What is "true" in internal realism?

Enrahonar 25, 1996 69-90 What is «true» in internal realism?' José Medina Northwestern University. Department of Philosophy Evanston, Illinois 60208-1315. USA. Data de recepció: 111011995 Abstract This paper is a criticai examination of Putnam's theory of truth as it evolves from metaphysical to interna1 reaiism. First, 1 analyze the model-theoretic argument that led Putnam to abandon the metaphysical concept of truth as correspondence and to adopt and epistemic view of truth. Though a powerful critique of the metaphysical realist conception of truth, this argument does not establish conclusively that the concept of truth has any epistemiccontent. Secondly, 1 discuss Putnam's ideaiization theory of truth, arguing that the identification of truth with nacceptability under ideal conditions)) is at odds with the claim that truth is context-transcendent, since the notion of justificarion is intrinsically context-dependent and no amount of idealization can reedeem its contextuai character. Finally, 1 suggest that the realist intuitions that Putnam's internai realism tries to capture cal1 for no more [han a deflationary view of truth. Acceptance of this view requires abandoning not only the idea that truth is an epistemic property, but also the idea that truth is a substantiueproperty that al1 true statements share and, therefore, a proper object of philosophical theorizing. Key words: truth, Putnam, internal reaiism, acceptability, metaphisical realism, justification, idealization. Resum. Quk és ((ueritatii en el realisme intern? Aquest article és un examen crític de la teoria de la veritat de Putnam en la seva evolució des d'un realisme metafísic a un realisme intern. En primer Iloc, presento una analisi de I'argurnent de la teoria de models que va dur a Putnam a abandonar el concepte metafísic de veritat com a correspondencia i a adoptar un concepte epistemic de veritar. Aquest argument, encara que constitueix una potent crítica de la concepció de la veritat del realisme metafísic, no estableix conclusivament que la veritat tingui algun contingut epistkmic. En segon Iloc, discuteixo la teoria idealitzada de la veritat de Putnam, tot argumentant que la 1. This paper is the product of many thought-provoking dicussions with Professor Meredith Williams in the winter of 1994. If there is any original idea in this paper, it is probably hers. 1 am also gateful for having had the oportunity to work with Arthur Fine and Michael Williams, which enabled me to read Putnam in a new light. The argument of this paper relies heavily on their philosophical work. 1 also want to thank ~rofessor Cristina Lafont for insightful discussions on Putnam and theories of truth, and David Bullwinkle for helpful comments on earlier versions of this paper. 70 Enrahonar 25, 1996 Josk Medina identificació de la veritat amb ~acceptabilitat en condicions idealsn no s'adiu amb la tesi que la veritat no és contextual sinó transcendent, per tal com la noció de justificació és intrínsecament contextual i no hi ha idealització que la pugui redimir d'aquest caracter. Finalment, suggereixo que les intuicions realistes que el reaiisme intern de Putnam intenta captar no requereixen més que una perspectiva deflacionista sobre la veritat. Adoptar aquesta perspectiva implica abandonar no sols la idea que la veritat és una propietat epistkmica, sinó també la idea que la veritat és una propietat substantiva que totes les afirmacions verraderes comparteixen i que, per tant, és objecte d'una teoria filosbfica. Paraules clau: veritat, Putnam, realisme intern, acceptabilitat, realisme metafísic, justificació, idealització. Sumari 1. Introduction 111. Truth, justification and idealization 11. From metaphysical realism to internal IV. Does internai realism need a theory realism: the model-theoric argument of truth? 1. Introduction What is realist about <<interna1 realism))? As Michael Devitt has pointed out2, it is difficult to understand the nature of Putnam's intellectual change, and to see anything realist about his new view. An important aspect of that change concerns Putnam's conception of truth. Putnam's view has shifted from an extra-theoretic and metaphysical notion of truth (Le. truth as correspondence with the mind-independent world) to an intra-theoretic and epistemic notion of truth (Le. truth as idealized ju~tification)~. However, Putnam insists, truth, the internal realist truth, though intra-theoretic, is not simply conventional and context-dependent. Unlike radical relativism, internal realism contains a substantial notion of correctness; and, unlike metaphysical realism, it links truth to justification. Putnarn seeks a theory of truth that grounds realist intuitions without making truth epistemically inaccessible. Truth now becomes for 2. M. Devitt, <(Realism and the Renegade Putnam~), Nous 17 (1983), p. 291-301. See p. 295. 3. This shift takes place in Reason. Tmth and Histoy (Cambridge University Press: 1981) and Realism and Reaso~z. Philosophical Papers. Volume 3 (Cambridge University Press: 1983), where Putnam departs from the metaphysicai reaiism of his early writings (collected in the first mo volurnes of his Philosophical Papers, Cambridge University Press: 1975). There is a further and more recent shift in Putnam's thought whose direcrion is srill unclear. In Words and L+ (Harvard Universiry Press: 1994), he no longer holds the conceprion of rruth as ((rationai acceptability under ideal conditionsn rhat he put fonvard in Reasoiz. Ti.uth and Histo~ (see p. V for a short statement of his new position), but he still maintains that there is more to truth than a disquorational account can offer, that truth is aa substantive property)) (p. 326), and that there is a conceptual connection benveen truth and jusrification (see doe es the Disquotational Theory Solve All Philosophical Problems?,,, p. 264-78; and ((On Truthn, p. 315-29). What is <<rrue)) in interna1 realism? Enrahonar 25, 1996 71 him ((idealized rational acceptability)i. This idealization cheory of truth is what supports the minimalrealism that Putnam advocates. One virtue of internal realism that Putnam emphasizes is that it c a n explain the convergence of scientific knowledge (which is for him a crucial feature of mature science) without bringing in any metaphysical baggage. Putnam thinks that he can hold on to realism as an explanatory stance in philosophy of science without having recourse to inflationary metaphysics. But, as Arthur Fine and Michael Williams point out4, inflationism as such is not overcome in internal realism, for Putnam simply resorts to an inflationism of a different kind. In Fine's terms, Putnam becomes a ((semantic inflater or truthmonger)i. He inflates the notion of truth, not with metaphysical content, but with epistemic content, by exchanging truth for ideal rational acceptability. The problem is that that semantic inflationism will not give Putnam what he wants, or so 1 will argue. In the end Putnam will not get a substantial notion of correctness out of the identification of truth with justification. The trouble is that there is prima facie a strong asymmetry benveen truth and justification. As Putnam recognizes, justification is a contextual notion while truth is not. But Putnam thinks that he can bridge the gap between truth and justification by idealizing the latter, and so he concludes that truth should be treated simply ((as an idealization of justification»5. 1 will argue that once you tie truth to justification, no amount of idealization will make truth context-transcendent, for any idealization will have to be anchored to a specific epistemic context. We have no clue about what ideal justification can consist in independently of particular justification contexts. And when we actually determine the ideal justification conditions for a claim, such conditions are dependent on current justification standards in such a way chat, though ideal, they remain contextual, that is, parasitic on that epistemic context. After all, the idealization theory of truth fails to preserve the realist intuition that Putnam is so eager to nurse (Le. that truth is stable or convergent). 1 will conclude, therefore, that there is nothing specifically realist about Putnam's idealized notion of truth. In fact, 1 will contend, the clarification of its epistemic content, despite its idealization, makes internal realism indistinguishable from a pragmatism a la Rorty (from whom Putnam is at pains to distance himself) . Finally, 1 will suggest that a merely disquotational elucidation of truth may be more congenial than the idealization theory to the minimal realism that Putnam wants to retain. Internal realism might be better offwith a deflationary view of truth, which is to say that no theory of truth is needed. The strategy ((let's idealize in order to remain realistsn does not seem to work. So maybe we should just leave truth alone if a realism without inflation is what we want. 4. A. Fine, <<Trurhmongering: Less is True,,, Canadian Journal ofPhilosophy 19 (1989), p. 611-16, esp. p. 61 1; and M. Williams, {{Do We (Epistemolgists) Need a Theoiy ofTruth!)), Philosophical Topics 14 (1986), p. 223-42, esp. p. 237. 5. ((Referente and Truth)), in Realirm and Reason, ibid., p. 84. 72 Enrahonar 25, 1996 José Medina In my analysis of Putnam's view of truth 1 will be concerned with two questions: first, whether the idealization theory makes the contexttranscendence of truth intelligible; and, second, whether a theory of truth is needed at al1 in order to preserve the reaiist intuition that truth is contexttranscendent. 1 will argue that both questions deserve a negative answer. In order to analyze and evaluate Putnam's developing conception of truth a close examination of his central arguments is required. The argument that led Putnam to abandon the metaphysical notion of truth as correspondence and to refashion the concept of truth in epistemic terms in Reason. Tmth and Histo y (hereafier RTH) is to be found in «Models and ~ealityn~. In the next section 1 will examine Putnam's move away from metaphysical realism via the model-theoretic argument of that paper. Section 111 will focus on the epistemic content and the idealizations that Putnam bdds into the notion of truth. And the final section will discuss whether the concept of truth can be taken as an object of philosophical theorizing, that is, whether we need a philosophical theory of truth that goes beyond the mere logical analysis and pragmatic elucidation of the predicate ((true)). 11. From metaphysicai realism to interna1 reaiism: the model-theoretic argument In «The Meaning of "Meaning")) Putnam argues that since we use the notions of truth and extension in an extra-theoretic way, we have to accept the realist perspective to which these extra-theoretic notions belong7. In that paper Putnam's view was that it is the world itself that decides what belongs and what does not belong to the extension of our terms: e.g., it is the microstructure of the stuff ((out there), that we cal1 water that determines membership in the extension of the term «water». This realist notion of reference is tied to a correspondence theory of truth8. A statement is true just when, and because, there is a state of &airs corresponding to it, that is, when the appropriate referential relations between its component parts and reality obtain. These substantial concepts of truth and reference are said to play a cruciai explanatory role in scientzj5c realism, which, following Richard Boyd9, Putnam defines as an interpretative hypothesis that includes two central theses, namely: 1. Terms in mature science rypically refer. 2. The laws of a theory belonging to a rnature science are rypicaily approximately truelO. 6. In Realism and Reason, ibid., p. 1-25. Norice rhat <<Models and Reality (1977) was writren before RTH (1981), although ir was not published until 1983 in Volume 3 of Purnam's Philosophical Papers. 7. aThe Meaning of "~Meaning",,, in Mind, Language, and Reality. Philosophical Papers. Volurne 1, Cambridge University Press: 1975, p. 236. 8. See Meaningand theMoralSciences, Routledge & Kegan Paul: 1978, Lecture 11, p. 18-33. 9. Realism and Scientzjc Epistemolohy, Carnbridge University Press: 1978. 10. Ibid., p. 20. What is (([tuea in inrernal realism? Enrahonar 25, 1996 73 These two principles taken together are supposed to account for the behavior of scientists and the success of science, notably the convergence of theories in mature science. The reference of scientific terrns and the truth of scientific laws were understood by Boyd and Putnam in metaphysical realist terrns, that is, in terms of correspondence between scientific language and the mind-inde- pendent world. However, Devitt notices that 1 and 2 are compatible with any view of truth and reference, and that they do not, by thernselves, entd rnetaphysical realisrn". The possibility of substracting rnetaphysical realisrn from the interpretative hypothesis expressed in 1 and 2 that Devitt points out seems to have been precisely Putnam's thought. Realisrn as an explanation of scientific practice acquires a renewed interest for Putnam when he realizes that it can in fact be detached from any rnetaphysical thesis12. 1 and 2 constitute the realist skeleton that remains after Putnarn's transition from metaphysical to internal realisrn. The notions of truth and reference used in those principles are sirnply reinterpreted or, if you like, internalized13. The thought that moves Putnarn to ((internalize), his earlier view is that we can be realists without having to resort ro metaphysically-inflated semantic notions. The guiding intuition of internal realism seems to be the realization that we do not need extra-theoretic notions of truth and reference to explain the convergence of scientific theories. Devitt writes: Let T' be the successor of T in any rnature science. Then convergence holds if and only ifffom theperspective ofl', the terrns of T typically refer and the laws of T are typically approximately true14. convergence requires only truth and reference relative to a theory, not unique correspondence with a rnind-independent world. Realism needs nothing more than intra-theoretic sernantic notions. Moreover, Putnam now argues, the correspondence theory of truth ultimately fails to make sense of the convergence of scientific knowledge, for in explaining truth in purely nonepistemic terms, it makes truth inaccesible. 11. For Devitt this means that there is nothing specifically realist about the interpretative hypothesis that Boyd and Putnam put forth, and that we should look for realism elsewhere. Devitt argues that reaiism is simply the thesis that there exists a mind-independent world. He emphasizes that such chesis is devoid of semantic content. In particular, he insists that no doctrine of truth is constitutive of realism as he understands it ({[Realism and the Renegade Putnam)), ibid., p. 293). 12. See (<Realism and Reason)), in Meaningand the MoralSciences, ibid., p. 123-138. 13. Notice that acceptance of realism as an explanation of scientific practice by the literal truth of scienrific theories does not automatically commit one to any theory of truth. In fact, as Williams has argued, realism as an interpretarion of science does not seem to require a substantiai notion of truth at all: aeven if we agreed that the success of our theories is explained by their truth, we shall have no reason for adopting a richer-than-disquotational conception of truth itselfn (aDO We (Epistemologists) Need a Theory of Truth?,,, ibid., p. 228). 14. M. Devitt, <<Realism and the Renegade Putnam,,, ibid., p. 294. 74 Enrahonar 25, 1996 JosC Medina The idea of an abstract isomorphism between language and reality, an abstract mapping of concepts onto things in the world, is not in the end very helpful. The problem is that «too many correspondences exist» (RTH, p. 73) and to single out the right correspondence between words and things, the ultimate adequate reference relations, would require an epistemic access to mindindependent objects that we lack. But if al1 correspondences are equally good, if any model can make a theory true, then no real convergence will obtain. As Putnam puts it, «[i]f al1 it takes to make a theory true is abstract correspondence (never mind which), then incompatible theories can be true» (RTH, p. 73). But the metaphysical realist holds that truth is unique correspondence, that there is One True Theory on which al1 scientific knowledge will ultimately converge. This view, Putnam argues, is ultimately unintelligible. He thinks that its unintelligibility can be shown in purely logical terms. This is what Putnam attempts to do in ((Models and Realityn, which contains the most complete formulation of Putnam's logical argument. This argument is the cornerstone of Putnam's transition from metaphysical to internal realism15. 1 In ((Models and Realityn Putnam offers an argument based on set theory which is intended to uncover the logical incoherence of the correspondence theory of truth. The argument has two different, though interrelated, parts. l In the first part Putnam argues that the notion of correspondence between words and discourse-independent objects is highly problematic and the view l of truth as unique correspondence a non-starter. The second part of the argument is intended to show that a radically non-epistemic notion of truth is unintelligible. ~ In the first part of his model-theoretic argument Putnam claims that set theory teaches us that «the total use of the language (operational plus theoretical constraints) does not «fix a unique ((intended interpretation»16. Oversimplifying Putnam's argument and disregarding the technical niceties of set theory, the central idea seems to be that for any model M of a theory, we can always construct a model M' which satisfies equally well al1 the theoretical and operational constraints of that theory, by simply establishing a one-to-one correspondence between the elements of the universe of M (say 1, 2, 3,. . .) and the elements of the universe of M' (say l', 2', 3',. . .). No matter what operational and theoretical constraints you place on a theory, you can always find more than one model that makes the theory true. The upshot of the argument is that there are always infinitely many different reference relations which satisfy al1 of the constraints that our practice may impose on our use of a language. Putnam concludes, therefore, that no sense can be made of the metaphysical realist claim that there is a unique language-world c~rres~ondence'~. A causal theory of reference to which realists like Boyd or Putnam's former self typically resort cannot solve the ((toa-many-correspondences)) pro- 15. See footnote 6. 16. «Models and Realiy, ibid., p. 4, emphasis dropped. 17. Ibid., p. 16. Whar is (arruex in interna1 realisrn? Enrahonar 25. 1996 75 blem. A causal theory (or for that matter any theory of reference) cannot single out one unique correspondence because ir simply adds more theoretical constraints to the semantic interpretation of the language, and again these constraints can be satisfied by more than one model. As Putnam puts it, unless the word «causes)) «is already glued to one definite relation with metaphysical glue, this does not fix a determinate extension for "refers" at a11))18. What Putnam wants to conclude from this argument is that truth and referente are intra-theoretic notions. Putnam takes the argument to show that the semantic interpretation of a theory is crucially dependent on the ((conventions)) that we adopti9 hence the (<theory-dependence of meaning and truthn20. The problem of indeterminacy arises only when we attempt to pin down the semantic interpretation of a theory from an external standpoint. From the perspective of its user the theory comes with an intended interpretation. It is only when we attempt to assign to it a semantic interpretation from a detached perspective (from a standpoint external to t he theory) that we are confronted with an infinity of models. And then no amount of constraints of whatever kind will single out one model. If you separate language and reality there is no way you can put them back together again. So the trick is not to introduce a gap between langua e and the world in the first place. Reference and truth have to be internalized', i.e relativized to a theory, point of view, or ((conceptual scheme)). The result of this internalization is that the idea of a single true picture of the world, the God's Eye View of metaphysical realism, is abandoned. Internal realism urges us to give up the old ideal of the One True Theory. As Putnam puts it in RTH, <cwhy should there not sometimes be equally coherent but incompatible conceptual schemes which fit our experiential beliefs equally well? If truth is not (unique) correspondence then the possibility of a certain pluralism is opened up)) (p. 73, emphasis added). If one side of the model-theoretic argument shows that truth is intra-the- oretic, the other side is intended to prove that truth is an epistemic notion. The claim of this part of the argument is that the metaphysical realist view that truth outruns even idealized j~stifiabilit~ is incoherent. The intuition of the metaphysical realist is that truth depends exclusively on the way the world is, and the world might be very different from the best knowledge of it that we can possibly have. Correspondence with the world is not assured even for the best imaginable theory if truth depends only on the way the world is independently of us. But Putnam's claim is that we do not even understand a radi- 18. Ibid., p. 18. Devitt argues against Putnam that in a causal theory the predicate causally related is aglued to one definite relation)) by causal relations, not with ((metaphysical glue)) (~Realism and the Renegade Putnam,,, ibid., p. 298). Devitr conrends thar a semantic theory applies to al1 theories, including itself, and that such self-reference need not be problematic. 19. aModels and Reality)), ibid., p. 10. 20. Ibid., p. 1 l. 21. ~Models are not lost noumenal waifs looking for someone to name them; they are constructions within our theory itself, and they have names from birth,, (ibid., p. 25). 76 Enrahonar 25, 1996 José Medina cally non-epistemic concept of truth. The argurnent goes as f~llows~~. Imagine a possible forrnalization TI of ideal scientific theory, that is, the best theory of the world possible for us humans, a theory at the limit of scientific investigation which fits al1 possible observational evidence and satisfies al1 possible theoretical constraints. Putnam's contention is that no sense can be rnade of the claim that such an ideal theory might be false, for, by definition, that theory will come out true in al1 possible rnodels. The rnetaphysical conception of truth as correspondence with the language and mind-independent world makes truth inaccesible for us, and is ultimately ~nintelli~ible~~. Putnarn concludes that we need a new analysis of the concept of truth in episternic terms. He reads his model-theoretic argument as a vindication of the connection betwe- en truth and our epistemic access to the world: how we represent the world becornes deterrninant for truth, which is refashioned in terms of coherence relations between representations (see RTH, p. 49-50). But what is realist about this conception of truth? Note that the intuition that what is most justifiable to believe rnay nonetheless be false is not really the exclusive property of metaphysicalrealism. It is rather the core intuition of allibilism, which is an essential component of any 4f realism properly so called2 . That intuition is independent of what we consider the ((naturen of truth to be (whether a correspondence relation or something else); in fact it is independent of whether or not we think truth has a ccnaturen in the first place. What the intuition expresses is sirnply that our knowledge of the world (whether actual or ideal) cannot determine what is true, that is, that the concept of truth is nonepistemic. Metaphysical realism is not the only view that gives content to the fallibilist intuition. A purely disquotational account of the predicate ((true)) also captures that intuition. In fact, unlike a correspondence theory, a disquotational account gives a precise forrnulation to the fallibilist intuition by showing that the truth of p neither entails, nor is entailed by, our knowing that punder ideal conditions, that the predicate ((true)) is not equivalent to (cknowable at the end of inquiqn). According to disquotationalism, what is involved in a sentence's being true is fully captured by the sentence itself hence the disquotational scherne (("p" true if and only if p)). To say that a staternent is true is not to say that what it expresses counts as knowledge for us, or that it would under ideal epistemic conditions; it is just to make the statement over again (hence the redundancy property of «true»): «"p is true iff p)), not ((iff p is knowable at the end of inquiry or under ideal episternic conditions)). So, as Williams puts it25, 22. Models and Reality, ibid., p. 11-13. 23. «The metaphysicai reaiist's claim that even the ideal theory TI might be false "in reaiity" seems to collapse into unintelligibility)) (ibid., p. 13). 24. For a lucid examination of the complex conceptual connections benveen fallibilism, truth and realism, see C. Lafont, ,<Truth, Knowledge and Reality,,, Graduate Faculy Philosophy journal1812 (1995), p. 109-126 25. «Do We (Epistemologists) Need a Theory of Truth?n, ibid., p. 224. What is <(truen in interna1 realism? Enrahonar 25, 1996 77 Thin as it is, even disquotational truth is «realistic», to the extent that, on a disquotational view, indeed deflationary views generally, truth is no more an epistemic property than it is on a full-blooded correspondence theory. It is dificult to see how one can be a realist without being a fallibilist, that is, without keeping the concepts of truth and justification separate. A radical antifallibilism which identifies truth with knowledge is just pure idealism. 1s there any room left for fallibilism in Putnam's view? Putnam's' argument curtails significantly the force of fdibilism. In his view there is room only for a weak fallibilism: our knowledge is de facto fallible, but its fallibility lies in the contigent conditions of our epistemic position; could we adopt an ideal epistemic position, truth would coincide with knowledge, which is to say that our knowledge would be infallible. On this view, the defeasibility of knowledge stems from the contingent gap between what is true and what we find acceptable given our epistemic situation. The ongoing revision of our knowledge reflects our constant effort to fd in that gap. But the gap is supposed to close under ideal conditions, and therefore as we approximate those conditions, there should be an increasing convergence between truth and knowledge. For Putnam, it is a mistake to suppose, as a strong fallibilism does, that there is an in principle unbridgable gap between truth and knowledge. He argues that the strong fallibilist conception of truth and objectivity as knowledge-transcendent is unwarranted and ultimately unintelligible when it comes to an ideal theory. For, since an ideal theory of the world is a theory whose formalization comes out true in al1 possible models (i.e. in al1 possible semantic interpretations), it is senseless to think that it might turn out to be false. For Putnam, this is supposed to express a conceptual truth about ((ideal theories)). This is just what an ideal theory is: a theory that cannot be false. 1s this a definite refutation of strong fallibilism? In what follows 1 will argue that a strong fallibilist need not be bothered by Putnam's argument, for in her view there is no room for such thing as an «ideal theory of the world)). What is unintelligible about the supposition that an ideal theory might be false? 1s it the very idea that no piece of human knowledge is infallible that is unintelligible? How is it that this idea is not only intelligible but quite intuitive with respect to actual theories and beliefs, and yet highly problematic with respect to ideal theories? What is so special about an ideal theory? Why is it impossible to think that the world might be different from the way an ideal theory would describe it? Indeed if we build into the very notion of ((epistemically ideal theory)) the impossibility of error, then we cannot make sense of the possibility of that theory being wrong. But it is not clear that we understand what an epistemically ideal theory is in the first place. That an ideal theory is true can be said to be analytic, true by stipulation. But do we really understand that stipulation? Devitt points out that it is not clear at al1 what a theory that satisfied al1 theoretical and operational constraints would consist in2'. 26. ~Realism and the Renegade Putnam)), ibid., p. 297 84 Enrahonar 25, 1996 José Medina facie clear how this view is different from relativism. There is indeed a strong antirealist tendency in the notion of truth as coherence. Putnam's epistemic idealizations are supposed to counterbalance that tendency. Again it is idealization, Putnarn thinks, that does the work to keep us realists. However, it is not at al1 clear that a realist notion of truth, an objective and context transcendent notion of correctness, can be derived frorn Putnam's epistemic idealizations, for, as we have seen, they are ineluctably parasitic on particular acceptability criteria. Richard Rorty argues that in the end «idealized» adds nothing tocrational a~ce~tability)), for Putnarn does not show how we can understand eidealized rational acceptability), otherwise than as acceptability to an ideal ~ommunity*~. And if Putnam does not want to appeal to Peirce's notion of a community of inquirers at the ideal lirnit of inquiry, such an ideal community cannot be but «US at our best)). «Identifying "idealized rational acceptability" with "acceptability to us at our best")), Rorty contends, ((is just what 1 had in mind when 1 said that pragmatists should be ethnocentrists rather than relativi~ts»*~. However, Putnarn has repeatedly expressed his discomfort with such a radical pragmatism. In Realism with a Human Face Putnarn formulates five different principles of warranted assertability in order to distance himself frorn ~orty~('. The main point of disagreernent between them is that while for Rorty al1 norrnativity is contextual (i.e. warranted assertability is always for a time and place), Putnarn wants to preserve the idea of ((some kind of correctness which is substantial)). Rorty's reservations concern Putnam's first two principles of warranted assertability: 1. In ordinary circumstances, there is usually a fact of the matter as to whether the statements people make are warranted or not. 2. Whether a statement is warranted or not is independent of whether the majority of one's cultural peers would say it is warranted or unwarranted4'. For Rorty the only fact of the matter as to whether or not S is warranted in asserting p is a sociological one and has to be ascertained «by observing the reception of S's statement by her peers))48. Putnam, however, seems to allude in 1 to something more than merely a matter of sociological fact. In fact 1 suggests that epistemic warrants are utterly independent of sociological considerations. But then, Rorty argues, Putnam talks as if warrant could be viewed ((sub specie aeternitatis)), as if there were some natural order of reasons that determines whether S is warranted in asserting p. 44. R. Rorry, ~Purnam and [he Relarivist Menacen, Journal ofPhilosophy Vol. XC, N. 9 (1993), p. 443-61, esp. p. 451. 45. Ibid., p. 452. 46. See ~Realism wirh a Human Face,), in Realism with a Human Face (Harvard University Press: 1990), p. 21-6. 47. Ibid., p. 21. 48. <<Purnam and [he Relativist Menacen, ibid., p. 449. Whar is <<truen in inrernal realisrn? Enrahonar 25, 1996 85 With respect to (2) Rorty says that the majority can indeed be ~ron~~~: p might be true even if the majority of S's cultural peers does not consider her to be justified in holding p. But how can p be in that case warranted? Justifiability or warranted assertability is a matter of «solidarity» with an epistemic community, If we take the relevant epistemic community to be S's culture, then her claim is not warranted. But the relevant epistemic community does not always have to be the actual community S happens to be in. However, warranted assertability is not independent of some community o rother. And, again, appeal to an «ideal,> community would not make normativity les~ contextual, for even then correctness is anchored to an epistemic context, namely, the one provided by the «ideal,> community as we conceive It. Putnam's idealization theory of truth and warranted assertability does not succeed in establishing a more substantial notion of correctness than the contextual normativity of a pragmatism i la Rorty. The idealization strategy does not seem to work unless backed with an epistemological realism that Putnam clearly would not accept. Putnam's claim that there is a fact of the matter about justifiability which is independent of community standards seems to have two different readings. On the one hand, Putnam can be read as claiming simply that assertability is not confined ro institutionalized norms, that the acceptability of one's assertions should not be restricted to the norms of one's culture. But this claim does not really entail a context-transcendent notion of objective rightness, but only the idea that institutionalized normative contexts are not exclusive or privileged (and, as 1 just suggested, this seems to be congenial with RortyS view). On a stronger reading, Putnam can be taken to be saying that there is an objective notion of justifiability completely independent of the epistemic standards of any community whatsoever, a n objective order of reasons to which our ideas of ideal justification and ideal acceptabilitity refer. The weak reading does not seem to take Putnam where he wants to go, but the strong reading takes him much further than he seems willing to travel. Putnam's idealization theory of truth contains two key claims that are at odds with each other, namely: 1. that truth is independent of justification here and now, but not independent of al1 justification; 11. that truth is expected to be stable or ~conver~entn~~. If the truth of a statement is supposed to be dependent on its justification, on its acceptability by certain standards, truth cannot be expected to be stable or convergent across relevantly different epistemic contexts (no matter how 49. «S may be &e unhonored prophet of some social movement or intellecrual revolution whose time has not yet comen, ibid., p. 450. 50. <<Referente and Trurh,,, ibid., p. 85; and RTH, ibid., p. 56. 86 Enrahonar 25, 1996 Tosé Medina much we idealize the notions of justification and a~ce~tability). No strong convergence can obtain once we tie truth to justification. The problem comes from running ((warranted)) together with ctrue)). It is the identification of truth with ideal rational acceptability which is itself problematic. It is not clear how Putnam can combine a non-foundational and strongly contextualist view of justification with a realist concept of truth unless he keeps them separate. But to keep truth and justification separate is to deny that the concept of truth is epistemic. This is what a deflationary view of truth enables one to do. Why doesn't Putnam take the deflationary route? Why doesn't he limit his minimal realism to a disquotational account of truth? IV. Does interna1 realism need a theory of truth? Putnam thinks that there is something more to truth than what is captured by the disquotational schema «p' is true if and only if p». In a merely disquotational view such as Quine's, to say that p is true is simply to assent to p. But to reduce truth to assent is to abandon the idea that truth is a substantial notion, and for Putnam «this is to give up what is right in realism»jl. ((That truth is a pro erty», Putnam says, «is the one insight of "realism" that we should not jettison))y2. Recently Putnam has become dissatisfied with his idealizarion theory of truth and he no longer identifies truth with ideal a~ce~tability~~, but he still maintains that truth is «a substantive property))54 that has an epistemic dimension, and that there is more to truth than a disquotational account can provide. For Putnam truth is a substantive property that statements have, a property whose nature neither the disquotational schema nor remarks on the use of the pedicate ((true)) can capture. He thinks rhat the disquotationalist claim that the truth of a sentence is fully expressed by the very assertion of the sentence amounts to an identification of truth conditions with assertibility conditions5j. So he contends that according to disquotationalism «the only substantive kind of rightness our statements (. . .) possess is warranted assertibility))56. The problem is that ((nobody has ever given a theory of what ((assertibility conditions)) are»57, and the only candidates available in the literature for an account of assertibility conditions commit us, according to Putnam, to behaviorism, cultural relativism, or methodological solipsism: what makes my 51. See aThe Causal Structure of rhe Physical),, in Realism with a Human Face, ibid., p. 93. 52. ((A Defense of Inrernal Realism,), in Realism with a Human Face, ibid., p. 32. 53. See Words and Life, ibid., esp. Preface, <<Does the Disquotational Theory Solve Al1 Philosophical Problems?)~, and non Truth)). See foornore 3. 54. <<On Truth)), ibid., p. 326 55. He interprets the diquorational approach ro truth as the claim that arhe description of tlie assertibility conditions (in the sense of confirmarion condirions) for senrences containing the word rrue tells us al1 rhere is ro know about rrurh)) (aOn Truthn, ibid., p. 326). 56. <(Does the Disquotational Theory of Truth Solve Al1 Philosophical Problemsn, ibid., p. 264. 57. <<On Trurhn, ibid., p. 327. Whar is iirruex in internai reaiism? Enrahonar 25, 1996 87 statement assertible is either my behavioral dispositions, or the norms of my linguistic community, or my own sense data. But Putnam misrerpresents disquotationalism by characterizing it as just another tbeory of truth. As Williams remarks, (c[t]o approach truth in a deflationary spirit is emphatically not to think of 'true' as denoting a theoretically significant property, explicating which will illuminate what is involved in any sentence's being true~~~. As 1 emphasized above (see p. 1 l), what the Tarsluan biconditionals show is that al1 that is involved in a sentence's being true is captured by the sentence itself. As Quine puts it, ccone who puzzles over the adjective 'true' should uzzle rather over the sentences to which he ascribes it. 'True' is transparenti>P9 According to traditional theories of truth, however itruth)) denotes a property shared by al1 true sentences, a property which is the rightful object of philosophical theory. A deflationary account of truth is not a theory of truth in this sense, but rather an analysis of the truthpredicate. A deflationary approach to truth seeks no more than a description of the logical behavior of (ctrue)) and a pragmatic account of its utility. This is precisely what a disquotational account of truth offers. On this view the necessary and sufficient conditions for a sentence's being true are expressed by the sentence itself (cc'p' is true iff pn). So what is accomplished by appending the truth-predicate to a sentence? According to Quine: ((To ascribe truth to the sentence [«Snow is white))] is to ascribe whiteness to snow.. . Ascription of truth just cancels the quotation marks. Truth is disquotation))GO. A deflationist can be content with a disquotational analysis of the truthpredicate and a pragmatic elucidation of the contexts of use of that predicate61. But for someone who still strives for a theory of truth, for someone who thinks that the property that cctruth)) denotes needs theoretical clarification, the disquotational-cum-pragmatic analysis of «truex is bound to be deeply dissatisfying, for it says nothing about what makes al1 true sentences true. The contention of deflationism is precisely that there is nothing to sayó2, that there is no single thing that al1 true statements share, that truth does not have an essence and true statements do not constitute a ccnatural kind)) or a theoretically inceresting category. As Williams puts it: «On a deflationary view, then, true sentences constitute a merely nominal kind. We could even say that, for a deflationist, though there are endlessly many truths, there is no such thing as t~uth)'~. 58. Unnatural Doubts, ibid., p. 11 1. 59. W. V. Quine, Pursuit of Tvuth (Harvard University Press: 1990), p. 82. 60. Ibid., p. 80. 61. As Williams puts it, for a deflarionisr ((remarks on the behaviour and utiliry of the rrurh-pre- dicate say just abour everything there is to sap about rruthn ( Unnatural Doubts, ibid., p. 11 1). 62. The argurnent goes as follows: (([he more a purely disquorational accounr of "rrue" can be shown [o capture whatever we want out of [he truth-predicare, the less reason there is for thinking that there must be some "rruth-ma!4ng" properry thar al1 true senrences share: the invitarion ro apply Occam's Razor ought to be irresistible>, (M. Williams, Unnatz~val Doubts, ibid., p. 369 1131). It is mainly its parsimony that makes disquotationalism so attractive. 63. Ibid., p. 112. 88 Enrahonar 25, 1996 José iMedina Putnam fails to take seriously the main insight of disquotationalism, namely, the deflationist claim that truth is not a property at al1 (neither warranted assertibility nor ideal justification nor anything else). On a disquotational view, to say that a sentence is true is not to ascribe a property to the sentence but just to reaffirm the sentenceG4. So ((p is true» is expected to be assertible whenever «p» is. However, this does not mean that «p is true)) says that «p is assertible)), any more than ((p)) itself doesG5. And, a fortiori, it does not mean that c(p is truen is equivalent to «p is assertible by the norms of m y linguistic communityn. Because on a deflationary view there is only (cimmanent truth)), Putnam thinks that a disquotational account of truth identifies the meaning of «true» with ccassertible according to the norms of a particular community)), and he concludes that a deflationary attitude towards truth commits one to cultural relativismGG. But, as Williams has argued, adopting a deflationary view of truth is not itself sufficient to make one a relativist: «to say that community practice is the basis for assessments of rightness is not to say that the standards embodied in such practice constitute a linguistic straight-ja~ket))~'. Deflationism is perfectly compatible with a minimal realism that makes truth context-transcendent. Since a deflationist does not analyze truth in epistemic terms, he does not identi5 truth with the contextual notion of justifiability or assertibility. And precisely because on a deflationary view truth is not an epistemic property, a deflationist has no difficulty expressing a corrigibilist outlook. As Williams remarks, it is the threat of relativism that seems to prevent Putnam from taking seriously «the deflationist's claim that there is no reason to suppose that truth has a nat~re))~*. But that threat disappears when we realize that on a deflationary view truth has no epistemic content whatsoever. Being non-epistemic without being metaphysical, a deflationary notion of truth can accommodate perfectly well the realist intuitions that Putnam wants to preserve. In fact interna1 realism would be better off with a deflationary attitude towards truth, for, as we have seen, the identification of truth with 64. Any difference between «p is rruex and <(p>) must have a pragmaric character: ir must be due ro the performative force of arrue,>. However, it is an open question what <(pragmaric» value the truth predicate has. Presumably different approaches to pragmarics would propose different accounts; and it could even be argued that the pragmatics of the truth-pre- dicare can supporr more than a merely deflationary accounr of truth. For an argument to chis effect from the perspective of Habermas pragmatics, see C. Lafonr (1995), ~'Truth, Knowledge and Realityn, ibid. 65. As Williams puts it, <(p is rrue' says whatever p says and so would say p is assertable only if p itself did. Bur to say rhat p is meaningful because asserting p is a move in a rule-governed practice is surely not to say rhat p is therefore a commenr on the practice that sanctions it), (nDo We (Epistemologists) Need a Theory of Truth?)), ibid., p. 240). 66. As WLUiams points out, Putnam usudy presents &e deflationisr as <'offering a naive epistemic accounr of truth, which rhen provides [he perfecr foil for Putnam's sophisticared epistemic accounrn (aDo We (Epistemologists) Need a Theory ofTrurh?),, ibid., p. 240-1). 67. Ibid, p. 239. 68. Ibid., p. 240. What is «true» in interna1 realism? Enrahonar 25, 1996 89 justification in Putnam's idealization theory does not succeed in making truth context-transcendent. But Putnam stresses that «as thinkers we are commited to there being some kind of truth, some kind of correctness which is substantia1 and not merely "disquotational"»69. However, saying that truth is disquotational prevents philosophical inflation but it does not weaken our ordinary notion of truth. We are indeed ((committed to there being some kind of truth)). But there is nothing that commits us to think of true statements as a natural kind and to think of truth as havin an essence, as something for which phi- 7% losophical theories are appropriate . Putnam is right in protesting that a disquotational theory does n o t ((solve al1 philosophical problems)) concerning the notion of truth71. But the point is that disquotationalism does not even address those problems (other than indirectly, by showing that there is nothing wrong with our ordinary notion of truth, that the ordinary uses of «true» are in fact transparent). So to depict deflationary accounts as weak attempts to solve philosophical problems about truth is an imporatnt misrepresentation, for these are not genuine problems for the deflationist and if he addresses them at al1 it is to disolve them rather than to solve them. Putnam's complaint that deflationism ((contains no answer to the question What is the nature of the property truth?~~~ shows that he is still under the sway of problems inherited from traditional theories of truth. If we think of truth as a property then it becomes a question what kind of property it is, whether absolute or relative, whether metaphysical or epistemic, whether indefinable or reducible to something else, etc. Both the correspondence theory of truth of realism and the acceptance theory of truth of anti-realism are answers to these questions; they both try to give a theoretical elucidation of the concept of truth by offering a certain picture of what truth consists in. But, as we have seen, both the metaphysical and the epistemic picture of truth have their own special dificulties; they both render truth problematic. As Fine points out, «the way out [of these problems] is not to "construe" truth at all», to adopt «a no-theory attitude toward the concept of tr~th»~~. This deflationary attitude that Fine rtecommends comes from the recognition that «truth» is «a concept already in use» and that we should respect its customary ~~rammar))~~. If the concept of truth is a ((fundamental semantical concept)) (or family of concepts), we should not expect it to be amenable to explication and we should be ((inclined to reject a 11 interpretations, theories, construals, pictures, etc., of 69. «Why Reason Can't Be Naturalized)) (in Realism and Reason, ibid., p. 229-47), p. 246, ernphasis dropped. 70. See A. Fine, The Shaky Game. Einstein. Realism and the Quantum Theory (The University of Chicago Press: 1986), for a detailed argument as to why the philosophical struth-rnon- gering)) should be stopped (see especially p. 137-42). 71. See «Does the Disquotational Theory Solve Al1 Philosophical Problerns?)>, in Words and Life, ibid., p. 264-78. 72. «On Truth)), ibid., p. 326. 73. The Shaky Game, ibid., p. 9. 74. Ibid., p. 133. 90 Enrahonar 25, 1996 Jose Medina truthn as reductive or circ~lar'~. So, far from being an alternative theory of truth, a deflationist view of truth should be taken to be, as Fine puts it, the expression of a ano-theory attitude toward truthn. If internal realism is not just a conciliatory solution, a middle way betwe- en metaphysical realism and relativism, but something more radical, i.e. a way of scaping from the dilemma or a way of disolving the dilemma, it must reject the premise that makes the dilemma possible: that is, the assumption that truth is a property in need of clarification. As long as the internal realist keeps this premise and continues to strive for a theory of truth, he cannot stop the oscillation between realist and anti-realist pictures of truth; the only success he can obtain on this turf is to reduce the oscillation to a minimum, to achie- ve some stability in the center; but the tension will still remain. 75. Ibid., p. 149.