ARTICLES

1. Avowals, expression and truth

By analogy with expressivism about other areas of discourse, expressivism about avowals is the view that the meaning of typical self-ascriptions of psychological states – i.e., sensation, belief, desire, intention, and the like – is a matter of expressing an attitude, rather than describing (reporting, or asserting the existence of) a state of affairs. Traditionally, expressivism has been glossed as the view that, qua expressions, avowals are not truth-evaluable (Hacker 1986: 298; Wright 2001: 358-64). To be sure, the traditional gloss does not amount to the claim that avowals are not evaluable as to their correctness, for it is compatible with the idea that avowals can be genuine (sincere, truthful) or not. What is clear, though, is that for the traditional gloss avowals are not truth-evaluable, and therefore genuineness (sincerity, truthfulness) is not truth. This paper contests the latter claim.

The aim of this paper, then, is to shed light on the relationship between the following two issues: whether avowals are truth-evaluable; and whether truth and genuineness (sincerity, truthfulness) are two separate properties. According to the traditional gloss on expressivism, both issues are connected: for in so far as avowals are not truth-evaluable, and in so far as they can be genuine (sincere, truthful), then truth and genuineness are not the same property. Contemporary critiques of the traditional gloss by self-confessed expressivists like David Finkelstein and Dorit Bar-On (references below) have targeted the thesis that avowals are not truth-evaluable, leaving unscathed the thesis that truth and genuineness (sincerity, truthfulness) are two different properties of avowals. Therefore, these critiques are committed to an undertaking to clarify the expressive nature of avowals within which a defence of the truth-evaluability of avowals is separate from their genuineness.

This paper will contest the soundness of the undertaking, and ultimately argue for an expressivist view of avowals where truth-evaluability is not separate from genuineness (sincerity, truthfulness). Hence, the expressivist view to be put forward will be significantly different from both traditional and other contemporary expressivist accounts, as it will be argued that, unlike traditional expressivism, avowals are truth-evaluable; and that unlike other contemporary expressivist accounts, truth-evaluability is not separate from genuineness (sincerity, truthfulness). This alternative view of avowals will be defended on the basis of an argument from disagreement to be presented below (in section 5). Prior to this defence, the phenomenon of disagreement under discussion will be introduced by means of an example (in section 3), followed by an examination of the difficulties encountered by contemporary expressivist accounts in dealing with the phenomenon in question (in sections 3 and 4). Finally, some alleged objections to the expressivist view of avowals favoured in this paper will be put forward and answered, as a result of which the expressivist thesis defended here will be further sharpened (in sections 7 through 9). But first, let us take a quick look at the prospects of expressivism as a general theory of meaning.

2. Expressivism, old and new

As previously stated, expressivism, as traditionally glossed, is the claim that the meaning of a given area of discourse is a matter of expressing an attitude (rather than a matter of describing a state of affairs), and is therefore not truth-evaluable. Expressivism, thus conceived, has long been considered a hopeless account of meaning in such areas of discourse as ethics. The reason for this goes back to deflationism about truth. Thus, quasi-realists like Simon Blackburn have argued both (i) that ethical propositions are not descriptions, but rather expressions of some commitment; and (ii) that truth-talk in ethics (including fact-, correspondence-, or even representation-talk) is perfectly acceptable, for given deflationism, truth is not a substantive or explanatory notion, and therefore truth-talk is metaphysically innocuous. (See Blackburn 1984, chapter 6; and 2010, chapters 2, 9 and 11.)

Furthermore, what quasi-realists like Blackburn insist upon is that, even if deflationism makes all the semantic differences between different areas of discourse disappear at the surface-propositional level (for truth-talk is pervasive), there is a host of relevant differences (i.e., differences regarding the function of those propositions in our lives) to be found below the surface. It is at this deeper level that expressivism emerges as a serious contender in ethics (and elsewhere). As he puts it:

Our practical lives may take account of possibilities, numbers, properties and universals, ethics and things we bump into. And our semantic talk may make all these things look alike. But underneath [...] lie differences of function, differences of role in practice, and these differences make all the difference. (2009: 47)¹

Although Blackburn has mainly focused on the meaning of ethical discourse, more recently some defenders of expressivism about avowals have also questioned the traditional gloss, arguing instead that avowals are both expressions and truth-evaluable (Bar-On & Long 2001; Finkelstein 2003; Bar-On 2004). To support this view, they have maintained that the traditional gloss incurs in a non sequitur when it claims that, qua expressions, avowals are not truth-evaluable. The traditional gloss only follows if expressions are not assertions, and only assertions are truth-evaluable. However, contemporary expressivists about avowals have pointed out that expressing and asserting are not mutually incompatible (Finkelstein 2003: 95-6; 2010: 194-5; Bar-On 2004: 299ff). The crucial thought here is that avowals are linguistic acts, and therefore have a semantic structure similar to third-person mental ascriptions; therefore, if the latter are truth-evaluable, so are avowals.

Following Bar-On, and others, this view can be labelled “neoexpressivism”, to distinguish it from the traditional gloss, which she calls “simple” expressivism. What this suggests is that neoexpressivism is a novel thesis about avowals, although in fact it was previously formulated by Carnap in the 1930s (Carnap 1935: 28). However, let us not quarrel over labels, as a more substantial issue awaits us – namely, whether neoexpressivism about avowals is correct.

3. A puzzle stated: disagreement in expression

Neoexpressivists argue that avowals are truth-evaluable because they are linguistic utterances with a semantic structure, similar to other linguistic utterances where psychological states are ascribed to others. As a result, avowals differ from other nonlinguistic expressions, like winces, in that the latter are not truth-evaluable. In this respect, avowals and nonlinguistic expressions are both similar and dissimilar, though in different respects: they are both expressions, but only avowals are truth-evaluable. Neoexpressivists have seized on this dissimilarity to try to shed light on the different roles that avowals and winces play in our expressive practices.

Thus, David Finkelstein has introduced the following example²:

avowals may indeed contradict – not merely undermine – what others say. Imagine that while suffering from a migraine, I wince and say to you, “My head really hurts a lot.” Later, in an insecure moment, you accuse me of having lied about my condition in order to avoid spending the evening with you. Here, I might reply: “What I said was true. I had the worst headache I can remember.” This is a very different conversation from any we might have had if the pain in my head had been expressed only nonlinguistically – if, e.g., I’d moaned, massaged my temples, and winced. In such a circumstance, you couldn’t have accused me of lying about having had a headache – only of feigning one. And I would not have met such an accusation by saying that my moaning, wincing, and massaging were true. (2003: 94)

According to Finkelstein, what the example shows is that avowals are not treated in the same way as nonlinguistic expressions, like winces. Thus, avowals could be accused of being a lie, but not winces. Furthermore, when an avowal is regarded as a lie, the accusation amounts to a contradiction of the initial avowal; for instance, somebody asserts that her head aches, and someone else denies that it is true. However, Finkelstein submits, the same cannot be said about winces: they cannot be contradicted in a similar way, although they can be undermined. Neoexpressivists explain this difference in terms of the fact that, despite being a piece of expressive behaviour, nonlinguistic winces, unlike avowals, are not truth-evaluable.

Let us consider now the following variation on Finkelstein’s example. As I moan and wince, you say: “Come on. Who do you think you’re kidding! Your head doesn’t ache.” Subsequently, I take exception to what you have said and go off in a huff. Why do I take exception and go off? Simply perhaps because I have been found out – my winces were a ploy not to spend the evening with you, you have realized, and I play the wounded victim as a defensive strategy. But what if my head really had ached? Why then did I take exception and go off? The following is plausible: my winces and moans were genuine, and I am badly hurt by your denial.

Can neoexpressivists make sense of this example? Finkelstein (2010) has explicitly introduced the idea of disagreement between first- and third-person psychological ascriptions to support the view that avowals, while expressions, are truth-evaluable. Thus, he considers the disagreement that arises when somebody “describes me as ‘wanting to visit Colorado in the spring’ and I correct her, saying, ‘No; it’s Wyoming that I want to visit’” (192-3). Here, disagreement is understood in terms of contradiction: somebody puts forward as a true claim that I want to visit Colorado, and I deny it. But this cannot be the only model for disagreement, on pain of ruling out something initially plausible, namely that when I wince and you say “Come on. Your head doesn’t ache”, I take it that you are disagreeing with me, and therefore I take exception and go off.³ This must be the reason why Finkelstein explicitly acknowledges that winces can be “undermined” by linguistic utterances.⁴

The crucial point now, though, is that Finkelstein’s distinction between two types of disagreement (contradicting versus undermining) entails that what goes on when I say “My head hurts” and you reply “No, it doesn’t”, is different from what goes on when I moan and you similarly say “Come on; your head doesn’t ache.” For, although both situations involve some form of disagreement, the linguistic versus nonlinguistic nature of the initial expression of pain affects the nature of the disagreement. Thus, in so far as my initial (linguistic) pain avowal is truth-evaluable, it can be denied that it is true (i.e., contradicted) by your words; whereas in so far as my winces are not linguistic, hence not truth-evaluable, it cannot be denied that they are true, although they can be undermined.

Perhaps the difference can be glossed as follows: the first situation involves a disagreement about truth (whether or not I am saying something true), whereas the second situation involves a disagreement about expression (whether my wincing is a genuine or feigned expression). This appears to be borne out by Finkelstein’s distinction in the quoted text between accusing someone of lying and accusing them of feigning,⁵ for the target of the former accusation is the truth of what has been said, whereas the target of the latter is not truth, but the genuineness of the original expression. On the face of it, this distinction between truth and genuineness seems plausible enough, at least when applied to winces. But the crucial question is: does it apply to avowals, too? If when I avow my pain and you deny the truth of what I say, what you are doing is taking issue with the truth of what I say, does that mean that genuineness is not an issue between us (i.e., that the genuineness of what I do is beyond question); or does it mean that, unlike truth, genuineness is not an aspect of what I say? The latter is not plausible, for avowals are expressions, and so evaluable as to their genuineness. But in that case (turning to the former), it must be possible for the issue about the genuineness of what I say to be under dispute at sometime, and if that is not the target of the dispute in cases of contradiction, i.e. when the truth of what I say is denied, how could the genuineness of what I say be under dispute?

Finkelstein provides no answers here, although it is his framework for the discussion of the differences involving disagreement with avowals versus disagreement with nonlinguistic winces that gives rise to the questions. So, unless answers can be found, the existence of those differences, and the very phenomenon of disagreement here highlighted, remains a puzzle for neoexpressivists about avowals.

4. The puzzle explored: avowals as acts versus products

Bar-On has defended neoexpressivism by means of a distinction between acts versus products which could be helpful here, in so far as it allows for disagreement between my wincing and your words. As she puts it,

[t]he product of an act of avowing, unlike a smile or a wince, or even a verbal cry such as “Ouch!”, is a semantically articulate self-ascription, an item with semantic structure and truth-conditions. [...] But, in any event, I think that the expressivist insight regarding avowals should be understood, in the first instance, as a claim about the relevant acts, not about their products. The claim is that there are notable similarities between acts of avowing a state and the act of giving it a (so-called) natural expression [...]. And this claim could be true, even if there were systematic differences between the products of acts of avowing, and other linguistic acts, on the one hand, and the products of naturally expressive acts, on the other. (2004: 251-2)

Applied to our examples, Bar-On’s distinction amounts to the claim that winces and avowals like “My head aches” are similar qua expressive acts, but differ in that only the latter involve truth-evaluable products. And the initially helpful suggestion is that, on the one hand, the disagreement between my wincing and your words is a matter of the expressive act in question, rather than the product or “vehicle” (2004: 253) involved; in other words, it is a dispute about whether my wincing is a truly expressive (i.e., genuine, truthful) act, something I take you to deny with your words, hence my reaction. On the other hand, the disagreement between my avowal and your words is not so much a matter of the expressive act in question, but rather of the linguistic product or vehicle involved; in other words, it is a dispute about whether my utterance “My head hurts” is itself a true product. Hence, Bar-On’s distinction between acts and products apparently helps to solve the puzzle over the nature of disagreement faced by neoexpressivists.

A corollary of Bar-On’s distinction is that the relation between linguistic avowals, on the one hand, and (nonlinguistic) winces, on the other, should be conceived of as follows: avowals and winces are all, qua expressive acts, evaluable regarding genuineness, whereas avowals are also, qua linguistic products, truth-evaluable. But this spells trouble for Bar-On’s initially helpful suggestion. For, in so far as there are two independent dimensions along which avowals are evaluable – i.e., truth and genuineness – it would be possible for an avowal to be true qua product, but not genuine qua act, and vice versa. Bar-On endorses this when she accepts the existence of genuine, but false, avowals as “expressive failures” (2004: 320-35), as when “[a] freshman being initiated in to a fraternity and primed to think he will be harmed may, upon having a piece of ice pressed to his neck, scream, ‘Stop it! My neck hurts!’” (322). Assuming, as one may reasonably do, that the freshman is not in pain, this is a false self-ascription, but it may nonetheless count as genuinely expressive (rather than a case of pretence), for he “has successfully expressed pain, [though he] has not succeeded in expressing [his] pain” (323).

But even assuming for the sake of argument that the examples discussed by Bar-On are explained by the distinction between acts and products (more on this example below), the distinction does not apply to cases of avowals that are in fact expressive of the subject’s own mental life. For here, the uttered words are themselves the expressive act, rather than something one could, as it were, peel off from the expressive act, and evaluate separately. “Peel off” conveys the right image here. For if one could (conceptually) detach the linguistic product from the expressive act, one could easily conceive of the clash between my avowal (e.g., when I utter “My head hurts”) and your words, as a case of disagreement between expressive acts, regardless of the products or vehicles involved. But if the latter is hardly intelligible, the proposed detachment is conceptually suspect, and any attempt to explain such cases of disagreement on the basis of a distinction between acts and products will not work.

In sum, the problem with Bar-On’s suggestion is that if, in order to make room for disagreement between wincing and linguistic utterances, neoexpressivists separate acts from products, they deprive themselves of the thought that with avowals it is the linguistic act, qua linguistic act (rather than simply qua act), that is the expressive act.⁶ The suggestion is hardly intelligible in the nonlinguistic case (try separating the act from the vehicle when someone pulls a face in pain)⁷; so why should it be more plausible in the case of avowals? The outcome is that Bar-On’s strategy to deal with cases of disagreement between wincing and linguistic utterances means trouble for her expressivist conception of avowals at large. Therefore, the puzzle faced by expressivists over the nature of disagreement, including the differences involving cases of avowals versus winces, remains.

5. The puzzle solved: truth as truthfulness

The gist of the argument so far is that neoexpressivists are caught in a dilemma: either they make sense of cases of disagreement between winces and linguistic utterances, but pay the high price of viewing human acts, including avowals, as expressive regardless of the vehicle involved; or they avoid any such view of human expression, but fail to account for the very phenomenon of disagreement, including the differences between cases of linguistic avowals and cases of nonlinguistic winces. Section 3 has argued that Finkelstein falls prey to the latter horn of the dilemma; whereas section 4 has claimed that Bar-On flounders on the former. At this juncture, abandoning neoexpressivism in favour of traditional or simple expressivism is not a real option, either. So, what is the way forward? The key step is to recognize that expressivism about avowals has not received a full characterisation in recent discussions. In other words, neoexpressivists are right to say that avowals are expressions and truth-evaluable, but have not yet clarified what truth-evaluability amounts to here. A possible way to remedy this is to argue that truth and genuineness (truthfulness) are not two separate properties of avowals, something that follows from their expressive nature. Let us explore this in more detail.

There is an important similarity between Finkelstein’s original example and the variation supplied above. In the variation on Finkelstein’s example, I wince and when you say “Come on; your head doesn’t ache”, I react as I do, because I take it that you are denying that my wincing is genuine, thereby showing that a disagreement has arisen concerning the genuineness (sincerity, truthfulness) of my wincing. In Finkelstein’s original example, I avow “My head aches”, you say “No; it doesn’t. You just don’t want to spend the evening with me”, and I subsequently take exception to your words, and react accordingly. Here, what makes my reaction intelligible is the existence of a disagreement between my avowal and your words. More precisely, I react as I do, because you deny that my avowal is genuine, that I genuinely (sincerely, truthfully) express my headache with my avowal. So, the disagreement concerns the genuineness of the avowal.

Now, if both scenarios involve cases of disagreement regarding the genuineness of my behaviour, and in Finkelstein’s original example the disagreement concerns truth-evaluable linguistic acts, then the disagreement in Finkelstein’s original example concerns the genuineness of a truth-evaluable linguistic act. In other words, when I reply, as I might have done in response to your words, “But of course it’s true that my head aches”, I am insisting that my previous linguistic behaviour was genuine (sincere, truthful). Insisting that what I had said before was true is a relevant reply, when there is a disagreement over the genuineness of my linguistic behaviour. Therefore, making sense of disagreement in Finkelstein’s original example and the variation on it requires that in avowals, given their expressive nature, truth is truthfulness.

The previous argument from disagreement does not entail that there are no differences between Finkelstein’s original example and the variation on it. There is no denying the fact that the former involves linguistic utterances, whereas the latter does not; therefore, only in the former case is it right to say that the disagreement is over whether I lied or not. But despite these differences, if lying is a form of feigning or deception (one that is open only to linguistic beings), there is no denying either that when we disagree over whether I have feigned my headache (either through linguistic utterances or through wincing), what we disagree about is the genuineness of what I did.

To sum up, we need not (and must not) separate truth from truthfulness to accommodate for the differences between linguistic and nonlinguistic expression. Rather, accounting for the phenomenon of disagreement includes explaining that in cases of disagreement involving avowals and in cases of disagreement involving winces alike, there is a dispute about the genuineness of expression, and the latter requires that genuine (sincere, truthful) expression is what truth amounts to in the case of avowals.⁸

6. Interlude: quasi-realism revisited

Before proceeding to discuss some objections in the next three sections, it will be useful briefly to establish how the foregoing relates to current quasi-realism, an influential account of meaning in different areas of discourse, as seen in section 2. In this respect, it must be noted that the claim defended in this paper does not sit well with Blackburn’s distinction between a metaphysically innocuous level of truth-talk, that applies pervasively throughout discourse (including psychological self-ascriptions) and a deeper level of differences, that unveils crucial expressive components in some areas of discourse (e.g., in psychological self-ascriptions). Contrary to this division into different levels, the conclusion of this paper is that an exploration of the role of avowals in our lives not only shows a close relationship between their expressive function and such conditions of correctness as sincerity, but also that this close relationship (at the deeper expressive level, as Blackburn would put it) is what truth in avowals amounts to. There is, therefore, no room for a neat distinction of levels, as a proper understanding of the expressive nature of avowals goes hand-in-hand with an understanding of what it is for avowals to be truth-evaluable. So, even if Blackburn is right that deflationism about truth must be assumed from the beginning, this will not thwart a philosophical investigation into the truth of avowals; in particular, an investigation into, or perhaps simply a reminder of, the relations between such notions as avowal, expression, truth, and truthfulness.

7. Wittgenstein’s expressivism about avowals

It could be objected that the previous argument from disagreement proves too much when it concludes that truth in avowals amounts to truthful expression. For it seems plausible to suggest that one could sincerely avow something, without it being true. Perhaps pain is not a good example here, Bar-On’s freshman notwithstanding (see section 4); so let us consider belief instead. I may sincerely avow that I believe that p, but fail to say something true, because I am mistaken either about what I believe (I might believe that q, rather than p), or even that I believe it (it might be a case of desire, rather than belief).

The point of the preceding sections is that, in so far as this is a real possibility, it is not one left open by expressivism about avowals. For, if avowals are expressions of one’s mental life, and one’s avowals are sincere (genuine), there is no identity criterion for the existence of the mental state in question (say, belief) other than the avowal itself. So, there is no room for a concept of truth different from truthfulness. Furthermore, this applies equally to pain and belief, something that Wittgenstein saw clearly when he faced this issue.

Thus, in the course of his rejection of the Cartesian conception of sensations in particular, or mind in general, as a set of private objects, Wittgenstein explores the thought if one could have doubts as to, say, whether one is in pain (PI §§246, 288), or whether one might mistake “a sensation for something other than what it is” (PI §288). His reply is that these possibilities require “the abrogation of the normal language-game”, where human behaviour is the expression of sensation; for then “a criterion of identity for the sensation” (other than the avowal itself) is needed, “and then the possibility of error also exists” (PI §288). Cartesianism is just such a conception of the mind, where an avowal is a description of a set of private mental objects, which in turn promotes a view of the relation between avowals and mind as a matter of fit between mutually independent items. “Truth” might be a label for such a relation. And in so far as Wittgenstein is opposing a Cartesian conception of the mind, it might seem that his endorsement of expression must involve a rejection of truth, or truth-talk (as in traditional or simple expressivism). But things are otherwise.

In this respect, it is instructive to attend to the following passage from the second half of the Investigations, where Wittgenstein raises the possibility of making a mistake regarding the content of one’s own thoughts, together with his reply for the particular case of the content of one’s own reported dreams:

Let us assume there was a man who always guessed right what I was saying to myself in my thoughts. [...] But what is the criterion for his guessing right? Well, I am a truthful person and I confess that he has guessed right. – But might I not be mistaken, can my memory not deceive me? And might it not always do so when – without lying – I express what I have thought within myself? [...]

    The question whether the dreamer’s memory deceives him when he reports the dream after waking cannot arise, unless indeed we introduce a completely new criterion for the report’s ‘agreeing’ with the dream, a criterion which gives us a concept of ‘truth’ as distinct from ‘truthfulness’ here. (PI, II, p. 189)

Wittgenstein’s conclusion is not that viewing one’s own avowals as expressive entails that truth-talk is out of place. Rather, what is out of place is a certain conception of what truth-talk amounts to here – namely, a relation of fit between mutually independent items; as well as a related view of what a mistake is – namely, lack of fit. Instead, what truth in avowals amounts to is the confession of a truthful person, one who expresses his mental life by avowing it. Therefore, Wittgenstein is not only rejecting traditional or simple expressivism, but also a view of truth in avowals as distinct from genuine or truthful expression. This is how he puts it, in a paragraph omitted in the previous quote:

The criteria for the truth of the confession that I thought such-and-such are not the criteria for a true description of a process. And the importance of the true confession does not reside in its being a correct and certain report of a process. It resides rather in the special conclusions which can be drawn from a confession whose truth is guaranteed by the special criteria of truthfulness. (PI, II, p. 189)

8. Genuine, but false, avowals

Now, a possible objector may insist that, even if the former section gets Wittgenstein’s exegesis right, it is simply too counterintuitive to argue, on behalf of expressivism about avowals, that the latter does not leave open the possibility that one could sincerely avow something, without it being true. As mentioned previously, belief is a case in point, for I may avow that I believe that p, while failing to say something true. Intention provides another suitable example: I may avow that I intend to do A, but get things wrong (without lying or intending to deceive), as shown by the fact that others who know me well may say with truth that things are not so (perhaps I intend to do something else instead, or simply I do not intend to do A). So, at least for the cases of belief and intention, sincere avowal is compatible with absence of truth; and if expressivism about avowals claims, or entails, that this is not an open possibility, so much the worse for expressivism about avowals. Or so a possible objector might insist.

In line with the rest of this paper, the reply should be that the objector is confused in taking expressivism to rule out a priori (actual) cases such as making a mistake about the avowed content of one’s mind (say, beliefs or intentions). In so far as these are actual cases, any attempt to rule them out a priori amounts to dialectical suicide. But the expressivist is not so desperate, or befuddled. Rather, his point is that those cases are not examples of expression. In other words, when I avow that I believe that p, or that I intend to do A, but fail to say something true, my avowal is not an expression of the contents of my mind, although it may appear to be so. Perhaps I think that I am expressing my mind, and perhaps others think so too (at least those who do not know me well, and therefore do not know better), but I am wrong, and so are they. A neat way to formulate this idea would be the following: such avowals are not cases of real expression, only of apparent expression; where an apparent expression is not a different type of expression, but rather not an expression at all (although mistakenly taken to be so). But crucially, so conceived, the cases of avowal pointed out by the objector are no trouble for the expressivist thesis defended in this paper, according to which if avowals are expressions, truth is truthfulness. For in so far as they are not cases of expression, they fall outside the scope of the expressivist thesis of this paper.

Furthermore, this fits nicely with the interpretation of Wittgenstein’s texts above. It has just been claimed that sincere but false avowals are not expressions, but are mistakenly taken to be so (by oneself or others). Therefore, a proper understanding of sincere but false avowals rests on a prior conception of avowals as expressions. Now, this is a Wittgensteinian point, for as previously seen, he claims that the possibility of a mistake regarding the avowed contents of one’s mind requires the “the abrogation of the normal language-game” – i.e., requires that those possibilities not be expressions of one’s mind. However, this does not mean that avowals are never expressions. On the contrary, as shown, Wittgenstein argues against a conclusion to the effect that, given the possibility of mistaken avowals, the relationship between avowals and mind must be captured in terms of truth-as-fit, rather than expression. It is now clear why this is so: if mistaken avowals are apparent expressions (in the sense of not being expressions, despite being taken to be so), then the relationship between avowals and mind must sometimes be captured in terms of expression, on pain of not having a model for the conception of mistaken (sincere but false) avowals.

So far, so good; but the objector may want more. In particular, why should one concede that those cases of avowal of belief and intention, and others like them, do not involve expression? As an alternative, let us re-consider the case of the freshman, introduced in section 4. According to the example, a freshman primed to think that he will be harmed may avow to be in pain when an ice cube is pressed to his neck, although it may be safely assumed that he is in no pain. Similar to the cases of belief and intention considered above, on the face of it this is a case involving a sincere, but false, avowal. Furthermore, according to Bar-On’s analysis of the example, in so far as the freshman is not engaged in pretence, it is a case of genuine expression: not an expression of the freshman’s own pain (this is our safe assumption), but an expression of pain, nonetheless. And now the objector’s desire for more takes on a clearer shape: why should we think of all these examples (whether pain, belief or intention) as cases of apparent expression; instead of cases of real expression along the lines suggested by Bar-On?

In Bar-On’s freshman example, her distinction between expressing pain versus expressing somebody’s (the freshman’s own) pain is related to a distinction she is already committed to, that between avowals qua products and avowals qua acts. Her point is, then, that the freshman’s utterance has a (conventional) expressive force qua product, which is autonomous from the expressive force of the act itself.⁹ So, an avowal could be expressive qua product, without being expressive qua act; and this is precisely what happens in the example: the freshman’s utterance expresses pain, although his act does not (for he is not in pain).

It was argued in section 4 that the distinction between products and acts leads to a confused conception of the expressive nature of true avowals (those which are in fact expressive of the subject’s mental life), in so far as it denies that with avowals it is the linguistic act, qua linguistic act (rather than simply qua act), that is the expressive act. But the question was bracketed then whether the distinction between products and acts could explain examples of sincere but false avowals, like that of the freshman. It is important to note now that the example does not demand, or recommend, Bar-On’s explanatory strategy. The point of the initiation game in which the freshman is taking part is to make him cry out in pain, without being in pain; in other words, making him cry out as if he were in pain. But in that case, the conceptual equipment that makes the example intelligible involves only the notions of real versus apparent expression. In fact, the most that could be said for Bar-On’s distinction is that talk of expression of pain versus expression of somebody’s (the freshman’s) pain is in fact a fanciful way of saying that we can make people cry out in pain, through the use of (conventional) linguistic tools, without them being in pain. So, if the very distinction between products and acts is already suspect (in cases of true avowal), and it is not independently required or advisable in cases of sincere but false avowals, like that of the freshman, then there is no proper content to the objector’s question as to why we should think of the latter cases as involving apparent expression, rather than real expression alongside Bar-On’s line. In sum, the objector’s question dissolves, and the conception of sincere but false avowals as apparent expression stands, a conception that is compatible with the expressivist thesis defended in this paper.

9. Self-deception for expressivists

It could be further objected to the expressivist proposal of this paper that it leaves no room for self-deception; or perhaps simply that it does not accommodate self-deception in a suitable way. For either expressivism is committed to a form of infallibility that rules out failures of expression, such as cases of self-deception, where one sincerely utters false avowals; or if not simply ruled out, such cases are not suitably incorporated into the expressivist picture. Thus, it may be conceded that if sincere avowals are expressions, then truth and truthfulness are not two separate properties. But then, one could object to that concession, on the grounds that sincere avowals are expressions of one’s mental life, unless one is self-deceived. So, if lack of self-deception is a further condition for expression, isn’t the move from sincerity to expression too quick?

There is something right in this line of thought – namely, that expressivism about avowals in general, not only the expressivist proposal of this paper, would be a stronger thesis if it had a coherent account of self-deception. (Although it would be ludicrous to add that while a full answer is pending, expressivism is worthless.) So, here is a sketch of an answer: self-deception is indeed a failure of expression, but this in turn means that the former is dependent on the latter, and not vice versa. Thus, expressivists need not deny the possibility (or actuality) of self-deception. But, in reply to the objection of the previous paragraph, the crucial thought is that, whereas self-deception is a matter of failing to express one’s mind through sincere avowal, expressing one’s mind is (in the linguistic case) simply a matter of sincere avowal, not sincere avowal plus lack of self-deception; for expression through sincere avowal is prior to self-deception.

Now, this priority claim might look too weak a basis on which to hang the fate of expressivism, for it may seem that it amounts to no more than a claim of conceptual priority, in the sense that self-deception can be defined as a failure of expression, to counteract an alternative definition of expression as sincere avowal without self-deception. But, this line of thought continues, definitions are too cheap to come by, and by themselves lack sufficient force to provide support for expressivism in general, or for the expressivist proposal of this paper in particular.

This is a fair point to make against a claim of conceptual priority, understood as a matter of definitions; but this is not the point expressivists should be making. Rather, what they should be saying is that, by stressing the priority of expression through sincere avowal over self-deception, they are making both a conceptual and an ontogenetic point. Conceptual, in the sense that their aim is not the purely formal one of providing a new consistent way of juggling some concepts, but rather the more substantial one of promoting a different outlook on the very phenomenon of self-deception, and its relation to expression through sincere avowal. Ontogenetic, in the sense that part of this outlook is the idea that viewing self-deception as dependent on sincere expression, rather than the other way around, makes for a more plausible developmental story about how children acquire the ability to express their minds through avowals. On the view that expression is sincere avowal without self-deception, children who express themselves through avowals must be subject to possible self-deception from the start; otherwise, they would not count as expressing their minds. But it is much more intuitive to think of self-deception as requiring a more complex mind than that of children who begin to express themselves through avowals. Therefore, expression must be in the developmental picture before self-deception is, and not vice versa. But if things are so, then in reply to the objection that expressivism makes no suitable room for self-deception, expressivists are right to counsel that we view expression through avowal as prior to self-deception.

Admittedly, this is just a sketch but, to sum up and conclude, the main objective of this paper has not been the defence of expressivism about avowals in general. Thus, no attempt has been made to motivate expressivism from scratch, or to consider all the pros and cons. Rather, the core of this paper has been the defence of a claim missing from contemporary discussions of expressivism about avowals – namely, that avowals are, qua expressions, truth-evaluable. This claim contrasts with the traditional gloss on expressivism, according to which avowals are expressions, and therefore not truth-evaluable; as well as with more recent neoexpressivism, according to which avowals are expressions, and truth-evaluable besides. To repeat, it is a claim missing from contemporary expressivism, but not from the work of one of the acknowledged founding fathers of expressivism – Wittgenstein himself.¹⁰

Biographical note

Ángel García Rodríguez was educated in Spain and England, before receiving a Ph.D. from the University of Hull (UK). He has taught at the University of Murcia (Spain) since 1995, and currently holds the position of Senior Lecturer in Philosophy. His research interests cover an area where philosophy of language, philosophy of mind and epistemology meet.