead he emphasizes the

of falsification, i.e., of the test

whose outcome is negative and vhich therefore necessitates the rejection

of an estabrished theory.

Like myself he sees science as progressing

from one theory to another incampatible one, the pOints of transition beinS,the falsification episodes which demand abandonment of the theory
which had previously prevailed.

Clearly the role that Popper and his fol-

lowers attribute to falsification is very like the one I have here been
assigning to anomalous experiences, i.e., to experiences which, by evoking
crisiS, prepare the way for a new theory.

Nevertheless, anomalous expe-

riences must not be identified with falSifying ones, indeed, I doubt that
the latter exist.

As was em,phasized repeatedly in Section VII, no theory

ever solves all the puzzles with vhich it is confronted at a given time,
nor are the solutions already achieved often perfect.

On the contrary,

it is .1ust the incompleteness and imperfection of the existing data-theol"".)'

fit which, at any time, defines IIW.llY of the puzzles 'I1hich characterize
normal science.

If any and every failure-to ..fit were ground for theory

re.1ection, all theories ought to be rejected at all times.

On the other

hand, if only severe failure-to-fit Justifies theory-rejection, then the

Popperians will require same criterion of "improbability" or of "degree of
falsification. " In developing one they will almost certainly encounter
the same network of difficulties -..>hich have haunted the advocates of probabilistic verification theories.
Many of

the preceding difficulties can be avoided by recognizing that

both of these prevalent and opposed views about the underlying logic of
scientific inquiry have tried to compress two sellm-ate processeo into one.
Popper's anomalous experience is

to science because it evokes

153
.... ne,l candidates fOrparad1sili.B'.1t falsification, .thcni,gh it surely exists,
does not occ:urw:Lth nor simply because of' the emergence of' anomaly.

In-

stead '1t1s' a subseqttent"and separate process Wll,ichm1ght equally well be

Caul!d verification smce it cOnsists m the triumph of a new paradigm

. JiJezthE:01dcme:FuxthermOre,,ic

---

cation process that the probabilist' s C:Ompa.riSOD-of'-theoriea plays a centralrole..SUCh a two-stage fOrmulationMsi Ithinlt, the virtue of great
Verisilli1litude;'and1tmayalsoena.1)lllulltObegin:exp11catingthe role of
.agreement (Or diSagreement )beween fact and theory in. the \rerification
Process.

To the histOrian,at least,1t makes l1tt1esense to suggest

that'ver1f:l.cat1on isestabl1.sliing tl1eagreement of fact w:Lth theory. All

histor1cslly<slgnfi'1cant"theories . have agreed withthe . facts, but. only
more;;.or-less. 'l'!lken i1!d.i'V1duaJ.ly they

are

on aparw:l.threS,Pectto agree-

ment.ButtheY are not on a par when taken collectively or even in pairs.

Itliiilkesagreata.eal of sense to ask wbiebot
theories f1ts the tactsbetter.

and competing

Thooghne1ther Priestley' snor lavoisier' e

theory, tor examplei l/.8l"eedprecisely with existing observations, few contemporar1eshellitatedmorethan a decade in concluding that lavoisier's
theory provided the better fitotthe two.

As

yet;

howeVer, this formulation mskesthe task of choosing betvieen

paradigms look both easier and more familiar than it is.

If. there t.'ere

but one set of scientific problems, one world within 'llhiChto discover them,
andoIle set of' standards for their solution,.

might be

settled more or less routinely by some process like counting the number of
prOblems solved by each.
cOIlI.Pletely.

BUt, in fact, these conditiODsare never met

competing paradigms

a.,"'"e.

alwys at least

slightly at cross-purposes.

Neither side will grant all the

aSSUl!lPtions that the other needs in order to make it" cone.

and Berthollet arguing about the

of chemical

are bound partly to talk through each other.

Like Proust -

they

Though each may hope to con-

vert the other to his way of seeing his science and its problems, neither
may hope to prove his case.

The competition between :paradigms is not the

sort of battle that can be resolved by proofs.

We have already seen several reasons why the proponents of competing
paradigms must fail to make complete contact with each other's argu1ll9nts.

these reasons have been described as the incommensurability

of the pre- and post-revolutionary normal scientific traditions, and we

need only recapitulate them briefly here.

In the first place, tha propon-

ents at competing paradigms will often disagree about the list of problems
which

scientific paradigm must resolve.

Must a theory of motion ex-

plain the cause of the attractive forces between particles of IlI3.tter or

may it s1mply note the existence of such forces?

Nerirl;on's dynamics was

widely rejected because 1 unlike both Aristotle' 8 and Descartes' theories,

it 1mpl1ed the latter anewer 'to the question.

When Ii'emon'e theory had

been accepted, a question was therefore banished frOll! science.

That ques-

tion, hawev-er, was one that general relativity may proudly claim to have
solved.

Or

asain. as disseminated in the nineteenth century, Lavoisiez"s

chemical theory forbade chemists to ask why the metals were

00

much alike,

a question which phlogistic chemistry had both asked and answered.

The

tranoition to Iavoisier's paradigm had. like the traL,llitio71 to Newtoll's,

meant a loss not only of a permissible question but c-f au aChieved. Eloll.ltion.

That loss was not, hOll'ever, permanent eithel'.

In the twentieth

155
century questions

q1.1alit1es of chemical substances have entered

science again,together 1dthsome anmrers to them.

MClre is:. however ,involved tha:nthe "1ncommensUl'a'bility of standards.
Since new paradig!llS 8re born :rran old cries, they ordinarily incorporate

muCh 'Of the v()cabttJ:tiio.y andapplustus, both'

tbatthe traditional.

parad.iSm

bad previously employed.

But they seldan

employ these borrowed elements in qUite thetrad1tionalway.

,Tithin the

terms,concepts, and'expel'1ll1entsfall:into new relat1on-

One with the other. The inevitabJ.eresult:l.s

. ships

we mUst c:all,

thoUgh the term. is not qUiter1gllt, a misunderstandingbet".reen the tvro

com,pet1tlg schools.

The men who Bcoffedat EinS1:e1n'sgeneraLtheory be-

cause space could not be "curved"-"itwas.not.:thet .. sortof,th1ng,.-were not

s1l!lPly Wl'0llg

or mistaken.

WJiS;1: theymeant by Bpac:e, i.e., Newtonian space,

Was necessarily flat, homogeneous, 1sotrop1c, and:unai'fected by the pres

ofmtter.

I f it

bad not been, Newton1anphyeic8

would not have

To make the trSnsit10n to Einstein's universe,the wholeconcep-

worked.

"'tua1

space,

and so on, . bad to

be shifted and laid down a,plnoZ!. na.1:urewhole Only men. who had together
undergone or tailed

to undergo that transformation. 'Would' be able to dis-

cover what, 1n phYs1cs, they agreed or disagreed about.

Or l1$':1in, the men

who called copEirnicuBmad:tor proclaiming that. the earth. moved were not
e1therjustwr6I1g orqu1te wrong.
fixed position.
1ngly:

T'ae:Lr

. it vas

Il.

Part of what they meant by "earth" was

earth, at least, could not be moved. CorresPond..

innova.tionlmsnot simply to .mqye.,the earth.

whole

Ra-!;her

ofreg8.i'd1!lgthepl'oblems of r,hysics and. 8strOnOl!!'J.

one which necessarily cll.anged the mean1.ng of both earth and motion

and Huyghena could 1"e1.11ize tha.t the earth's motion

.1813

e questio:f.\ ",.-i-1;h no

content for science.

inCOl!mleDaurability of cOlllJieM.ng Jit'radigmsl.

In e. SGllse "nicn I am unable

to explicate further the :proponents of ccmpet:l.ng paradigms pl'<'.ctice their

slowly, the other pcmdultillls that r'epeat -their motions ng!l.in

one J

S2'S cOOIDotmds. in t!1.e other miXtures.

am

aga:!n.

In

One i3 embedded. in e

,rorlds, the tiro grell,ipS of Bcientir'Gs see different 'thille;1J 'll1el.'l. 'chey look
from the

!l8.lOO

point in the

they can see I!.nything

has not cb.anged.

direction.

please.

Again, tJ:ll.\t 1s not to oe.y,

Both moe looking at the ,rcrld, and 'the

But they do see 8olOO;;'la't d:tft'erent

see them in differen'c relatiolls one '.;0 'i;he other.

and they

2hat ie"hy a lew -chat

cannot even be demollstrsted to one may ocClteiom'.lly seem :tntu:l.t:!:'!ely ob-vioua to the ocher.

Equally it is ""Illy; before they can hope 'GO cO!i!im:rn.icc;ce

:fully. one or the other must ell:'.Qer:l.ence 'Ghe couvel'sicn tJ:w.t ira bo.ve b"en

157
not generally accepted, part:l.Clll.a1'lyon tbecontinent, f'or"morethan bcli'
a century after the Principia &]@eared.Priestleynever accepted.tbe oxy_
gen theory,nor II:lrd Kelvin the electromagnetic theory I end eo on.

The

difficulties Oi' conversion have often been noted. by scientists themselves.

Darwin. in a. particularly ;percept-;ve passage at

Species, wrote:

tbeepd

of biS O1'i gill o:t:t-----

"Although lam fully Convinced of the truth of the ...'ielTS

S:!.ven in thiS. volume .. , I by no means ex.pect to conv!nceexperienced

naturalists whose minds are stocked with amultituiiB of fa.ctsall viewed,
during a long course of years, from a. point of viewdirect!t opposite to
mine. [B]ut I look with confidence to the future,--to young and rising' naturalists, who will be able to view both aides of the question with

Ant'l. MIllt. Planck,

AutobiographY,

sacuy. remarkea.-that:

in. his Scientific

"Anew sC:l.en.t:i.fictrutb. does not tri-

UlII.Ph by convincing its opponents and making them see the liGht, but rather
because its opponents eventually die, and a new generation gl"CWS up that
it is fe.m1liar with it."
These facts and others l1kethem are .too commonly knOlo;ll,tO. need further emphasis.

But they do need reevaluation.

In the past they have Illoet

often been taken to indicate that scientists, be1ngonly hUl!llUl, cannot always admit their errors, even when confronted with strict proof . I would

argue, rather, that in these matters neither TJroof nOlO error is at issue.
The transfer of alleg1ance from paradigm to

rience that cannot be forced.

is a conversion expe-

Lifelong resistance, lJSl'i;icularly from those

whose productive careers have committed them to an older tradition of normal SCience, is not a violation of scientific stnndards but en index to the

nature of scientific research itself.

The source of resistance is the

158
, assurance that the older paradigm will ultimately solve e.ll its p!'oblems,
that nature can be shoved into the box the psradigJD pl"o1rides.

Inevitably,

at times of revolution, that assurance Beems stubborn and piBheaded as indeed it is.

But it is also something more.

That same assuranoe is what

makes normal or puzzle-solving science possible.

PJld it 1s only through

normal science that the professional community of scientists succeeds,

first, in exhausting the potential scqpe and precision of the older para_
digm and, then, in isolating the locus of difficulty whose study may ver mit a new paradigm.
Still, "to say that resistance is inevitable and legitimate, that
paradigm change cannot be Justified by proof, is not to say that no

ments are relevant or that scientists cannot be ;persuaded to change their

minds.

Though a generation or more is often required to effect the change;

scientific communities have again and again been converted to new

digms.

Furthermore, these conversions occur not despite the fact that

scientists are human but because I;heyare.

Though some SCientists, par-

ticularly the older and more experienced ones, may resist indefinitely,
most of them can be reached by one argument or another.

CO!l.veraiOllG will

occur a few at a time until, a:1'ter the last hold-outs have died, the whole
,

profession will again be practicing under a Single, but now a different,

Parad18111.

We must therefore ask

arguments can serve to induce con-

version, and how these are met by those who do not find them compelling.
Before approaching that question directly, however, two caveats nm.st
be introduced.

Just because we are asking about persuaSion, or about

11.r-

gument ani counterargument, in a situation in which there can be no proof,

our question is a ne.r one I demanding a Bort of study that has not previcuEl;il

159
beenUlldertaken.

In this place we shall have to settle far a very :partial

and impression1sticsurvey.

Second, what has already been said combines

to suggest that,'II'hen'saked about persuasion. rather .thanproof',the question of the naturecif sCientific argument
need. have no s1:ngle nOr uniform

answex. Scientists e!ubraee a new

for all. sorts .of' reasons, 8.Ddusually,fol:\ several. at once.

Some of' these

reasons--for example the sumrorship that helped make Kepler a. Copernican-lie outside' the'e.ppe.rentsphel'e of. science' entirely Others "must depend
upon idiosyncrasies of' autobiography and personality Even the prior reputation 01'. theiIlnOVator 8.Ddhis teachers often plays an extraol'dinarily
significant role.
c

Ultililately, therefore, w<shEill have to ask this whole

question dii'i'erently. ., OIlrconcernlfillthen be not,what arguments may

serve' to convert the 1nd1V'1dUalbut ratherwhatcircUmatancesmay serve to
convert the entire group.

That problem, however;. I postpone to the IIext

. section, elCalD1n1ng meanwhile.somet;ypical aspects ofthe.argwnentation

. that figures so prominently 1n the battles over paradigm-change.

Mfi.iiY 0: thE!' argl!lllents for

and against paradigm" cl:l8nge concern the rel-

at1ve abilitiesof' the twoparad1gms to solve. problems, . and. overall these

are probably themostefi'ect1ve.Usually the .proponents of the. new paradigm u111 emphasize 'particularly their ability.to resolve the problema
that had gerierateda cr1sisfor.theolder .paradilYll'

Thus Copernicus ca.tld

claim, erroneouslY as it turned out, that he had. at last solved the vexing
problem of the lellgth of' the calendar year., Iavoisier could clam to have
solved the problems' of
coiiidcis.:l.fut6
motion.

of. weight-relations.

And. Einstein

fii!.ve reconeiledelectl'odyna.mics ;n:ths"revised llcience of

TIlis sort of argument proves particularly effective if' the

160

differences between the two paradigms suggest (1r permit the design of a.
"crucial" experiment, one whose outcome seems immediately congruent with
one of the paradigms but which proves very difficult to reconcile with
anythiDg but an obviously ad

formulation of the other. French resist-

ance to the wave-theory of light, for example, collapsed quite rapidly after Fresnel demonstrated the eXistence of a white spot at the center of
the shadow of a circular disc.

That was an effect which he had not an-

ticipated but which Poisson, initially one of his

had shown to

be a necessary consequence of his optical equations.

Or again,occasion-

ally a particularly effective argument for a new paradigm can be derived

from its ability to resolve problems that neither party could previously
have thought relevant to it.

Copernicus, for example, did not knOll the

telescope, and neither did his first opponents. Yet sixty-six years after
his death that instrument began to provide the most
all the arguments f(1r his new astronomy.

effective of

Similarly, Einstein seems not to

have antiCipated that general relativity would explain the well-known

anomaly in the motion of Mercury's perihelion, and he experienced a particular triumph when it proved to do so. Just because they CamLot have
been deliberately ''built into" the new paradigm from the start, unanticipaj;ed successes of this sort are often particularly persuasive.

[The

pOints made in this paragraph must be considerably expanded in the final

version.

Probably that expansion vill necessitate reorganizati0!l of the

material which follows.]

Yet though arguments like thoce described above are often immensely
persuasive, particularly to those wilom experience (or lack of er.perience)
has prepared for paradigm-change, we must aleo recognize that in principle

161
.

. =ot. beconclusive.andthatin

<t'arely prove to be so.

Usual.ly one or another. of the'cr1sis-prollferated ....eI's1orlaofthe older

pa.r8lU.8IIIw111 bavepart1allysucceeaedineolv1ng theproblcms which the
Jll:Wparadigm' sproponents declare tba1;.theyalone have resolved.: Defend-

---------<v
......s'-.oiLtbeP1'101'-Sc1ent1t1e. vaii1M.CnC8ll thereforelegltl.DfaWl;y !l:t'gue

bave.alreadyhad somesuceessand.tbat.. they will ehortly be

pletelysuccess1'u.l.

COlD-

'l'he problem at isstie,they.ean properly point out,

w111notbe the first obdurate. puzZle .that re'sea.l"ch Ullder the old" per.adiglll
bas successfully resolved.

That form of counterattack proves parM.culatly

effective.becsuse usually, when.the new.. parali1gm 1s 1'irstadvanced, 1t has

nOt 11;gel quite 8\1CCeeded in solvins the crisls-maltiilgprobiem, and the
cJa1ms.made in. its . JIQIIIe:are. characteristically excessive
.. tem

sys

in tact,tobee1ther.s1mplerar mo:reaccurate thtl.n

Ptolel/ly's; only I<'epler'ssubsequcn.t t'Ol'tIlUl.ationof the Copernican sjrstem

. co1lldprqperly
IavoiBier.sflw.

had made .. Or again. when

tltheairitsel. entire, " .. his llew theory oi' combus-

tiont.";:1.ft:not able at anto:copeviththeproblems::presented by the proliferation.of ne'li'gases,apo1nt which Priestley made wlthgl..'Ce.t success in
his counterattack.
In... short. during a parad1gm-debate.boththe old.l:mrad:lgm afJd the new

cand1dnteare.ususlly undergo1I\6el;ea4;ydevelopment . : The. iesue ill the de. bate ,therefore, usually does not prove to be which problems .each paradigm
bas solved.

Instead. the arg-..uuent is about whieh p&raMgm should be used

1l1 .. atteck1ngproblema.that. ne1ther. canyet guite claimj;o have solved.

later,-after the profession has again been un1fiedaudwheuthere ie
one leftwilliIIg to.:dei:end. the tre.ditiorw.l ....

an apparently conclusive

case for the new paradigm can often be constructed.

But that te>:tbook

case seldom appears until a time when it is no longer needeo..

A!.1d. ,Then

it does appear, it usually rests on arguments and upon data that became
accessible only after the historic confirmation debate bad ceased.
Foucault's pendulum, which "proves" the earth's rotation, and Fizeau's
experiment on the relative velocity of light in air and

are only

particularly well-known examples of experiments now used to make pOints

that no longer needed making when the experiments in question were first
performed.

Undoubtedly they are e:f'fecti'fe experiments.

If they had been

performed sooner, the confirmation-debate to which they seem relevant

would. probably have been shorter.

But there is no reason to think that

there would not still have been a debate.

Besides, these particularly

convincing demonstrations (and even they could be countered) usually share

one very significant characteristic.

Ordinarily they demand eguipment and

concepts that wre not available until after the debaw.

USllS.lly, they

are products of a normal scientific tradition that became possible only

when the

had almost ceased.

Even these lines of argument fail to display the heaviest artillery

available to the defenders of the scientific tradition.

otten they can

and do admit that, for application to the crisis-making problem, the new

candidate for paradigm has significant advantages.

But, simultaneously,

to a host of problems that the old paradigm did

its cha.llenger cannot.

but that

Thus, until the work of Ga.lileo, the Copernicana

could not give nearly so good an account of terrestrial motion as the one
Ptolemy and his succeSBors had given.

laVOisier .:as never aDle to el:plain

the common properties of the metals, nor, in the early stage>] of his

theory's development, could he el.'Plain either theproducUon of 1later by

oxygen..hydrogen. explosion or .the production. of aninf'lamluable .gas from
carl:!on .. Me.xwell!. s .. tb,eory,.despite.... its . author'.s.. expressed intent, made

.&IW solutioll to the problem of a mechanical aetherseem :f.mlIossibly dis. To generiilize, no

new .cendidatef()rparad1gmevel" .solves all the. traditional and .novel proble!Ds that confront. it, andi.ts apparent failures do provide a basis for
entirely l,egitimate resistan.ee.Some of' those f'ailures become problems
f.or normal research under the. new paradigm;: others, like the cause of
grIlyity, tb,es1m1larityof.themetals,. or the mechanical aether, are accOIIIIIIodatedouly by c:ha.tIging1;he standards of the. science itself
. . , That IlSedtO:change s1;andards' opens. up a whole. new line ofargumen\

tation,and one that is. of a rather d1fferentsort.. Not. all the. argumeIl1;s . l"ele:vant1;()ach()1ce of. pare,41gm .lle.e.d..
at. all.

prc:itllelD..solv1ng

Theopponents"of.;a;new;parad1gmC8Iland llometimes.doprocla1m

that. thecballenger, hQ'!leyermany prollle!DB;it can solye,. must be rejected

because it s1m.Ply is not... seience.

Iloris.that.aceusation. entirely un-

tounded .......The.. meJl .who. make it. areOIlly.de:t:ending. e1;Mdards ..tll!l.t have been
closely' assoc1ated.w1th the'

Or

again;. given anything like a ;. stand-oi'i'on problem..solVing ability, the

proponents of

paradigmmayclaim;tbat .their way of solving prob-

lems is more reasonable, or that. it.iss:iJDpler, or that.it fits better

with the findings of some other science, or of philoscpby, or of religion.
Occasionally the appeal maybe atraigh1;torwardlyto e,uthority, or it may
even be to the personality of' the paradigm's proponents . When this occurs, paradigm-debates. may. grow.

bitter.

Men hll:\Te lost

164
jobs because of them.; promising careers have been abruptly halted.
Obviously reactions like these are excfll3siva and
retrospect, even those scientists who feel their side
ally asb.llmed of their participation in them.

In

right are usu-

But that does not mean that

these extreme reactions are irrelevant to the issues of the confirmationdebate.

Just because logic fails, because the issue is conversion, be-

cause men must be made to see a world toot they can now look at w:!.thout
seeing, words ,like ''blind,'' "pigheaded," and "fool" may finally be the
only appropriate terms remaining to the participants in such a debate.
Sometimes, in fact, new converts apply the term "fool" to themselves for
failing sooner to see the light.

The men who use these terms do not, by

doing so, cease to behave as scientists.

There is a full continuum from

arguments based on data, through those involving standards, to those which

employ vitriol.

.And any argument in that entire arsenal can work.

Note,

however, a point to be emphasized in the next section: though vitriol is

occasionally a natural and relevant ingredient of SCientific controversy,
there

be no science i f it were the only ingredient or even the most

effective one.
Everything said so far should indicate that, insofar as they are
merely logical, debates over paradigm-choice must lead to stalemate.

That

they do not do so indicates that SCientists are reasonable human beings,

at least as

So

group.

An argument need not be logically compelling in order

to persuade most of them that they have made a mistake.

Only there turns

out to be no single argument that does (or ought to) persuade them all.
InsteaO., what occurs is an increasing shift in the stfJ.tistical distribution of professional allegiances.

At the start a new candidate tor

may

Mve'veryfew sUpporters, and the motives of soine of thel:l are often

suspect.

Copernicus wOUld almost certa:l.llly havebeenhorr1.fied by Bruno,

But often it makes ,little

. difference wtililt.motivesbriIIg its flrst sJlpp91"ters to alJ3.radigm.

are

competent they will improve it, explore :l.tspossibilities, and

If they
SilOl(

whatitWOUl.dbel:l.keto belong to ascientificcOl!llinmity guided by it

. '. And as that goes on, if' the parad:l.gmis one destined to. u1n its fight,

thenumbEir and streDgth of the persuaS1vearg\imentstavor:l.ngitw:l.ll :l.ncrease.' More sc:l.entists will then be converted, 8.ncl. the explora.tion of
the new J?S.rSil.:I.!,!DIw:l.ll

. Graduall;y the number. of' experiments, instru-

ments, articles, and books based upon the paradigm w:l.llmul,tiply;

more

Still

men, conrlnced ot the new view'sfruitflllhess, will undertake the new

modeforpracticiIIg norlllal science, until at last only a few elder holdouts remain.

And even they. we cannot . say, ...were.lIl'.ong .. Though .the his-

toriancan alwaysf1nd men-.Priestley.for instance-,..who were foolish to

resist for as long 8stheydid,he<w111.not find a point at which resistaDCe suddenly becameunscierititico

At most he may w:l.sh to saytllat the

man who continues to resist after his whole profession has been converted

ceased tabea scientist.

XII.

Progress through Revolutions

[This section has as yet received much less attention than any of

those which precede.

It still seems to me far from rig,.'1t, particularly

in its later pages, and I plan a complete revision for the final version.
But other

are now pressing, and I must therefore postpone

further work on the manuscript until the summer.

Since I

at that

time also to complete the other (p?'Ssumably less dras-;;ic) l-evisions that
will

needed before publication, I have decided not to delay further

but to ask for reactions to the draft in its present form.

Though they

may be unrecognizable in the final version, the pages which :follow should
at least indicate the main directions in which the conclusion to that version will proceed.
I am by no means clear about all the revisions that will be needed.
Many

will be stylistic and organizationa.l.

In addition, though it will

remain a sketch, the discussion must be expanded at several points.

"special" cha.ra.cter of the authority exercised by the scientific

The

CO!JllliUO-

ity must be considered in more detail. Also, the parallelism between

progress-through-normal-science and
made more apparent.

must

be

(In the eyes of the profession, what else but prog-

ress could the outcome of a successful revolution have been?) Fil":.ally,

the subject of truth-by-ukalle need6 both fu.ller and more t<)1!lp3rate tree.tment.

Doub-Uess still other paints \,1.11 emerge wen I face up to this

part of the job.]

l"c.e preced:l.ng pages = y my 8chelretic descriptio;; of.'

de-

as far as it here can go. Nevertheless, they cannot quite couclude this monograph.

If my

c.eilcription seems at all to catch the essentl_a:'

166

structure of a science's development. it will inevitably also seem to

a specialllroblem: .Why

____ w

the. outcome of the process sketched above

i:t;Y,0:L:. $l"8J1:t;illg thllt

C&n .. theente;t'prise's

_ _,

pOOl>

.sCience. how

apparent cumulativeness. be reconciled with the funda

v.,

.,

. mental roleplayed b9l'evoiutloUSiU. its develoPtJH:tltY. IihY. it. there is no

set of rules for scientific behavior and no fixed .mode for scien-

tific verification. ,should science seem to move ahlllsd as . say, art. philosophY, or theology doellnot'l

These questions pose what,! shall call the

problem of progress in the sciences.

.

TIro things IllUst at once be said about it .. First, sCien,ce undoubtedly

dOes progress in ways which. 1f not different in kind, are at least vastly
more apparent than the progress Visible in other creative. pursuits.
ond,. there
ought no longer be. a problem about
why this is so.
"--,
,-

.'

".-;

cQUld

" ' - - -

' "

Sec-

What but

from ,tile,

in

.the .preceding .pages? This is not to say, that the view... of science here developed presents no, problem, but only that it presents no, special one.
may ask,.as

questions like

b.E!suClllltllinglls.llcleJlcE!t.What !Jl1l,st
order that

beposllible at allY

nat urE!'

follcming:

Rev can there

including man, be like in

Furthermore, though we shall ask

those questions. we shall n()t even begin to BUggest an anmrer.

questions will remain as open as they were before.

that they are .. the only questions.

sort of enterprise

Those

They are not, ho;7ever,

questionsposedl>Y any special features of this monogl"aph.

of science necessarily raises them too.

One

Any

other viell

Let me, theTefore, try to shmf

If nature permits the e::istence of the'

in these pages, then, wh:i.le it E'.xists and euc-

ceeds, that enterprise cannot help but make prOgreBs.

We shall Bee this

168
i1' we

learn to accept as cauaea lmat have ordinarily been taken to be ef-

fects.

With that inversion the phrases "scientific progl'ess" and even

"scientific obJectivity" may come to seem redundancies.

What aspects of the scientific enterprise make for progress? There
are, I think, several, all rooted in the nature of the particular proi'easional groups that the evolution of mature science permits.

After the

pre-paradigm period and except briefly during revolutioca, a scientific

group yorka at any given time from a single paradigm or front a closely
related set. Very rarelY are two different schools to be found working
on the same problema, and in those few exceptional CBaes the competing
schools always share several major paradigms. Viewed from within any
single school, however, whether scientific or nonscientific. the result
of successful creative work is progress. HO""" could it poss:!.bly be anything else?
The artist, for

who extends the technique accessible to his

school or who applies its ex1sting body of technique to a new subject inevitably contributes to the progress of his group.

In fact, as Gombrich

has brilliantly shown, throughout classical antiquity and again in early

modern t1!lles, painting and sculpture were

cumulative disciplines.

Critics and historians, including both Pliny and Vasa\'1, recorded ;lith
veneration the series of inventions 'from foreshortening thl'ough chiaroscuro that had made possible successively more perfect representations of
nature.

other creative fields display progress of the sa.me sort.

theologian

articulates dogma or the rhiloaqpher who refines the

The

contributes to progress if only to the progress of those

share hie premises. No creative school has, I think, ever recognized a

category of workl!h1ch
was

on the one hand,a creat1ve succesebut which

Oll.theother, a contribution to ,the progrees of' the group.

we

If

progJ:'ess, . that cannot

be because . individual schools make.. nOml..Ratheritmuat.be., because there

cempe-'Iling-sehoele each. of-,which constsntlJ .qnestlons

foundations .of .the others.
ple, .hes .. made. no

liel'Y

The man'llhoSl"gllEls .thatphiloso,phy,for examemphasizelilthat there .arestill.4ristotelians,

not that. Ar1stotel1a,n1sm hal!

,\;oplogress.

But those doubts about progress ariaein the ,sciences too.

out the

I;he

Through-

are a multiplicity of competing

schools. evidence Of prOgress, exCePt

is very bard to find.

This istheperlod, as Isuggestedin.SectionII,:invlhich iudividuals

practice science but in 'll'hich the. result of''\'heir ente%'prise does not add
up to sclence as we know 1t.

And again, during perlods of revolution when

.. the.. f)'ndameTltal .. :tenets. of'.lItfleld.are....OllCe.more . '. at . . issue, 40ubts are repeatedly.expressed.about the verypossibllit;rof continued. pl"ogreas if' the
p8ra,digmopposed by.the

who. 0llPosed Newton,

1:l:1s

... Dark Ages.

Thosel!ho

to tt.-e

held that the rejec-

tion of' chemical "princ1ples" .1n favor of' laboratory elemnts

'!laS

re-

jection of achieved
chemicalexplanatlon. by those wnowould take. refuge
.
,

:,'..

'"

in a mere name . A similar,

moderately expressed, feeling that

1tmust end scientific progress seems to underlie. the c::pposit1Oll of Einstein, .Boehm, and others to .thedominent pr()babitlstic. interpretation of
quantum mechanics.

In short, 1t is only during perioos of nomal science

that progress seems.both obvious and assured . During those periods,

170

however, the scientific cOmmunity could view the fr-l1its of its work in no
-other way.
So far I must seem to be saying that progress is in the eye of the beholder and that science seems to progress only because there is ordinarily
no competitor to the scientific community's view of its task.
must be part of the answer to the problem of progress.

Indeed,

Our sense of the

uniqueness of sCientific progress does result from appearance alone.

There

is progress in other disciplines though the multiplicity of schools makes

it harder to see. But that is only eo part of the answer and by no means
the most important part.

I have, for example, already suggested that, once

the reception of a cammon paradigm frees the scientific community from the
need constantly to reexamine its first principles, the scientific community
can concentrate exclusively upon the subtlest and most esoteric of the phenomena that concern it.

Inevitably that does increase

ness and the efficiency with which it solves new problems.

the effectiveO'"...her aspects

of professional scientific life enhance this efficiency still further.

Some of these result fram what I can only call the ul1pl:'ecedenteo. insulation of the scientific community from the demands of the laity and of
everyday life.

Obviously that insulation has never been neal'ly complete,

and, even more obviously. it has broken down further since World War II,

possibly with resulting dangers for science.

Nevertheless, even today

there is no other community in which individual creative vork is eo exclusively addressed to and
most esoteric of poets

01'

by other memoera of the profession.

the most abstract of theologi&ns iu far more

concerned than the scientist mth lay approbation of !lin erective work
though he may be even less concerned with approbation in general. The

The

171
diffeI'ence proves' conseilucnUal;JUs1fceca.usehe 'is 'Work1ilgonly for an
audience of colleagues, an audience, that is, which shares hlsOW:tJ. st8nd, ards,.,the"scientistcan.talte

a single' set ofstands.rd.sfor granted..

He

'abou,1:l1hatsClIIIe othersr:OilP'"orschoolwiU .thinltand can

" theretore diepose of' one. problem and' set

those who workforca 'more

on. to the next more quickly than

, Even mOl'e'1mportant and ho-.r-

'eVer:incomplete ,theinsulation

ty from society

permits the 1nd.1V1dualbasic scientist to concentrate his attention upon

problems that' he hils go,od reason to be110'1ehe will be able to solve.

Un-

like the, engineer, many doctors, and 'most theologians ,the sCientist need
not choose problems because they ursently Ileed solution and inapparent
'disregard' of the' tOQls'ths:t:hehas:at','hand'withwhichto''solvethem.

In

.. tb1s respect Iblive myself found the contrast between physlcietsand IIl9.!lY
social sc1entists particuJ.arlyinstructive. 'Many of the latter tend, as
the>:f'ormer.almoet'never'dof to' defend their choice ofa research' problem-e.s., the effects ()f raciSldiscr1m1nationor the causes of the business
, ,cycle--chiefly'intepms of.the;.s'Ocialimportance of achieving a solution.
Which .group wciuldone.thenex,pectto solve problems at a more rapid rate?
'.rhe effects of insulation from ,the larsersoclety are sr:eatly intensified by another characteristic of the professional scientific community,
the nature 'of:,lts educational

In muSiC, the graphic arts, and

literature the pra.ctitioner; gains his education by exposure to the works

of other artists, principa.llyearlierartists. Textbook!), e:c:ept compendia of orhand.bool\:'sto originalcreatiOlls, have only

&

secondaloy role.

InhiS'i;ory, philosophY, and. the social sciences tex'cbook litez'a"l:;u-""e has e

greater significance. ',But in these fields even the elementary college

172

course employs parallel readings in original sources, sorne of them the

"classics" of the field, others the contemporary reBearch reports tha.t
practitioners write for each other. As a result the student in any one
of these fields is constantly made aware of the immense var1ety"of prob-

lems which the members of his future group have, in the

COurll6

of time,

attempted: to solve. Even more important, he has constantly before him a

number of competing and incommensurable solutions to these problems, solutions which he muet ultimately evaluate for himself.
Contrast this situation '!lith that in at least the contemporary natural sciences.

In these fields the student relies mainly on textbooks

until, in his third or fourth year of graduate work, he begins his cnm
research. Many science curricula do not even ask graduate students to
read in works not written speCially for students.

The few that do assign

supplementary reading in research papers and. monographs restr:!.ct such asSignments to the most advanced COUl"SeS and to materials that take up more
or less where the available texts leave 01'1'.

until the very last stages

in the education of a scientist textbooks are systematically 8ubstit"u:ted

for the creative scientific literature that made them possible.

Given

their coxi1'idence in the paradigms vhich make this educational technique

possible, few scientists would w"ish to change it. Why, after all, should
the student of, say, physics read the works of Nel.'ton, Faraday, Einstein,
or Schrtldinger when everything he needs to know about these works is recapitulated in a far briefer, more preciee, and more systematiC form in
a number of up-to-date textbooks.
Hithout '!Iishing to defend the excessive lengths to u:J.ich this
of education has occasionally been carried, ua cannot help

notina in

173
general it has been immensely effective.

rigid education, pr0llably more se ,than

dox

fOf

Of . course it ill a narrow and

other exceptparbs.ps in

Ol"thO-

wC)rk .. for. puz:;;le..;solving

the tradition. that the textb()Okll define, the scientist ia .. a.most. perfectly

Ji!w.othel'lll9l'e, he 1s well equipl'lld for .anotiteI!;ask as llell, the

generation through normal. science of significant crises.

Wnen.a crisis

arises, the scientill1;.is no1;.ofcourse,. so well. equipped.

Even though

prolonged crises are probably rei'le.cted .in. ,less rigid e.ducationaJ. practice,
. scientific

is n01; well designed. to produce the. man who will easily

afl'l!shapp,;ODach',But so :j.onSa.1I sOlllebody appearawitha new

candidate for
"

'

',;'

'

a youngman. or one new to the i'ield--the

' "

-,'

',-

,,'

10Slil due to. rigidity accrues onlyto.theindiv1dual.

Given a generation

in, which to _
effect
.the :.change,
individual rigidity is . compatible with e.
' , , , .>_
',.'
Co

.:.

-"

"

cOllllllUIlity that ,can switch from paradigm to paradigm when the occaSion demands . PartiC\1la1."ly .i.t 1.s. cCl!Dpa;i;illle.1lP./iln that.. very rigidity provides the
cOllllllUIlity wi.th a sensitive indicator.tha"t; something haagone
So far we
_

In-Ollg.

seemed to,cOlls1der only those.characteristics of the

scientific group which make for rapid and cumulative. progress within a
,-,"

__ "',

_ __

__

nOl'll!f;].. scientific

__

"" __

._,>."'.',,''',

,._

" . _ , " " __ , ' , ,

_.,_". _ _ .".

..

But ',hese same characteristics make for a

rather different sort of progress across scientific revolutions as well.

In part ase.if"l;he.t progress in only appearance.

tirely cumulative.
--

RevolutiOns are not

(In-

Like revolutions in. pOlitical life, theology, and ar-

'-'-"

-,>'

--

"

tistic taste, they involve loss as well as gain, but in the sciences this
loss is largely hidden by the
lows each revolution.
past.

In all

reconstruction 0t hill "cory tbs.t fol-

But revolutions are never a total

with ,the

some part of past achievemen'c is carried over, and

174
the nature of the scientific community ensures that ecieutieta will = y
over more of past achievement
l1kely to do.

practitioners of other fields are

Everything said so far should indicate that the sCientist,

trained to l(Ork within a single normal-scientific tradition, will have a

far clearer and more stubborn conception of the number and nature of his

field's achievements than will his colleague in a nonscientific field.

That being the case, he will be harder to move, less likely to find an alternate paradigm attractive for reasons irrelevant to his field.

In par-

ticular, though the individual scientist may, his C=nity is unlikely to

undergo a paradigm shift unless two conditions are met.

F:I.rst, the new

paradigm must seem to resolve aome problems that can be resolved in no

other knO".m way.

Second, the new paradigm lIr.lst ;prese:rve a reletively

large amount of the problem-solving ability which had accrued to science

through its predecessor.

Though new paradigms seldom or never do every-

thing that their predecessors could do, they usually preserve a good deal
of

that paat achievement and permit more besides.

I am not at all sure that there is more to be said about :progress in

the sciences. Given the sort of professional community described in thl<l

(

earlier sections of this monograph, science could not help but progress.
Yet I am also sure that most readers will find ,one element missing and
another distasteful in this account.
and then conclude with a brief
someone else to ask.

tHO

in order

on the question that remains for

The missing element is, I take it, aciEintific truth.

Tae process described as science

tion

let me discuss these

this monograph is a process of evolu-

primitive beginningS tward an increasingly detailed aIla. re,?illecl

understanding of nature.

But nothing that has here been said makes it a

175
- process of evolution. towards . -This isalmost-,inevitablytroubleeome (certainly i t has been so to me) .because weare"iieeply accustomed toseeiDg

as the

neare,r. to 1ts goal

In ,some e.ensewhich we

whether.or n01;that goal is ever to :

-------,afalrel'lt-Wwi'iollH1l1ng-te let the ph1lesophel'

we are sure

cessive stage of scientif;cknO\lledge.is.more nesrlytrue or more probably

true. or dmply closer.

,truth tbantheonebefore . Almost every form

of our usual 1magemakestruth the goal toward 'Which science strives and

with respect. towh1ch,its success must be. measured.

:But need there .b.e any suchgoal'l . Will not evolutioD from .do just as

,',

. C . _ " "

__ '

"",','-,

Vellas evolution toWards?, ,If it will,:l.tsurelycreates fewer problems.

Thesequestions,it

a;re,very .lilteones asked just a

century ago in a rather different. connection,and it may be 'Chat we need

to make. the same about-face that
. century
,

"

"

many
-

then.

By the. middle of the last

studen,ts Qfbiology andgeo},oSY1fere .collvinced that the spe-

, _"

_,._

cies of flora and

,,,

"',,

,,""

_'" - ._'

'"

,'. ,,"

"

'

0",'

, , ' _ ' : ' . __ "

'

,,"

"

of the globe in the nineteenth cen

tury had evolved from a, smaller Ilumber of"> more primitive

forms most of
__ c
'
, ".

, '

, " . , , , ,', .-'_ c.

which bad since ceased .to exist.

>' -,'

','

_,-.,_,'"

'

""/.

,'0'

But they saw nowsy to explaill this evo-

.:

. lution.wi thout supposing that

,',,',

,'"

,:"'.,-,

,,,'

,;

had
-'....

,;._",

,',;'

_'

__

,.',

some form

or plan toward. 'Which more primitive species bad striven. as their goal or
end.

teleological explanations

that sort were anathema to most

segments of the nineteenth, century scientific community, the few men ,,:ho
expressed this evolutionary, view usre ,not, taken so seriously as they might
now be.

In 1859 that.

<:han!Z;!!dfor a.n\llliber of reasoDeamong

which the most important was the,simultaneouB announcement by De.lHin and

thatevolut10n to existing forms and on beyond them could be

understood without recourse to a plan or goal laid dew at the start, perbaps in the mind of God.

Natural selection, survival of the fittest, would.

produce the same results including such marvelously adapted organs as the
eye and hand of man, organs which had previously provided immensely powerful
arguments for the existence of a Supreme Artificer.
Only this use of the analogy to natural selection is my own, and even
I do not want to insist upon its applicability in detail.
important respects the analogy is extremely close.

Nevertheless, in

The process described

in Section XI as the resolution of revolutions is the selection by conflict

within the scientific community of the fittest way to practice future
science. The net result of a sequence of such revolutionary selections,
separated by periods of normal research, is the wonderfully adapted instrument we call modern scientific knowledge.

And the entire process may bave

occurred, as ve now suppose biological evolution did, without the benefit

of a set goal, a permanently fixed scientific truth, of which each successive stage of scientific development was a better exemplar.
Reference to Darwin also serves more effectively to raise the specter
tbat I anticipate many readers of this monograph will find distastefUl.
The animals which survived natural selection were those llhich, partly by
"tooth and claw" and partly by superior fertility crowded cut their neighbors.

By the same token, when I say that paradigm selection is settled by

argument among SCientists, I may seem to be eaying that in the SCiences,

too, might makes right.

I bave, in fact, been told that on my view of

science Hitler would have been right if only he had succeeded in subjugating the world.

177
Two things seem Joo me to. be wrong with that interpretation.. First,
if' we can dispense with. "truth" when describing scientific development, we.
can dispense with "right" as .well . ' I have surely said. no more than that
the paradigm which wine will be the new :PSI'E1dtgni.
rephrased.

But it can be,ana then its

seem to have . . sa1d:tlJ.at .i.t.. makes no

'!be obJection must at

among scien-

tists. is settled, that.H1tler,torexample, could .. have . beenallowed to

'!'hat, 'ot course.wouldbetotally,absurd',:The.continued ex

istence of science depends upon the continued existence. of avery special

sort of COmmunity, one. that, among other. things, is committad to solving
problems about nature and knows very .ve:j.l wllatproblems it hass.olved
.' When....

is the

arbiterofparad1gm'debatestheprocessWiii!iBye.ceaaed.1;obe science
Just how special the community has to beif' .science is to survive and grow
may be indicated by the very tenuousness

ot

hold on the enter-

prise . Every civilization of whi.ch we hf\ve any l'ecordsat.sllhas

POB-

sesseds technology" an art, arel1gion, a. political . system, laws, and eo

on.

In

our own.

case!!. those.:f'acets of Civilization hI!,:ve been. as developed as

But only the Civilizations which.descend from}Iellenic Greece

have possessed more than., the most rudimentary science. The overwhelming
majority of sCienti:f'icknowledge is a product of . Europe and of the last
three centuries.

What

Euro.PSan society have

made it uniquely able to support a scientific community is e. problem beyond the scope of this monograph.

But recognizing that sciencehi<.s been

vastly more localized in space and time than any othel"human creative pursuit may help to persuade us of the special nature of the community that
produces it.

178
Is there any other question?
this section began.

Obviously there is--the one from which

What must nature, including man, be like in order

that science be possible at all?

Why should the scientific community be

able to reach a firm consensus unattainable in other fields?

'!hy should

consensus endure across paradigm-change after paradigm change? And 'Why

should the result of paradigm-change Within the scientific community be
progress?

From one point of View those questions, excepting the first,

have already been answered.

were before.
Cial.

But from another they are as open as they

It is not only the scientific community which muet be spe-

The nature of which that community is a part must also possess

quite special characteristics, and we are no closer to lmowing what they

must be.

That, however, is not this monograph's problem.

longs to the sciences themselves or to metaphysics.

quire only one remark about it.

Instead it be-

At this point we re-

Any conception of nature which is com-

patible with the growth of science by proof is compatible with the evolutionary conception of science developed here.

The problem of Dature is

made no more difficult by any contention advanced in this mOlloglaph.

the contrary, the problem of nature may be made easier.
conception of science here

On

For the looser

may be compatible with a conception

of nature that will not provide a base for science through proof'.

That

possibility of greater freedom combined \T.lth its vastly greater verisimilitude proVides strong reasone for supposing that science is the cart of
enterprise which this. monograph takes it to be.