Fictionalism in Metaphysics

Philosophy Compass 6/11 (2011): 786–803, 10.1111/j.1747-9991.2011.00442.x Fictionalism in Metaphysics Frederick Kroon* University of Auckland Abstract This is a survey of contemporary work on ‘fictionalism in metaphysics’, a term that is taken to signify both the place of fictionalism as a distinctive anti-realist metaphysics in which usefulness rather than truth is the norm of acceptance, and the fact that philosophers have given fictionalist treatments of a range of specifically metaphysical notions. 1. Introduction Suppose that you are an atheist, but you don’t think we should drop discourse ‘about’ God. Although you think that God-talk is to be taken at face value, as a description of the ways and commands of a transcendent being, and although you think that, so construed, its central claims are false, you think that engaging in such discourse provides the best way for people, including yourself, to be motivated to live moral lives and choose worthwhile life-goals. If so, you are a fictionalist in the modern philosophical sense of that word: you think it is important to accept religious claims – speaking and behaving as if you believe those claims – because you take such a commitment to a religious form of life to be extremely useful; but you don’t actually believe the claims. (Note that fictionalism in this sense is different from another way in which someone might count God-talk a useful fiction: a ‘propagandist’ view about how society benefits if other people are encouraged to believe religious claims, claims that the propagandist herself takes to be lies.)1 The religious fictionalism just mentioned is not, of course, a particularly reputable example of fictionalism. Most philosophers would reject the claim that moral motivation can, or should be, subserved by God-talk, and it is in any case hard to see how a motivation of this kind could succeed without genuine belief in God (although some have suggested that Voltaire comes close to such a view with his remark that if God did not exist, it would be necessary to invent him (Voltaire 1768)). As a general philosophical strategy, however, fictionalism is thriving. Indeed, ‘fictionalism’ has become a new buzzword in philosophy, with fictionalist treatments of this or that topic now rife in metaphysics.2 The reason is not hard to find. Metaphysics – in many of its manifestations – can look inflationary in the worst of ways. Metaphysical theorising (as well as metaphysical reflection on theorising in other disciplines) seems to reveal a commitment to numbers as Platonic entities, to possible worlds that exist outside of the normal causal order, to subjectindependent moral facts, to propositions, …. And yet attempts to render such talk harmless through paraphrases that try to remove such commitments are widely agreed to face deep problems. Fictionalism promises to reconcile the robustness and linguistic irreducibility of such apparent commitments with a rejection of the move from apparent commitment to genuine commitment. I begin this paper by discussing how best to characterise fictionalism as a distinctive metaphysical view before briefly looking at some precursors of fictionalism thus ª 2011 The Author Philosophy Compass ª 2011 Blackwell Publishing Ltd Fictionalism in Metaphysics 787 characterised. I then discuss how this characterisation applies to a number of well-known contemporary varieties of fictionalism, and draw a number of crucial distinctions before turning to the topic of fictionalism about specifically metaphysical notions. The final section briefly considers the problems and prospects of fictionalism as a strategy within metaphysics. 2. Fictionalism vs. Realism In order to understand fictionalism as a position in metaphysics, it is perhaps best to begin with a familiar opponent, realism. Realism itself can be understood in a wide variety of ways. On one influential account (Devitt 1991), realism about Xs is the view that Xs exist and do so independently of minds, theories, concepts, and the like. By analogy, we might take fictionalism about Xs to be the view that Xs exist as useful fictions rather than as genuine mind (etc.)-independent entities. Such a characterisation is quite problematic, however. To call something a fiction in this sense is presumably to say that it is a purely made-up thing, much like a character of fiction. But that leaves the characterisation open to unintended interpretations. Realists about fictional characters, in particular, will take such a characterisation to imply that fictions do exist (say, as abstract objects), and that is surely not what fictionalists have in mind.3 Take the familiar cases of fictionalism about numbers and about possible worlds. A fictionalist about numbers doesn’t think that numbers are abstract objects, not even man-made abstract objects, since her fictionalism is likely to be shaped by her thinking that numbers as abstract objects don’t have a proper place in the natural world-order. And a fictionalist about possible worlds will be careful to distinguish her view from the views of what Lewis called ersatzers – those who take possible worlds to be abstract representations rather than fully fledged worlds. A better way to understand fictionalism is by considering a less entity-focused account of realism, but one that still brings out the connection to fiction. We might say that realism about a region of discourse D (mathematical discourse, the discourse of physical science, moral discourse, and so on) claims that (i) sentences of D have a representational semantics (they purport to represent things and their properties) and so have truth-conditions, and further, that (ii) aiming at truth is the proper goal of assertive utterance and acceptance when engaging in D: in accepting a sentence of D (and so being disposed to make full use of the sentence in one’s theoretical and practical reasoning) one should believe that it is true.4 In short, the proper aim of participation in D – and the norm for acceptance – is truth. Fictionalism, by contrast, claims that while the sentences of D do indeed have a representational semantics, condition (ii) doesn’t apply. Instead, fictionalism about region of discourse D claims that those participating in D should not have (and perhaps, in the case of certain types of discourse, typically do not have) truth as their aim when they accept a sentence from D. The norm for acceptance is not truth – one can accept a sentence of D for the best of reasons, justifiably acting on it, using it in one’s theorising, drawing inferences and acting on those, and so on, but not actually believe it, since there are benefits other than truth for whose sake one should accept such a sentence.5 (Similarly, one can assertively utter a sentence of D without actually using it with genuine assertoric force: one can quasi-assert it.) Call this characterisation F. F is put forward as a characterisation of fictionalism. The connection to fiction lies in a certain analogy, although this time not one based solely on the role of fictional characters (Sainsbury 2009 offers a good account of the analogy). When Jane Austen writes in Chapter 1 of Pride and Prejudice that ª 2011 The Author Philosophy Compass ª 2011 Blackwell Publishing Ltd Philosophy Compass 6/11 (2011): 786–803, 10.1111/j.1747-9991.2011.00442.x 788 Fictionalism in Metaphysics Mr Bennet was so odd a mixture of quick parts, sarcastic humour, reserve, and caprice, that the experience of three and twenty years had been insufficient to make his wife understand his character. Austen displays a sentence whose semantics is representational: it purports to represent or describe the world as being a certain way, in this case ascribing a complex property to someone named ‘Bennet’. Still, we understand that Austen was not aiming to report that someone so-called had this property. At best, she was pretending that this was the case, or aiming to get readers to pretend or make-believe that this was the case. We, in turn, can fully accept the sentence (being prepared to cite it, e.g. when challenged to say what Bennet was like) without thereby showing that we believe what the sentence says. After all, we don’t believe there was such a person; at best we pretend to believe. So in fiction we find the combination of a representational semantics and a notion of acceptance that is not truth-normed. This analogy with fiction is no mere analogy, in fact. Those who promote fictionalism about some region of discourse D think that even though sentences from D have a standard representational semantics, assertive utterances of sentences from D should not express beliefs. How, then, are we to understand such utterances? The sort of appeal to pretence alluded to above constitutes one way. According to the pretence or makebelieve account of fictional discourse (especially as developed by Walton 1990), speakers engage in pretend-assertion when uttering fictional sentences, and in accepting such sentences they are pretending that they are expressing beliefs rather than actually expressing beliefs. Similarly, many fictionalists think that acceptance in a fictionalist spirit is a kind of pretend or make-believe endorsement of the content of a sentence: participants in the discourse immerse themselves in the pretence or make-belief that the world is as the fictionalist hypothesis portrays it to be.6 Following some remarks by Lewis (2005: 315), we might call this a preface form of fictionalism – the utterance of sentence S is preceded with a preface (implicit or explicit) that cancels the assertoric force of what follows, in this case by declaring that it should to be taken in a pretend-spirit. But this is not the only way in which to understand a disavowal of assertoric force. The analogy with fiction suggests another way. Many theorists of fiction think that to accept a fictional sentence S is to have the genuine belief that in the story, S (see, e.g. Lewis 1978). The implicit or explicit addition of the ‘in the story’ prefix allows a speaker to disavow the assertoric force of her utterance of S by generating another sentence (‘In the story, S’) whose assertoric force is not disavowed. The same strategy has been used by fictionalists, many whom think that to accept a sentence S in a fictionalist spirit is to believe that in the relevant metaphysical fiction, S. Call this a prefix form of fictionalism.7 3. Historical Precedents We have talked of fictionalism as an important contemporary phenomenon in philosophy. But most trends in philosophy have their precursors, and fictionalism is no exception. Even a brief examination, however, is enough to show that finding clear examples of fictionalism in the modern sense is not easy (for more discussion, see Rosen 2005; Sainsbury 2009, Ch. 7). The reasons are instructive. One often cited example of fictionalism is that of 16th century thinkers and astronomers like Andreas Osiander and Nicholas Ursus who thought that astronomy was in the business of ‘saving the appearances’ rather than describing reality, and argued that this was the best way to understand Copernicus’s heliocentric hypothesis: one could rely on it ª 2011 The Author Philosophy Compass ª 2011 Blackwell Publishing Ltd Philosophy Compass 6/11 (2011): 786–803, 10.1111/j.1747-9991.2011.00442.x Fictionalism in Metaphysics 789 when doing astronomy without deeming it true. Such a view does conform to F, but only because it counts astronomical discourse as a distinctive form of discourse. The evidence suggests that astronomical fictionalism was relatively short-lived, and this was probably because of its restriction to astronomy – a restriction that seemed increasingly arbitrary as astronomy was integrated into the rest of physical science, understood as an inquiry into the nature of physical reality.8 (Given the widely acknowledged interconnectedness of our beliefs, one of the lessons in this for contemporary fictionalism is that it may be hard to prevent fictionalism about one discourse from infecting others.) It is surprisingly difficult to find other clear historical examples of fictionalism until the 19th century (for a discussion of some contenders, see Rosen 2005; Sainsbury 2009, Ch. 7). Because of his injunction to ‘think with the learned, but speak with the vulgar’ (Berkeley 1710, §51), Berkeley is sometimes classified as a fictionalist about talk of physical objects, causation, and so on. But for Berkeley, to think with the learned was to think of such entities and relations as fundamentally mind-dependent; his claim was that ordinary (‘vulgar’) talk of physical objects, their properties and the relations among them requires a new immaterialist interpretation (the materialist interpretation is incoherent), not that such talk is for the most part false. In other cases, it is just difficult to be sure. Jeremy Bentham’s theory of fictions in language holds that even the most useful forms of language are full of terms that grammatically purport to name real things (singular terms like ‘rest’, ‘motion’, ‘obligation’, and so on, as well as common names like ‘plant’), even though they really only name ‘fictitious entities’ (Bentham 1843: 195). Assuming that talk of ‘fictitious entities’ is best taken as a colourful way of saying that the corresponding terms are fictitious or lack reference, this certainly sounds like a version of fictionalism. But Bentham thinks that we are often able to clarify or explain what we really intend to say by giving paraphrases. If his view is that such paraphrases look beyond grammar and the images associated with fictitious terms to give the literal meaning of such sentences – and in some ways this seems the most likely construal of Bentham’s words – then we don’t have a modern-style fictionalism (Rosen 2005: 52–6). Otherwise we do. There are other 19th century philosophers who come close to advocating varieties of fictionalism, for example, Nietzsche.9 But the philosopher who in some ways is the most likely example of a genuine fictionalist prior to the resurgence of fictionalism in the late 20th century is also the person who was most responsible for giving fictionalism the bad press it had for much of the 20th century: Hans Vaihinger, who argued in The Philosophy of ‘As If’ (Vaihinger 1924) that we construct models of the world and then behave ‘as if’ the world matches our models because of the utility of behaving this way. Vaihinger himself saw fictions in virtually every domain of enquiry. He did not always clearly distinguish, however, between the thought that a false but simple theory may be useful because its explanations and predictions are close enough to the truth, and the thought that in certain domains of enquiry acceptance of even the best available theory is not truth-normed. It is the latter thought that is distinctive of modern fictionalism.10 Armed with their verificationism about meaning, the logical positivists would have regarded that thought as unintelligible. In the end, their views won the day. (Fine 1993 is a sympathetic discussion of Vaihinger’s fictionalism.) 4. Some Varieties of Fictionalism It is time to look at the way F classifies some familiar contemporary forms of modern fictionalism, to underline what is distinctive about such forms. Earlier I suggested that ª 2011 The Author Philosophy Compass ª 2011 Blackwell Publishing Ltd Philosophy Compass 6/11 (2011): 786–803, 10.1111/j.1747-9991.2011.00442.x 790 Fictionalism in Metaphysics reluctance to take religious fictionalism seriously is partly due to the fact that few believe that moral motivation needs an appeal to God. It is more common to think that morality requires that some actions ought to be performed regardless of the ends people have. But that thought has itself given rise to a new and currently more reputable form of fictionalism: moral fictionalism. John Mackie famously argued (Mackie 1977) that such categorical imperatives are indefensible (there are no intrinsically motivating moral properties). As a result, he took everyday moral thought to be substantially in error, a perspective he called an ‘error theory’ of moral discourse. But despite holding such an error theory, Mackie thought that moral discourse could continue with the status of a ‘useful fiction’.11 Mackie did not develop this suggestion, and it is easy to think that such a moral fictionalism is subject to an objection that also applied to religious fictionalism: how can we be motivated to act in accordance with this fiction if we believe it to be a mere fiction? An excellent account of the resources available to the moral fictionalist in answering this and other questions facing moral error theorists is found in Nolan et al. (2005). But the most fully worked out response to this motivational worry is contained in Richard Joyce’s defence of moral fictionalism (Joyce 2001, 2005). After arguing that an important practical benefit of having moral beliefs is that they diminish the possibilities for rationalisation, he argues that morality as a ‘mere fiction’ could provide a similar benefit if the decision to adopt morality is thought of as a kind of pre-commitment. The resulting attitude to morality will then be an attitude of make-believe with much the same pull as genuine belief. (Joyce’s view is therefore what we called a preface form of fictionalism. He thinks a prefix version – one that, for instance, counsels participants to hold the belief that lying is wrong according to the moral fiction – is unable to provide moral thought with the right kind of pull.) Moral fictionalism has as its central task explaining why eliminating moral discourse is not the best response to moral error theory. One of the most widely discussed fictionalisms is Hartry Field’s mathematical fictionalism (Field 1980), which faces a similar kind of task, although one that is technically far more daunting. Field thinks that, as it stands, mathematics is largely false since it is committed to a domain of abstract mathematical entities and Field doesn’t think that there are such entities (he is here influenced by the argument in Benacerraf 1973 that such entities would be unknowable). But of course science uses mathematics, and in fact delivers one of the most powerful arguments for thinking that there are mathematical entities: the Quine-Putnam indispensability argument that we ought to believe in their existence since they are indispensable to our best scientific theories (cf. Quine 1948; Putnam 1975). Field’s distinctive response to this argument is to suggest how mathematics might be dispensed with and how it could nonetheless be used so fruitfully in science. He first shows how a typical scientific theory such as Newtonian gravitational theory can be reconstructed without quantifying over mathematical entities (without, e.g. including statements that imply that there are real numbers corresponding to masses, forces, and distances). Because it doesn’t quantify over numbers or other abstract mathematical entities, such a reconstruction is known as a nominalisation of the theory in question. Assuming such nominalisations are available for other parts of science, that leaves us free to adopt an error theory of mathematics. But what then justifies using mathematics? Field’s answer is that mathematics, though false, is conservative: enlisting it does not yield any claims about the world that couldn’t have been derived without it, although it typically provides an easier route to such claims. That is what makes mathematics so useful. Field’s fictionalism is sometimes called ‘hard-road’ fictionalism, because of the difficult nominalising task it sets itself. There are other mathematical fictionalists who think that, ª 2011 The Author Philosophy Compass ª 2011 Blackwell Publishing Ltd Philosophy Compass 6/11 (2011): 786–803, 10.1111/j.1747-9991.2011.00442.x Fictionalism in Metaphysics 791 even if mathematics is indispensable to empirical science, we can still understand such applications in a fictionalist way: nominalisation is not needed. For somewhat different versions of such an ‘easy-road’ fictionalism, see Balaguer (1996a,1998, 2008) and Leng (2010). Note that if scientific realism requires scientific theories to aim at truth, or (even stronger) requires mature scientific theories to be for the most part true (cf. Putnam 1975), then ‘easy-road’ mathematical fictionalists cannot be scientific realists since they allow mature scientific theories to make claims using numerical terms and quantifiers while denying that there are numbers that could make these claims true. Instead, a fictionalist like Balaguer embraces nominalistic scientific realism, the view that what empirical science entails about the observable and unobservable physical world (the nominalistic content of empirical science) is mostly true, while its platonistic content (what it entails about the realm of abstract mathematical entities) should be treated in a fictionalist spirit (Balaguer 1998: 131). To give a sense of the richness and variety of contemporary fictionalism, our final example is an influential view about part of this nominalistic content – the part concerned with science’s talk of unobservables. On the (broadly) fictionalist view of science that van Fraassen advanced in The Scientific Image (Van Fraassen 1980),12 science does not aim at truth about the unobservable aspects of the world. Although the sentences of the theories advanced by scientists should be interpreted at face value, as genuine statements capable of truth or falsity concerning both observable and unobservable aspects of the world, science itself only aims at empirical adequacy, the accurate representation of observable phenomena. According to this position (known as constructive empiricism), a scientist is entitled to accept a theory that makes claims about the unobservable if the theory is empirically adequate; but if she also believes it to be a true description of the world ‘s ⁄ he adopts beliefs going beyond what science itself involves or requires for its pursuit’ (Van Fraassen 1998: 214). So the acceptance of theoretical scientific statements is not truthnormed on the constructive empiricist conception of science (so far that is no different from what moral and mathematical fictionalism claim about moral and mathematical statements), but this time there is no error theory to condemn genuine belief in the truth of such statements – there is only the philosophical conviction that science does not sanction belief beyond the observable. 5. Some Distinctions This latter feature makes for a striking difference between van Fraassen’s constructive empiricism and the mathematical and moral forms of fictionalism described earlier. But the similarities between these various positions are no less striking, and worth emphasising. All three appear to conform to characterisation F. In particular, all three take the language of the discourse in question at face value, espousing what many call ‘factualism’ (an unfortunate term when fictionalism is the topic!): the view that the language in question has a representational semantics and that its sentences have truth-conditions. Van Fraassen’s constructive empiricism, for example, is to be distinguished from an instrumentalism that takes the content of a scientific theory to be no more than a certain sort of inferential potential, just as mathematical fictionalism is to be distinguished from a formalist view that lets mathematical sentences count as little more than strings of symbols with no content. Especially significant on this score is the contrast between moral fictionalism and a non-factualist account of moral language like expressivism, according to which moral utterances do not purport to offer descriptions of moral reality but instead express attitudes of approval and disapproval (Boo to lying! e.g. might be the expressive content of saying ‘Lying is wrong’). Moral fictionalists point to expressivism’s troubles with the ª 2011 The Author Philosophy Compass ª 2011 Blackwell Publishing Ltd Philosophy Compass 6/11 (2011): 786–803, 10.1111/j.1747-9991.2011.00442.x 792 Fictionalism in Metaphysics Frege-Geach problem – the fact that it seems unable to capture the role of moral sentences when these occur in complex sentences such as ‘If lying is wrong, then …’ (‘If Boo to lying!, then …’ makes no sense). They claim that moral fictionalism, with its commitment to factualism about moral language, has no such problem (see especially Kalderon 2005b; Nolan et al. 2005). Factualism is a thesis about semantics and content. Cognitivism, by contrast, is a thesis about the epistemic attitudes one should have towards (the content of) sentences from a region of discourse, given the proper aim of enquiry focused on the discourse. Cognitivists think that those who engage in the discourse should adopt the attitude of belief towards sentences that they accept for purposes of theoretical and practical reasoning, given the proper (truth-normed) aim of enquiry focused on the discourse. Fictionalists, by contrast, think that one should accept, but not believe, sentences that meet the (nontruth-normed) standards for acceptance appropriate to that discourse. So for fictionalism, the theses of factualism and cognitivism come apart: fictionalists are non-cognitivist factualists. Realists, on the other hand, embrace both cognitivism and factualism, while expressivists reject both. There are two more distinctions that are relevant to fictionalism, this time distinctions internal to fictionalism.13 We have already met the first of these: the distinction between preface and prefix forms of fictionalism, introduced in these terms by David Lewis (Lewis 2005). Fictionalists tend to be divided over whether their views are best seen in prefix or preface terms. As Joyce points out, by its very nature moral fictionalism is probably best seen as a preface fictionalism. Modal fictionalism, to be discussed in the next section, is usually presented as a form of prefix fictionalism. Mathematical fictionalism can arguably be presented in either way, although Field presents his version as a prefix fictionalism (‘[T]he fictionalist believes that 2 + 2 = 4 only in the sense that standard mathematics says that (or, has as a consequence that) 2 + 2 = 4’ [Field 1989: 2].) The second important distinction between forms of fictionalism is the hermeneutic (or descriptive) vs. revolutionary (or prescriptive) distinction. Using the terminology of Burgess and Rosen (1997), a fictionalist construal of a given discourse is revolutionary just in case it involves a prescriptive ‘reconstruction or revision’ of the aims of the original discourse. The claim is that we ought to use the discourse in such a way that we are not aiming to affirm something that is literally true when we wholeheartedly affirm a statement, even though affirming something that is true is the normal, uncritical point of engaging in the discourse in question. The hermeneutic fictionalist, by contrast, puts her fictionalism forward as a claim about how the discourse is in fact used, as an interpretation of the attitude of those participating in the discourse. (A paradigm example is fictionalism about fictional discourse, the kind of discourse we engage in when telling or recounting fictional stories, knowing them to be fictional. People who engage in fictional discourse do not believe for a moment that they are talking about real people or situations, that they are talking of a real person called ‘Elizabeth Bennet’ as they describe the plot of Austen’s Pride and Prejudice, say – and not because they have grown to disbelieve the commitments to such entities found in earlier uses of fictional language.) In terms of this distinction, Field’s mathematical fictionalism and Mackie’s (and Joyce’s) moral fictionalism clearly count as revolutionary; and although the kind of ‘easy-road’ mathematical fictionalism of Balaguer and Leng can perhaps be formulated in hermeneutic terms, it too is intended to be taken in a revolutionary spirit. Earlier, I classified van Fraassen’s constructive empiricism as a form of fictionalism about theoretical scientific discourse. Just how we should classify it on the hermeneuticrevolutionary scale is not an easy question. Unlike a revolutionary fictionalist like Field, ª 2011 The Author Philosophy Compass ª 2011 Blackwell Publishing Ltd Philosophy Compass 6/11 (2011): 786–803, 10.1111/j.1747-9991.2011.00442.x Fictionalism in Metaphysics 793 van Fraassen doesn’t advocate reform of his target subject area. But neither does he argue, in the manner one would expect from a hermeneutic fictionalist, that practitioners of theoretical science do not have truth as their goal. He admits that many do. In an exchange with Rosen (1994), van Fraassen clarifies his position by saying that he takes empirical adequacy to be the ‘intentional’ goal of science, not scientists (Van Fraassen 1994: 181). We might, then, view constructive empiricism as a hermeneutic form of fictionalism if we are prepared to see the participants in scientific discourse as idealised scientists: scientists qua scientist, as it were (which would take the view a step away from ordinary hermeneutic fictionalism). Rosen tries to capture the fictionalism of constructive empiricism in a rather different way: he thinks that constructive empiricism is itself the fiction, that it is a view ‘best seen as a fiction about science put forward not as true, but rather as adequate to the phenomena of scientific activity’ (Rosen 1994: 153). It is clear, in any case, that van Fraassen’s view is very different from the revolutionary forms of fictionalism discussed above for mathematics and morality. But mathematics and morality have also become the subject of hermeneutic-fictionalist treatment, and the contrast with the revolutionary forms is instructive. Mark Kalderon, for example, considers the phenomenon of fundamental disagreements between epistemic peers to argue that the ordinary practice of moral judgement is different from that of ordinary assertion and belief (Kalderon 2005b). To have a moral attitude, he argues, is to have a desire in what Tim Scanlon calls the ‘directed attention’ sense, and this is a form of make-believe because it involves structuring our consciousness as if we really believed the moral sentence. Where Kalderon focuses on what happens in fundamental moral disagreements, Yablo argues for his version of mathematical fictionalism, which he calls figuralism, by comparing mathematical language to the case of metaphor. Following the lead of Walton (1993), Yablo thinks that sometimes a metaphor invokes merely pretended things as representational aids to allow us to say something about the world (cf. the way ‘Sally made friends with the butterflies in her stomach’ can be used to make a claim about Sally’s learning to cope with her nervousness, even though this latter claim does not give the literal meaning of the sentence but only the content that is really conveyed through its utterance: the real content, for short). Yablo sees the same kind of phenomenon in mathematics. We engage in make-believe according to which there literally are things like numbers, and use it to make claims about the real world. The real content of an utterance of ‘The number of planets is 8’, for example, is There are exactly eight planets, which quantifies over planets, not numbers.14 (Note that this account of real content strikes trouble when applied to statements about these new representational aids, such as ‘The number of even primes is 1’; its real content would be There is exactly one even prime, which quantifies over numbers.15 Yablo sees this as a case when ‘descriptive needs arise [with respect to], not the natural world, but our system of representational aids as so far developed’ (2005: 96), and his figuralism deals with it by invoking more general pretences in which mathematical objects like numbers feature as representational aids for already posited mathematical objects.)16 There are many other hermeneutic fictionalisms on the market, some mentioned below. It is fair to say, however, that hermeneutic fictionalisms tend to face an objection that is arguably more damning than that faced by revolutionary forms of fictionalism: there is usually no evidence that participants in the discourse in question take themselves to be engaging in any kind of pretence. I briefly return to this issue in the final section when we discuss the problems and prospects for fictionalist approaches in general, but ª 2011 The Author Philosophy Compass ª 2011 Blackwell Publishing Ltd Philosophy Compass 6/11 (2011): 786–803, 10.1111/j.1747-9991.2011.00442.x 794 Fictionalism in Metaphysics before doing so I want to turn to the topic of forms of fictionalism about metaphysical notions. 6. Fictionalising Metaphysics Fictionalism is a distinctive metaphysical view of a region of discourse that stands opposed to realism about that region of discourse – discourse that need not itself concern metaphysical notions. But fictionalism may also target notions that have been the focus of active metaphysical debate, as we have already seen in the case of mathematical and moral forms of fictionalism. In this penultimate section, I will provide brief accounts of metaphysically focused fictionalisms, beginning with the ones already discussed, to give a sense of the impact that the fictionalist ‘research programme’ has had, and is continuing to have, on metaphysics. 6.1. RELIGIOUS FICTIONALISM This paper began by presenting religious fictionalism as a hypothetical and unpromising example of the genre. But religious fictionalism, admittedly in a more sophisticated form, is now seen by some as a viable kind of revolutionary fictionalism, and a worthy alternative to more standard anti-realist accounts of God-talk. For discussion, see Eshleman (2005, 2010) and Cordry (2010). 6.2. MORAL FICTIONALISM Joyce (2001), Kalderon (2005b) and Nolan et al. (2005) argue the case for moral fictionalism over moral cognitivism and competing versions of moral anti-realism, including Blackburn’s quasi-realism. Lewis, by contrast, argues that quasi-realism can simply be construed as a form of (preface) fictionalism (Lewis 2005), a view questioned by Nolan et al. (2005) and rejected by Blackburn (2005). Much of the discussion of Kalderon’s hermeneutic moral fictionalism has focused on its touted advantages over expressivism and its repudiation of moral error theory (see, e.g. Chrisman 2008; Friend 2008; Lenman 2008; Eklund 2009b). And, not surprisingly, much of the critical reaction against revolutionary moral fictionalism has focused on its moral error theory (for discussion, see Blackburn 2005; Lewis 2005; Finlay 2008, 2010; Joyce 2010). Hussain (2004) provides a more general critical perspective on the task facing moral fictionalism. 6.3. MATHEMATICAL FICTIONALISM This has been one of the most widely debated fictionalisms. The discussion following Field’s initial work has focused on concerns about the prospects of the kind of wholesale nominalisation of scientific theories that is required for Field’s programme to work (Malament 1982; Balaguer 1996b; Bueno 2003), as well as concerns about Field’s argument for the conservativeness of applied mathematics (Shapiro 1983; Field 1985; Melia 2006). As we saw earlier, the kind of ‘easy-road’ fictionalism defended by Balaguer (1998, 2008) and Leng (2010) sidesteps such problems, but has received its share of criticism (see, e.g. Colyvan and Zalta 1999; Chihara 2007, Appendix B). Yablo’s figuralism (2001, 2002, 2005) offers a very different slant on the issue of the application of mathematics. Not surprisingly, much of the discussion of Yablo’s view has centred on its hermeneutic ª 2011 The Author Philosophy Compass ª 2011 Blackwell Publishing Ltd Philosophy Compass 6/11 (2011): 786–803, 10.1111/j.1747-9991.2011.00442.x Fictionalism in Metaphysics 795 character. Burgess (2004) is a classic statement of the objection that all the evidence indicates that mathematicians literally assert what they say. 6.4. FICTIONALISM ABOUT FICTIONAL CHARACTERS (AND OTHER NONEXISTENTS) Talk of ‘nonexistent’ entities like Sherlock Holmes, Zeus and the planet Vulcan has often been thought to commit speakers to a domain of genuinely nonexistent entities, the most (in)famous such view being Meinongianism. Walton offers a pretence account of such talk (see Walton 1990, 2000), not only for claims that are put forward as truths of fiction, myth, etc. (‘Holmes is a detective’, say) but also for claims external to fiction and myth such as ‘Holmes doesn’t exist’, ‘Zeus is a mythological object’ and ‘Holmes is more widely admired than Poirot’. Other authors have stressed various refinements or other advantages possessed by pretence theories of fictional and other empty names; see, for example, Crimmins (1998), Recanati (2000), Kroon (2004a), Everett (2005). (See also Friend 2007 for a good overview of the debate between a pretence-theoretic account of talk of fictional characters and a realist account like that of Thomasson 1999.) Note that Walton’s pretence-theoretic account is a preface form of fictionalism about fictional characters. Brock (2002), by contrast, offers a prefix version of fictionalism about fictional characters in which a sentence like ‘Holmes is a fictional character’ is given the novel paraphrase ‘According to the realist hypothesis concerning fictional characters, Holmes is a fictional character’. 6.5. FICTIONALISM ABOUT PHYSICAL OBJECTS Berkeley was an immaterialist, not a fictionalist about physical object talk. But some modern-day fictionalists do take a fictionalist line. On one construal, van Inwagen is a kind of hermeneutic fictionalist about such talk (Van Inwagen 1990; Van Inwagen himself doesn’t accept the fictionalist label). A statement like ‘There is a table here’ is strictly false on this view (since it holds that living organisms are the only composite things there are), but what ordinary people manage to assert through such a claim — that there are simples arranged table-wise here — may well be true. An explicit fictionalism along these lines, but focusing on all composite things, including living things, has been defended by Dorr and Rosen (2002) (cf. also Dorr 2002). Dorr and Rosen argue that we do not know which of various competing accounts of when simple objects compose to form other objects is true, and, in particular, whether there really are such things as tables, chairs, water molecules composed of hydrogen and oxygen atoms, and so on. But in their view that is no reason to give up talk of such things. We can regard the idea of composition as a fiction to live by. On such a view, a claim like ‘Water molecules are composed of hydrogen and oxygen atoms’ is (quasi-) assertible because it is true on the assumption that composite things exist, even if it is not strictly and literally true. 6.6. FICTIONALISM ABOUT CERTAIN KINDS OF PROPERTIES Dorr (2002) even argues for fictionalism about complex properties (more generally, complex attributes), not just complex or composite objects. Properties are also the focus of a number of other fictionalisms on the market. Thus, moral fictionalism is often characterised as a fictionalism about absolute moral properties. Not surprisingly, some have put a fictionalist spin on the special case of colour properties. Colour fictionalists think that, although no properties meet the condition necessary for a property to be a colour (they ª 2011 The Author Philosophy Compass ª 2011 Blackwell Publishing Ltd Philosophy Compass 6/11 (2011): 786–803, 10.1111/j.1747-9991.2011.00442.x 796 Fictionalism in Metaphysics are error theorists about such talk), the world is as if there are colours, and we should keep on using colour-talk because of its utility in characterising the way things look (Joyce 2005: 291–8; Maund 2006; Gatzi 2010). Another fictionalism about a philosophically puzzling property, this time hermeneutic in form, is James Woodbridge’s truth-theoretic fictionalism (Woodbridge 2005). This holds that we talk as if there are substantive properties of truth and falsity (in realty there are none) so that we can make certain serious assertions indirectly, including certain otherwise inexpressible generalisations (such as the one expressed by saying ‘Everything the Pope says is true’).17 Woodbridge thinks that such a fictionalism offers significant advantages over the main deflationist accounts of truth-talk (see also Armour-Garb and Woodbridge 2010b). 6.7. FICTIONALISM ABOUT PROPERTIES, PROPOSITIONS, AND OTHER ABSTRACTA One can be a fictionalist about certain kinds of properties; one can also be a fictionalist about all properties – one might take property-discourse to be worth preserving for its expressive advantages while taking it to be false that there is an abstract ‘G-ness’ somehow held in common by all Gs. Indeed, the problematic nature of abstract entities (especially the problem of epistemic access; see Field 1989; Balaguer 1998) has been part of the motivation for a number of different fictionalisms, beginning with Field’s mathematical fictionalism. But where Field understood the motivation in revolutionary terms, others think that such problematic features show that abstract object talk is (often) not fully literal and that we need a hermeneutic understanding of such talk (see especially Yablo 1998, 2000, 2001, 2002). For a recent defence of a fictionalism of this kind applied to the case of propositions, see Armour-Garb and Woodbridge (2010a), which attempts to vindicate proposition talk by explaining it as an indirect, but practically indispensable, means of making claims about various features of the use of linguistic and mental items. 6.8. MODAL FICTIONALISM Alongside mathematical fictionalism, the most widely discussed fictionalism about an important metaphysical notion is modal fictionalism (fictionalism about possible world talk), especially as formulated by Gideon Rosen (Rosen 1990).18 Modal fictionalists seek to defuse the outrage that greets David Lewis’s idea that there literally are other possible worlds as well as merely possible objects (talking donkeys, say). They reject, or profess agnosticism about, theories committed to the existence of merely possible worlds and objects, but nonetheless take these to be eminently worth using because of the expressive advantages that possible world talk gives us in systematising claims of necessity and possibility. These advantages they try to secure by appealing to the fiction that there are such worlds, and then using general bridging principles such as: P iff according to the fiction of possible worlds PW, P*, where P is a proposition and P* its possible-worlds ‘paraphrase’ (Rosen 1990: 335), yielding such claims as ‘It is possible that there are talking donkeys iff according to PW there is a possible world in which there are talking donkeys’. There is vigorous debate about the content of the most appropriate modal fiction PW. Rosen himself prefers to base the fiction on Lewis (1968), after initially proposing to base it on a variant of Lewis’s (1986) theory of possible worlds. (The change of mind was in response to the technical debate on the Brock-Rosen objection that modal fictionalism is self-defeating.)19 But whatever the most appropriate modal fiction turns out to be, it is clear that the task of selecting and then justifying such a fiction is of the highest importance. After all, according to modal fictionalism the real facts about modal status – what is ª 2011 The Author Philosophy Compass ª 2011 Blackwell Publishing Ltd Philosophy Compass 6/11 (2011): 786–803, 10.1111/j.1747-9991.2011.00442.x Fictionalism in Metaphysics 797 really possible or necessary – depend on the choice. More precisely, they depend on the choice if the modal fictionalism on offer is strong modal fictionalism, according to which the truth of modal claims is to be explained in terms of the content of the fiction about possible worlds. Timid modal fictionalism denies this strong dependence (Rosen 1990: 354). Although it thereby foregoes being an explanatory theory of modality, timid modal fictionalism does have other advantages (for one thing, it escapes the problem of circularity facing strong modal fictionalism’s attempts to explain the operator According to the fiction PW of possible worlds; see Rosen 1990: 344; Divers 1999). These are only some of the issues, some philosophical, some technical, canvassed in the literature on modal fictionalism. (Nolan 2008 is an excellent survey.) It should be emphasised that the attention lavished on modal fictionalism is not a sign that this is a particularly fragile form of fictionalism, but rather that it seems particularly clear why such a fictionalism is worth having and what we should expect from such a fictionalism, given the sophisticated state of formal and philosophical reasoning using the notion of possible worlds. (This is also a case where there is a well-known and respected kind of realist alternative to the bloated modal realism that modal fictionalism has in its sights, namely ersatzism, the view – held by some of the main players in the field – that concrete possibilia should be replaced with abstract surrogates that play the same theoretical roles.)20 7. Concluding Reflections Even if we disagree with this or that version of fictionalism, the significance of fictionalism as a programme in metaphysics cannot be overstated. For one thing, the availability of fictionalist options shows the need for care when trying to establish metaphysical conclusions. Philosophers often derive ontological conclusions from considerations of semantics, as in the argument: ‘2 + 2 = 4’ is [clearly] true; the truth of this sentence requires that ‘2’ and ‘4’ refer to numbers; so there really do exist numbers. More cautious metaphysicians might instead appeal to the indispensability of numbers to our best theory of the world. Both kinds of arguments overlook a fictionalist alternative. Perhaps a sentence like ‘2 + 2 = 4’ is true only within the fiction of numbers (so that the argument is no better than: ‘Holmes is a detective’ is clearly true, …, hence Holmes really does exist’); and perhaps what is truly indispensable is not literal quantification over numbers, but the availability of a (conservative) fiction that allows such quantification. Of course, the mere availability of such fictionalist proposals is not enough to show that they are in the end genuine contenders in the relevant debates. What, in general terms, are the problems and prospects of fictionalist treatments of regions of discourse? I’ll conclude this paper by highlighting a number of the most salient issues. Note first of all that fictionalist projects face some quite general objections. Given its central place in such projects, it is not surprising, for example, that the belief – acceptance distinction has come under special critical scrutiny, with a number of philosophers questioning the distinction on the grounds that the propensity to act as if a certain sentence or theory is true is precisely the mark of belief (see, e.g. O’Leary-Hawthorne 1994; Horwich 2004; for a response, see Daly 2008). A general worry of a very different kind concerns the persistence of the notion of an abstract object in fictionalist projects. Propositions, stories and fictions are generally thought to be types of abstract object, and so to the extent that fictionalism appeals to such objects it seems to be committed to abstract objects. This is a potential problem for hermeneutic fictionalists like Yablo who think that talk of abstract objects is (often) best construed as non-literal. But it is even more worrying for ª 2011 The Author Philosophy Compass ª 2011 Blackwell Publishing Ltd Philosophy Compass 6/11 (2011): 786–803, 10.1111/j.1747-9991.2011.00442.x 798 Fictionalism in Metaphysics revolutionary fictionalists who see error in commitments to things like numbers or possible worlds, but then propose theories explicitly committed to entities that, by the theorists’ own reckoning, should also count as problematic. (Cf. Daly 2008; §4. Nolan 1997 and Kim 2005 discuss the problem for the case of modal fictionalism.) There are also problems associated with particular types of fictionalism. Revolutionary forms of fictionalism, for example, face the potential problem that the thought of massive underlying error in claims from the discourse in question may be hard to take seriously. Take the case of revolutionary moral fictionalism. Many take the thought that moral properties are intrinsically motivating to be deeply confused (Finlay 2008, 2010); others insist that there can be no reason to think that morality is wrong rather than that moral theorists are wrong (Lewis 2005: 315–8). Revolutionary mathematical fictionalists in turn have been criticised for their ‘comically immodest’ error theory of mathematics, their hubris in thinking that philosophy can dictate to a genuine science like mathematics (Burgess 2004; Burgess and Rosen 2005; for a response, see Leng 2005; Daly 2006). Note that such complaints against revolutionary forms of fictionalism work only on a case-by-case basis. (Assuming that Lewis is right is that possible world talk is ontologically committed to possible worlds understood in a realist way, one suspects that few would complain that modal fictionalism has a ‘comically immodest’ error theory of possible worlds talk.) The main problems faced by hermeneutic forms of fictionalism seem rather more severe, because they seem systemic, affecting virtually every version of the doctrine. One such problem is that there is no systematic, compositional account of the real content expressed by sentences understood in the hermeneutic-fictionalist way (Stanley 2001: 41ff.; see Yablo 2001: 91–2; for a brief response). But the most pressing problem is a problem mentioned earlier, involving the phenomenology of pretence. As descriptive accounts of how certain regions of discourse are in fact understood, hermeneutic forms of fictionalism seem to assume that participants are engaged in a form of pretence. Burgess, for example, remarks that for Yablo ‘mathematics is like novels, fables, and so on in being a body of falsehoods not intended to be taken for true’ (Burgess 2004: 23). Explicit games of make-believe aside, however, all the evidence suggests that participants in the discourses in question take themselves to be speaking soberly and literally (which in turn suggests that someone’s engagement in such a pretence is, for the hermeneutic fictionalist, inaccessible to that person; see Stanley 2001: 46ff.). These difficulties for hermeneutic forms of fictionalism are certainly troubling, but I doubt that they are fatal. In the face of the phenomenological problem, for example, hermeneutic fictionalists should insist that theoretical rather than phenomenological reflection may be needed to decide whether agents are engaged in the relevant kind of pretence.21 More generally, there is strong evidence, some already found in Bentham (1843), for the claim that much of what we say is figurative rather than literal, and that we treat any commitments we thereby incur (‘butterflies in one’s stomach’, say) with an appropriate lack of seriousness. Extending this insight to metaphysical topics has been an important theme in the work of Yablo, with his argument that our discourse ‘about’ abstract objects shares many of the features of such non-literal forms of discourse (just as we think it silly to wonder about the size of the butterflies in your stomach, for example, so we think it is silly to wonder about the intrinsic nature of the quality of wisdom or the number 17). We surely need some theoretical way – via an appeal to pretence, or similar – of capturing such insouciance.22 Whatever the final judgement of the philosophical community on this and other issues facing current forms of fictionalism, there is every reason to believe that, in one form or another, fictionalism will long continue to be a force in metaphysics. ª 2011 The Author Philosophy Compass ª 2011 Blackwell Publishing Ltd Philosophy Compass 6/11 (2011): 786–803, 10.1111/j.1747-9991.2011.00442.x Fictionalism in Metaphysics 799 Acknowledgement Thanks to Chris Tucker and an anonymous referee for useful comments. Short Biography Frederick Kroon teaches at the University of Auckland. Following early work in computability theory, and, more recently, on paradoxes of rationality, he now works mainly in philosophical logic and the philosophy of language. He has authored papers in these areas for a range of journals, including the Australasian Journal of Philosophy, Philosophy and Phenomenological Research, Philosophical Review, Journal of Philosophy and Noûs, while his work on paradoxes of rationality has appeared in Analysis and Ethics. Current research involves the application of hermeneutic fictionalism to a range of problems and puzzles in the philosophy of language. He has held visiting fellowships at the Australian National University, Princeton University and the Pittsburgh Center for the Philosophy of Science, and is currently a subject editor for the Stanford Encyclopedia of Philosophy. He holds the MA from the University of Auckland and the PhD from Princeton University. Notes * Correspondence: Department of Philosophy, University of Auckland, 11 Grafton Rd, Auckland 1024, New Zealand. Email: f.kroon@auckland.ac.nz. 1 Joyce (2005: 299) uses the term ‘propagandism’ for a related view about morality. I don’t deny that there is a sense in which a propagandist view might count as ‘fictionalist’, but because this sense doesn’t capture what is distinctive about modern forms of fictionalism I am here putting it aside. 2 Eklund (2009a) provides a good overview, while Kalderon (2005a) is a widely cited anthology of recent articles. For a more critical account of a range of fictionalist positions, see Sainsbury (2009). 3 At least it is not what most fictionalists have in mind. Bueno has recently proposed the ‘truly fictionalist strategy’ of taking mathematical entities to be fictional entities, construed as abstract artefacts created by the act of theorising (Bueno 2009: 66). Since abstract artefacts genuinely exist, however, it is not clear why we should think of this as a ‘truly fictionalist strategy’. 4 On this understanding of the attitude of acceptance, belief is an instance of acceptance. Others take acceptance and belief to be distinct attitudes (see, e.g. Sainsbury 2010). Nothing in what follows hinges on this difference. 5 The concept of acceptance at play here is full acceptance, not tentative acceptance (see Kalderon 2005a: 2). One may tentatively accept a claim C of Newtonian mechanics, say, on the grounds that it is a good approximation to the truth and approximate truth is all that is required for the purposes at hand. So the norm for tentative acceptance is not truth. But accepting C in this tentative way is compatible with not acting on C or using it in one’s theorising in other contexts, say when doing high-energy physics. Full acceptance is not subject to such variation. 6 Not all philosophers who acknowledge the acceptance-belief distinction and its importance to fiction and fictionalism think that [mere] acceptance should be understood in terms of pretence. Some think the attitude is sui generis (cf. Sainsbury 2010). 7 Lewis writes: ‘There are prefixes or prefaces (explicit or implicit) that rob all that comes after of assertoric force. They disown or cancel what follows, no matter what that may be’. The difference is that ‘[w]hen the assertoric force of what follows is cancelled by a prefix, straightaway some other assertion takes place… Not so for prefaces’ (Lewis 2005: 315). 8 Duhem (1969) argues that astronomical fictionalism was in fact widespread in antiquity, but the evidence does not appear compelling; see Rosen (2005: 26–36). 9 Nietzsche thought that our ordinary discourse about the world may have been instrumentally useful but involved massive falsehood. But unlike modern fictionalists, he took this to be reason to reject the discourse. Hussain (2007) argues that Nietzsche’s account of created value may be fictionalist in the modern sense, but see Leiter (2008) for reservations. 10 The distinction is important. Fictionalism is sometimes described as the claim that a theory doesn’t have to be true to be good. But unless more is said about the scope of ‘good’, such a characterisation is in danger of trivialising fictionalism; after all, many false theories are good to at least some degree – Newtonian mechanics qualifies, no doubt, but even phlogiston theory looks a good theory if we restrict its range of application (cf. Schurz 2011). ª 2011 The Author Philosophy Compass ª 2011 Blackwell Publishing Ltd Philosophy Compass 6/11 (2011): 786–803, 10.1111/j.1747-9991.2011.00442.x 800 Fictionalism in Metaphysics 11 Vaihinger (1924) also proposes a version of moral fictionalism, based on a fictionalist reading of Kant’s moral philosophy. 12 Most, but not all, accounts of fictionalism regard van Fraassen’s view as a type of fictionalism. See, for example, the introduction to Kalderon (2005a), Rosen (2005: 17) and Eklund (2009a). We canvass some difficulties with this classification in Section 5. 13 Eklund (2009a) further distinguishes between meaning and use fictionalism, and content and force fictionalism. 14 There are some interesting similarities between Yablo’s account and Berkeley’s account of the genesis of mathematical language in Berkeley (1710, §118ff.). But we should not count Berkeley as a mathematical fictionalist in the modern sense since there is no evidence that for Berkeley number talk is, from the point of view of the fiction, genuinely about something: ‘In Arithmetic, … we regard not the things but the signs, which … are not regarded for their own sake, but because they direct us how to act with relation to things …’ (§122). 15 Nolan and O’Leary-Hawthorne (1996) point out that this is a special case of a quite general problem for fictionalism (the earliest instance was the so-called Brock-Rosen objection to modal fictionalism). 16 Balaguer (2009, §2.3) thinks Yablo’s emphasis on real content makes him a paraphrase nominalist rather than, as I interpret him, a kind of (preface) fictionalist. I agree that at the very least Yablo shows himself a somewhat reluctant fictionalist (cf. his worry that in uttering a mathematical sentence it may be indeterminate whether the real meaning is the literal content or the real content, since ‘I want to be understood as meaning what I literally say if my statement is literally true … and meaning whatever my statement projects onto … if my statement is literally false’; Yablo 1998: 257). 17 See also Burgess (forthcoming), which argues that fictionalism can be characterised without appeal to the notion of truth, and then embeds this claim in a revolutionary truth-theoretic fictionalism focused on semantic paradox. (Beall 2004 and Kroon 2004b also defend forms of truth-theoretic fictionalism focused on semantic paradox.) 18 Armstrong’s combinatorial theory of possibility offers a variant form of modal fictionalism (Armstrong 1989). Nolan (1997: 261–2) briefly considers a modal fictionalism directly targeting claims involving the modal operators of necessity and possibility. 19 For the Brock-Rosen objection, see Brock (1993), Rosen (1993). Rosen (1995) accepts the response of Noonan (1994), but there continues to be debate about the objection; see, for example, Divers and Hagen (2006) and Liggins (2008). 20 Ersatzism, famously criticised by Lewis (1986), has been defended by, among others, Plantinga (1976), Stalnaker (1976), Prior and Fine (1977), Linsky and Zalta (1994), Nolan (2002), Ch. 5, and Sider (2002). Brogaard (2006) is an excellent comparative overview of the choice between fictionalist and ersatzist competitors to modal realism. Her own preference is for timid modal fictionalism over what she sees as the best available version of ersatzism, the kind of holistic ersatzism defended by Prior ⁄ Fine, Nolan, and Sider (her actual argument focuses on the ersatz pluriverse account defended by Sider). 21 See, for example, Joyce (2005: 292). See also Armour-Garb and Woodbridge (2010a), which argues for a hermeneutic-fictionalist treatment of proposition talk that does not attribute pretence to speakers but locates it within the theorist’s explanation of how a statement ends up with the serious content it does. 22 Armour-Garb and Woodbridge (2010a) contains a good discussion of this don’t-care attitude in the case of talk about propositions. Eklund (2005) argues that acknowledging such a don’t-care attitude need not take one in the direction of hermeneutic fictionalism about the discourse in question (he advocates a position he calls indifferentism). Works Cited Armour-Garb, B. and J. Woodbridge. ‘The Story about Propositions.’ Noûs (2010a). doi: 10.1111/j.14680068.2010.00794.x. —— and ——. ‘Why Deflationists Should be Pretense-Theorists (and Maybe Already are).’ New Waves in Truth. Ed. N. Pedersen. Basingstoke and New York: Palgrave Macmillan, 2010b. 59–77. Armstrong, D. A Combinatorial Theory of Possibility. Cambridge: Cambridge University Press, 1989. Balaguer, M. ‘A Fictionalist Account of the Indispensable Applications of Mathematics.’ Philosophical Studies 83 (1996a): 291–314. ——. ‘Fictionalism in the Philosophy of Mathematics.’ The Stanford Encyclopedia of Philosophy (Fall 2008 Edition). Ed. Edward N. Zalta, 2008. [Online]. Retrieved on 5 August 2011 from: http://plato.stanford.edu/archives/ fall2008/entries/fictionalism-mathematics/. ——. ‘Fictionalism, Theft, and the Story of Mathematics.’ Philosophia Mathematica 17 (2009): 131–62. ——. Platonism and Anti-Platonism in Mathematics. New York: Oxford University Press, 1998. ——. ‘Towards a Nominalization of Quantum Mechanics.’ Mind 105 (1996b): 209–26. Beall, J. C. ‘True and False — As If.’ The Law of Non-Contradiction. Eds. G. Priest, J. C. Beall, and B. ArmourGarb. Oxford: Clarendon Press, 2004. 197–216. Benacerraf, P. ‘Mathematical Truth.’ Journal of Philosophy 70 (1973): 661–79. Bentham, J. The Complete Works of Jeremy Bentham. Ed. J. Bowring, vol. 8. Edinburgh: William Tait, 1843. ª 2011 The Author Philosophy Compass ª 2011 Blackwell Publishing Ltd Philosophy Compass 6/11 (2011): 786–803, 10.1111/j.1747-9991.2011.00442.x Fictionalism in Metaphysics 801 Berkeley, G. A Treatise Concerning the Principles of Human Knowledge. Dublin: Aaron Rhames for Jeremy Pepyat, 1710. Blackburn, S. ‘Quasi-Realism no Fictionalism.’ Fictionalism in Metaphysics. Ed. M. Kalderon. New York: Oxford University Press, 2005. 322–38. Brock, S. ‘Fictionalism about Fictional Characters.’ Noûs 36 (2002): 1–21. ——. ‘Modal Fictionalism: A Response to Rosen.’ Mind 102 (1993): 147–50. Brogaard, B. ‘Two Modal–Isms: Fictionalism and Ersatzism.’ Philosophical Perspectives 20 (2006): 77–94. Bueno, O. ‘Is it Possible to Nominalize Quantum Mechanics?’ Philosophy of Science 70 (2003): 1424–36. ——. ‘Mathematical Fictionalism.’ New Waves in Philosophy of Mathematics. Eds. O. Bueno and Ø. Linnebo. Hampshire: Palgrave MacMillan, 2009. 59–79. Burgess, A. ‘Truth in Fictionalism.’ The Oxford Handbook of Truth. Ed. M. Glanzberg. Oxford: Oxford University Press. forthcoming. Burgess, J. ‘Mathematics and Bleak House.’ Philosophia Mathematica 12 (2004): 18–36. —— and G. Rosen. ‘Nominalism Reconsidered.’ The Oxford Handbook of Philosophy of Mathematics and Logic. Ed. S. Shapiro. New York: Oxford University Press, 2005. 515–35. —— and ——. A Subject With No Object. New York: Oxford University Press, 1997. Chihara, C. S. A Structural Account of Mathematics. Oxford: Oxford University Press, 2007. Chrisman, M. ‘A Dilemma for Moral Fictionalism.’ Philosophical Books 49.1 (2008): 4–13. Colyvan, M. and E. Zalta. ‘Mathematics: Truth and Fiction? Review of Mark Balaguer’s Platonism and AntiPlatonism in Mathematics.’ Philosophia Mathematica 7 (1999): 336–49. Cordry, B. ‘A Critique of Religious Fictionalism.’ Religious Studies 46 (2010): 77–89. Crimmins, M. ‘Hesperus and Phosphorus: Sense, Pretence, and Reference.’ Philosophical Review 107 (1998): 1–48. Daly, C. ‘Fictionalism and the Attitudes.’ Philosophical Studies 139 (2008): 423–40. ——. ‘Mathematical Fictionalism –No Comedy of Errors.’ Analysis 66.3 (2006): 208–16. Devitt, M. Realism and Truth. 2nd ed. Princeton: Princeton University Press, 1991. Divers, J. ‘A Modal Fictionalist Result.’ Noûs 33 (1999): 317–46. —— and J. Hagen ‘The Modal Fictionalist Predicament.’ Identity and Modality. Ed. F. MacBride. Oxford: Oxford University Press, 2006. 57–73. Dorr, C. and G. Rosen. ‘Composition as a Fiction.’ The Blackwell Guide to Metaphysics. Ed. R. Gale Oxford: Blackwell, 2002. 151–74. Dorr, C. S. The Simplicity of Everything. Unpublished PhD diss., Princeton University, 2002. Duhem, P. To Save the Phenomena: An Essay on the Idea of Physical Theory from Plato to Galileo. Trans. E. Doland and C. Maschler. Chicago: University of Chicago Press, 1969. First published as Le Système du Monde; Histoire des Doctrines Cosmologiques de Platon à Copernic, Paris: A. Hermann (1913). Eklund, M. ‘Fiction, Indifference, and Ontology.’ Philosophy and Phenomenological Research 71 (2005): 557–79. ——. ‘Fictionalism.’ The Stanford Encyclopedia of Philosophy (Spring 2009 Edition). Ed. Edward N. Zalta, 2009a. [Online]. Retrieved on 5 August 2011 from: http://plato.stanford.edu/archives/spr2009/entries/fictionalism/. ——. ‘The Frege-Geach Problem and Kalderon’s Moral Fictionalism.’ The Philosophical Quarterly 59 (2009b): 705– 12. Eshleman, A. ‘Can an Atheist believe in God?’ Religious Studies 41 (2005): 183–99. ——. ‘Religious Fictionalism Defended: Reply to Cordry.’ Religious Studies 46 (2010): 91–6. Everett, A. ‘Against Fictional Realism.’ Journal of Philosophy 102 (2005): 624–49. Field, H. ‘On Conservativeness and Incompleteness.’ Journal of Philosophy 81 (1985): 239–60. Reprinted with a postscript in Field (1989), 125–46. ——. Realism, Mathematics and Modality. Oxford: Blackwell, 1989. ——. Science Without Numbers. Princeton: Princeton University Press, 1980. Fine, A. ‘Fictionalism.’ Midwest Studies in Philosophy 18 (1993): 1–18. Finlay, S. ‘The Error in the Error Theory.’ Australasian Journal of Philosophy 86.3 (2008): 347–69. ——. (2010). ‘Errors upon Errors: A Reply to Joyce.’ Australasian Journal of Philosophy. doi: 10.1080/ 00048402.2010.510531. Friend, S. ‘Fictional Characters.’ Philosophy Compass 2.2 (2007): 141–56. ——. ‘Hermeneutic Moral Fictionalism as an Anti-Realist Strategy.’ Philosophical Books 49 (2008): 14–22. Gatzia, D. E. ‘Colour Fictionalism.’ Rivista Di Estetica 43.1 (2010): 109–23. Horwich, P. ‘On the Nature and Norms of Theoretical Commitment.’ From a Deflationary Point of View. Ed. P. Horwich. Oxford: Oxford University Press, 2004. 86–104. Hussain, N. ‘Honest Illusion: Valuing for Nietzsche’s Free Spirits.’ Nietzsche and Morality. Eds. B. Leiter and N. Sinhababu. Oxford: Clarendon Press, 2007. 157–91. ——. ‘The Return of Moral Fictionalism.’ Philosophical Perspectives 18 (2004): 149–87. Joyce, R). ‘The Error in ‘The Error in the Error Theory’.’ Australasian Journal of Philosophy (2010). doi: 10.1080/ 00048402.2010.484465. ª 2011 The Author Philosophy Compass ª 2011 Blackwell Publishing Ltd Philosophy Compass 6/11 (2011): 786–803, 10.1111/j.1747-9991.2011.00442.x 802 Fictionalism in Metaphysics ——. ‘Moral Fictionalism.’ Fictionalism in Metaphysics. Ed. M. Kalderon. New York: Oxford University Press, 2005. 287–313. ——. The Myth of Morality. Cambridge: Cambridge University Press, 2001. Kalderon, M., ed. Fictionalism in Metaphysics. New York: Oxford University Press, 2005a. ——. Moral Fictionalism. Oxford: Clarendon Press, 2005b. Kim, S. ‘Modal Fictionalism and Analysis.’ Fictionalism in Metaphysics. Ed. M. Kalderon. New York: Oxford University Press, 2005. 116–33. Kroon, F. ‘Descriptivism, Pretence, and the Frege-Russell Problems.’ Philosophical Review 113 (2004a): 1–30. ——. ‘Realism and Dialetheism.’ The Law of Non-Contradiction. Eds. G. Priest, J. C. Beall, and B. Armour-Garb. Oxford: Clarendon Press, 2004b. 245–63. Leiter, B. ‘Nietzsche’s Moral and Political Philosophy.’ The Stanford Encyclopedia of Philosophy (Winter 2008 Edition). Ed. Edward N. Zalta, 2008. [Online]. Retrieved on 5 August 2011 from: http://plato.stanford.edu/archives/ win2008/entries/nietzsche-moral-political/. Leng, M. Mathematics and Reality. Oxford: Oxford University Press, 2010. ——. ‘Revolutionary Fictionalism: A Call to Arms.’ Philosophia Mathematica 13 (2005): 277–93. Lenman, J. ‘Against Moral Fictionalism.’ Philosophical Books 49.1 (2008): 23–32. Lewis, D. ‘Counterpart Theory and Quantified Modal Logic.’ Journal of Philosophy 65 (1968): 113–26. ——. On the Plurality of Worlds. Oxford: Basil Blackwell, 1986. ——. ‘Quasi-Realism is Fictionalism.’ Fictionalism in Metaphysics. Ed. M. Kalderon. New York: Oxford University Press, 2005. 314–21. ——. ‘Truth in Fiction.’ American Philosophical Quarterly 15 (1978): 37–46. Liggins, D. ‘Modal fictionalism and Possible-Worlds Discourse.’ Philosophical Studies 138 (2008): 151–60. Linsky, B. and E. Zalta. ‘In Defense of the Simplest Quantified Modal Logic.’ Philosophical Perspectives 8 (1994): 431–58. Mackie, J. Ethics: Inventing Right and Wrong. Harmondsworth, New York: Penguin, 1977. Malament, D. ‘Review of Field’s Science Without Numbers.’ Journal of Philosophy 79 (1982): 523–34. Maund, B. ‘The Illusory Theory of Colours: an Anti-Realist Theory.’ Dialectica 60 (2006): 245–68. Melia, J. ‘The Conservativeness of Mathematics.’ Analysis 66.3 (2006): 202–8. Nolan, D. ‘Modal Fictionalism.’ The Stanford Encyclopedia of Philosophy (Fall 2008 Edition). Ed. Edward N. Zalta, 2008. [Online]. Retrieved on 5 August 2011 from: http://plato.stanford.edu/archives/fall2008/entries/fictionalismmodal/. ——. ‘Three Problems for ‘Strong’ Modal Fictionalism.’ Philosophical Studies 87 (1997): 259–75. ——. Topics in the Philosophy of Possible Worlds. New York: Routledge, 2002. ——, G. Restall, and C. West. ‘Moral Fictionalism Versus the Rest.’ Australasian Journal of Philosophy 83 (2005): 307–30. —— and J. O’Leary-Hawthorne. ‘Reflexive Fictionalisms. ’Analysis 56 (1996): 26–32. Noonan, H. ‘In Defence of the Letter of Fictionalism.’ Analysis 54 (1994): 133–9. O’Leary-Hawthorne, J. ‘What Does van Fraassen’s Critique of Scientific Realism Show?’ The Monist 77 (1994): 128–45. Plantinga, A. ‘Actualism and Possible Worlds.’ Theoria 42 (1976): 139–60. Prior, A. N. and K. Fine. Worlds, Times and Selves. London: Duckworth, 1977. Putnam, H. ‘What is Mathematical Truth?’ Historia Mathematica 2 (1975): 529–43. Reprinted in H. Putnam, Mathematics, Matter and Method: Philosophical Papers Vol. II, Cambridge: Cambridge University Press (1979). Quine, W. V. O. ‘On What There Is.’ Review of Metaphysics 2.5 (1948): 21–38. Reprinted in From a Logical Point of View, Cambridge, MA: Harvard University Press (1953). Recanati, F. Oratio Obliqua, Oratio Recta. Cambridge, MA: MIT Press, 2000. Rosen, G. ‘Modal Fictionalism.’ Mind 99 (1990): 327–54. ——. ‘A Problem for Fictionalism about Possible Worlds.’ Analysis 53 (1993): 71–81. ——. ‘Modal Fictionalism Fixed.’ Analysis 55 (1995): 67–73. ——. ‘Problems in the History of Fictionalism.’ Fictionalism in Metaphysics. Ed. M. Kalderon. New York: Oxford University Press, 2005. 14–64. ——. ‘What is Constructive Empiricism?’ Philosophical Studies 74 (1994): 143–78. Sainsbury, M. Fiction and Fictionalism. London and New York: Routledge, 2009. Sainsbury, R. M. ‘Fiction and Acceptance-Relative Truth, Belief, and Assertion.’ Truth in Fiction. Ed. F. Lihoreau. Heusenstamm: Ontos Verlag, 2010. 137–52. Schurz, G ‘Structural Correspondence, Indirect Reference, and Partial Truth: Phlogiston Theory and Newtonian Mechanics.’ Synthese 180 (2011): 103–20. Shapiro, S. ‘Conservativeness and Incompleteness.’ Journal of Philosophy 80 (1983): 521–31. Sider, T. ‘The Ersatz Pluriverse.’ Journal of Philosophy 99 (2002): 279–315. Stalnaker, R. ‘Possible Worlds.’ Nous 10 (1976): 65–75. Stanley, J. ‘Hermeneutic Fictionalism.’ Midwest Studies in Philosophy 25 (2001): 36–71. ª 2011 The Author Philosophy Compass ª 2011 Blackwell Publishing Ltd Philosophy Compass 6/11 (2011): 786–803, 10.1111/j.1747-9991.2011.00442.x Fictionalism in Metaphysics 803 Thomasson, A. L. Fiction and Metaphysics. Cambridge: Cambridge University Press, 1999. Vaihinger, H. The Philosophy of ‘As If’. Trans. C. K. Ogden. London: Kegan Paul, 1924. First published in German as Die Philosophie des Als Ob, Berlin: Verlag von Reuther & Reichard (1911). Van Fraassen, B. ‘The Agnostic Subtly Probabilified.’ Analysis 58.3 (1998): 212–20. ——. ‘Gideon Rosen on Constructive Empiricism.’ Philosophical Studies 74 (1994): 179–92. ——. The Scientific Image. Oxford: Oxford University Press, 1980. Van Inwagen, P. Material Beings. Ithaca, New York: Cornell University Press, 1990. Voltaire. ‘Epı̂tre à l’Auteur du Livre des Trois Imposteurs.’ Œuvres Complètes de Voltaire. Ed. L. Moland. Paris: Garnier, 1877–85, 1768. Vol. 10, 402–5. Walton, K. ‘Existence as Metaphor?’ Empty Names, Fiction and the Puzzles of Non-Existence. Eds. A. Everett and T. Hofweber. Stanford: CSLI Publications, 2000. 69–94. ——. ‘Metaphor and Prop-Oriented Make-Believe.’ European Journal of Philosophy 1 (1993): 39–57. ——. Mimesis as Make-Believe: On the Foundations of the Representational Arts. Cambridge: Harvard University Press, 1990. Woodbridge, J. ‘Truth as a Pretense.’ Fictionalism in Metaphysics. Ed. M. Kalderon. New York: Oxford University Press, 2005. 134–77. Yablo, S. ‘Abstract Objects: A Case Study.’ Philosophical Issues 12 (2002): 220–40. ——. ‘Does Ontology Rest on a Mistake?’ Proceedings of the Aristotelian Society 72 (Suppl.) (1998): 229–83. ——. ‘Go Figure: A Path Through Fictionalism.’ Midwest Studies in Philosophy 25 (2001): 72–102. ——. ‘The Myth of Seven.’ Fictionalism in Metaphysics. Ed. M. Kalderon. New York: Oxford University Press, 2005. 88–115. ——. ‘A Paradox of Existence.’ Empty Names, Fiction and the Puzzles of Non-Existence. Eds. A. Everett and T. Hofweber. Stanford: CSLI Publications, 2000. 275–312. ª 2011 The Author Philosophy Compass ª 2011 Blackwell Publishing Ltd Philosophy Compass 6/11 (2011): 786–803, 10.1111/j.1747-9991.2011.00442.x