Answering the conceptual challenge: three strategies for deflationists

Abstract

We defend deflationism about truth against a pressing challenge, which is to explain how deflationists can understand the role that the concept of truth appears to play in accounts of several other philosophically important concepts. We provide three strategies that deflationists can employ in response to the specific challenge regarding assertion that has been raised in several recent articles, viz., that the truth concept plays an ineliminable explanatory role in an account of assertion. We then show how to extend our strategies to accounts of other central philosophical concepts, by applying them to accounts of belief, knowledge, and logical validity. The result is a set of recipes for deflationists about truth to employ in developing responses to worries that might be raised about the explanatory role of the truth concept.

1 Introduction

Our aim here is to defend deflationism about truth (henceforth, simply deflationism) against a pressing challenge regarding how deflationists can understand the role that the concept of truth appears to play in accounts of other philosophically important concepts, for example, assertion, belief, knowledge, and logical validity (cf. Jamin Asay, 2021; Dorit Bar-On & Keith Simmons, 2007/2021; Richard Heck, 2021; Andrew Howat, 2018; Simmons, 2018).¹ We develop three different strategies that a deflationist can apply in addressing this issue, focusing initially on the specific challenge pertaining to the truth concept’s putative role in an adequate account of assertion. After explaining ways in which deflationists can answer this specific challenge, we indicate how our responses to it can be extended to similar challenges pertaining to the other concepts noted above. If we are right, then deflationists have several strategies for responding to these worries regarding the putative explanatory role of the truth concept.

The plan is as follows. Section 2 lays out our understanding of deflationism and explains a clarifying three-part distinction, along with a priority thesis holding between its facets. Section 3 introduces the conceptual challenge to deflationism by noting a distinction between metaphysical ‘why’-explanations and conceptual ‘what’-explanations and rehearses some familiar cases where the concept of truth appears to factor in an account of the ‘what’ of something else. Section 4 turns specifically to the contention that the concept of truth plays a central role in any adequate account of what assertion is and introduces a first deflationary strategy for dealing with this challenge based on an “assertion-first” approach. Section 5 presents a second deflationary strategy for relating these concepts that views assertion as amenable to substantive analysis, just in truth-free terms. Section 6 presents a third deflationary strategy for relating the concepts of truth and assertion in a tighter fashion, Sect. 7 explains how our deflationary strategies extend to challenges stemming from other concepts, and Sect. 8 concludes.

2 Deflationism

On our understanding of deflationism, it is a different kind of approach to the topic of truth from the traditional one because a deflationary account is not actually a theory of truth. Deflationary views are best understood as various theories or accounts of truth-talk, which is that fragment of discourse that employs the alethic locutions (centrally, ‘is true’ and ‘is false’), rather than as theories of a property of truth (or of falsity).² Deflationists reject the standard assumptions and presuppositions about truth-talk (and about the truth concept) that generate so-called inflationary theories of truth (e.g., correspondence theories, coherence theories, pragmatist theories, and primitivist theories), and as a result they resist the traditional move to positing a substantive truth property. Deflationists instead offer novel understandings of the functioning of the alethic locutions and of the concepts they express.

Truth theorists of any stripe will accept some sort of equivalence between a sentence (or utterance, proposition, etc.) and an ascription of truth to it, an equivalence expressed by the instances of the theorist’s preferred version of the “neutral” truth schema,

(TS) [p] is true iff p,

where the square-brackets stand for some unspecified nominalizing/naming device.³ But in contrast with inflationists about truth, who claim that these equivalences are substantive and hold in virtue of the nature of the truth property attributed by the use of the truth predicate, deflationists maintain that the instances of (TS) are in some sense immediate, fundamental or brute. They take the instances of their favored version of (TS) to be “conceptually basic and explanatorily fundamental” (cf. Paul Horwich, 1998a, p. 21, fn. 4), or direct consequences of how the truth predicate operates, logico-linguistically.⁴ Some deflationists (e.g., Hartry Field, 1994) even go so far as to endorse a kind of meaning (or “cognitive”) equivalence between the left- and right-hand sides of their favored version of (TS).

The different versions of deflationism are best understood as meta-theories of truth, that is, as theories of truth-talk and (derivatively) of the truth concept; they diverge from one another in the alternative accounts they provide of the operation of the alethic locutions. But there are common "deflating" themes that unify the many versions of deflationism.⁵ One useful way to organize these themes is in terms of a three-part division, introduced by Bar-On and Simmons (Ibid., pp. 61, 68) and discussed by Asay (Ibid.), Heck (Ibid.), and Howat (Ibid.), between different facets of deflationism: metaphysical deflationism, linguistic deflationism, and conceptual deflationism.

Metaphysical deflationism is the thesis that there is no substantive property of truth expressed by the alethic locution ‘is true’ or determined by the concept of truth. If deflationists countenance a truth property at all, they take it to be a “thin”, disunified (or “fragmented”) one, as there is nothing that all of the true things have in common in virtue of which they all count as true.⁶Linguistic deflationism is a general thesis about the operation and function of the alethic locutions, which is sometimes explained as holding that these expressions operate merely as formal devices, playing a kind of logical role, rather than a descriptive one (cf. Field, Ibid.; Horwich, Ibid.; Leon Horsten, 2011). As we would elaborate this thesis, it amounts to the view that the fundamental linguistic functioning of the truth predicate⁷ is exhausted by its role in implementing a kind of semantic descent, where this notion is generalized to cover a variety of ways of denominalizing different kinds of nominalizations of content-vehicles,⁸ thereby transforming a mention of some content-vehicle into something equivalent to a use of it (or of its content). Being such a device, when coupled with other resources (e.g., ordinary quantifiers), the truth predicate has certain uses that have it provide a natural-language surrogate for quantification into sentence positions, thereby performing a special kind of generalizing role. These uses of the truth (and falsity) predicate enable speakers to express agreement or disagreement with a potentially infinite body of claims. Finally, conceptual deflationism, as Bar-On and Simmons (Ibid., pp. 61–2) (and following them, Howat, Ibid.; Asay, Ibid.; Heck, Ibid.) explain it, is the view that the truth concept is “thin” in the sense that what it takes to grasp that concept is not much more than a disposition to accept all instances of some version of (TS), or to accept all of the inferences from a truth-ascription to its target and vice versa, or to accept their intersubstitutability (in non-opaque contexts). The idea is that the concept emanates from the roles that the truth predicate plays, which for a deflationist are merely logical and expressive. As a result, conceptual deflationism maintains that the truth concept has no explanation in terms of any other concepts, i.e., that it admits of no complex/reductive analysis, and holds that the truth concept never plays a substantive role in explanations or accounts of any other concepts. (It is this last point that has become the focus of certain recent challenges.)

While we take the identification of these facets of deflationism to be clarifying, we reject Bar-On and Simmons’s (Ibid.) “separation thesis”, according to which they are both separate and independent of one another. On our understanding, the facets are ordered in a kind of hierarchy, with linguistic deflationism having priority over the others. To understand our resistance to the separation thesis, consider what it would take to violate either metaphysical or conceptual deflationism. A violation of metaphysical deflationism would involve there being a substantive property of truth, which would make the truth predicate a means for attributing that property. But this would amount to the truth predicate functioning to make a substantive description of something, which is in violation of linguistic deflationism.⁹ A violation of conceptual deflationism would involve the truth concept involving more than logical or expressive aspects, something that would be shown, for example, if it ineliminably factored into an adequate account of some other concept. But this would make the use of the truth predicate in a statement of that account serve to express a substantive explanatory factor, which is more than what linguistic deflationism can allow.

It follows from these considerations that if either conceptual or metaphysical deflationism is incorrect, then linguistic deflationism is incorrect. Contraposing (and assuming that we are not intuitionists), if linguistic deflationism is correct, then so, too, are conceptual and metaphysical deflationism.¹⁰ So linguistic deflationism functions as a sufficient condition for metaphysical and conceptual deflationism, and their conjunction functions as a necessary condition for linguistic deflationism.¹¹¹² Hence, it seems that one cannot consistently endorse linguistic deflationism while denying either conceptual or metaphysical deflationism, and, as we claimed above, linguistic deflationism has a kind of priority over metaphysical and conceptual deflationism. It is for this reason that we reject any separation thesis and offer our “priority thesis”, regarding the facets of deflationism.

3 Deflationism and the conceptual challenge

One consequence of our priority thesis is that an argument against either metaphysical or conceptual deflationism will serve inter alia as an argument against deflationism in toto, since it will, via modus tollens, provide a reason for rejecting linguistic deflationism, which is the primary or central thesis of any deflationary view. Thus, deflationists must answer purported challenges to any of the facets of deflationism. Our aim here is to respond to certain newer challenges focused specifically on conceptual deflationism.¹³

As we have seen, conceptual deflationism maintains that the truth concept does not play any substantive explanatory role. When deflationism first emerged, the original challenges to it attempted to establish an explanatory role for truth and were directed primarily at metaphysical deflationism (see, for example, Field, 1986; Kitcher, 2002; Putnam, 1978).¹⁴ As Peter Achinstein (1983), Sylvain Bromberger (1966) and Carl Hempel (1965), and others have argued, answers to why-questions serve as explanations of various phenomena.¹⁵ If the truth predicate figured essentially in an answer to a why-question, such as “Why is electron theory successful at predicting observations?”, then that would provide a reason for thinking that there is a property of truth that has a causal-explanatory role to play in an account of why the cited phenomenon occurs. This would be incompatible with metaphysical deflationism and, as we noted above, would also violate linguistic deflationism, which would thus undermine deflationism as we and others understand that view.

The kind of explanatory role at issue in some more recent challenges is not the same as that posited in the earlier ones directed at metaphysical deflationism. Alongside answers to why-questions, answers to what-questions provide a different sort of explanation, one that explains what something is, or consists in, or amounts to, i.e., an analysis or account of it. If the truth predicate figured essentially in an answer to a what-question about some other concept, for example, as part of an answer to the question, “What is knowledge?”, this would violate conceptual deflationism as it would provide us with a reason for thinking that the truth concept has an explanatory role to play in our conceptual scheme, and this would be more to the concept of truth than conceptual deflationism can allow.¹⁶ But, again, if the truth concept did this, then the predicate that expresses it this concept would specify a genuinely explanatory factor instead of performing a merely expressive role. If the truth predicate did this, then it would not be functioning merely expressively, as a kind of logical device. This would constitute a violation of linguistic deflationism and would thus undermine deflationism. The more recent critiques of deflationism from Asay (Ibid.), Bar-On and Simmons (Ibid.), Heck (Ibid.), Howat (Ibid.) and Simmons (Ibid.) that we have cited here all focus on challenging the conceptual facet of the view. In what follows, we consider a range of such challenges to conceptual deflationism, beginning with the challenge stemming from the putative role that the truth concept plays in an account of assertion.

4 Deflationism and assertion

One prominent methodology in theorizing about truth that has emerged in the past 30 years emphasizes the importance of satisfying various platitudes pertaining to truth (see, for example, Frank Jackson et al., 1994; Michael Lynch, 1998; Crispin Wright, 1992). One such platitude taken to be of central importance is that to assert is to present as true (cf. Wright, Ibid., p. 34). In their more recent papers, Asay (Ibid.), Bar-On and Simmons (Ibid.), Howat (Ibid.) and Simmons (Ibid.) have elevated this particular platitude to the status of a what-explanation of assertion.¹⁷ They attribute this stronger position to Frege and take it to go beyond the weaker reading of the platitude as claiming merely that what sets assertion apart from other speech-acts is that when we assert, we end up presenting what we have asserted as true. They further claim that if we follow this supposedly Fregean view on the concept of assertion, then one cannot understand what it is to assert without possessing the truth concept. This then becomes the basis of challenges they pose to conceptual deflationism, which, as we have explained, constitute challenges to deflationism as a whole.

Whether this is the correct interpretation of Frege or not, for the purposes of what we are calling “Deflationary Strategy #1”, we propose resisting the claim that one must employ the concept of truth to give an adequate account of assertion by adopting what we will call an “assertion-first” view, by analogy with Timothy Williamson’s (2000) knowledge-first view. This is not simply an ad hoc maneuver for defending deflationism; for one thing, a primitivist understanding of assertion could still maintain that assertion is a substantive notion (and thus not be a deflationary account). It is also worth noting that some have attributed this view of assertion to Frege himself,¹⁸ though he is ultimately not a deflationist. (For more on this last point, see Armour-Garb and Woodbridge, forthcoming.) Whatever Frege’s actual position on the nature of assertion, one reason for thinking that an assertion-first view is possible and does not simply assume deflationism is that it is analogous to the position that Williamson has developed on the nature of knowledge, a position that is neither a deflationary view of knowledge nor an account coupled with a deflationary understanding of truth-talk.¹⁹

According to Williamson, knowledge is prior to other cognitive or epistemic kinds, both conceptually and metaphysically. On his account, the concept of knowledge is sui generis and is a theoretical primitive. As a result, while Williamson takes knowledge to be a substantive concept (in virtue of its role in psychological explanations²⁰ and, as we discuss below, the role he assigns to knowledge in explaining assertion), he denies that any of the traditionally proposed conditions for knowing, e.g., that knowledge involves justified true belief (perhaps with some post-Gettier fourth condition), serves to characterize the nature of knowledge. By analogy, according to an assertion-first view, assertion is prior to any cognitive state kinds and to any other illocutionary acts, both conceptually and metaphysically. On this view, the concept of assertion is sui generis and is a theoretical primitive. As sui generis, there will not be any analysis forthcoming of the concept of assertion (but, as a theoretical primitive, the notion can still be substantive or “inflationary”, which it would need to be if it is going to play any explanatory role, as seems plausible). As a result, none of the proposed features of asserting, e.g., that asserting involves presenting as true, serves to specify what assertion is. Since an assertion-first theorist does not subscribe to the view that to assert is to present as true, she does not violate conceptual deflationism. After all, if assertion is unanalyzable, then the truth concept is not implicated as an answer to a what-question about assertion.

That said, even if one were to reject an identification of asserting with presenting as true, the weaker claim, that we can specify how asserting is different from the performance of other speech acts by pointing out that when we assert we present as true, still seems to express a platitudinous connection, as Wright, et al. have maintained. But a deflationist can accept this point by taking presenting as true to capture a distinguishing mark of asserting, where that amounts to it being a necessarily concomitant result of performing that act, which is unique to that act without it amounting to a constitutive feature of what it is to assert.²¹ Since this does not make an appeal to the truth concept as part of an account of what assertion is, accepting this poses no problem for conceptual deflationism. But it might seem to challenge linguistic deflationism, since, in the phrase ‘present as true’, which is short for ‘present as being true’, the “gerundized” truth predicate does not seem to be operating as a device of semantic descent, as deflationists claim it does in our uses of ‘is true’. So, the question that still presses for deflationists is whether taking this use of ‘as true’ in a specification of the distinguishing mark of assertion creates a problem for linguistic deflationism.

We contend that it does not. What we want is a way of distinguishing what we do when we assert from what we do when we perform other speech-acts, like questioning or commanding, and, for those purposes, it seems natural (if not required) to bring in the expression ‘as true’. What a deflationist who accepts the unanalyzability of assertion needs, then, is a suitably deflationary way of identifying a feature of assertion that distinguishes it from other speech-acts—in particular, a way of capturing the “as factual” aspect of assertion in contrast with the “as desired” or “as speculated” aspects of other sorts of speech-acts. Ideally, deflationists would like to be able to mark what we are doing when we assert without enlisting the truth predicate at all. It turns out that a certain deflationary approach provides a way to do precisely this.

A tempting first thought for such a deflationist is to appeal to the core deflationary idea of semantic descent and contend that, for an assertoric utterance of a specified individual sentence, such as ‘Snow is white’, specifying the distinguishing mark here is accomplished by claiming that when you assert that snow is white, you present snow as being white. Similarly, one might contend, when you assert that whales are mammals, you present whales as being mammals. And so on, for all of the sentences that speakers can assertorically utter.

One might object to this first thought, however, since one can assert, for example, that snow is white even when there is no snow in one’s vicinity to present, which would seem to put pressure on the claim that this serves as an explanation of assertion.²² We can agree with the objector on this, but that is because our present goal is not to explain what it is to assert, since the strategy in play assumes that assertion is unanalyzable. Rather, the task at hand is to explain what one does when one asserts—that is, the results, or the consequences, of asserting, rather than what “manner of doing” asserting is. And this point about the possibility of asserting in conditions of absence does nothing to undercut our contention that one can accomplish this task in suitably deflationary terms. In particular, regarding the “absence” point, we would note that just as one can indicate things that are not in one’s vicinity by using language (e.g., indicating a conference one attended by talking about it), one can use language to present such things, in the sense of indicating them, as being various ways, e.g., talking about snow and presenting it as being white, even when there is none in one’s vicinity.²³

While our first “semantic-descent-like” thought may seem promising, it still leads to trouble because of a more serious and continual “absence” problem, since we sometimes, and sometimes unwittingly, assertorically utter sentences that involve vacuous expressions. When one assertorically utters ‘Vulcan is a planet’, she does not, and in fact cannot, present Vulcan as being a planet, since there is nothing to present (pace, Meinongianism). One possible response to this issue would be to say that, in the case of assertoric utterances of ordinary, vacuous-name-containing sentences, nothing gets asserted. This would enable our first thought for a deflationary response to stand, but we would not want to rest our defense of deflationism on so restrictive a thesis. Fortunately, we can avoid doing so.

Instead of employing the restrictive, bullet-biting strategy just mentioned, one can specify the “as factual” aspect of asserting, the consequence that distinguishes it from other speech-acts, without appealing to truth-talk or to the truth concept by deploying a particular deflationary approach.²⁴ The approach we have in mind derives from A.N. Prior’s (1971) analysis of truth-talk in terms (following Frank Ramsey, 1927, 1929) of sentential variables and quantifiers governing them, with Prior’s explicitly adverbial reading of these formal devices. Inspired by Ludwig Wittgenstein (1953, Sect. 134), Prior (1971, p. 38) took ‘This is how things are’ to be a sentential variable. In this vein, as a means for “improving standard English”, with respect to its “paucity of quantifiers”, Prior (1956, p. 201, 1971, p. 37) introduced a kind of adverbial quantification to govern variables for presentations of “how things are”. He coined the informal quantifier neologisms ‘anywhether’, ‘everywhether’, and ‘somewhether’, derived from the question-word ‘whether’, along with a corresponding ‘th-’ answer-word, ‘thether’, to serve as the sentential variable they govern. On this approach, we can (as Ramsey did) understand a sentence like ‘Everything Jamal believes is true’ quasi-formally in terms of ‘For all p, if Jamal believes that p, then p’ (cf. Prior, 1971, p. 24), which can be rendered non-formally using Prior’s neologisms as ‘If Jamal believes that somewhether, then thether’ or ‘For anywhether, if Jamal believes that thether, then thether’. A sentence like ‘What Isabel believed is true’ expresses ‘For some p, Isabel believed that p, and p’, which would be rendered neologistically as ‘Isabel believed that somewhether, and thether’.

We need not go all the way to employing Prior’s adverbial neologisms, however, in order to implement the kind of adverbial strategy that we are proposing here. Following Stephen Yablo’s (1996) initial appeal to ordinary “how-talk” in accounting for ways-talk or possible-worlds-talk, we can do something similar here in discussing assertion. One can see how our semantic-descent-inspired first thought for a deflationary strategy for specifying the distinguishing mark of assertion can be presented in a generalized form using how-talk, viz., in terms of “presenting things as being somehow (in particular)”.²⁵ Recognizing the problem that vacuous expressions present, the next move is to modify the account as follows: What distinguishes assertion from other illocutionary acts is a matter of this speech-act being a way of indicating how the world (or reality) is, at least from the perspective of an assertor. We do this, when we assert, by presenting the world as being somehow. Thus, and adding a bit more fineness of grain, our proposal is that when we assert, we present the world as being such that, in it, things are somehow.

For a specific individual case, such as asserting that snow is white, we can say that, when we do that, we present the world as being such that, in it, snow is white. This proposal avoids the problem that vacuous-name-containing sentences presented for our first thought. This is because the claim, that when we assert that Vulcan is a planet, we present the world as being such that, in it, Vulcan is a planet, does not require that there actually be anything that answers to ‘Vulcan’. The same goes for sentences that contain vacuous predicates.²⁶

So, to achieve the goal of specifying the distinguishing mark of assertion, we do not need to enlist the truth predicate. A possible worry one might have about our proposal, however, is that it smuggles in an appeal to a substantive notion or property of truth, just using different language. The worry is that the way we claim that a deflationist can capture the “as factual” feature of assertion, viz., in term of presenting the world as being (such that, in it, things are) somehow, really just amounts to an account of representing the world, and offering such a representation as capturing how the world is, that is, getting the world (or things, or “the facts”) right. But this seems a lot like an appeal to a notion of truth, now manifested in talk of representations getting presented as accurately representing the world.²⁷

To assuage this worry about our proposal, we note that the different qualifier phrases that get employed, in distinguishing different assertions, fragment what is involved in different acts of assertion in a way that undercuts the possibility of reconstructing a unified notion of truth. The adverbial specification of ‘as being such that, in it, snow is white’ does not involve the same factors as the adverbial specification of ‘as being such that, in it, Vulcan is a planet’. The difference is not just that ‘Vulcan’ is a vacuous term while ‘snow’ is not; there is also a difference in the factors involved in the first adverbial specification and that of ‘as being such that, in it, snow is cold’. This “disunity” point is related to the familiar deflationary claim that, while there may be a set of truths, and, in that sense, an extension of the locution ‘is true’, the elements of the set are not members of it for any common reason.²⁸ They form a set of truths only in the sense that, for different reasons, a nominalization of each can correctly have the truth predicate appended to it (to denominalize the nominalization, as it were). So, the existence of such a set would violate neither metaphysical nor linguistic deflationism. Similarly, different adverbial specifications are made in characterizing the distinguishing marks of different assertions, and these differences are integral in each case to what gets done when one makes that particular assertion. So, it is not the case that the very same thing results whenever one asserts, despite the possibility of using how-talk to “generalize” over the disunified results of asserting. As a result, there is no unified notion of the sort that one could consider a reconstruction of a substantive truth concept or truth property that is being smuggled in to play an explanatory role in our proposal.

A residual concern one might have at this point is that the “fragmented” deflationary understanding of what ‘somehow’ covers that we just explained will make the notion of assertion fragmented as well, meaning that there is no genuinely unified speech-act of assertion. But this particular fragmentation worry is not legitimate. Since Deflationary Strategy #1 takes assertion to be indefinable, our present proposal offers only a deflationary account of a consequence of what we do when we assert; it does not offer an analysis of what assertion is. So, even if Deflationary Strategy #1 fragments this consequence, that need not result in a fragmentation of assertion itself. Thus, generalizing on the distinguishing marks of particular assertions by using how-talk in the way we have indicated avoids violating linguistic, metaphysical, or conceptual deflationism.

The how-talk generalization used to specify what is unique about assertion in general employs a rough and not-quite-regimented kind of adverbial quantification already available in natural language (e.g., English) via its use of the expression ‘somehow’.²⁹ But, while the ‘some’ component of this expression involves a kind of quantification, this is not ordinary objectual quantification over a domain of “hows”, since there are no such entities. Moreover, since there is no class of linguistic items that serve as the substituends of any implicit “that is how things are” sentential variables–if any such variables are even detachable from the quantifier expressions–they also cannot be interpreted as substitutional quantifiers.³⁰ The generalizing here is essentially adverbial, rather than objectual (or substitutional). This makes it “clunkier”—less fine-grained and less flexible—than the kind of generalizing one gets with objectual quantification (or even with substitutional quantification).

To improve on this “clunkiness”, when theorizing about what we do when we assert, one might want to incorporate what the adverbial quantification covers into one’s account via a more rigorously formalizable framework, a more flexible one that can be easily integrated with other formal theorizing in first-order logic. The most direct way to do this is by having objects on hand for objectual quantification. With this in mind, one might introduce propositions as objects that get presented when we assert. However, once we have introduced propositions, we should recognize that we can distinguish a variety of ways of presenting them, which can indicate different types of speech-acts.³¹ We can capture the “as factual” aspect of asserting in this framework by saying, “When we assert a proposition, we present that proposition as true”. The question is whether deflationists can do all of this while reaping the benefits of replacing non-nominal quantification over adverbially specified presentations of the world with objectual quantification that can be regimented within first-order logic.

We contend that deflationists can do this. Indeed, this is where deflationary mottos like “the truth predicate is merely a formal device” have bearing. We can think of bringing in the framework of propositions as implementing a form of semantic ascent, from talk about the world and how it is being presented, to talk of these objects, posited as entities that specify the world as being somehow. And a deflationist can claim that the introduction of truth-talk here, to mark in propositional terms a unique consequence of asserting, just involves a use of ‘true’ that points to its core logical function of semantic descent, to undoing at a framework-level (or perhaps a meta-level) the semantic ascent implemented by introducing propositions as objects over which we can quantify, thereby returning our focus to the world.³²

We can even see a suggestion of this understanding in Prior’s views. In his adverbial sentential-variables account of truth-talk, Prior (1971, p. 21) follows Ramsey’s extended redundancy theory in maintaining generally that no instance of truth-talk is really about any proposition (being instead “about whatever the proposition is about”). However, Prior still allows that one might engage in what we will call “proposition-talk” (as we just quoted him as doing), including talk “about them” being true or being false. But he (Ibid., pp. 29–30, 98) considers this kind of talk just to involve a figure of speech about logical constructs or talk of “quasi-properties of quasi-objects”, rather than being some genuinely ontologically committing form of discourse. This is the way that deflationists should think of the introduction of the framework of propositions, with the role of truth-talk being to “logically unconstruct” applications of that framework.

In a particular case, such as an assertion that snow is white, Deflationary Strategy #1 takes the distinguishing feature of the speech-act to be its resulting in the speaker presenting the world as being such that, in it, snow is white. In this specific case, one can semantically ascend from talk of the world being presented to talk of the proposition that snow is white being presented. But now we need to capture the distinguishing “as factual” mode of presenting that asserting accomplishes. For a deflationist, this amounts to incorporating the detail that one is presenting the world as being such that, in it, things are somehow. This can be accomplished by semantically descending from the entity-talk of propositions introduced, back down to some adverbially specified talk of presenting the world. Since truth-talk implements semantic descent, it can transform talk of presenting the proposition that snow is white into talk of presenting the world. So, the adverbial expression, ‘as true’, recalling that this is short for ‘as being true’, indicates that assertion involves an adverbial modification of a presenting of the world. In this particular case, the claim, that when one asserts that snow is white, one presents the proposition that snow is white as true, captures the same aspect of assertion as the claim that when one asserts that snow is white, one presents the world as being such that, in it, snow is white. Generalizing from particular cases, the claim that when one asserts one presents a proposition as true captures the same aspect of assertion in general as the claim that when one asserts, one presents the world as being such that, in it, things are somehow.³³

A deflationist can take further inspiration from Prior here and claim that all that the proposition-talk and truth-locutions are doing is a form of “logical constructing” and “logical unconstructing”, or, to put it into more standard terminology, a kind of semantic ascent and semantic (re-)descent. One should not, therefore, take these ways of talking to reflect anything substantive in the world, or even in our conceptual scheme. Their central purpose is to serve as formal devices providing certain advantages in expressing something that one could, in principle, express without them, where these advantages include the fineness of grain and flexibility that comes with incorporation into the rigorous, formalized framework of objectual quantification and predication in first-order logic. Specifying what is unique to the speech-act of assertion, by employing talk of propositions and enlisting a truth predicate, is thus compatible with linguistic deflationism, where one has followed Deflationary Strategy #1 and retained conceptual deflationism by endorsing the assertion-first view of assertion as sui generis and unanalyzable. In fact, this Prior-inspired approach to preserving linguistic deflationism will apply throughout all of the strategies we offer deflationists in responding to the conceptual challenge.³⁴

5 Deflationism and assertion as analyzable

Endorsing an assertion-first position is not the only tack a deflationist can take in responding to the challenge we have been considering regarding the putative connections between truth and assertion. It is also possible for a deflationist to hold that assertion admits of an illuminating analysis, so long as she does not give the concept of truth a fundamental role in answering the what-question regarding assertion. This is what is involved in what we will call “Deflationary Strategy #2”. This strategy clearly makes assertion something explicitly substantive—a unified kind of speech-act that is distinct from other kinds of speech-acts and whose what-question admits of a substantive answer. Given the presumed substantiality of an answer to the what-question about assertion, one might worry that any linking of asserting with presenting as true makes the concept of truth expressed by this use of ‘true’ a substantive one, which would seem to violate conceptual and thus linguistic deflationism. But this strategy can avoid any violation of conceptual deflationism, since, as with Deflationary Strategy #1, a deflationist can simply take talk of “presenting as true” to indicate merely a distinguishing mark of assertion. So long as the account of assertion that the deflationist adopts is offered in terms that avoid any appeal to truth, the concept of truth will not function as part of the answer to the what-question about the nature of assertion.

The most developed “truth-free” analysis of assertion is that presented by Robert Brandom (1983, 1994). In his work on assertion, he explains this speech-act in terms of deontic scorekeeping and an inferential articulation involving the finer-grained normative statuses of commitment and entitlement.³⁵ One potential concern about his views, however, is that Brandom makes frequent use of truth-talk in discussing assertion and appears to identify asserting with presenting as true (or “putting forward as true” or “taking-true”). This might make his views seem potentially stronger than what a deflationist would want to include in her response to the conceptual challenge, perhaps to the extent of violating conceptual deflationism.³⁶ However, even though Brandom’s use of truth-talk to specify assertion does mean that the phrase ‘presenting as true’ picks out something substantive, his account of assertion is offered in terms that avoid any appeal to the notion of truth. As a result, the concept of truth does not function as part of his answer to the what-question about the nature of assertion. Brandom simply uses ‘true’-involving expressions as part of a label for something substantive, but his doing this does not render the concept of truth expressed by this use of ‘true’ incompatible with any aspect of deflationism. In a similar fashion, Brandom (1994, p. 329) allows the use of the locution ‘truth’ in the expression ‘truth-conditions’ as a label for what one indicates in a specification of the content of a declarative utterance. But, as with assertion, he explains content via an independent, prior, inferentialist account that makes no use of the concept of truth. So, with respect to his accounts of both assertion and content, there is no need to take any uses of truth-locutions in alternative labels for these concepts as expressing a separate, explanatory concept of truth.³⁷

Even if we are wrong about Brandom, and his official views do violate conceptual deflationism, there are other “truth-free” accounts of assertion that deflationists can employ, while implementing Deflationary Strategy #2. As mentioned above, any account that does not make the truth concept perform an explanatory role in an answer to a what-question is a candidate for incorporation into strategy #2.³⁸ With such an account on hand, a deflationist can accept that when one asserts one presents as true, again, as merely specifying a distinguishing mark of assertion.

As an example of this point, consider employing Deflationary Strategy #2 with the position on assertion championed by Williamson (1996, 2000). Williamson’s answer to the what-question about the notion of assertion is that assertion is the unique speech-act governed by a “knowledge rule” to the effect that.

(K-A) One must: assert that p, only if one knows that p.

He takes this speech-act to be “unique” in the sense that assertion is the only speech-act that is governed by this (K-A) rule. So, for Williamson (2000, pp. 238–266), (K-A) is “constitutive” of the speech-act of assertion and is akin to a rule of a game. He further claims that, by articulating (K-A), we describe our normal practice of assertion and that doing so is like articulating the rules for a traditional game. Breaking the rule does not result in the failure to make an assertion, but it does make one liable to criticism. For Williamson, (K-A) is a norm that provides the condition on which a speaker has the authority to make an assertion. Since, as we have explained above, Williamson also subscribes to a primitivist view of knowledge, his account of knowledge does not bring in the truth concept in its train up into his account of assertion. He provides a substantive answer to the what-question about assertion, but without taking the truth concept to perform any explanatory role.

6 Deflationism and identifying assertion with “presenting as true”

Both deflationary strategies considered thus far resist any identification of asserting with presenting as true. What we call “Deflationary Strategy #3” both takes the notion of assertion to be analyzable and accepts this identification. This strategy might be attractive to a deflationist who wants to accommodate platitudes about truth and who agrees with a strong reading of the platitude Wright has emphasized, “that to assert is to present as true”.

The basis of Deflationary Strategy #3 again involves analyzing the relevant instance of truth-talk, now taken as used to express what assertion is, as a formal device for expressing something adverbial, something that could be expressed without truth-talk by employing quantificational how-talk, as we described in accounting for truth-talk understood as employed to state merely a mark of assertion. On this how-talk analysis, the claim that to assert is to present a proposition as true gets re-rendered as the claim that to assert is to present the world as being such that, in it, things are somehow. A result of adopting this approach in the context of an identification of asserting with presenting as true, however, is that the fragmentation point we considered in our discussion of Deflationary Strategy #1 threatens to emerge with added strength. It now appears to apply as an objection to this deflationary approach, rather than as a response to an anti-deflationary challenge.

Above, the acknowledged fragmentary aspect of the relevant how-talk claim served to fend off any worry that a substantive truth concept (or property) was being smuggled in, and any further concerns that this would fragment assertion itself were addressed by noting that the how-talk claim expressed only a mark of assertion. Now, however, this instance of how-talk is being taken to express what assertion is. As a result, Deflationary Strategy #3 requires an account of this claim that balances making it out as fragmented enough to block any smuggling-in concerns, with making it out as unified enough that assertion still counts as a genuine kind of speech-act. This balancing act can be pulled off because this how-talk generalization involves an element of unity. The claim, that to assert is to present the world as being such that, in it, things are somehow, has it that, in every case of asserting, the speaker presents the same thing—the world—and does so in the same “presenting how it is” manner. It is just that specifically how the world gets presented as being will differ in each of the different cases, and these differences in what thereby gets done, again, are enough to block any sort of smuggling-in concerns. This may seem to be a narrow position to stake out, but we maintain it is a stable, defensible one.

As we noted above, the main reason for using truth-talk (and proposition-talk) to cover what the how-talk claim expresses is that the object-and-property combination of those discourses provides expressive advantages over employing how-talk.³⁹ What a deflationist must do is explain how her uses of these ways of talking are ontologically uncommitting. On this front, we pointed to Prior’s views of proposition-talk as a “figure of speech” introducing “logical constructs” or “quasi-objects” and of truth-talk as trafficking in a “quasi-property” in order to provide a surrogate for a flexible framework of adverbial quantifiers and sentential variables. But if one maintains that what truth-talk and proposition-talk together really express are these adverbial matters, which can be expressed in natural language directly with how-talk, this raises a question: How do these ways of talking, and specifically talk that appears to be about presenting certain objects as having a certain property, manage to express adverbial matters pertaining to how the world is presented?

One way to answer this question is to take this “presenting adverbial matters” operation of truth-talk and proposition-talk as simply brute facts about them. A different, and we think more satisfying, approach is to parallel how Yablo (1996, pp. 268–269, 275–279) connects possible-worlds-talk to a form of how-talk, and to explain Prior’s ideas of “figures of speech” introducing “quasi-objects” and a “quasi-property” by offering pretense-based accounts of proposition-talk and truth-talk (e.g., ones based on Kendall Walton’s (1990, 1993) uses of the idea of make-believe). These accounts would explain the instances of truth-talk and proposition-talk as exploiting the rules of particular pretenses they invoke, in order to express certain adverbial matters indirectly, by making “as if” to express matters pertaining to propositions and to a truth property.⁴⁰ The pretenses involved in this form of alethic fictionalism would also provide a further element of unity beyond what how-talk involves on its own, since the pretenses that underwrite proposition-talk and truth-talk will involve unified, albeit pretense-bound, answers to certain what-questions, such as what propositions are (e.g., abstract content entities) and what truth is (e.g., some sort of correspondence relation).

One might worry that if an application of Deflationary Strategy #3 involves an appeal to alethic fictionalism in the way just described, then this will include a substantive truth concept, which would be incompatible with conceptual deflationism. But this worry is based on a misunderstanding of the notion of a substantive concept and of the role that concepts can play via their applications within pretenses. A concept can be considered “substantive” by being a concept of something that is substantive, but this does not mean that the concept plays a genuinely explanatory role in an explanation of anything real, which is what it would take for a concept to be substantive in the sense that is relevant to challenges to deflationism.

Consider, for example, the concept of being magical. This seems to be a robust notion that even plays a significant role in certain genres of literature. But it is just made up; as a result, it never plays a genuinely explanatory role. A deflationist who endorses an application of Deflationary Strategy #3 that incorporates alethic fictionalism can say something analogous about the truth concept: It may be a concept of something substantive, but it is only a made-up one. Even if the fiction involved includes some “substantive story” about truth, those aspects of the fiction would not factor into a genuine answer to any what-question regarding any other concepts. Thus, while the truth predicate may be related to what seems to be a substantive concept (but only within the scope of a pretense), the concept never plays any genuine explanatory role.⁴¹ It remains a “thin” concept in the sense that the only kind of role the predicate expressing the concept ever really performs is a purely expressive one. As a result, these aspects of the view pose no threat to any facet of deflationism.

7 Our deflationary strategies and other challenges to conceptual deflationism

Having provided three strategies that deflationists can employ in responding to the conceptual challenge pertaining to assertion, we will now show how these strategies are available for dealing with other challenges that arise from the role the truth concept seems to play in accounts of other philosophically important concepts. We are not here endorsing any of the understandings of these concepts that we discuss; we are just citing them as options for deflationists to consider when applying one or the other of the three strategies we have provided for responding to various instances of the conceptual challenge.

Consider, for example, an account of belief as taking a proposition to be true with the aim of doing so only if it is true. (Cf. Velleman, 2000, p. 250.) To avoid violating conceptual deflationism, in virtue of this platitudinous-seeming claim about belief, one might follow Deflationary Strategy #1 and adopt a “belief-first” position by analogy with Williamson’s knowledge-first view, or with the assertion-first view that we described above, taking belief to be sui generis and indefinable.⁴² The appeal to the notion of truth typically employed in describing belief could then be taken as indicating a distinctive mark of belief, rather than as figuring into an analysis of its nature. The use of truth-talk in specifying this mark might then be understood as implementing a surrogate for sentential variables and quantifiers, something adverbial that could also be more directly, albeit more crudely, expressed in natural language via how-talk. (Cf. Armour-Garb & Woodbridge, 2022.) Thus, a deflationist applying this strategy might claim that when one believes something one takes the world to be somehow, with the aim of doing so only if how one is taking the world to be is how the world actually is.

Alternatively, one might follow Deflationary Strategy #2 and take belief to be analyzable without any appeal to the truth concept with the distinctive mark of belief that is typically stated in terms of truth getting presented adverbially with how-talk. For the required alternative analysis of belief, one might take inspiration from Brandom’s work and offer an account of believing that parallels his truth-free account of asserting.⁴³ Another option for someone following strategy #2 is to embrace a functionalist-style account of belief in terms of an internal state’s role in response to sensory inputs (as well as other internal states) and in the generation of action (and other internal states), provided the account of any representational element does not involve the truth concept. (Cf. Field, 1978, for this sort of account.) Finally, one could follow Deflationary Strategy #3, identifying believing with what the truth-involving description expresses (understood in terms of the adverbial re-rendering into how-talk), perhaps bringing in pretense accounts of both proposition-talk and truth-talk. The latter accounts would explain how those ways of talking end up expressing what the relevant how-talk claims express, but, again, with greater flexibility, and, as in the case of assertion, they could further help to fend off any fragmentation worries.

Next, consider how a deflationist might address a traditional account of knowledge in terms of justified true belief (perhaps plus some post-Gettier fourth condition). In this case, there is already a well-developed basis for a deflationist who wants to apply strategy #1 to avoid violating conceptual deflationism: Williamson’s (2000) knowledge-first approach. While Williamson is no deflationist, a deflationist could endorse a knowledge-first view and then take the truth-involving “JTB + ” description as just indicating distinguishing marks of knowing. The truth-involving specifications of these marks could also be re-expressed (reversing their order) adverbially using how-talk via something like the following: When one has knowledge, one takes the world to be somehow, how one is taking the world to be is how it is, and the reasons one has for taking the world to be how one is taking it to be are such that they are conducive to taking the world to be somehow only when that is how it is (perhaps with some post-Gettier condition integrated in).

For an application of Deflationary Strategy #2 to knowledge, what is required is an analysis that does not give the truth concept an explanatory role in an answer to the what-question regarding knowledge. To this end, a deflationist might appeal to Joseph Tolliver’s (1989) truth-free account of knowledge in applying Deflationary Strategy #2. Tolliver (Ibid., pp. 40–45) explains cognitive states non-propositionally, in terms of proper functioning with respect to different roles in the control of behavior, and he analyzes knowledge states in terms of a belief matching a state of affairs that both causes it and is treated as positively relevant to the belief’s role in controlling behavior, the match being with respect to the amount of information each contains regarding the state of affairs obtaining. Since Tolliver (Ibid., pp. 40–1) explains something’s containing information regarding something as a matter of the nomological/causal effect of the former on the probability of the latter, he (Ibid., p. 45) claims that his view does not require the attribution of any semantic properties to account for knowledge states. Tolliver (Ibid., p. 48) also argues that, while his approach is compatible with a truth requirement for knowledge, this could be met by a deflationary understanding of truth-talk. Thus, his account is one available to someone who wants to apply Deflationary Strategy #2 to knowledge.

A deflationist who thinks that knowledge admits of an analysis and who (perhaps in virtue of embracing certain alethic platitudes) accepts the usual truth-involving claims as part of an answer to the what-question regarding knowledge will want to apply Deflationary Strategy #3, understanding the relevant instances of truth-talk to express just what the how-talk re-renderings of them express. However, as we have seen, truth-talk provides a tidier and more flexible means for expressing what the how-talk claims express, so there is motivation for bringing in the notions of truth and proposition to express this. A deflationist might stop here, but, once again, the gap between the surface appearances of the truth- and proposition-talk and what the how-talk claims express is neatly bridged by a pretense-based analysis, here of talk of justified true belief (plus some post-Gettier fourth condition, if necessary).

Understanding the truth- and proposition-involving claims in this way might also allow their use to express additional unity regarding the concept of knowledge, conveying further intuitions about knowledge’s unity by indicating that those intuitions are part of the relevant pretenses. As in the cases of applying this strategy in accounting for assertion and belief, the use of both truth-talk and proposition-talk in this sort of what-explanation for knowledge does not indicate that any substantive explanatory role is played by the concepts of truth or of proposition. So, it appears that all three strategies provide suitable responses to the challenge of how deflationists might address the traditional account of knowledge.

Another central concept in which the truth concept appears to play a central role is that of logical validity. As we all learned in logic class, an argument is valid if and only if the truth of its premises guarantees the truth of its conclusion. Alternatively, one might follow Alfred Tarski (1941) and maintain that a zero-premise argument (i.e., a statement), A, is valid if and only if A is true under all reinterpretations of its non-logical constituents. Either option would seem to explain validity in terms of truth, which would be in tension with conceptual deflationism. However, all three of the deflationary strategies that we have presented can also be applied to solve the problem that the classical accounts of validity appear to present for conceptual deflationism.

As a basis for strategy #1, a deflationist might follow Jeffrey Ketland (2012) and adopt a “validity-first” proposal. Ketland’s leading idea involves endorsing a primitive-validity view that would be an analog to the “assertion-first” view and to Williamson’s knowledge-first view. As Ketland (Ibid., p. 421) notes, this view endorses “validity as a primitive notion rather than one defined in some standard manner” (italics original). More specifically, his idea is to treat ‘x is valid’ as a primitive unary predicate.⁴⁴ If successful, this would enable a deflationist to make use of the concept of validity without compromising her conceptual deflationism, taking the truth-involving claims associated with validity, again, as indicating a distinguishing mark of an argument having this feature. A deflationist can then say that what is expressed by a claim like ‘When an argument is logically valid, it follows that if all of its premises were true, then the conclusion would have to be true’ is just that when an argument is logically valid, it follows that if the world were how all of the premises together present it as being, then it would also have to be how the conclusion presents it as being. Or, in Tarskian terms, if A is logically valid, then how A presents the world as being is how it actually is under all reinterpretations of A’s non-logical constituents. Logical consequence could then be explained as the relation between the premises and conclusion of a logically valid argument, and an argument’s soundness would be a matter of it being logically valid with the world actually being how the premises present it as being.

Deflationary Strategy #2 can also be applied to the case of logical validity. This approach requires an appeal to a truth-free analysis of logical validity, and one can turn to Brandom for such an account. He (1994, pp. 104–107, 114–115) explains logical validity in terms of a prior notion of good material inference, where this is explained in terms of incompatibilities among the sui generis normative statuses of commitment and entitlement. (Ibid., p. 115) The account of logical or formal validity then proceeds in a way that parallels Tarski’s account, so that an argument A is logically valid if and only if A involves a good material inference and every reinterpretation of A’s non-logical constituents (that is, any substitutions for any of its non-logical vocabulary) will maintain the goodness of the material inference (i.e., no substitution for any of A’s non-logical vocabulary will turn the good inference A involves into a bad one). Brandom provides the required truth-free means of specifying the logical vocabulary via his (Ibid., pp. 108–113) reading of Frege (1879), according to which the logical vocabulary is demarcated in terms of its role in making explicit “the inferential involvements in virtue of which nonlogical claims have the conceptual contents they do”. Given this truth-free account of logical validity, a deflationist following Deflationary Strategy #2 can, once again, read the truth-involving claims typically made regarding validity as just specifying a distinguishing mark of an argument being logically valid, explaining the notion of logical validity (and thus the related notions mentioned above) and how the truth concept relates to it, all without violating conceptual deflationism.

If a deflationist follows Deflationary Strategy #3 in providing an account of logical validity, she will take the truth-involving claims made about it to express the core of what logical validity is in general. She can then say that what is expressed by the claim, that an argument being logically valid is it being such that if all of its premises were true, then the conclusion would have to be true, is just that an argument being logically valid is it being such that if the world were how all of the premises together present it as being, then it would also have to be how the conclusion presents it as being. Or, in Tarskian terms: Argument A being logically valid just is the world actually being how A presents it as being and remaining so under all reinterpretations of A’s non-logical constituents. It is also possible to incorporate fictionalist accounts of truth-talk and proposition-talk here, again taking these ways of talking as providing a more flexible and expressive means of saying (indirectly) what the relevant how-talk claims express directly, as well as bridging the gap between what the truth-involving claims appear to be about and what the how-talk claims express and perhaps expressing further unity and substantiveness regarding the concept of validity.

We take these considerations, presented via the examples examined in this section, to illustrate how to extend our deflationary strategies to other cases where the truth concept appears to play a role in answering a what-question about some philosophically important concept.

8 Concluding remarks

We have provided three strategies that deflationists can employ to respond to the alleged challenge that a substantive concept of truth is required in an explanation of the illocutionary force of assertion, and we have shown how these strategies can be extended to other supposed challenges to conceptual deflationism, which involve the concepts of belief, knowledge, and validity. As we noted above, while these challenges are focused on conceptual deflationism, they actually pose a challenge to deflationism as a whole.

We have thus shown that there are a number of different ways a deflationist could uphold conceptual deflationism, along with linguistic and metaphysical deflationism. A deflationist might employ strategy #1 and adopt a primitivist view of some concept, taking it to be unanalyzable, contrary to some challenger’s claim that it must be analyzed in terms of truth.⁴⁵ Alternatively, a deflationist could employ strategy #2, which takes the concept in question to be analyzable but maintains that this can be done via a truth-free analysis.⁴⁶ Finally, a deflationist could employ strategy #3 and take some truth-involving claim to express an answer to a what-question about the relevant concept, but maintain that in doing so the truth-talk plays a merely expressive, rather than an explanatory, role. For example, we have shown how, whether a deflationist applies strategy #1, holding that assertion is unanalyzable and that “presenting as true” picks out merely a distinguishing mark of asserting, or she takes assertion to be analyzable, and then either applies strategy #2, taking there to be an analysis in truth-locution-free terms (with “presenting as true” again indicating just a distinguishing mark of assertion), or applies strategy #3, accepting an identification of presenting as true with asserting, she can still uphold conceptual deflationism, along with linguistic and metaphysical deflationism. Thus, deflationism is safe from the particular challenge pertaining to providing an account of the illocutionary force of assertion, as well as to the other putative challenges that we have considered, which supposedly arise from the truth concept’s role in accounts of the concepts of belief, knowledge, and validity. Our hope is that once deflationists understand how these three strategies can be implemented, as a means for answering the alleged challenges to conceptual deflationism that we have considered here, they will also see how these strategies can be applied to resolve challenges that arise for the other concepts that appear to put pressure on conceptual deflationism and, thus, as we maintain, on deflationism as a whole.