Expressing Our Attitudes

Expressing Our Attitudes 1. Introduction This volume collects nine of Mark Schroeder’s essays on expressivism, two of which are previously unpublished, along with a substantial introduction that helpfully ties them all together.1 The essays work very nicely as a collection. They are mutually illuminating, and together they make a ‘cumulative case’ for a particular conclusion – namely, that expressivist theories are best understood in terms of their ‘surprising and novel views’ about the nature of propositions and propositional attitudes (vii). As Schroeder sees things, the most promising way for expressivists to develop their view is for them to ‘get over’ their hesitation about propositions (71) and to embrace an expressivist-friendly account of what propositions are. This requires going beyond the ‘merely deflationary talk about propositions’ that is typically endorsed by expressivists like Blackburn and Gibbard and developing a view on which propositions can ‘actually do some theoretical work’ (75–6). The volume is divided into four parts. The essays in Part 1 (Chs. 1–2) serve to mark out places where Schroeder now thinks he was mistaken about expressivism. The mistakes were: thinking that expressivists had to reject the ‘same-content’ account of the expression relation, and thinking that the van Roojen problem was fatal for higher-order attitude versions of expressivism; Schroeder now thinks that both problems can be avoided by embracing propositions. The essays in Part 2 (Chs. 3–5) explore the consequences, for expressivism, of understanding both normative and non-normative beliefs as instances of the same type of attitude that are distinguished by their contents. (If you’re wondering how this sort of view could count as a form of expressivism, read on.) The essays in Part 3 (Chs. 6–7) explore how conservative an expressivist theory can be in its commitments about the nature of the mind. Roughly, they ask whether expressivists can maintain a familiar ‘Humean’ picture of the mind, on which the mind can be decomposed into two different kinds of mental state with different ‘directions of fit’. The essays in Part 4 (Chs. 8–9, both previously unpublished) explore the limits of purely formal methods in semantics. Chapter 8 argues that the formal systems we use when doing semantics are not ‘autonomous’ vis-à-vis the philosophical interpretations we give to those systems, and it explains how expressivism differs from relativism. Chapter 9 considers some of the obstacles facing expressivism about epistemic expressions, and it argues that they call for philosophical reflection about the kinds of things that contents need to be. This selective summary mentions only some of the topics that Schroeder treats in this volume with characteristic rigor and elegance. There are many others as well. Rather than trying to survey the volume as a whole, though, what I want to do in this critical notice is to focus on one part of the conclusion that Schroeder takes his essays collectively to support – the part which says that ‘we should understand expressivism as coming with particular commitments about the nature of propositions’ (15). I’ll mostly set aside the part which says that we should understand expressivism as coming with particular commitments about propositional attitudes, since that is in some ways a more familiar and less controversial idea. We are well accustomed to thinking of expressivist views as having something interesting to say about the nature of, say, moral belief; we are far less accustomed to thinking of them as having (or as requiring) an interesting theory about the nature of the moral proposition.2 Schroeder thinks that they need both, and it is the latter part of this claim – and the motivations for it – that I’ll be primarily interested in here. 2. Schroeder's understanding of expressivism Let me begin by raising a very general concern about Schroeder’s approach. All of the sophisticated propositional theories that Schroeder develops on behalf of the expressivist here have highly unintuitive consequences. Schroeder himself acknowledges some of these consequences (see, e.g. 71–2, 96 n. 24, 204–5), and I will discuss some others below (see also Richard 2015). This raises an obvious question: What is the point of developing these views, and why should we be interested in them – given that they are, as Schroeder seems to recognize, highly unintuitive? There would clearly be no point in offering up a new theory about moral language simply for the sake of having something new or surprising to say. Such a theory would need to be motivated by something else, something besides the fact that it is ‘surprising and novel’ (vii), in order to be worth developing and discussing. Schroeder motivates the development of his unconventional propositional theories by tying them to a philosophical tradition that is independently interesting, namely expressivism. To that end, he frequently says things like this: his interest is in showing us how expressivism ‘pushes us into the project of completely rethinking the nature of thought and language more generally’, and in showing us how expressivist theories are ‘best understood’ (vii); he is interested in ‘where [the noncognitivist] revolution stands today’ (1); he is considering a set of dialectical pressures that ‘forces expressivists into deep commitments about the mind’, and that ‘results in commitments about propositions’ (15); he is offering arguments that show that expressivists ‘need to believe in propositions’ (16) and so on. In short, he presents his propositional theories as things that we ought to be interested in, insofar as we are interested in the expressivist project, because they are the kinds of theories that an expressivist will need to endorse if she is going to make her project work. As he says in the Introduction: The central theme of the essays is that though expressivism, the contemporary heir to the noncognitivism of the 1930s, may be developed in a range of different ways, and may have intriguing and surprising applications, all of these require rich commitments about the mind and about the nature of propositions, truth, and propositional attitudes. (1) Unfortunately, the essays consistently fall short in this respect. The closest Schroeder comes to showing anything like this is in Chapter 2, where he shows that a particular strategy suggested by Blackburn in the 1980s – one that Blackburn has since disavowed (Blackburn 2010: 1–2; 2015: xvi) – requires taking on a controversial view about what propositions are. In general, though, the essays manage to generate the appearance of success on this front by underdescribing – or in some cases misdescribing – the commitments of ‘traditional expressivism’, as well as the views that Schroeder would have us put in their place. Schroeder takes expressivism to be a view about meaning. He emphasizes Gibbard’s (2003: 6–7) familiar claim that expressivists attempt to explain the meaning of a certain kind of term indirectly, by saying what state of mind it can be used to express. And he takes this strategy to involve the provision of a distinctive kind of semantics for the relevant domain, one that differs from traditional, truth-conditional approaches. As Schroeder puts it: [T]he expressivist idea is to give a compositional semantics which generates, for each sentence, ‘P’, what it is to believe that P. (116)3 So, for example, the metaethical expressivist tries to give a compositional semantics for moral language that tells us what it is to believe that M, for any moral sentence ‘M’. Traditionally, Schroeder thinks, expressivists have tried to do this without invoking propositions. That’s where they went wrong. Rather than trying to do without propositions, or trying to get by with ‘merely deflationary talk’ about them (75), expressivists ought to embrace propositions and give an expressivism-friendly theory about what they are like. This is a controversial way of understanding expressivism. It has become increasingly common in recent years to think of expressivism as a view in ‘meta-semantics’ – i.e. roughly, as a view about why normative sentences have the semantic values they do, rather than a view about what those semantic values are (Charlow 2014; Chrisman 2012 ; Ridge 2014; Silk 2013). Others (Blackburn 2013; Price 2011, 2013) think that expressivism is best understood in terms of a ‘pragmatist’ approach to theorizing about meaning – i.e. roughly, one that attempts to explain how we get from certain kinds of doings to certain kinds of meanings. It is disappointing that Schroeder does not engage with these alternative ways of understanding the view, since he is, after all, attempting to illustrate commitments that are supposed to be required by expressivism, regardless of how the view is developed. In any case, I think we can see that there’s something off about the way that Schroeder characterizes expressivism without having to wade into those debates. Notice, to begin with, that we could construct the kind of compositional theory Schroeder describes above by simply assigning every sentence ‘P’ to the belief that P. ‘Murder is wrong’ gets assigned to the belief that murder is wrong; ‘Kindness is good’ gets assigned to the belief that kindness is good; and so on. Even the truth-conditional theorist can tell this kind of story, e.g. by associating every sentence ‘P’ with the proposition that P, and then insisting that each such sentence gets associated (derivatively, as it were) with the belief that P. Crucially, in telling this kind of story, the truth-conditional theorist does not end up saying anything that an expressivist would need to reject. The problem with this sort of story, from the expressivist’s point of view, is not that it says anything false, but that it is unilluminating in certain respects (see, e.g. Blackburn 1998: 50–1, Gibbard 2003: 62–3). This suggests that the expressivist’s chief concerns arise at a different level from the one we are operating at when we are developing a compositional semantics for moral language. Consider the following remarks from Blackburn and Gibbard, respectively: I no longer think of the predicates involved in evaluation … as semantically different from others. But this is only because semantic categories are too coarse-grained to do the necessary work: they lull us into remaining ‘unconscious of the prodigious diversity of all the everyday language-games’. … Semantics tells us only where we have ended up; my explanatory interest is in suggesting why and how we have ended up where we did. (Blackburn 2010: 2). I claim … that we can explain much obliquely, characterizing normative concepts by describing independently the states of mind that employ them. We can of course also characterize normative beliefs … in terms of their content. We can identify a normative judgment as the belief that such-and-such – say, as the belief that one ought to shun pain, or that politicians are all sleazy. But this, I am saying, is not the only way to characterize normative beliefs; we can also explain them as plan laden judgments, or as carrying plan-laden presuppositions. This yields further insight. (Gibbard 2003: 185) I take these remarks – and others like them – to suggest something like the following sort of ‘two-levels’ picture. At one level of description, we can say – quite rightly – that moral beliefs are beliefs in propositions. The belief that stealing is wrong is a belief whose content is, precisely, that stealing is wrong. But if that’s all we say, we’re likely to face certain problems and puzzles – e.g. puzzles about the practicality of moral belief, or about how we could ‘be in touch with’ moral properties. We can work toward dissolving these puzzles by asking – at another level of description – what it is to have a belief with this sort of content. The answer to that question will tell us something about the practical state of mind that constitutes having a moral belief – i.e. belief in a moral proposition. If that’s right, Schroeder’s characterization of the expressivist as trying to give a compositional semantics without appealing to moral propositions seems inapt. Schroeder is aware that there is an issue in the offing here, and he goes on to add, after the characterization quoted above, that ‘interesting or essentially expressivist views go further, and specify that for some sentences ‘P’, to believe that P is not simply to have an ordinary descriptive belief that the world is a certain way’ (116). But Schroeder’s conception of expressivism as a project in compositional semantics leads him to look for reasons that are, as it were, internal to semantics to explain the motivations for being ‘essentially expressivist’ in this sense. He says: [Traditional noncognitivist expressivism in metaethics] is essentially expressivist, because it can’t be simply derivatively generated by any truth-conditional semantic theory. Though there is something it is for grass to be green, on this view, there is nothing it is for stealing to be wrong. So a semantic theory for ‘wrong’ can’t work by telling us what it is for stealing to be wrong; it must work by telling us what it is to believe that something is wrong, which is quite different from having an ordinary descriptive belief about how things are. (117) But this is seriously misleading. Even in the ordinary descriptive case, truth-conditional theories do not work by telling us what it is for grass to be green. (That is the job of the philosopher or the colour scientist.) What such theories do is (roughly) to associate the sentence, ‘Grass is green’, with the conditions under which it would be true. So the motivation for going ‘essentially expressivist’ can't be that in the ordinary descriptive case, semantic theories work by ‘telling us what it is’ for the relevant state of affairs to obtain, whereas they can’t do this in the moral case. Rather, the thought would have to be that in the moral case a semantic theory can’t work by associating truth conditions with moral sentences because there are no conditions under which such sentences are true. But, of course, that cannot be right either, since the expressivist does think that such sentences can be true. What has gone wrong here, I think, is that Schroeder has failed to fully appreciate the consequences of the expressivist’s deflationism. One thing we have learned from quasi-realists over the years is that deflationism about truth tends to go hand in hand with deflationism about beliefs, facts, propositions, and other inter-related notions, and it is a familiar point that once expressivists embrace deflationism, it becomes increasingly difficult to characterize their view straightforwardly in semantic terms (see, e.g. Dreier 2004; Price 1994). And there has been much debate about how deflationism interacts with the project of giving a truth-conditional semantics. Unfortunately, Schroeder’s essays do not sufficiently engage with these issues. Insofar as Schroeder’s arguments depend upon casting the expressivist project in terms of a particular strategy within compositional semantics, they thus seem at best incomplete. He doesn’t show that expressivism ‘result in’ or ‘require’ surprising and novel views about propositions and propositional attitudes. What he shows, at best, is that the expressivist will not be able to develop an unorthodox approach to compositional semantics without embracing such views. But that may be just another reason to stop thinking of expressivism as a view within compositional semantics. 3. What do we get from ‘the propositions theory’? In this section, I want to look at Schroeder’s propositional approach in a bit more detail. I’ll start by explaining a central distinction and illustrating it with one of the propositional views Schroeder advances on behalf of the expressivist. Then I’ll consider one of the main arguments he offers in support of this approach – one that would seem to apply even to expressivists who do not think of their view in terms of the provision of a distinctive kind of compositional semantics. Schroeder thinks that expressivists have been largely mistaken in trying to do without propositions – or in trying to get by with ‘merely deflationary talk’ about them (75). The key to seeing this is to distinguish between propositions and representational contents. This distinction shows up in a number of places in the volume, but it receives its most thorough development and discussion in Chapter 3, which Schroeder takes to be ‘the most central essay of the volume’ (16). The rest of my discussion will focus primarily on this essay, since it helpfully summarizes many of the points that Schroeder makes elsewhere in the volume, and since it provides his most direct argument for the claim that expressivists need (non-deflationary) propositions. Here is the distinction. Propositions are the objects of propositional attitudes, they are the primary bearers of truth and falsity, and they are the meanings of declarative sentences in context. Representational contents, by contrast, are more metaphysically committal types of entities: they serve to carve up the world; they correspond to distinctions in reality; they are associated with metaphysical commitment; and they are the appropriate objects of excluded middle (17, 27, 69, 85, 88). Schroeder’s proposal on behalf of the expressivist is then to insist that beliefs can be carved up at two different ‘levels’ (89). At one level, we just say that beliefs are relations to propositions. Some, but not all, propositions determine representational contents. Ordinary descriptive beliefs are relations to these propositions, and so we can also carve them up at a further level of description – one that makes explicit the fact that the belief is (also) a relation to a representational content. Moral beliefs, by contrast, are relations to propositions that don’t determine representational contents, and so they can’t be carved up at this further level. That is all rather abstract, and it is easier to see what Schroeder has in mind here by looking at how this distinction figures in his own development of a propositional expressivism. Consider, then, the view that Schroeder calls biforcated attitude semantics (first advanced in Schroeder 2008; developed here in Chs. 3 and 4). According to this theory, propositions are pairs of entailing properties that an agent might have, and beliefs are complex attitudes of being for each of these properties (92–4). For example, the proposition that grass is green, on this view, is identified with the following pair of properties: ⟨proceeding as if grass is green, not proceeding as if grass is not green⟩ To believe this proposition is to be for proceeding as if grass is green, and to be for not proceeding as if grass is not green. Being for some property is a functional state that ‘leads one to acquire that property, other things being equal’ (92). Someone who believes that grass is green will thus be in a functional state that leads him to acquire the property of proceeding as if grass is green, which involves being related in a particular way to a representational content – namely, grass’s being green. By contrast, the proposition that murder is wrong, on this view, is identified with the following pair of properties: ⟨avoiding murder & disapproving of murder, disapproving of murder⟩ Someone who believes that murder is wrong will thus be for avoiding murder and disapproving of murder (the second member of the pair is redundant here; it’s included for technical reasons that we can safely ignore for present purposes). This involves being related to murder, avoidance, and disapproval. But, since this proposition does not determine any representational content in the way that the proposition that grass is green does, believing it does not involve bearing any relation to a representational content. It simply involves bearing a relation to the proposition – i.e. to the relevant pair of properties. Schroeder’s view – particularly his insistence that beliefs can be carved up at two different ‘levels’ – might seem similar to what I described above as a ‘two-levels’ picture. But in fact the two views are very different. One crucial difference, for our purposes, is that Schroeder wants to explain the non-representational character of moral beliefs by appealing to a substantial view about the kinds of things that moral propositions are, whereas deflationary expressivists like Blackburn and Gibbard want to maintain a principled silence about the latter.4 Why, then, do we need a substantial theory of the moral proposition? Some of what Schroeder says here seems to stem from a misunderstanding of the deflationary view – one that construes the deflationist as not believing in ‘real propositions’ (70), but merely talking ‘as if’ there are propositions (68). But the deflationary expressivist is not trying to talk as if there are moral propositions even though there aren’t any. She simply thinks that there’s not much to say about moral propositions, beyond the fact that they are a ‘focus’ for a certain kind of practical thought and discussion (Blackburn 1998: 50). Though we can’t say much about them directly, we can nevertheless bring them into view by letting them ‘emerge’ (Blackburn 2009: 208, Gibbard 2015: 273) from a story about practical thought that does not appeal to such propositions at the outset. Schroeder thus comes closer to his target when he characterizes the deflationist as simply not appealing to propositions for certain theoretical purposes (75). He then lists five problems that are supposed to arise for an expressivist who adopts such a deflationary view, all of which, he claims, ‘are easily solved by any view which postulates propositions across the board’ (75). Three of these problems bring out the difficulty of giving a compositional semantics without appealing to propositions – problems in accounting for negation (78–9); problems in accounting for quantification into the complement position of verbs like ‘believes’, and identification of the things we are quantifying over in such cases (e.g. when we say things like ‘Jill said the very thing that Jack believes’) (79); and problems in accounting for modals with moral prejacents (e.g. in claims like ‘It might be that killing is wrong’) (82–4). I have already raised some concerns about the extent to which these problems should be seen as damning to expressivism per se – that depends on whether we take the expressivist to be trying to give a compositional semantics of the sort Schroeder describes – so I will set them aside here. Let me note, though, that Schroeder’s discussion here offers valuable lessons for anyone who does want to carry out this sort of project. The other two problems are ones that would apparently arise even for an expressivist who is not interested in this kind of compositional project, for they are supposed to stem from the expressivist’s insistence that moral and non-moral beliefs are, at the most fundamental level, two different kinds of attitude; and this is something that even the ‘two-levels’ expressivist will want to say.5 So it is worth looking at these two problems in a bit more detail. The first problem is to explain why moral and non-moral beliefs turn out to ‘have a striking array of features in common’, given that they are fundamentally different kinds of state (80). The second problem is a more general version of the first one; it is, in effect, is to explain why there are both moral and non-moral hopes, suppositions, desires, and so on ‘for every one of the other attitudes’, as well as why all of these pairs of moral and non-moral attitudes ‘always have so much in common’ (82). Since Schroeder takes the first problem to be a particular instance of the second one, I will treat the two together. The features that Schroeder takes moral and non-moral beliefs to have in common are Moral beliefs and non-moral beliefs ‘come in varying levels of confidence’ (80). Moral beliefs and non-moral beliefs ‘share the same disagreement properties’ (i.e., believing that x is wrong disagrees with believing that x is not wrong, just as believing that x is green disagrees with believing that x is not green) (80). Moral beliefs ‘combine with desires to produce motivation to action in the same way as non-moral beliefs do’ (80).6 Moral beliefs ‘share much of the same phenomenology as non-moral beliefs – phenomenology that is not shared by hoping, assuming, preferring, wondering, or any of their ilk’ (80). He then says that for the propositional theorist it is ‘no mystery’ that these two kinds of belief have so much in common: According to this view … these two states … share a common core … the common attitude of belief that merely takes different contents between the belief that stealing is wrong and the belief that grass is green. It is no wonder, according to the theory of propositions, that believing that stealing is wrong and believing that grass is green have so much in common – it derives from their common core. (81) By contrast, expressivists ‘who allow only for a deflationary way of talking about propositions’ have to accept that ‘it is essentially a coincidence that two states which are fundamentally so different turn out to share so many of the same features’ (81). Here, I think Schroeder moves much too quickly. Surely there can be different kinds of states – different in something like the way that the deflationary, ‘two-levels’ expressivist claims – that nevertheless share a range of important features. Nor is it obvious that the best or the only explanation of why such states share these features will have to appeal to the fact that they both consist in bearing a certain relation to a proposition. Certainly there could be other explanations. I can see why there might seem to be a problem here if all that the deflationary expressivist had to say about moral and non-moral beliefs is that they are very different kinds of state. And that is indeed how Schroeder characterizes such views here. But that is not all that theorists like Blackburn and Gibbard have to say. Gibbard (2003), for example, goes to great lengths to explain why moral and non-moral beliefs have so much in common. In the ordinary descriptive case, he says, there is a distinction between ‘mere apprehension’ and ‘full-fledged judgment’, and he insists that there is a corresponding distinction in the practical case between ‘straight attitudes’ and ‘attitudinal judgments’ (2003: 79–82). One of Gibbard’s aims in telling this kind of story is precisely to explain why, in both the ordinary descriptive case and in the practical case, we end up with beliefs that involve a kind of judgment. As Gibbard says: Once a being is capable of agreeing and disagreeing with possible states of mind, both factual apprehendings and straight attitudes become members of larger classes: factual apprehendings become a special class of factual judgments, and straight attitudes become a special class of purely attitudinal judgments. Factual judgments and attitudinal judgments, moreover, are special cases of judgment in general: some of these judgments are attitude-laden … and the remainder are purely factual. (81–2). It seems plausible – surely it is at least possible – that the kind of story Gibbard tells here would have the resources to explain why both moral and non-moral beliefs have so many features in common. Indeed, we can think of this story as offering another way of identifying a ‘common core’ between moral and non-moral belief, without simply saying that they both consist in bearing a certain relation to a proposition. Of course, in the end, this sort of story may not be able to explain the things that Schroeder wants to explain. My point is just that this is not obvious; it’s something we would need to argue for. And we would need an even stronger argument for the claim that no such story could be told, either by Gibbard or by anyone else. But Schroeder does not provide either of these. He simply asserts that all deflationists will have to take it as a coincidence that moral and non-moral beliefs turn out to have so much in common. And that seems unwarranted. As I say, the issue here stems partly from the fact that Schroeder underdescribes the deflationist views he is arguing against. There is a parallel problem, I think, for the view that Schroeder takes himself to be defending – the so-called ‘propositions theory’. To see this, turn now to Schroeder’s suggestion that the problems he is discussing here are all ‘easily solved’ by any view that posits non-deflationary propositions (75). And consider again his view (qua biforcated attitude semanticist) about what propositions are. The proposition that grass is green is a certain pair of properties, the proposition that murder is wrong is a very different pair of properties, and belief in both of these propositions consists in being for each of the properties in the pair. Schroeder says that ‘it is straightforward how biforcated attitude semantics explains the features that moral and non-moral beliefs have in common – they will be explained by their common core, the fact that they are both biforcated attitudes’ (95). However, while this view may have a straightforward explanation of the fact that moral and non-moral beliefs share the same disagreement properties, and perhaps also of the fact that they share much of the same phenomenology, it is not at all straightforward how it explains the other two properties Schroeder mentions. How does it explain, for example, the fact that moral and non-moral beliefs come in varying degrees of confidence? What is it to be confident in my attitude of being for certain pairs of properties? Again, my point is not that Schroeder can’t answer these questions, but just that they do need to be answered, and that Schroeder doesn’t tell us how to do this. The natural way to answer them, for someone who accepts Schroeder’s view, would seem to be something like this. In the ordinary descriptive case, the believer is conceived as being for (e.g.) the property of proceeding as if something is the case, and that is something we can do more or less tentatively. In the moral case, the believer is conceived as being for (e.g.) the property of avoiding and disapproving of a certain kind of act, and that too seems like something we can do more or less tentatively. This notion of tentativeness seems to correspond to some intuitive notion of confidence, so the fact that we can have varying degrees of confidence in both moral and non-moral beliefs might be explained by the fact that both of these are states of being for, together with the fact that we can be more or less tentative in our states of being for. There are two things to note here, though. The first is that this sort of explanation has nothing to do with the fact that both of these states are conceived as relations to propositions. All of the explanatory work is done by the story about being more or less tentative in one’s practical dispositions. And the second point is that this kind of story is equally available to Schroeder’s opponent. A deflationist about propositions can similarly say that both moral and non-moral beliefs can exhibit varying degrees of confidence because we can be more or less tentative in, for example, our representations and our plans. So it’s hard to see how Schroeder’s propositional view gains any advantage here. Consider next the claim that moral beliefs ‘combine with desires to produce motivation to action in the same way as non-moral beliefs do’ (80). On Schroeder’s view, both moral and non-moral beliefs will dispose you to acquire certain pairs of properties, but beyond that it is far from clear that they will have very much in common when it comes to their motivational profile (indeed, Schroeder seems to say as much on p. 95). To believe that grass is green is to be disposed to acquire (e.g.) the property of proceeding as if grass is green, and to believe that murder is wrong is to be disposed to acquire the property of avoiding and disapproving of murder. But how will these beliefs interact with, say, the desire to look at something green, or the desire to do something wrong? It is not at all obvious what to say here, in part because it is not at all obvious how Schroeder’s beliefs even could interact with desires in the usual way. This brings us to the second problem I mentioned above – the problem of explaining why there are moral and non-moral desires, hopes, suppositions, etc., and why all of these pairs share so many features in common. Schroeder says that any theorist who posits non-deflationary propositions can provide ‘the most elegant explanation possible’ here (82). The explanation is that so long as there are such things as the proposition that stealing is wrong and the proposition that grass is green, those propositions may be taken as the objects of any of the attitudes, and any attitude toward the proposition that stealing is wrong will have as much in common with the same attitude toward the proposition that grass is green… (82) Once again, though, it is not clear how this explanation works once we plug in Schroeder’s own view about what propositions are. Consider, for example, the proposition that stealing is wrong, which is identified with the pair of properties ⟨avoiding stealing & disapproving of stealing, disapproving of stealing⟩. It is far from obvious that, so long as this proposition exists, it may be taken as the object of any of the attitudes. I have no idea what it would be, for example, to bear the relation of hoping toward this proposition. It doesn’t seem like the right sort of entity to be an object of hope. Clearly this cannot be a matter of hoping that I will avoid and disapprove of stealing, since that would be hoping that an ordinary descriptive proposition is true (one that concerns my future doings, and not one that concerns the wrongness of stealing). It’s hard to know what else we might say, though, and Schroeder doesn’t offer much help. He suggests that all of the explanations here will be straightforward once we’ve got propositions in the picture, but I take these remarks to show that this is not actually the case. 4. Conclusion Schroeder covers a lot of ground in these essays, and his arguments are ambitious. He attempts to show that all versions of deflationism fail; that all versions of the ‘propositions theory’ solve his problems with ease; that every way of carrying out the expressivist project requires taking on substantial views about the nature of propositions; and so on. These are very difficult claims to argue for. I have been pointing to some ways in which Schroeder’s arguments for these strong claims seem to me to fall short. These criticisms notwithstanding, Schroeder’s essays raise a number of interesting and important challenges for anyone interested in carrying out the expressivist project, and there is much to be learned from working through his arguments. Deflationists, in particular, will have a rich set of issues to grapple with here. Footnotes 1 Expressing Our Attitudes: Explanation and Expression in Ethics, Volume 2, by Mark Schroeder. Oxford University Press, 2015, xii + 266 pp. 2 For concreteness, I will focus throughout on metaethical expressivism, but it’s worth noting that Schroeder takes his arguments to apply to nondescriptivist views more generally. 3 Schroeder says that this is only a ‘first pass’, but his reservations here come from considerations about slurs that are immaterial to the present discussion. 4 Another way of seeing the difference is to notice that the ‘two-levels’ expressivist has something further to say about moral belief at the ‘other level of description’, whereas Schroeder’s claim is precisely that moral beliefs can’t be carved up at any further level of description (while representational beliefs can). 5 Or at least this is something the two-levels expressivist may want to say. The issue depends in part on whether she wants to go ‘global’ with her expressivism (see Blackburn 2013; Gibbard 2015; Price 2013). 6 Here, Schroeder seems to me to be overlooking one of the key motivations for traditional expressivism, which is precisely that ordinary beliefs don’t produce motivation in the same way as moral beliefs. References Blackburn S. 1998. Ruling Passions . Oxford: Oxford University Press. Blackburn S. 2009. Truth and a priori possibility: Egan’s charge against quasi-realism. Australasian Journal of Philosophy  87: 201– 13. Google Scholar CrossRef Search ADS   Blackburn S. 2010. Practical Tortoise Raising and Other Philosophical Essays . Oxford: Oxford University Press. Google Scholar CrossRef Search ADS   Blackburn S. 2013. Pragmatism: all or some? In expressivism, Pragmatism and Representationalism , ed. Price H., 67– 84. Cambridge: Cambridge University Press. Google Scholar CrossRef Search ADS   Blackburn S. 2015. Apologia pro vita sua. In Passions and Projections: Themes from the Philosophy of Simon Blackburn , ed. Johnson R.N., Smith M., xv– xix. Oxford: Oxford University Press. Charlow N. 2014. The problem with the Frege-Geach problem. Philosophical Studies  167: 635– 65. Google Scholar CrossRef Search ADS   Chrisman M. 2012. On the meaning of ‘ought’. In Oxford Studies in Metaethics , vol. 7, ed. Shafer-Landau R., 304– 32. Oxford: Oxford University Press. Dreier J. 2004. Meta-ethics and the problem of creeping minimalism. Philosophical Perspectives  18: 23– 44. Google Scholar CrossRef Search ADS   Gibbard A. 2003. Thinking How to Live . Cambridge, MA: Harvard University Press. Gibbard A. 2015. expressivism and the truth in representation. In Meaning Without Representation , ed. Gross. S., Tebben N., Williams M., 210– 23. Oxford: Oxford University Press. Google Scholar CrossRef Search ADS   Price H. 1994. Semantic minimalism and the Frege point. In Foundations of Speech Act Theory: Philosophical and Linguistic Perspectives , ed. Tsohatzidis S. L., 132– 55. New York: Routledge. Price H. 2011. Naturalism Without Mirrors . New York: Oxford University Press. Price H. 2013. expressivism, Praagmatism and Representationalism . Cambridge: Cambridge University Press. Google Scholar CrossRef Search ADS   Richard M. 2015. What would an expressivism semantics be? In Meaning Without Representation , ed. Gross. S., Tebben N., Williams M., 137– 59. Oxford: Oxford University Press. Google Scholar CrossRef Search ADS   Ridge M. 2014. Impassioned Belief . Oxford: Oxford University Press. Google Scholar CrossRef Search ADS   Schroeder M. 2008. Being For: Evaluating the Semantic Program of expressivism . Oxford: Oxford University Press. Google Scholar CrossRef Search ADS   Silk A. 2013. Truth conditions and the meanings of ethical terms. In Oxford Studies in Metaethics , vol 8, ed. Shafer-Landau R. 195– 222. Oxford: Oxford University Press. © The Author 2017. Published by Oxford University Press on behalf of The Analysis Trust. All rights reserved. For Permissions, please email: journals.permissions@oup.com http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Analysis Oxford University Press

Expressing Our Attitudes

Analysis , Volume 78 (1) – Jan 1, 2018

Loading next page...
 
/lp/ou_press/expressing-our-attitudes-ZgNMVkPWh0
Publisher
Oxford University Press
Copyright
© The Author 2017. Published by Oxford University Press on behalf of The Analysis Trust. All rights reserved. For Permissions, please email: journals.permissions@oup.com
ISSN
0003-2638
eISSN
1467-8284
D.O.I.
10.1093/analys/anx090
Publisher site
See Article on Publisher Site

Abstract

1. Introduction This volume collects nine of Mark Schroeder’s essays on expressivism, two of which are previously unpublished, along with a substantial introduction that helpfully ties them all together.1 The essays work very nicely as a collection. They are mutually illuminating, and together they make a ‘cumulative case’ for a particular conclusion – namely, that expressivist theories are best understood in terms of their ‘surprising and novel views’ about the nature of propositions and propositional attitudes (vii). As Schroeder sees things, the most promising way for expressivists to develop their view is for them to ‘get over’ their hesitation about propositions (71) and to embrace an expressivist-friendly account of what propositions are. This requires going beyond the ‘merely deflationary talk about propositions’ that is typically endorsed by expressivists like Blackburn and Gibbard and developing a view on which propositions can ‘actually do some theoretical work’ (75–6). The volume is divided into four parts. The essays in Part 1 (Chs. 1–2) serve to mark out places where Schroeder now thinks he was mistaken about expressivism. The mistakes were: thinking that expressivists had to reject the ‘same-content’ account of the expression relation, and thinking that the van Roojen problem was fatal for higher-order attitude versions of expressivism; Schroeder now thinks that both problems can be avoided by embracing propositions. The essays in Part 2 (Chs. 3–5) explore the consequences, for expressivism, of understanding both normative and non-normative beliefs as instances of the same type of attitude that are distinguished by their contents. (If you’re wondering how this sort of view could count as a form of expressivism, read on.) The essays in Part 3 (Chs. 6–7) explore how conservative an expressivist theory can be in its commitments about the nature of the mind. Roughly, they ask whether expressivists can maintain a familiar ‘Humean’ picture of the mind, on which the mind can be decomposed into two different kinds of mental state with different ‘directions of fit’. The essays in Part 4 (Chs. 8–9, both previously unpublished) explore the limits of purely formal methods in semantics. Chapter 8 argues that the formal systems we use when doing semantics are not ‘autonomous’ vis-à-vis the philosophical interpretations we give to those systems, and it explains how expressivism differs from relativism. Chapter 9 considers some of the obstacles facing expressivism about epistemic expressions, and it argues that they call for philosophical reflection about the kinds of things that contents need to be. This selective summary mentions only some of the topics that Schroeder treats in this volume with characteristic rigor and elegance. There are many others as well. Rather than trying to survey the volume as a whole, though, what I want to do in this critical notice is to focus on one part of the conclusion that Schroeder takes his essays collectively to support – the part which says that ‘we should understand expressivism as coming with particular commitments about the nature of propositions’ (15). I’ll mostly set aside the part which says that we should understand expressivism as coming with particular commitments about propositional attitudes, since that is in some ways a more familiar and less controversial idea. We are well accustomed to thinking of expressivist views as having something interesting to say about the nature of, say, moral belief; we are far less accustomed to thinking of them as having (or as requiring) an interesting theory about the nature of the moral proposition.2 Schroeder thinks that they need both, and it is the latter part of this claim – and the motivations for it – that I’ll be primarily interested in here. 2. Schroeder's understanding of expressivism Let me begin by raising a very general concern about Schroeder’s approach. All of the sophisticated propositional theories that Schroeder develops on behalf of the expressivist here have highly unintuitive consequences. Schroeder himself acknowledges some of these consequences (see, e.g. 71–2, 96 n. 24, 204–5), and I will discuss some others below (see also Richard 2015). This raises an obvious question: What is the point of developing these views, and why should we be interested in them – given that they are, as Schroeder seems to recognize, highly unintuitive? There would clearly be no point in offering up a new theory about moral language simply for the sake of having something new or surprising to say. Such a theory would need to be motivated by something else, something besides the fact that it is ‘surprising and novel’ (vii), in order to be worth developing and discussing. Schroeder motivates the development of his unconventional propositional theories by tying them to a philosophical tradition that is independently interesting, namely expressivism. To that end, he frequently says things like this: his interest is in showing us how expressivism ‘pushes us into the project of completely rethinking the nature of thought and language more generally’, and in showing us how expressivist theories are ‘best understood’ (vii); he is interested in ‘where [the noncognitivist] revolution stands today’ (1); he is considering a set of dialectical pressures that ‘forces expressivists into deep commitments about the mind’, and that ‘results in commitments about propositions’ (15); he is offering arguments that show that expressivists ‘need to believe in propositions’ (16) and so on. In short, he presents his propositional theories as things that we ought to be interested in, insofar as we are interested in the expressivist project, because they are the kinds of theories that an expressivist will need to endorse if she is going to make her project work. As he says in the Introduction: The central theme of the essays is that though expressivism, the contemporary heir to the noncognitivism of the 1930s, may be developed in a range of different ways, and may have intriguing and surprising applications, all of these require rich commitments about the mind and about the nature of propositions, truth, and propositional attitudes. (1) Unfortunately, the essays consistently fall short in this respect. The closest Schroeder comes to showing anything like this is in Chapter 2, where he shows that a particular strategy suggested by Blackburn in the 1980s – one that Blackburn has since disavowed (Blackburn 2010: 1–2; 2015: xvi) – requires taking on a controversial view about what propositions are. In general, though, the essays manage to generate the appearance of success on this front by underdescribing – or in some cases misdescribing – the commitments of ‘traditional expressivism’, as well as the views that Schroeder would have us put in their place. Schroeder takes expressivism to be a view about meaning. He emphasizes Gibbard’s (2003: 6–7) familiar claim that expressivists attempt to explain the meaning of a certain kind of term indirectly, by saying what state of mind it can be used to express. And he takes this strategy to involve the provision of a distinctive kind of semantics for the relevant domain, one that differs from traditional, truth-conditional approaches. As Schroeder puts it: [T]he expressivist idea is to give a compositional semantics which generates, for each sentence, ‘P’, what it is to believe that P. (116)3 So, for example, the metaethical expressivist tries to give a compositional semantics for moral language that tells us what it is to believe that M, for any moral sentence ‘M’. Traditionally, Schroeder thinks, expressivists have tried to do this without invoking propositions. That’s where they went wrong. Rather than trying to do without propositions, or trying to get by with ‘merely deflationary talk’ about them (75), expressivists ought to embrace propositions and give an expressivism-friendly theory about what they are like. This is a controversial way of understanding expressivism. It has become increasingly common in recent years to think of expressivism as a view in ‘meta-semantics’ – i.e. roughly, as a view about why normative sentences have the semantic values they do, rather than a view about what those semantic values are (Charlow 2014; Chrisman 2012 ; Ridge 2014; Silk 2013). Others (Blackburn 2013; Price 2011, 2013) think that expressivism is best understood in terms of a ‘pragmatist’ approach to theorizing about meaning – i.e. roughly, one that attempts to explain how we get from certain kinds of doings to certain kinds of meanings. It is disappointing that Schroeder does not engage with these alternative ways of understanding the view, since he is, after all, attempting to illustrate commitments that are supposed to be required by expressivism, regardless of how the view is developed. In any case, I think we can see that there’s something off about the way that Schroeder characterizes expressivism without having to wade into those debates. Notice, to begin with, that we could construct the kind of compositional theory Schroeder describes above by simply assigning every sentence ‘P’ to the belief that P. ‘Murder is wrong’ gets assigned to the belief that murder is wrong; ‘Kindness is good’ gets assigned to the belief that kindness is good; and so on. Even the truth-conditional theorist can tell this kind of story, e.g. by associating every sentence ‘P’ with the proposition that P, and then insisting that each such sentence gets associated (derivatively, as it were) with the belief that P. Crucially, in telling this kind of story, the truth-conditional theorist does not end up saying anything that an expressivist would need to reject. The problem with this sort of story, from the expressivist’s point of view, is not that it says anything false, but that it is unilluminating in certain respects (see, e.g. Blackburn 1998: 50–1, Gibbard 2003: 62–3). This suggests that the expressivist’s chief concerns arise at a different level from the one we are operating at when we are developing a compositional semantics for moral language. Consider the following remarks from Blackburn and Gibbard, respectively: I no longer think of the predicates involved in evaluation … as semantically different from others. But this is only because semantic categories are too coarse-grained to do the necessary work: they lull us into remaining ‘unconscious of the prodigious diversity of all the everyday language-games’. … Semantics tells us only where we have ended up; my explanatory interest is in suggesting why and how we have ended up where we did. (Blackburn 2010: 2). I claim … that we can explain much obliquely, characterizing normative concepts by describing independently the states of mind that employ them. We can of course also characterize normative beliefs … in terms of their content. We can identify a normative judgment as the belief that such-and-such – say, as the belief that one ought to shun pain, or that politicians are all sleazy. But this, I am saying, is not the only way to characterize normative beliefs; we can also explain them as plan laden judgments, or as carrying plan-laden presuppositions. This yields further insight. (Gibbard 2003: 185) I take these remarks – and others like them – to suggest something like the following sort of ‘two-levels’ picture. At one level of description, we can say – quite rightly – that moral beliefs are beliefs in propositions. The belief that stealing is wrong is a belief whose content is, precisely, that stealing is wrong. But if that’s all we say, we’re likely to face certain problems and puzzles – e.g. puzzles about the practicality of moral belief, or about how we could ‘be in touch with’ moral properties. We can work toward dissolving these puzzles by asking – at another level of description – what it is to have a belief with this sort of content. The answer to that question will tell us something about the practical state of mind that constitutes having a moral belief – i.e. belief in a moral proposition. If that’s right, Schroeder’s characterization of the expressivist as trying to give a compositional semantics without appealing to moral propositions seems inapt. Schroeder is aware that there is an issue in the offing here, and he goes on to add, after the characterization quoted above, that ‘interesting or essentially expressivist views go further, and specify that for some sentences ‘P’, to believe that P is not simply to have an ordinary descriptive belief that the world is a certain way’ (116). But Schroeder’s conception of expressivism as a project in compositional semantics leads him to look for reasons that are, as it were, internal to semantics to explain the motivations for being ‘essentially expressivist’ in this sense. He says: [Traditional noncognitivist expressivism in metaethics] is essentially expressivist, because it can’t be simply derivatively generated by any truth-conditional semantic theory. Though there is something it is for grass to be green, on this view, there is nothing it is for stealing to be wrong. So a semantic theory for ‘wrong’ can’t work by telling us what it is for stealing to be wrong; it must work by telling us what it is to believe that something is wrong, which is quite different from having an ordinary descriptive belief about how things are. (117) But this is seriously misleading. Even in the ordinary descriptive case, truth-conditional theories do not work by telling us what it is for grass to be green. (That is the job of the philosopher or the colour scientist.) What such theories do is (roughly) to associate the sentence, ‘Grass is green’, with the conditions under which it would be true. So the motivation for going ‘essentially expressivist’ can't be that in the ordinary descriptive case, semantic theories work by ‘telling us what it is’ for the relevant state of affairs to obtain, whereas they can’t do this in the moral case. Rather, the thought would have to be that in the moral case a semantic theory can’t work by associating truth conditions with moral sentences because there are no conditions under which such sentences are true. But, of course, that cannot be right either, since the expressivist does think that such sentences can be true. What has gone wrong here, I think, is that Schroeder has failed to fully appreciate the consequences of the expressivist’s deflationism. One thing we have learned from quasi-realists over the years is that deflationism about truth tends to go hand in hand with deflationism about beliefs, facts, propositions, and other inter-related notions, and it is a familiar point that once expressivists embrace deflationism, it becomes increasingly difficult to characterize their view straightforwardly in semantic terms (see, e.g. Dreier 2004; Price 1994). And there has been much debate about how deflationism interacts with the project of giving a truth-conditional semantics. Unfortunately, Schroeder’s essays do not sufficiently engage with these issues. Insofar as Schroeder’s arguments depend upon casting the expressivist project in terms of a particular strategy within compositional semantics, they thus seem at best incomplete. He doesn’t show that expressivism ‘result in’ or ‘require’ surprising and novel views about propositions and propositional attitudes. What he shows, at best, is that the expressivist will not be able to develop an unorthodox approach to compositional semantics without embracing such views. But that may be just another reason to stop thinking of expressivism as a view within compositional semantics. 3. What do we get from ‘the propositions theory’? In this section, I want to look at Schroeder’s propositional approach in a bit more detail. I’ll start by explaining a central distinction and illustrating it with one of the propositional views Schroeder advances on behalf of the expressivist. Then I’ll consider one of the main arguments he offers in support of this approach – one that would seem to apply even to expressivists who do not think of their view in terms of the provision of a distinctive kind of compositional semantics. Schroeder thinks that expressivists have been largely mistaken in trying to do without propositions – or in trying to get by with ‘merely deflationary talk’ about them (75). The key to seeing this is to distinguish between propositions and representational contents. This distinction shows up in a number of places in the volume, but it receives its most thorough development and discussion in Chapter 3, which Schroeder takes to be ‘the most central essay of the volume’ (16). The rest of my discussion will focus primarily on this essay, since it helpfully summarizes many of the points that Schroeder makes elsewhere in the volume, and since it provides his most direct argument for the claim that expressivists need (non-deflationary) propositions. Here is the distinction. Propositions are the objects of propositional attitudes, they are the primary bearers of truth and falsity, and they are the meanings of declarative sentences in context. Representational contents, by contrast, are more metaphysically committal types of entities: they serve to carve up the world; they correspond to distinctions in reality; they are associated with metaphysical commitment; and they are the appropriate objects of excluded middle (17, 27, 69, 85, 88). Schroeder’s proposal on behalf of the expressivist is then to insist that beliefs can be carved up at two different ‘levels’ (89). At one level, we just say that beliefs are relations to propositions. Some, but not all, propositions determine representational contents. Ordinary descriptive beliefs are relations to these propositions, and so we can also carve them up at a further level of description – one that makes explicit the fact that the belief is (also) a relation to a representational content. Moral beliefs, by contrast, are relations to propositions that don’t determine representational contents, and so they can’t be carved up at this further level. That is all rather abstract, and it is easier to see what Schroeder has in mind here by looking at how this distinction figures in his own development of a propositional expressivism. Consider, then, the view that Schroeder calls biforcated attitude semantics (first advanced in Schroeder 2008; developed here in Chs. 3 and 4). According to this theory, propositions are pairs of entailing properties that an agent might have, and beliefs are complex attitudes of being for each of these properties (92–4). For example, the proposition that grass is green, on this view, is identified with the following pair of properties: ⟨proceeding as if grass is green, not proceeding as if grass is not green⟩ To believe this proposition is to be for proceeding as if grass is green, and to be for not proceeding as if grass is not green. Being for some property is a functional state that ‘leads one to acquire that property, other things being equal’ (92). Someone who believes that grass is green will thus be in a functional state that leads him to acquire the property of proceeding as if grass is green, which involves being related in a particular way to a representational content – namely, grass’s being green. By contrast, the proposition that murder is wrong, on this view, is identified with the following pair of properties: ⟨avoiding murder & disapproving of murder, disapproving of murder⟩ Someone who believes that murder is wrong will thus be for avoiding murder and disapproving of murder (the second member of the pair is redundant here; it’s included for technical reasons that we can safely ignore for present purposes). This involves being related to murder, avoidance, and disapproval. But, since this proposition does not determine any representational content in the way that the proposition that grass is green does, believing it does not involve bearing any relation to a representational content. It simply involves bearing a relation to the proposition – i.e. to the relevant pair of properties. Schroeder’s view – particularly his insistence that beliefs can be carved up at two different ‘levels’ – might seem similar to what I described above as a ‘two-levels’ picture. But in fact the two views are very different. One crucial difference, for our purposes, is that Schroeder wants to explain the non-representational character of moral beliefs by appealing to a substantial view about the kinds of things that moral propositions are, whereas deflationary expressivists like Blackburn and Gibbard want to maintain a principled silence about the latter.4 Why, then, do we need a substantial theory of the moral proposition? Some of what Schroeder says here seems to stem from a misunderstanding of the deflationary view – one that construes the deflationist as not believing in ‘real propositions’ (70), but merely talking ‘as if’ there are propositions (68). But the deflationary expressivist is not trying to talk as if there are moral propositions even though there aren’t any. She simply thinks that there’s not much to say about moral propositions, beyond the fact that they are a ‘focus’ for a certain kind of practical thought and discussion (Blackburn 1998: 50). Though we can’t say much about them directly, we can nevertheless bring them into view by letting them ‘emerge’ (Blackburn 2009: 208, Gibbard 2015: 273) from a story about practical thought that does not appeal to such propositions at the outset. Schroeder thus comes closer to his target when he characterizes the deflationist as simply not appealing to propositions for certain theoretical purposes (75). He then lists five problems that are supposed to arise for an expressivist who adopts such a deflationary view, all of which, he claims, ‘are easily solved by any view which postulates propositions across the board’ (75). Three of these problems bring out the difficulty of giving a compositional semantics without appealing to propositions – problems in accounting for negation (78–9); problems in accounting for quantification into the complement position of verbs like ‘believes’, and identification of the things we are quantifying over in such cases (e.g. when we say things like ‘Jill said the very thing that Jack believes’) (79); and problems in accounting for modals with moral prejacents (e.g. in claims like ‘It might be that killing is wrong’) (82–4). I have already raised some concerns about the extent to which these problems should be seen as damning to expressivism per se – that depends on whether we take the expressivist to be trying to give a compositional semantics of the sort Schroeder describes – so I will set them aside here. Let me note, though, that Schroeder’s discussion here offers valuable lessons for anyone who does want to carry out this sort of project. The other two problems are ones that would apparently arise even for an expressivist who is not interested in this kind of compositional project, for they are supposed to stem from the expressivist’s insistence that moral and non-moral beliefs are, at the most fundamental level, two different kinds of attitude; and this is something that even the ‘two-levels’ expressivist will want to say.5 So it is worth looking at these two problems in a bit more detail. The first problem is to explain why moral and non-moral beliefs turn out to ‘have a striking array of features in common’, given that they are fundamentally different kinds of state (80). The second problem is a more general version of the first one; it is, in effect, is to explain why there are both moral and non-moral hopes, suppositions, desires, and so on ‘for every one of the other attitudes’, as well as why all of these pairs of moral and non-moral attitudes ‘always have so much in common’ (82). Since Schroeder takes the first problem to be a particular instance of the second one, I will treat the two together. The features that Schroeder takes moral and non-moral beliefs to have in common are Moral beliefs and non-moral beliefs ‘come in varying levels of confidence’ (80). Moral beliefs and non-moral beliefs ‘share the same disagreement properties’ (i.e., believing that x is wrong disagrees with believing that x is not wrong, just as believing that x is green disagrees with believing that x is not green) (80). Moral beliefs ‘combine with desires to produce motivation to action in the same way as non-moral beliefs do’ (80).6 Moral beliefs ‘share much of the same phenomenology as non-moral beliefs – phenomenology that is not shared by hoping, assuming, preferring, wondering, or any of their ilk’ (80). He then says that for the propositional theorist it is ‘no mystery’ that these two kinds of belief have so much in common: According to this view … these two states … share a common core … the common attitude of belief that merely takes different contents between the belief that stealing is wrong and the belief that grass is green. It is no wonder, according to the theory of propositions, that believing that stealing is wrong and believing that grass is green have so much in common – it derives from their common core. (81) By contrast, expressivists ‘who allow only for a deflationary way of talking about propositions’ have to accept that ‘it is essentially a coincidence that two states which are fundamentally so different turn out to share so many of the same features’ (81). Here, I think Schroeder moves much too quickly. Surely there can be different kinds of states – different in something like the way that the deflationary, ‘two-levels’ expressivist claims – that nevertheless share a range of important features. Nor is it obvious that the best or the only explanation of why such states share these features will have to appeal to the fact that they both consist in bearing a certain relation to a proposition. Certainly there could be other explanations. I can see why there might seem to be a problem here if all that the deflationary expressivist had to say about moral and non-moral beliefs is that they are very different kinds of state. And that is indeed how Schroeder characterizes such views here. But that is not all that theorists like Blackburn and Gibbard have to say. Gibbard (2003), for example, goes to great lengths to explain why moral and non-moral beliefs have so much in common. In the ordinary descriptive case, he says, there is a distinction between ‘mere apprehension’ and ‘full-fledged judgment’, and he insists that there is a corresponding distinction in the practical case between ‘straight attitudes’ and ‘attitudinal judgments’ (2003: 79–82). One of Gibbard’s aims in telling this kind of story is precisely to explain why, in both the ordinary descriptive case and in the practical case, we end up with beliefs that involve a kind of judgment. As Gibbard says: Once a being is capable of agreeing and disagreeing with possible states of mind, both factual apprehendings and straight attitudes become members of larger classes: factual apprehendings become a special class of factual judgments, and straight attitudes become a special class of purely attitudinal judgments. Factual judgments and attitudinal judgments, moreover, are special cases of judgment in general: some of these judgments are attitude-laden … and the remainder are purely factual. (81–2). It seems plausible – surely it is at least possible – that the kind of story Gibbard tells here would have the resources to explain why both moral and non-moral beliefs have so many features in common. Indeed, we can think of this story as offering another way of identifying a ‘common core’ between moral and non-moral belief, without simply saying that they both consist in bearing a certain relation to a proposition. Of course, in the end, this sort of story may not be able to explain the things that Schroeder wants to explain. My point is just that this is not obvious; it’s something we would need to argue for. And we would need an even stronger argument for the claim that no such story could be told, either by Gibbard or by anyone else. But Schroeder does not provide either of these. He simply asserts that all deflationists will have to take it as a coincidence that moral and non-moral beliefs turn out to have so much in common. And that seems unwarranted. As I say, the issue here stems partly from the fact that Schroeder underdescribes the deflationist views he is arguing against. There is a parallel problem, I think, for the view that Schroeder takes himself to be defending – the so-called ‘propositions theory’. To see this, turn now to Schroeder’s suggestion that the problems he is discussing here are all ‘easily solved’ by any view that posits non-deflationary propositions (75). And consider again his view (qua biforcated attitude semanticist) about what propositions are. The proposition that grass is green is a certain pair of properties, the proposition that murder is wrong is a very different pair of properties, and belief in both of these propositions consists in being for each of the properties in the pair. Schroeder says that ‘it is straightforward how biforcated attitude semantics explains the features that moral and non-moral beliefs have in common – they will be explained by their common core, the fact that they are both biforcated attitudes’ (95). However, while this view may have a straightforward explanation of the fact that moral and non-moral beliefs share the same disagreement properties, and perhaps also of the fact that they share much of the same phenomenology, it is not at all straightforward how it explains the other two properties Schroeder mentions. How does it explain, for example, the fact that moral and non-moral beliefs come in varying degrees of confidence? What is it to be confident in my attitude of being for certain pairs of properties? Again, my point is not that Schroeder can’t answer these questions, but just that they do need to be answered, and that Schroeder doesn’t tell us how to do this. The natural way to answer them, for someone who accepts Schroeder’s view, would seem to be something like this. In the ordinary descriptive case, the believer is conceived as being for (e.g.) the property of proceeding as if something is the case, and that is something we can do more or less tentatively. In the moral case, the believer is conceived as being for (e.g.) the property of avoiding and disapproving of a certain kind of act, and that too seems like something we can do more or less tentatively. This notion of tentativeness seems to correspond to some intuitive notion of confidence, so the fact that we can have varying degrees of confidence in both moral and non-moral beliefs might be explained by the fact that both of these are states of being for, together with the fact that we can be more or less tentative in our states of being for. There are two things to note here, though. The first is that this sort of explanation has nothing to do with the fact that both of these states are conceived as relations to propositions. All of the explanatory work is done by the story about being more or less tentative in one’s practical dispositions. And the second point is that this kind of story is equally available to Schroeder’s opponent. A deflationist about propositions can similarly say that both moral and non-moral beliefs can exhibit varying degrees of confidence because we can be more or less tentative in, for example, our representations and our plans. So it’s hard to see how Schroeder’s propositional view gains any advantage here. Consider next the claim that moral beliefs ‘combine with desires to produce motivation to action in the same way as non-moral beliefs do’ (80). On Schroeder’s view, both moral and non-moral beliefs will dispose you to acquire certain pairs of properties, but beyond that it is far from clear that they will have very much in common when it comes to their motivational profile (indeed, Schroeder seems to say as much on p. 95). To believe that grass is green is to be disposed to acquire (e.g.) the property of proceeding as if grass is green, and to believe that murder is wrong is to be disposed to acquire the property of avoiding and disapproving of murder. But how will these beliefs interact with, say, the desire to look at something green, or the desire to do something wrong? It is not at all obvious what to say here, in part because it is not at all obvious how Schroeder’s beliefs even could interact with desires in the usual way. This brings us to the second problem I mentioned above – the problem of explaining why there are moral and non-moral desires, hopes, suppositions, etc., and why all of these pairs share so many features in common. Schroeder says that any theorist who posits non-deflationary propositions can provide ‘the most elegant explanation possible’ here (82). The explanation is that so long as there are such things as the proposition that stealing is wrong and the proposition that grass is green, those propositions may be taken as the objects of any of the attitudes, and any attitude toward the proposition that stealing is wrong will have as much in common with the same attitude toward the proposition that grass is green… (82) Once again, though, it is not clear how this explanation works once we plug in Schroeder’s own view about what propositions are. Consider, for example, the proposition that stealing is wrong, which is identified with the pair of properties ⟨avoiding stealing & disapproving of stealing, disapproving of stealing⟩. It is far from obvious that, so long as this proposition exists, it may be taken as the object of any of the attitudes. I have no idea what it would be, for example, to bear the relation of hoping toward this proposition. It doesn’t seem like the right sort of entity to be an object of hope. Clearly this cannot be a matter of hoping that I will avoid and disapprove of stealing, since that would be hoping that an ordinary descriptive proposition is true (one that concerns my future doings, and not one that concerns the wrongness of stealing). It’s hard to know what else we might say, though, and Schroeder doesn’t offer much help. He suggests that all of the explanations here will be straightforward once we’ve got propositions in the picture, but I take these remarks to show that this is not actually the case. 4. Conclusion Schroeder covers a lot of ground in these essays, and his arguments are ambitious. He attempts to show that all versions of deflationism fail; that all versions of the ‘propositions theory’ solve his problems with ease; that every way of carrying out the expressivist project requires taking on substantial views about the nature of propositions; and so on. These are very difficult claims to argue for. I have been pointing to some ways in which Schroeder’s arguments for these strong claims seem to me to fall short. These criticisms notwithstanding, Schroeder’s essays raise a number of interesting and important challenges for anyone interested in carrying out the expressivist project, and there is much to be learned from working through his arguments. Deflationists, in particular, will have a rich set of issues to grapple with here. Footnotes 1 Expressing Our Attitudes: Explanation and Expression in Ethics, Volume 2, by Mark Schroeder. Oxford University Press, 2015, xii + 266 pp. 2 For concreteness, I will focus throughout on metaethical expressivism, but it’s worth noting that Schroeder takes his arguments to apply to nondescriptivist views more generally. 3 Schroeder says that this is only a ‘first pass’, but his reservations here come from considerations about slurs that are immaterial to the present discussion. 4 Another way of seeing the difference is to notice that the ‘two-levels’ expressivist has something further to say about moral belief at the ‘other level of description’, whereas Schroeder’s claim is precisely that moral beliefs can’t be carved up at any further level of description (while representational beliefs can). 5 Or at least this is something the two-levels expressivist may want to say. The issue depends in part on whether she wants to go ‘global’ with her expressivism (see Blackburn 2013; Gibbard 2015; Price 2013). 6 Here, Schroeder seems to me to be overlooking one of the key motivations for traditional expressivism, which is precisely that ordinary beliefs don’t produce motivation in the same way as moral beliefs. References Blackburn S. 1998. Ruling Passions . Oxford: Oxford University Press. Blackburn S. 2009. Truth and a priori possibility: Egan’s charge against quasi-realism. Australasian Journal of Philosophy  87: 201– 13. Google Scholar CrossRef Search ADS   Blackburn S. 2010. Practical Tortoise Raising and Other Philosophical Essays . Oxford: Oxford University Press. Google Scholar CrossRef Search ADS   Blackburn S. 2013. Pragmatism: all or some? In expressivism, Pragmatism and Representationalism , ed. Price H., 67– 84. Cambridge: Cambridge University Press. Google Scholar CrossRef Search ADS   Blackburn S. 2015. Apologia pro vita sua. In Passions and Projections: Themes from the Philosophy of Simon Blackburn , ed. Johnson R.N., Smith M., xv– xix. Oxford: Oxford University Press. Charlow N. 2014. The problem with the Frege-Geach problem. Philosophical Studies  167: 635– 65. Google Scholar CrossRef Search ADS   Chrisman M. 2012. On the meaning of ‘ought’. In Oxford Studies in Metaethics , vol. 7, ed. Shafer-Landau R., 304– 32. Oxford: Oxford University Press. Dreier J. 2004. Meta-ethics and the problem of creeping minimalism. Philosophical Perspectives  18: 23– 44. Google Scholar CrossRef Search ADS   Gibbard A. 2003. Thinking How to Live . Cambridge, MA: Harvard University Press. Gibbard A. 2015. expressivism and the truth in representation. In Meaning Without Representation , ed. Gross. S., Tebben N., Williams M., 210– 23. Oxford: Oxford University Press. Google Scholar CrossRef Search ADS   Price H. 1994. Semantic minimalism and the Frege point. In Foundations of Speech Act Theory: Philosophical and Linguistic Perspectives , ed. Tsohatzidis S. L., 132– 55. New York: Routledge. Price H. 2011. Naturalism Without Mirrors . New York: Oxford University Press. Price H. 2013. expressivism, Praagmatism and Representationalism . Cambridge: Cambridge University Press. Google Scholar CrossRef Search ADS   Richard M. 2015. What would an expressivism semantics be? In Meaning Without Representation , ed. Gross. S., Tebben N., Williams M., 137– 59. Oxford: Oxford University Press. Google Scholar CrossRef Search ADS   Ridge M. 2014. Impassioned Belief . Oxford: Oxford University Press. Google Scholar CrossRef Search ADS   Schroeder M. 2008. Being For: Evaluating the Semantic Program of expressivism . Oxford: Oxford University Press. Google Scholar CrossRef Search ADS   Silk A. 2013. Truth conditions and the meanings of ethical terms. In Oxford Studies in Metaethics , vol 8, ed. Shafer-Landau R. 195– 222. Oxford: Oxford University Press. © The Author 2017. Published by Oxford University Press on behalf of The Analysis Trust. All rights reserved. For Permissions, please email: journals.permissions@oup.com

Journal

AnalysisOxford University Press

Published: Jan 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off