yes, but. .. some skeptical remarks on realism and anti...
TRANSCRIPT
Yes, but. ..Some Skeptical Remarks on Realism and Anti-Realism
SummaryThis paper argues that the much discussed issue between "scientific realism" and "instru-
mentalism" has not been clearly drawn. Particular attention is paid to the claim that only real-ism can "explain" the success of scientific theories and - more especially - the progressivelyincreasing success of such theories in a coherent line of inquiry,This claim is used to attempt toreach a clearer conception of the content of the realist thesis that underlies it; but, it is here con-tended, that attempt fails, and the claim itself hangs in the air. A series of increasingly sophis-ticated versions of the "instrumentalist" thesis is considered, and both these and the contentionsof realism are placed in relation both to particular examples of scientific development and posi-lions historically maintained by philosophers and by scientists. The author's conclusion is that,••·hen the positions are assessed against the background of the actual history of science, (a) eachor the contrary doctrines, interpreted with excessive simplicity, is inadequate as a theory of thedialectic of scientific development; (b) each, so interpreted, has contributed in importantinstances to actual damage to investigations by great scientists (Huygens, Kelvin, Poincare);whereas (e) in both the theoretical statements and the actual practice of (in the author's opinion)the most sophisticated philosophers/scientists. important as~ts of realism and instrumentalismare pt~entll'sethtl: in sll\:h :to __ ~ Ih:t.t the tllegt\! C>..'nlndktion t-ctwt-en them \:lI\ish~.
ResumeOn oppose souvenlle realisme scientifique etl'instrumentalisme. Mais l'alternative n'est pas
c1airement formulee. On insiste particulierement sur la these que seulle realisme peut «expliquer»Ie succes des theories scientifiques et plus specialement Ie succes grandissant avec Ie temps de cestheories dans UDe IigDe coherente de recherches. Cet argument est utilise dans Ie but de clarifier Iaposition realiste, mais - ainsi que Ie montre I'auteur - ce but n'est pas atteint et la these elle-meme reste suspendue en I'air. Une serie de versions de plus en plus sophistiquees de la these «ins-trumenlaliste» sont examinees et ces versions - ainsi que les pretentions du realisme - sont con-rrontees d'une part A des moments particuliers du developpement scientifique, d'autre part A despositions qui ont ete dHendues par des philosophes et par des scientifiques. L'auteur conclut que,lorsque les positions sont evaluees sur la base de I'histoire reelle de la science a) aucune des deuxdoctrines contraires, interpretee avec une simplicite excessive, ne rend adequatement compte de ladialectique du developpement scientifique; b) chacune d'elle, interpretee ainsi, a eu des conse-quences nHastes sur les recherches de grands savants (Huygens, Kelvin, Poincare); alors que c)aussi bien dans les enonces theoriques que dans la pratique des philosophes/savants considerespar I'auteur comme les plus subtils, des aspects importants de realisme et de I'instrumentalismesont simultanement presents de telle maniere que la soi-disant contradiction entre eux s'evanouil.
• The University of Chicago, Department of Philosophy, 1050 East 59th Street, Chicago -II1inios 60637 USA
ZusammenfassungIn diesem Papier wird dafUr argumentiert, dass der vieldiskutierte Streitfall zwischen «wis-
senschaftlichem Realismus» und «Instrumentalism us» nicht klar beschrieben worden ist. Beson-dere Aufmerksamkeit wird der Behauptung geschenkt, dass nur Realismus den Erfolg wissen-schaftlicher Theorien und insbesondere den sUlndigwachsenden Erfolg solcher Theorien in einerkohllrenten Untersuchungsfolge «erklllren» kann. Diese Behauptung wird verwendet, urn eineklarere Konzeption yom In halt der ihr zugrundeliegenden realistischen These zu erreichen; aberes wird hier behauptet, dass der Versuch fehlschlllgt und die Behauptung selbst in der Luft hllngt.Eine wachsende Reihe hochentwickelter Versionen der «instrumentalistischen» These wirdbetrachtet und sowohl diese Versionen a1s auch die Behauptungen des Realismus werden zuspeziellen Beispielen wissenschaftlicher Entwicklung, sowie zu historisch yon Philosophen undvon Wissenschaftern vertretenen Positionen in Beziehung gesetzt. Die Schlussfolgerung desAutors ist, dass - wenn die Positionen vor dem Hintergrund der aktuellen Wissenschafts-geschichte beurteilt werden - (a) jede der beiden gegenslltzlichen Doktrinen bei einer Ubermllssigvereinfachten Interpreiation als Theorie der dialektischen Wissenschaftsentwicklung inadllquatist; (b) jede - so interpretiert - massgeblich zu tatsllchlichen Schllden an den Untersuchungengrosser Wissenschafter (Huygens, Kelvin, Poincare) beigetragen hat; wohingegen (c) sowohl inden theoretischen Aussagen a1s auch in der aktuellen Praxis der (gemllss Meinung des Autors)tiefsinnigsten Philosophen/Wissenschafter Aspekte des Realismus und des Intrumentalismusgemeinsam vorhanden sind, und zwar so, dass der anaebliche Oegensatz zwischen Ihnen ver-schwlndet.
By the word "skeptical" I do not mean to suggest, primarily, disbelief; myideal skeptic is Socrates, not Pyrrho. Among the claims put forward in recentyears in the name of "scientific realism" there are many things I agree with;but there is also an admixture of what seems to me unclear in conception, orunconvincingly argued. This is not so very different from what I thought, inmy student days, about the doctrines of logical empiricism - which have sincebeen pretty harshly dealt with. In the latter proceeding, I believe that somerather valuable philosophical lessons have been (at least partly) lost orobscured; and I fear that unless a sufficient ferment of Socratic skepticism iscultured within the realist brew, it will go stale and its vogue too will soonpass. Indeed, it may be late in the day to express such a fear; the pendulum hasby now swung perceptibly in the other direction; and yet I hope that it may bepossible to regain hold of what was of positive worth in positivism, and toretain possession of what is of real value in realism.
One of the critical points on which I feel myself at odds with the contend-ers for realism concerns the characterization of the central contrast under dis-cussion - that between a "realist" and an "instrumentalist" view of theories.It seems to be widely assumed that this contrast is tolerably clear. ~, on theother hand, do not see the issue between realism and instrumentali$m.as well-joirled-;J ;ll.This· opi~io~- m-a-ystrike y;u as-havi;;g ~out it a smack ofCarnap aii'd of Ernest Nagel; and that, in turn, may suggest that I myself,after all, ~m an instrumentalist (since I believe that realists do so classify bothCarnap and Nagel). Well, I have already told you that logical empiricism hasstrong merits in my eyes; but part of my plea is: let us not be hasty in applyinglabels. ..- -.
I shall invoke another honored empiricist name (and nostalgic remi-niscence): Friends of mine, when I was a student, had taken a course ofHempel's at Queens College in New York; and they reported that Hempel haddiscussed a pair of related fallacies, which he called "the fallacy of nothingbut" , and "the fallacy of something more" . My friends did not elaborate; butthis terse fragment seemed intellectually nourishing, and I still snack upon it.
How, then, is one to understand the position that theories are "mereinstruments" - that they are "nothing but" instruments? First of all, instru-ments for what? One not infrequently receives, from physics students whohave been indoctrinated with a crude version of positivism, the formulation- intended to debunk philosophical questioning - that a theory is "nothingbut" an instrument for calculating the outcomes of experiments. The tacitimplication here is that the experiments in question form a rather specialdomain: both restricted - so that theories are quarantined from any com-merce with our views of "what the world really is"; and technical - so that(to put it in its most banal form) to understand a theory is "nothing but" to beable to calculate the solutions to examination exercises. But the tacit implica-tion is false. I once heard Eugene Wigner make the point most elegantly, in alecture on philosophical problems of quantum mechanics and quantum field-theory: He remarked that it is often claimed that aii OReever does in quantumtheory is calculate the results of scattering processes. But, said Wigner, while
\
this is a/most true, it is not quite true; one also usei' quantum theory, forexample, to calculate the density of aluminum.
The trite form of debunking instrumentalism makes a false estimate of thescope of a theory as an instrument. There are, of course, broader and nar-ro~er t~o!jes;_ b~Un principle, the ~~i;ged by the~retic;}~at.;rl!1science as a whole has been, and is still,'"aiithe phenonie~li"-~fnature. Such a
-;;pe was en;i;aged-:-ror instai'iCe,exp"ii~itly(a~d with-eloq~ence) by Newton;and in our own time an enormous advan~e has been achieved towards its r a1i-~zation, since the quantum theory has (made possible an understandin , in Kfundamental terms, of the quantitative and qualitative properties of ordinary f\D)matter: the density of aluminum, the magnetizability of iron, the colors of f l'
natural bodies. 'c...! c.), j-;)/\ \
(
~)
Now, the expansion of scope beyond that of "experiments" as narrowlyconceived - to all of what in quasi-Kantian language may be called "the
~fj world of experience" - raises a certain difficulty "for the formulation of aninstrumentalist view; for it is no longer appropriate to speak of simply "cal-culating the outcomes". In point of fact, the problem arises already in the nar-rower domain, as critical discussion of the "observational/theoretical" dicho-tomy has made abundantly clear. But this difficulty can certainly be dealt with_._-~------------
~"J.l(v'';' IJ
~\.'.;'
~ - since it is in/act dealt with in scientific practice (although perhaps not withthe formal-exactness that philosophers of science tend..!o_.prefer). i shouldlike, for the-present purpose, simply to assume that a reasonable account ofthe relation of a theory to phenomena is available, and to use the phrase "thetheory affords a representation of the phenomena" for that relation. Then theinstrumentalist's claim can be read as: A theory is "nothing but" an instru-ment for representing phenomena.
From the foregoing schematic formulation of a somewhat liberalizedinstrumentalism, I turn to the realist and ask, what is it that you demand of atheory beyond this? - what "more" should a theory afford than a representa-tion - in a suitable sense correct, and in a suitable sense adequate - of phe-omena? The standard response to this question seems to me, I am afraid, noless trite and shallow than the trite and shallow version of instrumentalism: itis that the terms (predicates, function-symbols) of the theory should"genuinely refer", its quantifiers have associated domains, and its sentencesconsequently be true or false (its asserted sentence preferably true). Thetrouble with this can be seen from two sides. First, there is the classical philo-sophical problem of the epistemology of transcendent metaphysics - theproblem that Kant (we may say) posed to the ghost of Leibniz, and that ledKant himself to the view that all theoretical knowledge is 0/ the phenomenalworld: the world of experience. The nub of the-.matter can be put very simply:it is, How can you know that things-are as you say they aref'iithe claimed
! ,"refe,r~f_the,theO~Y~i,s.some~hing.b~Y. O,ndits correctness and-adequacyin representing phenomena - if, that is, for a given theory, which (we maysu-p·pose)_doesrepreseJrt phenomena"correctly lmd adequately, there are still
; two possibilities: (a) that it is (moreover) true, and (b) that it is (nonetheless)jalse - then how in ttie-~orld could we ever tell what the actual case is,?
I shall return to this question; but first I want to describe what I have re-ferred to as the other side of the trouble with the appeal to reference. It issimply that one can "Tarski-ize" any theory that is constructed in the standardway - that is, with the help of a metalanguage of suitable strength, and alsocontaining the extra-logical resources of the language of the theory T underdiscussion, one can supplement T with a theory of reference for it, T'; and cando so in such a way that whenever e is a theorem ofT, the statement that e istrue isa theorem of T'. ~, whereas the.KID1tiancritical ob'ection seeme~to put reference and truth l!tte!.ly:beypnd ktl0w!~c!8..~the Ta!.~kian theory ofreference and truth in a rather serious sense trivia/izes the desideratum putforwarcibYthe realists. (There ii;-of ·~ou~se, no paradox -in this contra~t:Tmld"s-uleory was never intended to explicate the Kantian "transcendent" or"noumenal" realm.)
I believe that we are dealing here with something that has gone seriouslywrong in the philosophical tradition of recent decades. Some thirty-eight yearsago Carnap offered his distinction between questions of existence "internal"to a linguistic or theoretical framework, and such questions construed (as onemight say) "absolutely" ("external questions of existence"); and discussedthat distinction in a way that made really quite clear the situation I have justbeen speaking of. But at about the same time there appeared Quine's attackupon Carnap's approach to the problem of "meaning"; and the trend of phi-losophical opinion went increasingly with Quine and against Carnap, in par-ticular in preferring "reference" to "meaning" as a clearer and more reliablebasis for philosophical considerations, and also in rejecting Carnap's wholenotion of a clearly defined "linguistic framework" (a rejection perhaps rein-forced by the suggestion that such a notion has a taint of "relativism" and anaffinity with the views later advanced by Kuhn). I cannot do more here thanbarely allude to this history, and suggest that it is one of the developments inwhich analyses of genuine value by the logical empiricists have gotten lost.(Hilary Putnam evidently noticed some years ago that there is merit in thosearguments of Carnap.)
,SoJa!:! then, !maintain,_.n~!~.~uehas ~r0oined: refer~.£~~ferencehas failed to stake out any claim to which a sGphisticated.lnstrumentalistsho~ld fee~ need toobject. The pointc~n be restated this way: Th~ se~anticsof reference and truth-(for agiventheory)is itself a theory. One wFo consid-ers th~orie; in general to be "mere" instruments nee<fnave no hesitation, andr--- - . -'_.' ... ~. - --_ ..- -. -'-'-.--. --_ - -no qualms of intellectual conscience, in accepting this theory as an "instru-m~nt"(th~ the innocuQu'"5ness~f;:(;allsm-in its"internal" aspect). If, on the~the;-ha~d, th~-realist objects th;t this is not what he meant by his realisticsemantics; that he meant the semantics not as a mere instrument, but asgenuinely giving the real or true reference of the theoretical vocabulary; thenthe attempt to explicate a clear issue has fallen into a clear circle: we are backin the problem of the noumena, or the fallacy of "something more" .
But I have left hanging the question how we could possibly tell whether atheory, beyond its presumed satisfactory relation to phenomena, is
\ "moreover" true or not (more havin&.b.e.epdeIJlanded - JJUt.anotclearly char-I, ~)\ a~.!.~zed'-:rnorci.:....=_~~~to._teIUvhether the desire~eeIl-deliv-
--k5- / ' It c:.~l. On this question a certain proposal has-been made recurrently, and hasJ-;- I'\.;,\ been pursued with energy in recent years by Richard Boyd: namely, .L~.::~t: p~ tha~onsideration, not simply.of individual theories i~heir se~rat_e
Jtf ,\,:l" w.•.,rel~l!.o.n.tophenomena,_but o.Uhe. connections, of theori~s_~ith ~~!:..~ngoillg, .\" process of scientific inquiry, can - and indeed does - afford evidence thatr:" l'(" \l.. our theories_have not merely instrumental but also "real"_~ali~ity:--
I a 1I, •• '
t ~l' ~." r f \""
\.f 'i ./7 '*,,,,...<.:,' I(~"'-' . Irt';~vre.w J: '" <;..U'" .••.•.. ..\.:wl ,\vh,.lOl't \ ~ \.....J-
52 Howard Stein
I ~ ~L I._I\' v h ()I':t .\. L l.<- v..l h..v- L 't>M.,. V ',..r 'f'<-.tJ ;
Yes, but. .. - Some Skeptical Remarks on Realism and Anti-Realism
My own opinion is thaUhis appeal to what may be described as the.histori-, -.--"'- ....- - -",.-\ r cal dialectic of inguiry.:::;:-and_with,~'historical" explicitly understood to imply
'\Ith~t th~~~~ ...~.:.~,~~p~~e_dt~. ~e~e~ ..P!r.-i£Ol·in~nature~ does indeed,I direCt phi!.~lil~a.tt~~<!~ .~o ~~~. ~.~. tl.t.!~~rtanussues:--Wlth 'much that\ Bo<Ycrhassaid in detail on this tOPiCI strongly agree. But as a mode of argu-
ment for "realism" over "instrumentalism", the attempt seems to me to failcompletely. What the move does accomplish so far as this ideological issue isconcerned - and accomplish without any need to argue in detail - is to pointout that the instrumentalist himself, if he wishes to do justice to the role
\ actually played by theories in science, had better extend still further his con-
Iception of what theories are instruments for, to include their role as resourcesfor inquiry; especially as sources of clues in what Peirce called "abduction":the search for good hypotheses.
Can an instrumentalist do this? Boyd appears to believe that he cannot doso successfully - that there are methodological principles of inquiry, fruitfulin practice, that are in some sense available to the realist but not to the instru-mentalist. But this clalm Is caught In a dilemma: Bither the methodologicalp~~qp'~~d thllacLQi.!!!~jUucc~.~~J~_~§jar-in!he !!!~!~ry~finquiry, are susceptible of clear. f-"E!n!J.lation...in.terms.of.Jh~u.elati~_ns ..~~~ies to pheno...nenljaoothis: I think, is what Boyd intends) - in this case,
iithe instrumentalist ought to adopt those principles, on the grounds of theII evidence for their instrumental success; or they are not susceptible of such for-
{?rjl mulation - but then, where is the argument? In short, the instrumentalist'siJ. I own principles entail no prohibitions: what the realist can do, the instrumen-
I talist can do also.. To be sure, I have distorted Boyd's central argument in an important way.
His own claim is that what the realist can do and the instrumentI!H~t_nQ.tjs;;plai~th~'~~c~;s-;cl"th~~ethod-ojoglcal-principlesihai he t!ikesJ(lbe cQl
as, or all similar to, the "real" color of the grass. It may even be urged thatwithout such an explanation, the qualitative character of our perceptionsstands as a miracle.
This mode of argument has from time to time made a powerful impres-sion, and I think there is something psychologically very tempting in it. But Isubmit that it is in fact a snare and a delusion. (Note, by the way, its totallyunacceptable conclusion that so-called "color-blind" people see false colors.)- !"ere is perhaps morOCth~o one thing wrong_witht~e-",gument; htlLON..hat_jIsee as its fatal defect I,'Sthis: that the Pos,tulate in qUestion,doesn't explain any- J
thmg. o~~-=- ,whatevet:.-!heti.utILinaib~_al?om the g~a~s- ....!!!!!~~ Ia~tu~~see wh~~ c~r.!,a.!~I!~~p.!ocess~_.occ~r" af!d thaUhese_PIQ~~ e.~e I~ate,!U!y_exchangeLQf ..eIl~rgy".and ,momentum with tJ!~~~ent.11Iherefore, ,;~!!..if w;e_~o~sum~,::-_ whate~er it maY,m!~n_- t~grass I v
re~!Y..haL.~E.0lor bke...!!teqUal~~~~_J:?,e.~~elv~!..!,~~,~~.~~of_0l!!.perccp.!,~!1dOlj ~.not follow from the postulatedJ?r_()perty,o( the,gr.?ss: bet~een the grass itS<:.!f!tan_~LQiiLii~.LCeptfo~QfJtih~(e,stilUntervene !he very PhYSi~aiZP)ly'SJOIO&~alJ/'pr~£e.sseswhose connection with the quality is alleged to be unintelligible. I
I <ifithis'case"oirectly'relevanfbecause it is one inwhich a~ion isint~d as "explanatory":'which-~d~; examination'is seen to-b~ i~-~ff;~t---- -- - _.~_.~- - ._. -.- - ----piscqnnectf?d.f!..Qm its....expl'!-'!..a-'Hj~. (The positivists liked to speak of pseudo-questions; here we have a pseudo-answer.)
(2) The second example comes from the history of science: it is theattempt by a very great physicist, Christiaan Huygens, to "explain" the pheno-menon of weight, in accordance with the metaphysical principles shared bymost "progressive" natural philosophers of the seventeenth century. (Huy-gens himself refers to "the philosophy of the present time" - or, as he alsocalls it, "the true philosophy" - according to which all natural effects are tobe explained as results of bodies in motion impinging upon one another, andthereby changing one another's motions; and he expresses the opinion that ifwe cannot provide such explanations, all hope is lost of understanding any-thing in physics.)
The theory Huygens proposes, as in his opinion the only possible one toaccount on such principles for the weight of bodies towards the earth, is this:There exists a very subtle - Le., finely divided - matter surrounding theearth, and moving very rapidly in circulation about it. Ordinary terrestrialbodies do not participate in this circulatory motion; therefore the tendency ofthe swirling matter to recede from the center has the effect upon such ter-restrial bodies of tending to sweep them towards the center - that is, towardsthe earth. But the tendency towards the earth is spherically symmetric,
whereas an ordinary vortical swirl is about an axis rather than a center; andmoreover, if the swirl were all in one direction, it could not fail eventually tobe communicated to ordinary bodies. For both these rea~ons, Huygens says, itis necessary to posit that there is not one single swirling motion, but rather acomplex of such in which swirling occurs in constantly changing directions, sothat the average state is spherically symmetric, and so that ordinary bodies donot have time to be swept one way before the direction changes and they areswept another way, The picture is of a kind of rotational Brownian motion;and Huygens mentions the analogy of heated droplets that are observe toexecute rotatory oscillations about constantly changing axes, After he hadread Newton's argument in the Principia for the conclusion that weighttowards the earth (and towards the sun and all the other planets) varies inmagnitude inversely as the square of the distance from the center - an argu-ment he found fully convincing - Huygens added to his hypothesis that thevelocities of rotation of his vortices would have to vary with distance in such away as to yield the Newtonian inverse-square variation of weight.
Now, it is really quite extraordinary that a man of Huygens's great intel-lectual clarity and power should have been in any way satisfied with this"explanation". Let us set aside any qualms about the dynamical effects that amotion of the kind proposed would actually have upon ordinary bodies; con-ceding all that Huygens asserts about this, I~,~ 'ust ask wEat in the worldcpuldcause'the matter_~[l1.!.1yge1!s's~_rtice~_to behave in such an extraor-dinary way? - One thinks of Byron's satirical lines on Coleridge:------
Explaining metaphysics to the Nation:I wish he would explain his explanationi,
Here, then, we have a case where - at least if WI': make some (possiblyunjustified) concessions - a proposed explanatory contraption might in facthave the consequences required of it; but where a ju..,dgent not s~ed ~ythe great desire to feel that it understands (an example ~hil.!J._be1ie~!s~~times called-"fiilse consciousness") will regard the explanation as quite~bit;:ary: as,soto·speak':a:-quasi-pseudo-answci;-. -------- -- - - ,
+~-"N '[/ W~Tis 'C'ommontl?~-the.!..~c~~~~ati~ of intellectual standardsJ Y:':j under what I.may call ideological p!,essure: Let me cite, by way ofconfrast, ait
. ' example of a problem that does genuinely call for explanation by a naturalisticand evolutionary epistemology such as Boyd advocates; a problem that isalmost surely not ripe for attack, but on which one may hope for greatenlightenment some day, This is the question how it is that our own naturalendowment - which has evolved for its "instrumental" value in coping withfar more immediate aspects of the world - has also proved to be an "instru-
ment" capable, under favorable circumstances, of (e.g.) discovering quantummechanics. One reason I think it imP<?Ftl!-~!J<LQ!LC~~Pf.P~.~1Jd.(L-exp1ana-if "tions i~isel~~!~~y_t~I!(Hc,t~~k rt:~_p"r.?..~!~IJl~!an~ tQ.<!1}llthemtd9.I.! JIwhat real soliillon!-~!~lk~.:__ ,
I have'said thiU think there are genuinely interesting and important issues-- ----_ .. ----'-- ---------- ---'-- -- '--- --concerning the role of t~Qri~jn_the Q.ngQiQg..m".9cess.QL inqu!!:>,. I_do ~otbelievethatth~ are issues that divide naturally,along the line "realist" /"in-~~e-;}talist". Years i.go, I wrote 'a paper - which I have not p~biWie(r; Du'tha;;;;-~a number of audiences - discussing the work of Poincare onelectrodynamics, and its relation to the special theory of relativity. In thatpaper, I pointed out what I take to have been striking failings in Poincare'sview of the proper role oftheories in inquiry. Now, Poincare was certainly aninstrumentalist; and in my discussion, I took him to task for unwillingness toregard certain things as "real". At the end of the paper, how~er:,) r~mar~d __that it should be evident to theaucIience 'that: - desPite my use of such_"iire'taphysicalT"iiotions- a;--"~ea1 ;;~.the -;hole iriatys'is 'couiCi"'-be~ec;t -in'methodologicaflerms=-jn~effeCi-,_thaUbe true issue was nof'ielilism·vs.instrumenTaJiSlil.'fhe-;ltuation in question is a doubly ironic one. 'To explainthis, let m~ single out just one of the points of physical theory involved. Thecontested reality was that of the ether; and Poincare, becauselt~,rega!ded theether as a fiction rather than a reality, was unwilling to take very seriouslyfaithough..~_~;;::;ii'ii~g:tO:PlllY'-~itl1)'the ide'a tliat cKarged pirticiese;;h~ngem;~tum with the ether. This, ho~ever, is a very odd p~~'itio~fo'r-~~instru-I~§,U9 t~~(t.he fi;;U!£!!yj;fur_[hereIs~arr~~tai"aiffn-the iqstru-m~l!Jj~t _yi~w.(orgrading_.!..h..e en~itiesof a theorY.in degrees of reality or ficti-t~~ss :- ..regarding particles as more real than the ether. This is a doublestandard that recurs in the empiricist tradition, but it is not itself a genuinepart oc'empiricism or of instrumentalism - it is, on-the co"ntrary, a bit of
____ •. ~- ---- _.' - ~ .-".,..••_,. ~_ ~_ - --'-0' • - '..• __
unregenerate realism, doing the work of the Devil among the empiricists andinsrrumentaiis~~""-- ..-' '--, ----
'the-second irony I had in mind is the circumstance that Hilary Putnam,writing as an exponent of realism, has used the notion of the ether, in contrastwith that of atoms, as an example of a scientific term that failed to "refer" -there is no such thing as the ether, whereas there are atoms (note that thisendorses the discrimination committed by Poincare); and that Larry Laudanhas exploited this as one example among others in an argument against real-ism - namely, against the view that there is any strong connection between"instrumental success" (here, of the wave theory of light and of Maxweli'selectrodynamics, both of which posited an ether) and "real reference". Formy part, I throw up my hands at this: Why should we say that the old term
"ether" failed to "refer"? - and that the old term "atom" did "refer"? Why,that is, except for the superficial reason that the word "atom" is stili used intext-books, the word "ether" not? - It would be possible to do a lengthy dia-
!rlectical number on this; but in brief: our own physics teaches us that there is. / I i \no0i.~~J!t~t.~.as~I !~~~rope.I!ies ~~i~e~.!'y ninet~~nt~~~~!~ry pt;Y;i~is!~J£f.
,,
i.\A i'" ~j'iI J~he_ethe..r:0r,(~~_ato"!s; b!;!tthat, ~n..!hc:_~!h~~_hand,.J.n bOJhJn~.ta!!c~s,!~~.er{-' ,. \ I,mportant par:t~.of. the ninet!:t:!1th-c~ntu!X!heories are co.~r~c~~For instance,
so far as Poincare's conviction that momentum -con~ervation must holdamong the "particles" taken by themselves is concerned, it is established nowbeyond a doubt that ordinary bodies do exchange momentum with "theether" - Le., with the electromagnetic ~nd even that this field has to be ~ _regarded as the seat of a distribution qf mass, ,nd as participating in gravita- \ - ~)_I tional interactions. The two cases - th:aHrl'the ether and that of atoms -'> ~vSJ
I are, in my view, so similar, that the radical distinction made between them by
I the referential realists confirms in me the antecedent suspicion that thisconcern for reference - and associated with it, another Quinean motif, the
I concern for what is called the "ontology" of theories - is a distraction from! -! what really matters.I
; I agree wholeheartedly with Boyd that we have learned - that is to .SJlY,scientists fiave-inaeecnearn;ci, in th~ir practice; ~~. in Our phU9SQ~hicalre-flec!lons_uPQ.ii:sd~[lc~;-we'should by.now have learned explidtly_- that suc-cessful scientific theories are to be taken _very.seriously as clues to the deeper
11----- --~ •. ~ ._- _. - - ~_ •.• - "'"' •• - • ~
(""II I ~;r·~nders..!~~~ing .of~hen~tnen~: Le.,,_..c;lues,inthe search for better and m~.:.-~ i fund~rpental theorIes. But thIS process of development, from the less to the
, more fundamental (Tie!erlegung derFuniamente, as, acc~rding to HerrO,ann
IWeyl, HiTbert Iiked-tZ;call it) has v~y little to do with the "referential" rela-~9.; tlons of theories or with the r~latlonships of'their_~~ntolo8Ies~3~ the sens~-of\/ , Quine. To borrow from the ancient philosophical tradition, what I believe thel1istoryof science has shown is that on a certain very deep question Aristotlewas entireiy wrong, and Plato - at least on one reading, the one I prefer -remarkably right: namely, our science comes closest to oomQrehending "the
I, '1~ea!:J.n009j.1SJ!GCJ)uo.tOe~l!Q§ta.!!~~s'~-and.thejrJcinds,Jrnt in itsJ!.~CQU!1LQ.f-,', -{ or I t.~.~"f.9tmS'~.which phenomena "imitate" (for ~Eorms~read_~~t!t~oretical, .'/ ~tructur~s", fo~ "imitate"~ "31re repr.esente.dby"). In mathematics its~lf--
concerned with Forms without regard to phenomena - the deepest generaldiscovery of the whoie modern epoch, initiated by Gauss and notablyadvanced through the work of Galois, Riemann, Dedekind, Lie, Klein,Poincare, Hilbert, and Emmy Noether, to an enormous flowering in our owntime, is that in the Hilbertian "Tieferlegung" deeper' understanding of theForms or structures of mathematics characteristically consists in their being
seen in unexpected relations to one another, or to newly discovered structures,which playa unifying or generalizing or expanding role in our understandingof the more familiar ones (the Forms "participate in one another"). And ~the development of physics in the same period - but above all in our own cen-tury =an-analogous~- in smne ways even' more astortishing discovery hasemerged; more astonfshingbecause-it involves our understanding.o~ ordinaQ!!fam;Tiar tnings, ana-because the key 'to this understanding -turns out to lie inm'athemaHcal structures'~ a~d improved understanding in structur~/uaee;;~~i~g","Ofjust su'chkind ~~haracteristically occur in pure math~maticsitself:-Aiid, I snoiilcCidd, more astonishing not leasCbecause ofthe-Cii:cumstance Iam here concerned to emphasize, that !!!J!ti~,~!:.uctu.riltde~p~ningwhat tends to persist_-=_to remain, as, i~~ere, quas~:!nvariant th~ough thetransformation of theories - is on the whole (and especially in wha:twe t~ink G..of as the -"deepest" - or most "revolutionary" - transformations) not the 1,1featuresmosCc'orispicuous in referential semantics: the substances or "enti- 'fc,/'
ties" an'! !heir own "basic".J)roperties and relations,bu~,th~._~;re ~~t~~t( ~,'I\J1rnatMmatical forms. 1 N(' ..._---,,'--
What Ihave just said is pretty general. Let me give two examples of what Imean: the first a very simple one, which Ihope will serve for clarification; thesecond, one of contemporary relevance.
(1) Newton shocked the natural philosophers of his time by his claim tohave discovered, and proved, that sunlight is a mixture of physically distinctkinds of light. His formulation of this result was extremely careful, andremains entirely cogent today. The basic "entity" in his analysis is what hecalls a "ray" - by which he means not (as we ordinarily do) a line of propaga-tion of light, but rather an "elementary part of light", conceived as spatiallylocalized and as propagated along a definite trajectory: sunlight, he says,contains rays of indefinitely many distinct kinds, which may be separatedfrom one another, and which prove to have distinct and immutable physicalproperties; among these distinguishing properties are distinct indices of re-fraction (for a given medium), powers of stimulating sensations of distinctcolors, and a certain quantity having the dimensions of a length (which weknow as the wave length).
Now light, as we have come to understand it, whether we "represent" it bymeans of the classical wave theory of Fresnel, the electromagnetic theory ofMaxwell, or (now most fundamentally) by the quantum theory of the electro-magnetic field, does not have "parts" that are propagated along definite tra-jectories. One ca~~<?L.th~JightJhat,reaches a certai~point P atA<;er-tain time t, that at an earlier time t' this same bit of light was at a point_Q;Jh~,
. ,.....,.-- .-- .-._-------_.
notion of identitY..21Jg.fQ!i!.@p(}rts tlJ.rough time is simply inapplicable at allt~!-,-
~~fore, _~~e~i~~~r~as to_~~at~~'Y.~on~tler~[aLf~i1s to "refe~",or one has to strain to find some way to associate that term with an "entity"~E_~<?f the l_a!e!.and_d_e_~pe~,theories-= at b~st_~~~!anti.~and unre~ardi;g'E0~J On the other hand, it is anessentiaftasko-f physical understanding, '"i'Oda most instructive one, to exhibit the formal relations of the Newtonian theoryof rays to the later theories. One finds, when one does this, that there areseveral different structural aspects of the later theory, each of which has a sig-nificant bearing upon the success of the notion of a ray in serving to representthe optical phenomena dealt with by Newton. Thus the principle of "cuttingat the joints" - a motif in recent realism that Iheartily applaud - seems toindicate not using referential semantics as the basis of comparison of theories(at least in the case cited; but Ioffer it as typical).
(2) A number of attempts have been made to give a so-called "realisticinterpretation" of quantum mechanics. These have characteristically takenthe form of specifications of "realistic semantics" for the theory. The res~ts
() seem to me, in each_~ase,n<!,tonlIJI.l!~su~iv~,-P!lL_::.}Yltat iLI1!,l,l~hw.Q.fse-\f fundamentally uninteresting. (The author of one of those theories once') eX'J)ressed-iQ'me'the "app~~hension that his attempt might be subject to the
criticism said to have been pronounced by Pauli on another proposal: Pauli'sremark was, "That theory isn't even wrongl") My reason for regarding theseessays in quantum realism as uninteresting is this: All the "realist" interp!:rta-tion~akejlJor.grant~~t~~~t ce,!ai!!_H~ecifi<;_~nstit!lentsof the mJ!t.P.sDli\ti~alstructure of the theory - the ro' ection 0 erators - are to be in!enm~t~d-as
I \ \ 1 standing f?r_~!:H!_'!!!!.!?f c:~_a.!n~asj~':e..n!it!e~":'.<parti~k~l;_bec.ause..1haLis~\~ \' ~hat they correspond t~!n the classisal theory.jJutJ! examines carefully
,,)'. ' 01" \ho~ phe":~m!n~ -~presen.!~E.:J>.J_~~<J!l1!-.!1t_~m_theory -_a_nd ttljs\. ,.. requires attention not only to what hysicists say but to how the physics• \ t .___ _-.... --~ _
\.,i jf_ \ ~ctually wo~hen p~eci§ely.,tbisjnj~rp.ret!ill_Qnjn)e..rmsof "entities"~nd, '\'1 \, '2, "attributes':~an be seen to beE!ghly d~bi~~. I do not claim to have a defini-nl,\",r, \ tive formulation of the meta-physics of quantum mechanics; but I believe
)[rather strongly that the difficulties it presents arise from the fact that theJ1 mode in which this theory "represents" phenomena is a radically novel one. ~
\-:: other words, I think the Iive_p!:'oblemsconcern the relation of the For~ -indeed, if you lik~~heJm1.!:!lment - to phenomena, rather than the rela-tion of (putative) attribu es to (putative) ~ntities, and that the ideological~Oiiv'esol'j'iealism" ha.y.e...here,_s~s_a kin of scholasfic distraction,-- - - -- ---turning attention away from what is "real" in the subject.- . -- -.~-----'---~. -
I have so far discussed at far greater length what I see as ideological anddelusive in the claims of recent realism than I have the corresponding faults ofanti-realism; the latter I have only touched on in speaking of the shallowerversions of instrumentalism. Yet this last - a shallow instrumentalism -seems to me no less inimical to sound philosophic understanding. Since I havementioned the case of Quantum mechanics in the former connection, it is onlyfair to cite it again here. I have just referred to the "live problems" of philo-sophic understanding posed by Quantum mechanics. There are those whoh~e argued that no such probJems exist in fact - on the grounds that, inconcrete cases, rhysicistsagrec::,Q,p'-hq'Y.::l)}tftlieoryiStOoeappliea "topl1eno-mena, and!liai is all we are entitled to expec~Q.fii]Ji.~Qry.--~- -- -
t is certainly not a contemptible argument - not one to be dismissed outof hand. But it is necessary to use some circumspection in evaluating it. Therelevant Question is closely related to the criticism I have earlier made of the
-LL.,.J:hysicsstudegt":'s~,!, against philosophy: How !!!.<ug.eQIY.JUQP~t!l[lp.Ued)Vto which henomena? (There is a wonderful story - whether true or_._~
apocryphal I do not know - about a discussion between a distinguishedmathematician and a very distinguished physicist on the problem ofaxio-mati zing Quantum field theory. The physicist said, so the tale has it: "Why dowe need axioms? The job of physics is to calculate the results of experiments.If you describe an experiment for me, I can perfectly well calculate itsoutcome." To which the mathematician replied - in a Polish accent thathelps the story -: "Dot's very nice, Dick; but vot viii ve do after you?" [asituation in which, lamentably, we now find ourselves).) The point is thatQuantum mechanics - or its more fundamental, but, to date, from amathematical point of view less clearly formulated modification, relativisticQuantum field theory - is one of the two basic and universal theories of con-temporary physics (general relativity is the other); its presumptive scope,therefore, is all the phenomena of nature. But we genuinely do not under-stand, not merely how in detail to apply this"theory to all natural phenome;a,....••.-------_ ...._.-. --..--~~t in principl~. how such app ~~~y be.p'o~sible; for instilffce;--nol5iie!really knows how Quantum mechanics could be applied to the universe as awhole; and yet such application seems to be necessary if there is to be a uni-fication of Quantum theory with general relativity. And this difficulty, whichmay well strike you as after all something Quite technical that ought to J "',concern physicists rather than philosophers, is intimately connected with thfamous problem of what is called the "reduction of the wave packet", whichas been at the center of philosophical discussion of the theory from it~beginning - and which is precisely the problem that the instrumentalist moveI have deprecated is intended to exorcise. II
Note again that - as in the case I have mentioned of Poincare - thecriticism I have just made need not be put in "realist" terms. For myself, I dotend to think in such terms - to ask what the theory implies about the world'sreal constitution. But the roblems I have criticized the instrumentalist a!.&u-
tvJ\ment for evading are if I am r~ti!!J:~garding thell.l.as~j;.t~-1X \ "!.entalistproblems. Indeed, I said at the outset that the issue between realismand instrumentalism seems to me not to be clearly posed; and what I reallybelieve is that between a cogent and enlightened "realism" and a sophisticated
\~
"instrumentalism" there is no significant difference - no difference thatmakes a difference. I shall try, in my remaining remarks, to clarify this claimand make it plausible.
Let us briefly consider the idealism of Berkeley: the doctrine that what isreal is just minds and their perceptions, and that all our beliefs about thephysical world are just "instruments" for organizing and anticipatingexperience. Can this doctrine be refuted? Need this doctrine be refuted?
It is an interesting point that we are unable to fotmulate our informationabout the world in a way that restricts reference to that which is - or eventhat which is possibly - perceived (the failure, more or less recognizedeyentually by Carnap, of phenomenalism as a "basis" for a scientific lan-guage). But although this is a fact of great importance for the methodology ofscience and of philosophy, it does not refute Berkeley. It can even be .construed as implying a peculiarity of our conceptual faculty, rather than of"the world" (cf. Kant).
Can we argue that only some further contraption behind the appearancescould - or could with any likelihood - produce regularities in phenomena of ))the kind represented in our beliefs? Why is such an argument any mor~~lplausible than that to a "mechanical" cause of gravitation, in the sense of :;.ritVyHuygens; or to a material ether supporting the electromagnetic field? I hav ~.\~alluded to both these scientific issues earlier; although it may seem to be far\U"fo>fremoved from the Question I have raised about Berkeley, I should like to dis-\ ~'1~ Xcuss the electromagnetic case in just a little more detail now. ~: ~Il
When Maxwell proposed his great theory, he called it a dynamical theory ~.\ \of the electromagnetic field; and explained that he used that term because the '6.J; r
theory postulated the existence of matter in motion in all the regions within V\ ("r"which electrical and magnetic interactions occur - in the regions, that is, be- (tween and around the electrically charged bodies and electric currents. In hisfirst attempts to work out a theory of this kind, he made, tentatively, detailedassumptions about this matter, that is about what wes called the "ether"; buthe regarded these hypotheses as mere aids to the development of the theory,not as plausible proposals about the true constitution of the ether. When the
~. theory was fully worked out, Maxwell uncoupled it from the earlier detailed\ ether model, and proposed instead - with some success - to subsume his
~
theory under the principles of the generalized dynamics of Lagrange, which. made it possible to state dynamical laws for a system whose internal constitu-
tion was not fully specified. (I say with some success: Maxwell carried out thisprogram only partially; his work was later extended, first by G. F. FitzGerald,and then, to its full completion, by H. A. Lorentz.)
Now, Maxwell's own view of this situation was unquestionably realistic.Here is what he says in the last paragraph of his Treatise on Electricity andMagnetism, after reviewing a number of theories, alternative to his own,which postulated the transmission of energy in some form with a finite veloc-ity through space from particle to particle:
But in all of these theories the question naturally occurs: - If something is transmittedfrom one particle to another at a distance, what is its condition after it has left the oneparticle and before it has reached the other? ... In fact, whenever energy is transmittedfrom one body to another in time, there must be a medium or substance in which theenergy exists after it leaves one body and before it reaches the other, for energy, as Tor-riceiii remarked, 'is a quintessence of so subtile a nature that it cannot be contained inany vessel except the Inmost substance of material things'. Hence all these theories leadto the conception of a medium in which the propagation takes place, and If we admitthis medium as an hypothesis, I think it ought to occupy a prominent place in ourinvestigations, and that we ought to endeavour to construct a mental representation ofall the details of its action, and this has been my constant aim in this treatise.
didn't know proved to be impossible to construct at all in a satisfactory wayon the basis of a mechanical ether in the traditional sense; but those moreabstract principles of dynamics under which' Maxwell subsumed the laws ofthe field have been taken over, first into the relativistic dynamics thatessentially established the impossibility of a traditional one, and then intoquantum electrodynamics. Of the principles of Lagrangian dynamics - thatForm which has survived these radical transformations of physics - White-head was moved to declare that "the beauty and ~\most divine simplicity ofthese equations is such that these formulae are worthy to rank with thosemysterious symbols which in ancient times were held directly to indicate theSupreme Reason at the base of all things" .
Now, 1 have defended the "reality" of the ether - in a certain sense -against the claims both of Putnam as a realist and Laudan as an anti-realist;and 1have quoted the undeniably realist sentiment of Maxwell on the subject.Yet, as 1 have said, the physicist who completed Maxwell's subsumption ofthe laws of the field under the Lagrangian principles was H. A. Lorentz.Lorentz did this in his celebrated memoir of 1892 on electrodynamic pheno-mena in moving bodies - where the motivation he cites for carrying out Max-well's program can fairly be described as an "Instrumentalist" one. He refersthere to the theoretical work of Heinrich Hertz on ~he same subject; notes thatHertz hardly concerns himself with a mechanical account of processes in thefield (it should be remembered that Hertz said, in this connection, that "Max-well's theory is Maxwell's system of equations"); comments hat this methodhas its advantages, but that on the other hand "one is always tempted to revertto mechanical explications"; and then makes the crucial remark that his ownpurpose, which was to develop the theory for bodies that move through theether without any entrainment of the latter, "seemed to me difficult to achieve... without having as guide some theoretical idea. The views of Maxwell[namely, on the Lagrangian dynamics ofthe field) may serve as foundationfor the theory sought".
How does this "instrumentalist" view of the useiulness of theoretical ideasas guides for research - guides, it may be noted, to be used opportunistically:Lorentz in no way implies that mechanical representations are a sine qua non,and on other occasions he denies this explicitly - how does this view squarewith the "realist" commitment of Maxwell? Here is a methodological state-ment of the latter, taken from his article on "Attraction" in the ninth editionof the Encyclopaedia Britannica:
[W)e must bear in mind that the scientific or science-producing value of the effortsmade to answer these old standing questions is not to be measured by the prospect theyafford us of ultimately obtaining a solution, but by their effect in stimulating men to a
A bit confusing perhaps, since 1 have said that Maxwell eliminated thedetailed model of the ether from his developed account of the theory. But the"details of its action" that he means, in this statement, are those detailed c~~ "-----". - -- '._-cepts by which he represen s testate of the electromagnetic field itself. -concciY.ed.-J.rrti~.as..:representing the -stat;-o-f th~her:b~'illout a
I ~led account of how the fieW'QliiOtitles are determined-bY'a1run~ng-mechamca:tfi1e(ffum - and those detailed laws tnatgovern the behavior of the~-------field, and that express the kinetic and potential energy of the presumedmedium as functions of the field quantities. This purgation of the theoryfrom other details about the medium, this rende~it in a more "abstract"orgeneral form, was d~;by Maxwell because o'f anothe~ princil'le_he cher-iShea:thatOf dlstmgUlsning, 'asfil-ras onecan, bet~~eri 'what j;'k-nown with~me security, or held at I~wiih-some probability, and ;hati~;:e-and
'\evenImpiausible conject~~~(this- principle he explicitly associates WithINewt6n:f---'--' .- ---_ ,-_.-_. ,.'
1 have already intimated, in my earlier remarks about Poincare, and aboutPutnam and Laudan, what the sequel was: The details that Maxwell knew he
thorough investigation of nature. To propose a scientific question presupposesscientificknowledge,and the questionswhichexercisemen's mindsin the presentstateof sciencemayvery likelybe such that a littlemore knowledgewould shewus that noanswer is possible. The scientificvalue of the question, How do bodies act on oneanother at a distance?is to be found in the stimulusit has givento investigationsintothe propertiesof the interveningmedium.
This dialectical tension, as it may be called, between a realist and an instru-mentalisrattit~e, _~"iisting-togetherwithQuTcoiltradiction:seems to me char-acteristic of the deepest scientists. B;sides Maxwell, whom I have quoted oniioth ;;iOe-s~tis-very-cl~ll,rfymarufeSted in Newton, for example - and in Ein-stein. I have already mentioned Poincare as an example of a scientist - agreat scientist, and a philosopher - whose work in physics suffered from animbalance on the instrumentalist side. Two examples of great physicists whosework suffered to a degree from the opposite imbalance are Huygens - whocould not accept Newton's theory of gravitation because he saw no possibilityof a satisfactory "mechanical explanation" of universal attraction, and Kelvin,-- Maxwell's good friend, closely associated with Maxwell's own develop-ment of his electromagnetic theory, who to the end of his life - in 1907,morethan forty years after the first publication of Maxwell's theory, maintained avacillating and somewhat negative attitude towards that theory because nosatisfactory mechanical model of the electromagnetic ether had been found.
Back then, once more, to Berkeley. Can one argue that if the natural worldconsisted in nothing but the perceptions of minds, it would be a "miracle"that these perceptions follow rules formulated in terms of the "entities" conc-eived by science - or by ordinary common sense? Why should this seem any I !more miraculous than the universal gravitational attraction of bodies (which I'was, indeed, rejected by Huygens, and by Leibniz as well, on just such Igrounds)? Or how is it more miraculous than the fact that such an apparently I'"abstract" thing as the electromagnetic field just does satisfy the equations of i 'Maxwell and the dynamics of Lagrange, without anything more substantial Ithat "underlies" it? Indeed, if one examines the explanation offered by Iscience today for the existence and properties of aluminum (to revert to anearlier example), it is hard not to feel that this explanation, at least as much asBerkeley's, grounds ordinary things upon a miracle. But the simple fact is thatwhatever our science adopts, perhaps provisionally;lls 'its "ultimat<~'''prin-ciples, just because they have no further ground, remain "inexplicable"; andthe farther they are from the familiar, the more they will seem "miraculous".
What we are left with is that other, provisiomlIly, "ultimate" or unex-plained fact, that we do find ourselves compelled to formulate our beliefs innli:!-phenomenalistic terms; and in this process, atoms, electrons, fields, and
the like are in a case quite analogous to that of chairs, tables, and the like in\ Berkeley. The justifiable claim to "reality" possessed be those "theoreticalI entities" is of the same kind as the justifiable claim - not after all denied by
;\ II Berkeley - of these ordinary objects. To hold this is to reject the faulty-em pi--)1 dcist double standard; and "realism" in this sense I endorse unreservedly.V Berkeley, at this point, in my view, has been, not refuted, but seen to be irrele-
\
vant to any real issue in the understanding of science. And my general conclu-sion is: Realism - Yes, but ... instrumentalism - Yes also; No only to anti-realism.