As Barry Stroud sees it, the journey from engagement to metaphysical dissatisfaction has four stages.1 First, we embark on the project of acquiring some especially metaphysical insight into how the world is with respect to some region of our thought about it – say evaluative, or modal or causal thought. The insight that we seek is especially metaphysical when we aspire to go beyond the insight that we gain about the world when we make warranted and true judgements that are the characteristic of that specific way of thought itself. Thus, no especially metaphysical insight is delivered by realizing that it is better to let the guilty go free than to punish the innocent, or that there could not be a true contradiction or that consumption of alcohol alters the mood. What metaphysicians are after, rather, are insights along the lines that there is (or there is not) something evaluative, or something modal or some causal connexion in the world itself and independently of us. Secondly, a special threat to the feasibility of this distinctively metaphysical project arises wherever a specific region of our thought has the special combination of features of irreducibility (of a certain kind) and indispensability (of a certain kind). Thirdly, that threat emerges in the first instance from the attempt to sustain the metaphysical position that many would characterize as anti-realistic (although Stroud does not use that term). The problem is that a certain aporia is implicit in such an attempt to combine the thought that the world is a certain way – with respect to value, modality or causation – (as we must, in line with indispensability) with the further thought that it is not ‘really’ that way. To be clear, the suggestion is not that the conjunction of these thoughts is inconsistent. Nor is it the case that any consideration at our disposal makes it the case, and far less does any show it to be the case, that the relevant metaphysical thesis is false. The predicament is of the theorist rather than of the theory. It has the character of a Moorean quasi-inconsistency, as were we to attempt to claim, ‘It is not the case that P, but I believe that P’ (137), or as intimated by Stroud’s epigrammatic quotation of Wittgenstein, ‘If there were a verb meaning, “to believe falsely”, it would not have a meaningful first person present indicative’. The anti-realist position about value, causation or modality is not properly assertible by anyone who is also engaged in the intellectual practice of thinking about things in evaluative or causal or modal terms. Fourthly, the predicament of the would-be anti-realist does not confer victory, by default or otherwise, on her realist opponent. As with the anti-realist position, and even after the limitations of the anti-realist position have been demonstrated, no available consideration makes it the case, and far less does any show it to be the case, that the contrasting realist thesis is true. As with the anti-realist position, we have a metaphysical claim that has a truth-condition, and one that for all we know is satisfied, but which is never legitimately available to us to assert. And so, at the end of our journey, we find that no thesis of the especially metaphysical kind that we sought to establish at the outset is available to us (in a significant range of cases) to assert legitimately.
That is how I understand Barry Stroud’s dual project of characterizing the predicament that gives rise to metaphysical dissatisfaction, and in showing how those who would be metaphysicians are condemned to that predicament. The book begins (Ch. 1) with a preliminary sketch of the overall journey described above. Then, there are sketches of how the metaphysical game is played out to a stalemate in each of the cases of causation, necessity and value (Chs. 2–4). With the prima facie case made, the screws are then tightened in two places, as detailed explanation is offered of exactly what kind of indispensability it is that paves the way to metaphysical dissatisfaction (Ch. 5) and of that state of metaphysical dissatisfaction in which we arrive (Ch. 6).
In advance of critical commentary, I want to ensure that my positive appraisal of Engagement is recorded. This is a rich and fascinating project that looks to go deeper than we allow ourselves to go when we help ourselves to easy presumptions about the feasibility or the infeasibility of (as it were) first-order metaphysical theorizing. By thinking hard about Stroud’s picture of the underpinnings of such metaphysical theorizing, we give ourselves a chance of making secure progress that will be properly earned. This is work that is deeply serious, incredibly stimulating and of the greatest importance.
2. Initial observations
Some initial observations may help to convey something further to the potential reader about the style and orientation of the book.
Stroud’s project here is at the meta-metaphysical level of concern. This kind of concern is now, of course, perfectly familiar since it has frequently accompanied various attempts at metaphysical theorizing over the last 40 years. Thus, various ‘realism’ debates of the 1970s and 1980s were frequently attended by concerns about whether any non-metaphorical content can be associated with realist versus anti-realist disputes: and present debates about what is fundamental are attended by concerns about the place and relative weighting that the broadly theoretical virtues of conservativeness, economy and intuitive appeal should enjoy in our selection of best theory, in the particular case where the theory is metaphysical. The meta-metaphysical concerns of Stroud’s Engagement are much closer in spirit to the former phase of meta-metaphysics than to the latter.
A related observation is that there is an improved and reflective methodological understanding to be gained here of what was going on with certain reactions to various versions of moral anti-realism (error theories and non-cognitivisms) in that earlier period. For one can recognize, retrospectively, from the standpoint of Stroud’s chapter on Value, two distinct reactions to those anti-realist positions that were not always disentangled in the heat of the meta-ethical battle. Underlying both reactions was the thought that something (or other) about our engagement in evaluative practice itself undercuts the metaphysical stance that the anti-realist would occupy. (I will not attempt to elaborate that thought here). One reaction would be to share Stroud’s metaphysical caution, and settle for rather attenuated and cautious meta-ethical positions such as ‘anti-non-cognitivism’ (cf. Wiggins 1987). The other, more sanguine, reaction was to attempt to re-use the resources used in establishing the moral anti-realist’s predicament for the purposes of constructing a moral realism. Or, at least, that is to put the matter charitably. For on reading Stroud, we are reminded of what ought to have been evident anyway: namely, that there is no question of winning through to any kind of substantial moral realism simply by showing that the moral anti-realist position is embroiled in some kind of aporia.
Those who are familiar with Stroud’s immensely impressive, and rightly celebrated body of work (stretching back to Stroud 1968) will find that the present project recapitulates, and integrates, some familiar and central Stroudian themes. Stroud is gracious to his reader in avoiding the tiresome device of extensive self-reference, but there may be more to this than a display of good taste. Perhaps Stroud will feel that what I take to be familiar Stroudian themes are, in fact, significantly different from those displayed in his earlier approaches to the issues. However, other readers may join with me in finding that when the crucial discussion of Indispensability arrives in Chapter 5, it brings back to the table the familiar and powerful Stroudian appraisal of the status of transcendental arguments. Briefly, the three-part appraisal is this. First, the classical (Kantian) attempt to deploy transcendental arguments immodestly, to establish directly and substantively metaphysical conclusions, does – as Kant thought – require a commitment to some kind of idealism (transcendental idealism, of course, for Kant). That is, an incredibly powerful lemma in the form of some version of idealism is needed to generate any telling inconsistency between the propositions: (i) that certain thought about the world inevitably presents it as though it has certain aspects and (ii) that the world is not really so. Secondly, then, the metaphysical project is to be abandoned insofar as it does bring with it such an overarching commitment to idealism. Thirdly, however, there is significant philosophical value (maybe even as much philosophical value as we can hope for) in attempting to deploy transcendental arguments modestly: by deriving predicaments of the thinker (theorist) we can show that there are certain intellectual attitudes that we cannot take to our practices while we rely (as we must) on those practices. So while Stroud avoids the use of the term ‘transcendental argument’ – and even when he engages, as of old with Strawson (136, 152) – I hope it is helpful to signal that the book returns to these rich pastures.
3. General observations
Moving beyond initial observations, I want to engage with Engagement at two levels. At the level of the general, I think that there are two salient points at which the master dialectic is most in need of further support and I will note these in turn. At the level of the specific, I will comment in some detail on Stroud’s treatment of the case of modality.
The first general point concerns the question of irreducibility. One point that is clear and appreciable is that it simplifies many kinds of metaphysical discussion if we are in a position to put certain reductionist moves aside. In some sense, and in the right cases, we want to establish that what we say about facts or discourse at a certain level is in danger of being wasted if it can be agreed that an appropriate kind of reduction is in the offing. For, the thought is, where we can reduce, we should, and then train our further metaphysical efforts on the reductive base, or perhaps on the nature of the relations that obtain between the levels of what is reduced and of what it is reduced to. This is all very vague. But I hope that the vagueness might be forgiven since it is directed sympathetically in support of (what I take to be) Stroud’s drift: it is only once (further) reduction is ruled out that metaphysical push comes to metaphysical shove. But, having registered this point, there is something systematically puzzling about how Stroud goes about making the case for irreducibility and establishing its metaphysical significance. To put the point bluntly, the case for irreducibility seems typically to be made at the level of sense, while the reader is invited to draw conclusions about the realm of reference. The complaint against reducibility is typically that the reductionist simulacrum does not preserve our concepts: yet the subsequent suggestion is that this amounts to a failure to preserve the facts that are represented as obtaining by our original conceptualization. (Thus see, for example, the initial remarks on reduction of causation (24–25), the initial remarks on reduction of value (90–92), and throughout.) Once the complaint is put that way then, of course, no author will step forward to claim this account of what is intended (unless she is also prepared to repudiate – say – the reducibility of the chemical to the physical, and perhaps even informative identifications in general). And there is many a story that might be told about how we can bridge the apparent gap so that we can move legitimately to metaphysical conclusions (claims about the realm of reference) from conceptual premises (claims about the realm of sense). However, I register the worry that this, rather crucial, part of the master dialectic of Engagement is under-developed.
The second general point is more swiftly made and concerns the capacity of anti-realistic aporia to bring realism down with it. The point is well taken that anti-realist aporia confers no immediate victory on realism. And the point is also well taken that there is a certain sense in which anti-realist aporia ‘closes’ a question that was initially supposed to be open: given anti-realist aporia, we are no longer in a debate that the anti-realist can win. But it was not evident to me, on initial readings, exactly why (or how) this combination of considerations is supposed to show that the realist cannot win either.
In advance of my discussion of the case of modality (Ch. 3, Necessity) I want to acknowledge the author’s modest and careful claims about what he hopes to achieve in the ‘case studies’ (xi–xii). Even taking as given the form of the programme itself, application to any one case (such as modality) in a chapter length treatment is bound to be more suggestive than definitive, and so it is advertised. So in exploring Stroud’s discussion of modality by expanding into territory where the author does not take us, I do not mean to imply a criticism of omission per se. Consideration of this expanded territory shows us how, and where, the literature allows more advanced starting points for the prosecution of the Stroudian project than are given to the reader in Engagement. But we might take this as a vindication of the potential of the project Engagement to encompass many significant developments that are not explicitly mentioned in its pages.
Stroud is here concerned with necessity as a mode of truth or as a feature of whole statements and propositions: moreover it is necessity that is also alethic and absolute (59–60). Hereafter concern with such an absolute de dicto necessity will be taken as read. What is (so) necessary does not depend on how anything is. In line with the general programme, it is important to Stroud’s purposes that he put this modality in the frame of being both irreducible and indispensable, in order to show from there that metaphysical dissatisfaction ensues.
The case for indispensability is rather quick. We are frequently urged to accept that thinking modally is an inevitable feature of any (remotely sophisticated) thinking at all. More specifically, the claim is that any determinate structured enquiry, or sequential thought, must feature properly modal notions of logical consequence or incompatibility (77, and surrounding material). In looking to gauge anti-realistic reaction to this claim, it is important to distinguish two different kinds of anti-realist attitudes to (such) necessity. There is the kind of anti-realist position (occupied by Blackburn 1986) that accepts this appearance of indispensability as accurate (and more of which below). This is to be contrasted with the properly sceptical position – the Quinean position – according to which the appearance of any significant indispensability is misleading. Now, one thing that might be said is that the appearance of indispensability is significant in setting the terms of the dialectic. Thus, it might reasonably be held that it is evidence in favour of an indispensability position, and perhaps even that it establishes indispensability as the default position, that the modal idioms are naturally and pervasively found in roles relating to the characterization of good deductive inference (it must be the case that if the premises are true then so is the conclusion etc.). But the Quinean position consists in part in challenging the modal appearances and in offering non-modal surrogates: and once these are on the table, reliance on default indispensability will not do. Moreover, once one acknowledges (as Stroud does, 78–80) that certain Quinean non-modal surrogates for the modal logical notions (e.g. logical truth for logically necessary truth) are perfectly apt for certain purposes, the burden of proof requires the presentation of other legitimate purposes for which the Quinean surrogates are inadequate. Furthermore, to lean at this point on the alternate consideration of irreducibility is not satisfactory: that the Quinean surrogates are not (identical to) those truly modal notions that we know and love (80–81) is not the point. The indispensability claim must be independent of the irreducibility claim to the extent that different concepts cannot immediately be ruled to be unfit for the right purposes just because they are not the very concepts that we started with. In this matter, the game is considerably further on than Stroud’s discussion discloses. Ian McFetridge (1990) offers what appears to be precisely the kind of modest transcendental argument which is perfectly suited to establish much of what Stroud requires: namely, that we are at least committed to belief in logical necessity so long as we are engaged in certain inferential practices under minimal constraints of success. More specifically, McFetridge argues for the rationally unavoidable acceptance of some or other rules (and subsequently propositions) as logically necessary by anyone who relies on the notion of any rule being truth-preserving: and let us grant, to heighten the transcendental drama, that this would be anyone who goes in for inference at all. There are, then, two things to note about McFetridge’s stance. The first is that this stance is consistent with a consideration that Stroud thinks crucial to the understanding of the well-conceived transcendental projects. For what is argued by McFetridge is not that belief in logical necessity is an absolutely inevitable feature of any form of thought whatsoever: but the conditional claim that it is an inevitable feature of any thought that relies on (truth-preserving) inference. So the Spinozan thought remains alive: perhaps our need to modalize is a product of having the kind of sub-Divine capacities that require inference, in lieu of an ability to see the non-modal world entirely as it is and all at once. But, as indicated, far from being a defeater of Stroud’s aims, this conditionality is of a kind for which Stroud specifically provides within his conception of relevant indispensability (130–38). The second point is that the McFetridge argument is that it relies crucially on a lemma about what it would be to accept a rule as absolutely necessary: it would be to take it to be truth-preserving under any supposition whatsoever. It is at this point that the Quinean, in looking to resist commitment to belief in logical necessity, is well advised to make her move. For there are now sharp questions to be asked about how widely the class of all suppositions is supposed to range, about which kinds of supposition among these we have independent reason to regard as (themselves) indispensable and about whether it is really commitment to hold (as truth-preserving) under all suppositions of all kinds that properly generates a commitment to belief in absolute necessity. It does not matter for present purposes how matters might proceed from there. (Although, to give the hint, the two-fold Quinean thought will be that McFetridge’s case for belief in a kind of necessity relies on the inclusion of counterfactual supposition, and that such counterfactual supposition is itself entirely dispensable, see Divers and Elstein 2012). The point is that the McFetridge–Quine dialectic, even developed only this far, appears to offer considerable insight into how a case for the relevant (transcendental) indispensability of belief in logical necessity ought to be made and might be resisted. But I do not see why Stroud should not simply welcome this as an advanced prosecution of the problematic that he sketches in Chapter 3.
The case for the irreducibility of necessity might also have been made stronger. What Stroud is ultimately interested in establishing is: (a) that however we think exactly of reduction, there is no sense in which our modal idioms/notions/concepts/ideas are properly reducible to the non-modal, and (b) as advertised at the outset, this is a major step on the route to metaphysical dissatisfaction. I read Stroud in Chapter 3 (66ff.) as rehearsing something very much along the lines of a dilemma in support of (a) that is familiar (see, e.g. Blackburn 1986). But it is a dilemma in whose path significant obstacles have already been placed (for what follows, see Hale 2002). One half of the dilemma in support of (a) has it that if reduction is to a basis that is (held to be) contingent – whether it be our choices of how to use words, the reactions and attitudes we happen to have, etc. – this will not do. For what is necessary is really necessary and not really contingent. However, this strategy for reduction is not indefensible. For the commitment of the position is to the iterated modal claim that what is necessary is contingently necessary and not that what is necessary is (really) contingent. And to deny that the necessary is necessarily necessary is not to find oneself in the unplayable position of denying that the necessary is necessary. The other half of the dilemma in support of (a) has it that if reduction is of necessities to a basis that is (held to be) itself necessary, this reduction inevitably leaves some residual necessity unexplained. This strategy for reduction is also defensible. For such a reductionist who takes the base to be necessary (so accepting that Necessarily P) need not appeal to the necessity (but only to P) for the purpose of reducing (explaining) the original necessity. That is: it is necessary that Q, and it is necessary that P, but what explains the necessity of Q is P and not its being necessary that P. As with the McFetridge–Quine dialectic, this Blackburn–Hale dialectic appears to offer a starting point for the prosecution of Stroud’s project that is some way ahead of the starting point that we are offered in Chapter 3. (I note further that, in the handling of the irreducibility issue, since he rests his case on the power of the general dilemma against the reduction of necessity, Stroud does not see the need to engage at all with what is arguably the most eminent of specific reductionist proposals in the literature, namely that of Lewis 1986.)
The case against the quasi-realist position in the case of necessity illustrates a weakness which I think attends Stroud’s treatment of quasi-realism in general. Stroud, I think it is fair to say, is not at all sympathetic to the various Humean and neo-Humean positions that he discusses, whether concerning value, causation or modality. The agenda, as we know, is to deprive these positions of any claim that they have on issuing sustainable negative metaphysical verdicts (that the world in itself is devoid of causation, modality, value … ) by trapping them ultimately in aporia. So, the most effective and compelling way of pursuing that agenda would be to grant the Humean all the rope that he needs in order to hang himself. Yet Stroud’s irritation with the Humean frequently leads him to grant the rope only grudgingly and eventually, after criticising inessential details of the Humean position, and this is something of a distraction from the real prize. So, it is a pity that energy is devoted to attempts to saddle the Humean opponent with a view that he explicitly disavows: for example, a simple regularity view of causation or the view that Impossible-P must really be taken to mean the same as I-cannot-conceive-that-P. And while, to be fair, Stroud often does clearly and accurately represent the Humean position as doing without these crude and undesirable commitments, this rather feeds the regret that he does not always do so. As an example of how, and where, Stroud could afford to be more discriminating (and to do so in a spirit of enlightened self-interest), consider Blackburn’s own worries about the apparent limitations of the quasi-realist treatment of modality (Blackburn 1986). These worries advert to very general, and essential features of the quasi-realist position, and they appear to carry the threat of exactly the kind of aporia that Stroud takes to settle the non-assertibility of a negative metaphysical verdict. The quasi-realist predicament that Blackburn discerns might be sketched, briefly, as follows. According to the modal quasi-realist we make judgements of impossibility in reaction to the discovery of certain kinds of breakdown in our thought. But our reflective modalizing involves scrutiny of how our thought breaks down. In particular, when we can explain in naturalistic terms how our thought breaks down, and in the terms and conditions that are characteristic of naturalistic explanation, we will (typically) see that this is just how it is with creatures like us and refuse to apply the modal sanction of deeming impossible. And this is especially so when we (thereby) become able to make-something-of the troublesome thought at one remove by appreciating how a different kind of thinker would be in a position to make something of the thought although we (due to our known and non-mysterious limitations) are presently unable to do so for ourselves. Thus – for example – we do not insist on the impossibility of there being phenomenologically distinctive modes of perceiving the world other than those at the disposal of humans: it is perfectly possible that there is something that it is like to be a bat. To retreat, briefly, to old ground, there is no vicious circularity afoot in the consideration that we would have to regard such limitations on our thinking as contingent. For while the quasi-realist, along with the rest of the human race, is inclined to recognize them as contingent, it is not their contingency to which she appeals in basing her explanation of the conditions under which we properly acquire modal belief. So let us move quickly on to the serious problem. Let Naturalism be the hypothesis that every fact is naturalistically explicable, and assume that Naturalism is true. In that case, all thought-breakdown facts are naturalistically explicable and so, a fortiori, are all such facts in the special cases where they are judged presently to justify a verdict of impossibility. What is supposed to allow us a justified verdict of impossibility is that we are not at present in possession of the kind of explanation that would undercut that justification. But from an over-reaching, or dominant, Naturalistic perspective, it seems, we should nonetheless appreciate that all of our judgements of impossibility are false (or otherwise in error). Now it is far from clear whether this is the tightest formulation of the predicament, nor exactly which punch line is supposed to follow. But what is clear is that the quasi-realist who maintains commitments to impossibility is – at the very least – in the territory of Stroudian aporia: the quasi-realist position looks vulnerable to the accusation of being committed to avowing that she falsely believes that it is impossible that P. So, as in other cases illustrated, we already have in the literature the beginning of a fascinating problematic that seems to fit snugly into Stroud’s project. In this case, to be clear, the problematic in question is one that Stroud himself does introduce (86–87). But early skirmishes squeeze out detailed or extensive discussion of this (surely) crucial matter.
I have been critical, and perhaps excessively so, of Stroud in suggesting that one might wish that more had been done by way of taking advantage of the literature on modality in order to present his project in a more advanced stage of development. But I hope that this will be forgiven as the expression of impatient anticipation of how this excellent project might be developed. And I recommend whole-heartedly that philosophers of modality, and others, should engage with the project of Engagement as a matter of priority.