Realism without the Real Larry Laudan Philosophy of Science, Vol. 51, No. 1. (Mar., 1984), pp. 156-162. Stable URL: http://links.jstor.org/sici?sici=0031-8248%28198403%2951%3A1%3C156%3ARWTR%3E2.0.CO%3B2-I Philosophy of Science is currently published by The University of Chicago Press. Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/about/terms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/journals/ucpress.html. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. The JSTOR Archive is a trusted digital repository providing for long-term preservation and access to leading academic journals and scholarly literature from around the world. The Archive is supported by libraries, scholarly societies, publishers, and foundations. It is an initiative of JSTOR, a not-for-profit organization with a mission to help the scholarly community take advantage of advances in technology. For more information regarding JSTOR, please contact support@jstor.org. http://www.jstor.org Sat Oct 20 18:47:44 2007 http://links.jstor.org/sici?sici=0031-8248%28198403%2951%3A1%3C156%3ARWTR%3E2.0.CO%3B2-I http://www.jstor.org/about/terms.html http://www.jstor.org/journals/ucpress.html DISCUSSION: REALISM WITHOUT THE REAL* LARRY LAUDAN? Center for the Study of Science and Society Virginia Polytechnic Institute and State University 1. Background. Clyde Hardin and Alexander Rosenberg recently pub- lished a critique (1982) of an essay of mine (1981) in this journal. Where I had sought to raise some doubts about the ability of scientific realism to explain some key features of science which had been said (by Putnam, Newton-Smith and Boyd) to be uniquely explicable by realist epistemol- ogy, Hardin and Rosenberg maintain that realism is amply provided with explanatory resources for handling the sorts of cases I discussed. It ap- pears to me, however, that the Hardin-Rosenberg approach buys what explanatory power it can claim at a high cost; namely, by a more or less wholesale repudiation of much that the scientific realist holds dear. In- deed, it seems to me that the position they actually defend is so attenuated a form of 'realism'-if it be realism at all-that it is scarcely distin- guishable from that of an instrumentalist. In this brief reply, I want to show why Hardin and Rosenberg, despite their intentions to the contrary, seem to be cutting the ground out from under classical scientific realism. Before I can deal with their arguments in detail, however, some back- ground needs to be sketched in briefly. Scientific realism, at least as de- veloped by Sellars, Putnam, Boyd, and Newton-Smith, seems to rest on the following epistemic intuition: the theories of the "mature sciences" (e.g., contemporary physics) are sufficiently well-tested that it is reason- able to assume that the world is substantially the way those theories say it is. This realist construal extends not only to the 'directly observable' or 'directly testable' parts of our theories, but equally (one might almost say especially) to the explanatory and often 'unobservable' portions of our well-tested theories. I shall call this view the thesis of hard-core sci- entific realism. That thesis rests, in turn, on the legitimacy of a particular form of inference, often called 'inference to the best explanation'. Vir- *Received March 1983; revised August 1983. ?This research was made possible by a Sustained Development Grant from EVIST, supported by the National Science Foundation and the National Foundation for the Hu- manities. Philosophy of Science, 51 (1984) pp. 156-162. Copyright O 1984 by the Philosophy of Science Association 157 REALISM WITHOUT THE REAL tually all the arguments for the thesis of hard-core realism involve vari- eties of inference to the best explanation. Indeed, it is only because the realist sets such store by inference to the best explanation that he is able to motivate the characteristic realist inference from the well-testedness of a theory to the claim that there is a warrant for accepting it as true. In the essay of mine to which Hardin and Rosenberg take exception, I challenged the intuitions which motivate the realist enterprise by arguing (among other things) that many (now discredited) scientific theories of earlier eras exhibited an impressive sort of empirical support, arguably no different in kind from that enjoyed by many contemporary physical theories. Yet we now believe that many of those earlier theories pro- foundly mischaracterized the way the world really is. More specifically, we now believe that there is nothing in the world which even approxi- mately answers to the central explanatory entities postulated by a great many successful theories of the past. My approach in that essay, which we might call the historical gambit, was to show that these historical cases call into question the realist's warrant for assuming that today's theories, including even those which have passed an impressive array of tests, can thereby warrantedly be taken to be (in Sellars' apt image), "cut- ting the world at its joints". In the process of elaborating this historical gambit, I spent some time examining certain apparent flaws both in the realist theory of reference and in the realist semantics of approximate truth, flaws which it is not necessary to summarize here. Finally, I tried to show in a general way that the historical gambit exhibits the unrelia- bility of inference to the best explanation as a strategy for warranting truth claims on behalf of scientific theories. Hardin and Rosenberg reply by asserting that the realist has two plau- sible responses to the historical gambit: (a) the realist can readily concede, if he likes, that the central explanatory concepts of many superceded the- ories have failed to refer (in the denotative sense) and yet still maintain his belief that those theories were 'approximately true' and empirically successful; and (b) that the realist, by adopting a functional, non-deno- tative theory of reference, can even hold that the central explanatory con- cepts of past theories genuinely "referred", despite the fact that there is apparently nothing in the world which even approximately answers to the descriptions which those theories gave of their central concepts. I think that both of those suggestions are ill-advised, precisely because they undermine the very thesis of hard-core scientific realism which they were apparently intended to advance. The object of this reply is to show how (a) and (b) fail to enhance the case for scientific realism. 2. The Divorce of Reference and Success. Scientific realists have long maintained that, if a theory is a highly successful explainer and predictor 158 LARRYLAUDAN of phenomena, that is, if it has stood up to a demanding battery of em- pirical tests, then it is reasonable to assume that the world's structure is substantially the way that theory claims it to be. As Sellars once ob- served, the realist is one who says, "if the kinetic molecular theory is highly confirmed, then there really are molecules". Hardin and Rosen- berg are evidently prepared to jettison this key tenet of scientific realism, at least with respect to theories of the past. Specifically, they are ready to say that there can be theories which have been highly successful, even "approximately true", even though there is no reason to think that there is anything like the basic entities which those theories postulate. Their example is Mendel's theory of the gene and the phenotype. As they point out, "there is nothing like genes, or phenotypes, as Mendel or his immediate successors construed them" (1982, p. 606); ". . . genes can reasonably be asserted not to have existed" (1982, p. 606); "there is nothing, no one thing, like the gene" (1982, p. 606). Hardin and Ro- senberg say that Mendel's theory can be admitted by a realist to be non- referring and yet still be held to be "approximately true", on the grounds that Mendel's theory (including the laws of segregation and independent assortment) is "close to the truth". They thus see in Mendel's theory a counter-example to my claim that the realist must construe any successful and approximately true theory as a genuinely referring one. Quite what they mean by saying, let alone what grounds they have for holding, that Mendel's theory (as opposed to his laws) is "approximately true" is left opaque. But I do not want to repeat earlier arguments of mine about the technical difficulties confronting the semantics and criteriology of approximate truth. I think it would be more instructive to look at the far-reaching implications of the Hardin-Rosenberg claim that a realist can readily grant that many highly successful theories (such as Mendel's) failed to get things right at the level of basic structures. For my part, I would be the last to quibble with the claim that successful theories need not be genuinely referring; indeed much of my paper was devoted to docu- menting that claim. What I find curious is the belief of Hardin and Ro- senberg that a realist can, consistently with his general epistemic outlook, happily acquiesce in the divorce of empirical success and referential pre- sumption. It seems clear that if we once accept that theories can be, indeed that they have been, highly successful even when their ontological underpin- nings are very wide of the mark, then we have completely undercut the (Putnam-Boyd-Newton-Smith) claim against which my paper was di- rected that only by accepting a realist construal of theories, can we ex- plain why those theories (in particular, and science in general) are suc- cessful. If theories can be successful even when their stories about fine 159 REALISM WITHOUT THE REAL structure are strikingly at odds with what there is, then how, if at all, can realism "de-mystify" the "miraculously" successful character of science? The Hardin-Rosenberg answer is apparently this: by accepting contem- porary biology as true, and by showing that many of the consequences of contemporary biology are congruent with many of the consequences of Mendel's theories, we can thus 'explain' why Mendel's theories worked, even when they did not secure reference for their basic explanatory con- cepts. But that is precisely to beg the question. For where is the warrant for presuming that contemporary biological theory has got things right at the level of deep structure? Indeed, so soon as Hardin and Rosenberg acknowledge, as they are at some pains to do, that empirical success does not betoken genuineness of reference or truthlikeness, then it becomes unreasonable for them to take contemporary theories as benchmarks of what there is and how it behaves. Having once granted that (as they do vis a vis Mendel's theories) that inference to the best explanation is sus- pect, it appears unseemly to throw those suspicions to the wind when we apply that form of inference to the science of our own time. If empirical success (or its presumed epistemic counterpart, approximate truthlike- ness) is not co-variant with genuineness of reference, then there is no reason whatever to believe that the world is the way our most successful theories say it is. And that, in turn, makes the thesis of hard-core sci- entific realism a non-sequitur, for that thesis asserts the legitimacy of assuming that the world is pretty much as our best-tested theories say it is. Realists like Putnam and Boyd have resisted the Hardin-Rosenberg manoeuvre (viz., allowing that empirical success may not betoken gen- uineness of reference), and well they should; for if the realist once con- cedes that empirical success is not a reliable litmus for authenticity of reference, then he is no longer in a position to argue-as is the realist's wont-that the empirically successful theories of our time can warrant- edly be held to be genuinely referring. And without that baseline as a given, he is in no position to use contemporary theories as the touchstone for deciding on the referential status of the explanatory concepts of earlier theories. Curiously, Hardin and Rosenberg are quite happy to resort to the latter strategy (witness their use of contemporary genetics to adjudi- cate whether Mendel's concepts refer), without realizing that their divorce of reference and empirical success undermines the rationale for that strat- egy. For such reasons as these, it seems reasonable to conclude that, what- ever merits may attach to the divorce thesis (and non-realists are likely to find it attractive indeed), it does little to strengthen the case for sci- entific realism. 160 L A R R Y L A U D A N 3. The Functional Theory of Reference. The second approach to ref- erence which Hardin and Rosenberg explore is rather different. Here, one does not deny that a concept in a theory refers 'merely' because nothing satisfies the theory's description of the concept. Rather, one stipulates that the concepts in an outmoded theory can legitimately be taken to refer just in case we now accept the existence of entities which play many of the same causal and explanatory roles as the concepts of the outmoded theory. On this analysis, one says (for instance) that the Daltonian atom "referred" because, even though "nothing answers to the description of . . . the Daltonian atom" (1982, p. 613), it was invoked to explain many of the same things which our conception of the atom is used to explain. Similarly, Hardin and Rosenberg argue that we can legitimately say (counter to my claim) that the ether concept 'referred', since it was invoked to explain many of the same phenomena which we now explain by the elec- tromagnetic field. Indeed, as they would have it, "'ether' referred to the electromagnetic field all along" (1982, p. 614), because of the similarity of causal roles played by the ether in classical physics and the electro- magnetic field in contemporary physics. The point of elaborating this account of reference, which radically sep- arates reference from denotation, is that it allows Hardin and Rosenberg to refute my charge that many successful theories have had central ex- planatory concepts which were evidently non-referring. With this func- tional theory of reference, they are now in a position to say that, since many of my examples were drawn from the history of ether theories, one need not accept my characterization of ether theories as non-referring. (Incidentally, I took my characterization of ether as non-referring from the then-realist, Hilary Putnam (1978, pp. 20-22), who offered it as a prime example of a non-referential concept.) Two things are wrong with this functional theory of reference. In the first place, it proves to be far too tolerant for the realist's purposes, since it countenances as genuinely referring all manner of ill-developed theo- ries. It can be argued, for instance, that Aristotle's conception of natural place plays many of the same causal and explanatory roles as Newtonian gravitational forces. Cartesian vortices perform a similar function. Yet is anyone seriously prepared to maintain that Aristotle and Descartes, whose conceptual frameworks were so antithetical to action-at-a-distance, were really 'referring', had they but known, to gravitational attraction? What Newton, Descartes and Aristotle do have in common is a conviction that there is something which causes heavy bodies to fall. And in a trivial sense, they are thus all referring to "the cause of fall". But that establishes no interesting commonality of reference at the level of explanatory struc- ture. Secondly, and more crucially, the Hardin-Rosenberg account of ref- erence confuses a shared explanatory agenda (i.e., common problems to 161 REALISM WITHOUT THE REAL be solved) with a shared explanatory ontology (i.e., the characteristics of the postulated explanatory entities). The fact that (a) two theories address (many of) the same problems and (b) that those theories group the problems to be solved into similar 'bundles', is surely not sufficient to establish (c) that the explanatory on- tologies of the two theories 'refer' to the same entities. Especially for a realist, what a theory is thought to 'refer to' must have some connection with the existential claims the theory makes about deep or fine structure. To make reference parasitic on what is being explained rather than on what is doing the explaining entails that we can establish what a theory refers to independently of any detailed analysis of what the theory asserts. If, to save his theory of reference, the scientific realist is forced into such voodoo semantics, then realism scarcely seems worth the candle. But that is only the beginning of the problems confronting the func- tional theory. The major objection to this approach is that, like the di- vorce thesis, it repudiates what I earlier called the hard-core realist in- tuition. By insisting that a theory may be highly successful and that its central concepts may successfully 'refer', even when there is nothing in the physical world relevantly like the entities postulated by the theory in question, Hardin and Rosenberg have conceded that the deep-structural claims of a theory can be systematically wide of the mark, although the theory itself is highly successful at those levels where it can be tested. Which is to say that there is no co-variance between the empirical success of theories and the correctness of their deep-structural claims. But to grant that is just to concede that "inference to the best explanation", which is the realist's core inferential strategy, is badly flawed. It seems to me that Hardin and Rosenberg have not yet realized just how much of realism they are asking us to give up. They treat it as com- pletely unproblematic, for instance, that we can 'read off' a correct ac- count of the structure of the cell from contemporary biological theory. Indeed, it is precisely because they take modern genetic theories as cor- rect that they can argue that Mendel's account of the gene is so badly mistaken. Yet where is their warrant, as realists, for treating our under- standing of the gene as referentially kosher? So long as the realist could utilize 'inference to the best explanation', he could at least consistently argue that modem theories are more nearly right than their predecessors. But, having conceded that genuineness of reference is not a sine qua non for empirical success, Hardin and Rosenberg no longer have a license for treating our theories as any more likely to be referentially sound than those of Mendel. In sum, if the realist once gives up the claim that impressive empirical success is a warrant for presuming the existence of the explanatory en- tities postulated by successful theories, then scientific realism becomes a 162 LARRYLAUDAN semantics in search of an epistemology. And the realist's claim to possess an epistemology with some novel explanatory content becomes, in the Hardin-Rosenberg analysis, an ever more utopian pipe-dream. REFERENCES Hardin, C. & Rosenberg, A. (1982), "In Defense of Convergent Realism", P h i l o s o p h ~ ~of Science 49: 604-6 15. Laudan, L. (1981), "A Confutation of Convergent Realism", Philosophy of Science 4 8 : 1-49. Putnam, H. (1978), Meaning and the Moral Sciences. London: Routledge & Kegan Paul. You have printed the following article: Realism without the Real Larry Laudan Philosophy of Science, Vol. 51, No. 1. (Mar., 1984), pp. 156-162. Stable URL: http://links.jstor.org/sici?sici=0031-8248%28198403%2951%3A1%3C156%3ARWTR%3E2.0.CO%3B2-I This article references the following linked citations. If you are trying to access articles from an off-campus location, you may be required to first logon via your library web site to access JSTOR. Please visit your library's website or contact a librarian to learn about options for remote access to JSTOR. References In Defense of Convergent Realism Clyde L. Hardin; Alexander Rosenberg Philosophy of Science, Vol. 49, No. 4. (Dec., 1982), pp. 604-615. Stable URL: http://links.jstor.org/sici?sici=0031-8248%28198212%2949%3A4%3C604%3AIDOCR%3E2.0.CO%3B2-W http://www.jstor.org LINKED CITATIONS - Page 1 of 1 - http://links.jstor.org/sici?sici=0031-8248%28198403%2951%3A1%3C156%3ARWTR%3E2.0.CO%3B2-I&origin=JSTOR-pdf http://links.jstor.org/sici?sici=0031-8248%28198212%2949%3A4%3C604%3AIDOCR%3E2.0.CO%3B2-W&origin=JSTOR-pdf