Reading Putnam. By Maria Baghramian (ed.)

Reading Putnam, by Maria Baghramian (ed). Routledge, 2012. pp. 400. H/b £85.00. P/b £26.99. This review will appear in Mind, published by OUP. It is a pre-print. The definitive version will be available at Mind. Hilary Putnam is hard to read. This was the thought behind Reading Putnam, an excellent collection of papers, with responses by Putnam, published in 1994. In 2002 and 2005, two similar collections were produced, both called ‘Hilary Putnam’. These were buttressed by special editions of Erkenntnis (1991), Philosophical Topics (1992), Revue Internationale de Philosophie (2001) and European Journal of Analytic Philosophy (2008). As I write, we await Putnam’s Library of Living Philosophers volume. This most recent instantiation of the universal Reading Putnam belongs, then, within an impressive tradition. And, like Putnam’s own work, it is difficult, but rewarding. An immediate difficulty arises, because the papers in Maria Baghramian’s collection span so many different topics, and range across several different timeslices of Putnam. Tyler Burge and Alex Mueller focus on Putnam’s earlier work (mostly pre-1976). Richard Boyd and Michael Devitt are concerned with Putnam’s internal realism (roughly 1976–1990). David Macarthur, Charles Travis and John McDowell focus on Putnam’s natural realism (roughly 1990s onwards). Neither David Albert nor Stanley Cavell make any attempt to read Putnam (at any stage). And in both his introductory paper and his direct responses to the contributors, Putnam is as keen as ever to criticise his former self and break new ground. But Putnam is also keen to emphasise the unity in his views over time (pp. 19– 22). My aim in this review is to probe that unity, and I shall do this by focussing on themes that cut across a few of the papers in Reading Putnam. Would-be readers looking for a more comprehensive summary should consult Baghramian’s helpful introduction to the volume, which also includes an intellectual biography of Putnam. Idealism. Devitt has been criticising Putnam’s internal realism for three decades, and his contribution to this collection summarises his criticisms. One of Devitt’s main accusations (p. 111) is that internal realists are committed to saying: (a) If there had been no minds, then there would have been no stars. Although Putnam renounced internal realism in 1990, he continues to regard this accusation as utterly wrong-headed. His reasoning is as follows (pp. 123–4). It is a well-confirmed fact that what happens to humans on Earth has almost no effect upon the stars. So, if we regard (a) as a sentence of our best broad empirical theory, then (a) is obviously false. And that is exactly how the internal realist reads (a). Putnam briskly moves on, but we ought to pause for a moment. For, when he is not replying to Devitt, Putnam concedes that his internal realism was dangerously close to idealism (pp. 26, 29). Why is that not an admission that internal realists are committed to (a) after all? This is our first puzzle in reading Putnam. Here is how to resolve it. As an internal realist, Putnam offered a theory of understanding; an all-encompassing theory about how we understand empirical theories. And this theory of understanding entails idealism, albeit idealism of a subtle sort. The point is easiest to explain if we first consider idealistic theories of understanding. The simplest theories of this sort maintain that any empirical claim can be translated into a claim about my sense data (or similar). More complicated idealistic theories of understanding will concede that translation per se is impossible, but will retain the idea that the understanding of any empirical claim is to be explained purely in terms of sense data. But any such view – no matter how complicated – entails a subtle form of idealism. For, whilst someone who advances such a view might come out with (a) in an effort to convince us that she is not an idealist, she will be unable to understand herself as making a claim about mindindependent stars. In Reason Truth and History (1981, CUP, pp. 55–6, 121–4), Putnam explicitly rejected an idealistic theory of understanding. Instead, he explained understanding purely in terms of what my sense data would be under epistemically ideal conditions. But anyone who advances such a modalised theory of understanding must explain how we understand counterfactuals of the form ‘if conditions were epistemically ideal, then…’. Putnam now thinks that the internalist realist will have to say that such counterfactuals are themselves to be understood purely in terms of sense data. In sum: Putnam thinks that the modalised theory collapses into an idealistic theory of understanding, which entails a subtle form of idealism. (For more, see Putnam’s “Between Scylla and Charybdis”, in R.E. Auxier & L.E. Hahn (eds.) The Philosophy of Michael Dummett (2007, Open Court), pp. 161–3, and my The Limits of Realism (2013, OUP), chs. 5, 9 and 11.) Correspondence. The preceding discussion highlights the importance of semantic questions for the realism debate. Unfortunately, over the past three decades, Devitt has consistently and unequivocally denied precisely this point. To make his case vivid here, Devitt (p. 107) claims that it would not be ‘paradoxical or incoherent’ for a Realist to maintain that none of our words refer to anything. Putnam rightly takes Devitt to task over this (p. 125n3). If none of our words refer to anything, then when Devitt attempts to advance his Realist position by saying ‘there are electrons’, he fails to say anything about electrons. So, pace Devitt, it would be incoherent for him to attempt to advance Realism whilst rejecting the idea of any correspondence between words and reality. But now a second puzzle in reading Putnam emerges. For, only a few pages earlier, Putnam wrote ‘I would not say that I have a “correspondence theory of truth”’ (p. 97). What, then, prevents Putnam from falling afoul of his own criticism of Devitt? Fortunately, Putnam provides plenty of guidance on how to read him here (pp. 32–3, 97–99). Putnam does not think that the correspondence theory of truth is wrong, exactly. Rather, he thinks that the idea of a correspondence is often advanced in the wrong spirit. To understand what this amounts to – and to draw out a diachronic unity in Putnam’s views – it will help to consider an earlier timeslice of Putnam. Putnam’s model-theoretic arguments are often read as an attack on the very idea of a correspondence theory of truth (see, for example, Boyd p. 40). This is a misreading. As an internal realist, Putnam presented a correspondence theory of truth himself, albeit as a part of our best broad empirical theory. His objection to the idea of correspondence was just that it could form no part of a theory of understanding, in the sense just explained. (See Putnam “Realism and Reason” (1977), Proceedings and Addresses of the American Philosophical Association 50.6, pp. 483–9, and my 2013, chs. 8–9.) Putnam’s demand for a theory of understanding has gone the same way as his internal realism. But his unease with correspondence has not entirely dissipated. Whilst truth does involve correspondence, Putnam is reluctant to emphasise this point, since such an emphasis would probably mislead us into thinking that all true sentences must correspond to the world in the same way. And that, in turn, would lead us to neglect the diversity of linguistic phenomena. Existence. As an instance of this diversity, consider these sentences: (b) (c) Tables and chairs exist. Complex numbers exist. These sentences are grammatically similar, but they have very different uses. And, on roughly these grounds, Putnam maintains that existence is multivocal. Charles Parsons’s contribution to this volume is motivated by his ‘puzzlement’ with this argument (pp. 182, 194). Where Putnam traces the differences in (b) and (c) to ‘a difference in what we mean by “exist”,’ Parsons suggests that we should trace them to ‘a difference between tables and chairs and complex numbers’ (p. 194). Parsons goes on to maintain that existence is a univocal, formal concept In his reply to Boyd, Putnam hints that he could happily concede this point to Parsons (pp. 96–7). As with the idea that truth involves correspondence: there is nothing wrong, exactly, with the idea that existence is univocal; the problem is just that the univocity of existence is often advanced in the wrong spirit. That spirit is evinced when, for example, a metaphysician appeals to the univocity of existence in an effort to convince us to take seriously the question of whether or not mereological sums exist. But the formal univocity of existence no more entails that all ‘existence questions’ are worth debating (see Parsons pp. 193, 197), than the correspondence theory of truth entails metaphysical realism. Natural kinds. I have just linked the issues of existence and correspondence. Putnam himself once linked them explicitly, in the following Alarming Passage: ‘Objects’ do not exist independently of conceptual schemes. We cut up the world into objects when we introduce one or another scheme of description. Since the objects and the signs are alike internal to the description, it is possible to say what matches what. (Putnam 1981, p. 52; see also p. 54) The Alarming Passage smacks of idealism – of the not-very-subtle sort – and Putnam swiftly moved to distance himself from it. In particular, Putnam painstakingly and explicitly rejected the ‘cookie cutter metaphor’, implicit in the Alarming Passage, according to which the world is a doughy lump that we can slice into objects in various ways (see his “Truth and Convention” (1987), Dialectica 40.1–2, p. 70). It is surprising, then, that Boyd endorses the Alarming Passage even as he describes himself as a ‘metaphysical realist’ (p. 44–5). The threat of idealism continues to loom, when Boyd claims that ‘natural kinds are the workmanship of (women and) men established by (bicameral) linguistic legislation’ (p. 62). But the threat of idealism peaks when Boyd writes this: Because natural kinds are themselves artefacts of actual language use, reference and kind definitions (and ‘reality’) are aspects of a single phenomenon. Indeed the establishment of a natural kind and the establishment of reference to it by a natural kind term are basically the same phenomenon. Thus the problems of determinateness of reference raised by Putnam’s ‘modeltheoretic’ arguments do not arise. (p. 53; see also pp. 75–6) But perhaps a realist reading of Boyd is available. Boyd is not claiming that the existence of electrons somehow depends upon us. Instead, he is claiming that the naturalness of the collection of electrons depends upon us. This view is quite plausible, since it follows from the twin thoughts that naturalness depends upon the context of inquiry, and that we have some control over such contexts (see e.g. p. 67). However, this view also (sensibly!) entails that electrons existed before we humans ever coined the word ‘electron’. Accordingly, and pace Boyd, the view does not even begin to address the question of why the word ‘electron’ refers to all and only the electrons. Conceptual relativism. Boyd’s difficulty here highlights a tension that Putnam has wrestled with for decades. It is extremely tempting to claim that objects or kinds are (somehow) relative to frameworks, practices, or conceptual schemes. But, if we are going to avoid idealism, we need to accept that there is no sense in which we have the freedom to create (most) objects. Once we have faced up to this fact, we are left with only some very mundane freedoms. This is why Putnam’s most recent discussions of conceptual relativism (e.g. pp. 23–4, 28–9, 221) often seem to reduce to the truism that one can express the same truths in many different ways. That this is a truism does not show that it philosophically sterile. Crucially, the truism is only that there are many different ways to express the same truths. This leaves open the general question as to when two sentences express the same truths. (I say much more about this in my 2013, chs. 18–19.) Moreover, Putnam’s views on this last point have recently shifted, as demonstrated in his exchange with Block. In this volume, Block presents a barrage of stimulating arguments in favour of the possibility of inverted spectra. He first presented these at the ‘Putnam @80’ conference, in 2007, organised by Baghramian (indeed, most of the papers in this volume are descendants of that fantastic conference). At that conference, Putnam replied by denying ‘that sameness of qualitative character is well-defined both when we consider one person at different times and when we consider different people at the same or different times’ (p. 320). Putnam had made the same move against Boyd almost three decades earlier, when he recommended that we ‘abandon our metaphysical realism about sensations and about “same” (as applied to sensations)’ (1981, p. 94). In both cases, Putnam’s point was that equally adequate conceptual schemes might yield different verdicts on ‘when two people have the same qualia’; but, since these qualia-schemes are equally adequate, we have only a harmless case of conceptual relativism, rather than a serious metaphysical dispute. In this volume, though, Putnam announces a change of heart. He now thinks that an inference to the best explanation should lead us to adopt Block’s qualiascheme (pp. 320–1). And this change of heart raises some deep questions concerning one of Putnam’s most enduring criticisms of the subject of Ontology after Quine. Putnam has long argued – and continues to maintain – that the metaphysical debate about mereological fusions is a waste of time, since many different mereological-schemes are equally adequate. But many philosophers have sought to defend one mereological-scheme over another by invoking inference to the best explanation. So what distinguishes the debate about qualia from the debate about mereology? It will not help, simply to maintain that one of the rival qualiaschemes omits (or gets wrong) some important qualia-facts; after all, proponents of the mereology debate will say exactly the same about the rival mereology-schemes. The more general question facing Putnam is just: When do two schemes count as ‘equally adequate’? Putnam makes no serious attempt to answer this here (though see p. 23–4) and, whilst this might frustrate some readers, his relative silence is very wise. Only an answer that is neutral between all disputants could be satisfactory; but, ultimately, there can be no neutral answer to the general question. It is, after all, a truism that an ‘adequate’ scheme must accurately represent the (relevant) genuine phenomena. But, just what the genuine phenomena are, is precisely what is contested in any metaphysical dispute. Externalism and openness. For better or worse, Putnam now takes qualia much more seriously than he once did. He has no desire, though, for this to compromise his quite general externalism concerning mind and language. This externalism is the focus of the exchange between Travis, McDowell and Putnam. Putnam has long rejected the idea that concepts are governed by ‘algorithmic’ application- or recognition-conditions. The highpoint of Travis’s paper is a fascinating discussion of this ur-thought, and Travis’s discussion gestures towards a deep unity between Putnam’s (earlier) arguments for linguistic externalism and his (later) advocacy of perceptual externalism. This resonates poignantly with Burge’s complaint that, if the early proponents of ‘semantic externalism’ had only considered philosophy of perception, then ‘[t]he discoveries about language would have been made in a broader and more natural setting’ (p. 267). However, Travis pursues one strand of Putnam’s thought rather too far. (My diagnosis here draws on McDowell’s excellent reply to Travis, especially pp. 341–2.) Putnam now maintains that the conceptual schemes of the internal realist would constitute an ‘interface between us and the world’ (p. 26), which would prevent us from being properly open to the world. Running with this idea, Travis claims that our conceptual framework would constitute a similar interface, if we were to insist that perceptual experience has conceptual content (pp. 326–7, 335–9). To assess Travis’s suggestion, we need to understand how Putnam uses the word ‘interface’. For Putnam, an interface is something which screens off the objects of the world, preventing us from experiencing them or thinking about them. So an interface, for Putnam, is not a point where subject and world meet, but something which prevents any genuine meeting. Interfaces consequently raise the same problems as do (idealistic) theories of understanding: they leave us unable even to make sense of ourselves as experiencing, or thinking about, the objects of the world. Consequently, to postulate an interface is to embrace idealism (of the subtle sort). With this in mind, let us return to Travis’s line of thought. Perhaps I look at a star, and have a certain experience. Some philosophers think that I could not have had this experience, or thought my subsequent thoughts, if I had lacked certain concepts. However, all parties to this debate hold that what I see, and subsequently think about, is the star itself. So there is no threat of idealism, simple or subtle. Travis’s charge fails to stick, but we are left with a valuable lesson in how hard it can be to read Putnam. This review has focussed on only a few papers in Reading Putnam. I particularly regret that I have not discussed the papers by Macarthur, Mueller and Ruth Anna Putnam, all of which offer excellent critical reconstructions of particular aspects of Putnam’s thought (respectively: his frequent appeals to ‘common sense’ and his discussions of the analytic/synthetic and fact/value dichotomies). My aim, in focussing on just a few papers, is to highlight some deep connections that lie beneath the surface of this volume. These connections arise, not because the volume was assembled with a specific and narrow focus, but thanks to Putnam’s continual exploration of a series of profound, and profoundly interconnected, topics. It is as difficult and complex as it is rewarding to join him on that exploration. Acknowledgements. Many thanks to Hilary Putnam for fascinating correspondence concerning this review, particularly concerning the sense(s) in which theories of understanding might entail idealism. Thanks also to Maria Baghramian, Tim Crane and Rob Trueman. TIM BUTTON University of Cambridge St. John’s College, Cambridge, CB2 1TP, UK [email protected]