Assertion, Lying and the Norm of Truth

Abstract

In chapter four of Truth and Truthfulness Bernard Williams presents an account of assertion that relies heavily on the ‘psychological’ notions of belief and intention. In chapter five his definition of lying similarly relies on such notions. For Williams, insofar as there are norms governing assertion as such or norms broken by lying as such, these norms relate to saying what you think to be true, as distinct from saying what is true. I argue that this ‘psychologized’ account of assertion (and lying) is for various reasons mistaken. A consequence of Williams’s approach is that ‘Shall I tell the truth here?’ is presented as a much more open question for an agent than it possibly can be. Only by adverting to the language-game presupposed by that question’s having any sense at all can we arrive at a fair picture of when and how the answer ‘No’ might be reasonable.

In chapter four of Truth and Truthfulness Bernard Williams lays out ‘the standard conditions of A’s asserting that P’ as follows:

A utters a sentence “S”, where “S” means that P, in doing which either he expresses his belief that P, or he intends the person addressed to take it that he believes that P. (Williams 2002, 74.)

On the face of it, Williams’ account of assertion takes the notion of belief as logically prior to that of assertion. He is aware that very often the criterion for one’s believing that P is simply that one is disposed to assert ‘P’,¹ but he sees no problem in that for his view, and his reason for seeing no problem here is instructive. He points out that the phrase ‘expresses his belief’ shouldn’t be taken as implying that the belief was already there (though it might have been). In Williams’ words, ‘What makes [a given content] into a belief may be that we are asked about the matter or about the belief and then have to decide whether we are prepared to assert it or not’ (p. 82).² The grounds for such a ‘decision’, for Williams, will relate especially to such interpersonal factors as whether ‘I am concerned to be helpful’ or whether ‘I want my friend to understand how I see things’. The picture that emerges is one in which any norm of truth associated with assertion arises from personal and interpersonal phenomena: belief, trust, co-operation, etc. Such a norm of truth cannot be seen as built into the very idea of assertion. In other words, insofar as an assertion is meant to be true, that fact can’t be derived from the very idea of assertion, but relates to matters that are connected with assertion in ways that are (in a certain sense) contingent. What this all amounts to we shall see presently.

How are we to understand ‘”S” means that P’ in Williams’ statement of the standard conditions of assertion? Williams admits the necessity for including this clause a little earlier in the chapter but regards it as not part of his brief to explain the sense of the clause, writing ‘if we were trying to explain meaning, this would be unhelpful. But we are not trying to explain meaning, only to give an account of what assertion is’ (p. 71). Alarm bells ought surely to ring at this point. For it seems natural to say that for a given bit of language to mean that P is, centrally, for the rules governing the use of ‘P’ to require one’s saying ‘P’ in certain circumstances, not saying ‘not P’ in those circumstances, and so on. These various requirements³ provide criteria for linguistic competence—failing too often to conform to them casts doubt on whether one knows the meaning of ‘P’—but they also underlie, or even constitute, the norm of truth as it applies to assertions of ‘P’. For in the paradigm case the circumstances in which you are to assert ‘P’ are (only) when P: e.g. you’re to assert ‘It’s raining’ (only) when it’s raining. And to assert ‘P’ when P is to speak truly.

In short, a person couldn’t get as far as knowing that ‘S’ means that P without knowing that they are to say ‘P’ in certain circumstances, where to do so is to speak truly. They might indeed one day decide to break or flout the rules, and maybe with good reason. But they can only make such a decision because they already have a grasp of the rules. And a criterion for such a grasp is, in general terms, one’s being disposed (often enough) to abide by them. One doesn’t in the normal run of things decide to abide by such rules, though one can certainly decide to break them.

We of course need to be careful in talking of rules here. It would be wrong or at least misleading to say that among the rules the grasp of which constitutes linguistic competence are a specific set of rules governing the string of words ‘If you put that food there the dog is likely to eat it before we sit down to lunch’. Perhaps we should speak of rules governing words, rather than whole sentences—but can I really grasp any rules for ‘dog’ quite apart from ones requiring what to come out with by way of complete speech acts? These issues, knotty though they are, do not undermine the very idea of linguistic meaning as rule-governed, though they may undermine the aims and claims of micro-reductionist theories of meaning.⁴ And if I am right, the Achilles’ heel in Williams’ account of assertion is precisely his failure to recognise the way in which linguistic understanding presupposes a consciousness of various rules and norms, including what might be called the norm of truth.

Williams does indeed make room for the idea of a norm of truth attached to assertion as such. He does so, roughly, by arguing (a) that people are generally expected to express their beliefs (rather than things they don’t believe), and (b) that assertions are, by definition, appropriate tools for expressing one’s beliefs.⁵ The explanation of (a) relates partly to those needs and desires of humans living in society which Williams sums up with his phrase ‘the State of Nature’⁶ and partly to the nature of belief itself. Given that the ‘aim of belief’ is to have only true beliefs, and belief is not subject to the will, falsehood is a’fatal objection’ to a given belief in the sense that it is impossible to carry on believing what you take (= believe) to be false. The facts about the State of Nature together with the facts about belief give sense to the idea of a norm of truth attaching to the expression of belief. Meanwhile the explanation of (b) lies with Williams’ (quasi-)definition of assertion that I quoted at the start of this chapter.

Thus for Williams ‘You said what was false’ does not as such embody any complaint—it does not impugn what was said. Rather, it conveys information useful to the person addressed: it helps them adjust their belief system appropriately for interacting with the world and with other people. By contrast, ‘You said what you believed to be false’ does (prima facie) embody a complaint, insofar as it is tantamount to ‘You tried to deceive me/someone’. Williams’ account of what it is to lie is consistent with these ideas:

I take a lie to be an assertion, the content of which the speaker believes to be false, which is made with the intention to deceive the hearer with regard to that content. (Williams 2002, 96)

Note the lack of a condition that the assertion be false; for Williams, a true assertion can be a lie.

For one who is in the position of ‘deciding’ what to assert, the governing rule (according to Williams) is not ‘Assert what is true’ but rather ‘Assert what you believe to be true’. In the sphere of action a similar contrast can be made, between the rules ‘Do what is right’ and ‘Do what you believe to be right’. By the phrase ‘governing rule’ I mean to advert to such things as grounds for complaint and grounds for regret, such grounds residing in the breaking of a rule. Note that I cannot aim to follow the first rule in either of these pairs of rules without aiming to follow the second, and vice versa. If I aim to assert what’s true I will be aiming to assert what I think is true—and vice versa.⁷ And if I aim to do what’s right I will be aiming to do what I think is right—and vice versa. Nevertheless there is a difference between following one rule and following the other.⁸

This perhaps comes out more clearly when we consider the pair of rules relating to action. ‘Do what you believe to be right’ is the same beast as ‘Follow your conscience’, and there is indeed a tradition of making this rule out to be the prime rule of ethical conduct. Apart from anything else someone who followed that rule would have no grounds (stemming from the rule) for feeling regret or remorse if he came to believe that what he did was wrong, so long as at the time of the action he didn’t think it was wrong. But it is clear that regret is, typically, the appropriate response to realising you did something wrong. The fact that you only have your present self’s word for it that the act was wrong doesn’t detract from this obvious truth. We also need to point out that we cannot coherently equate ‘doing right’ with ‘doing what you think is right’, unless ‘right’ means something different in those two phrases. From this it seems to follow that ‘I did it in good conscience’ cannot automatically exculpate.

It can even be debated whether having done the wrong thing in good conscience is worse or better than doing it while taking it to be wrong. The latter case is that of akrasia or moral incontinence, which is certainly a species of defect. But was Himmler’s speech to his SS officers at Poznan in 1943 the better for being sincere?⁹ One might think that made it worse.

In using the words ‘right’ and ‘wrong’ I by no means intend to imply that ethical deliberation can adequately be thought of as deploying these, let alone only these, notions. These words seem to me to be a kind of promissory notes; roughly, ‘It would be wrong to do that’ amounts to ‘There are important reasons against doing that which outweigh any reasons for doing it’, and the reasons themselves, if the statement is true, will be more substantial and informative, as: ‘She’s your mother’, ‘You promised not to’, ‘That would be cowardly’, and so on. None of this, I think, affects the argument of the preceding two paragraphs.

But does all this carry over to assertion and truth? If I find out that what I said yesterday was false should I regret saying it, as opposed to regretting that I misled someone or appeared foolish or manifested an early symptom of failing powers? It is true that ‘regret’ might be an odd term to use in this context. Analogously, if I realise that in yesterday’s chess game I inadvertently made an illegal move half way through which neither of us noticed, though it didn’t stop the other fellow from beating me—should I in that case regret my move? After all, my opponent seems happy enough!

Well—but what I can surely recognise is the potential ground for a complaint by my opponent, made in response to that move. ‘Complaint’ isn’t here meant to signify whingeing or getting upset; ‘objection’ would perhaps be better, and the notion in question rides on the sort of stopping and forcing modals which Anscombe argued to be the paradigm expressions of rules: ‘You can’t…’, ‘You have to…’¹⁰ To state such an objection is a move within the game itself,¹¹ and the use of one of these modals obviously has an action-guiding function. To state the objection in the past tense, so to speak, is to employ terms internal to the game but in retrospect—e.g. ‘You couldn’t actually do that, it meant castling through check’, a statement that is not in itself action-guiding but which is parasitic for its sense upon the present tense, action-guiding objection.

All this, I think, goes for finding out that I asserted a falsehood. That I believed what I said yesterday is not the end of the matter. For the fact is I misspoke. I inadvertently broke a constitutive rule of the language-game, and there is the ground for the retrospective complaint or objection: ‘Your assertion was false; making it was an error’. That no one believed (or heard) me is no more relevant to this objection than was my opponent’s beating me at chess to the statement ‘You couldn’t actually make that move’.

Williams discusses the idea that falsehood is as such an objection to a person’s asserting something and the flaw in his critique of this idea is precisely that it fails to notice the special role played by the modals ‘You must’ etc. He writes:

No-one can hold that if A’s assertion is recognized to be false, it follows that he ought not to have made it. He may have made it in good faith, on convincing evidence, and so on. […] So will the objection of falsehood be fatal to an assertion [in the sense that] if a person recognizes that an assertion of his is false, he cannot go on making it? No, since it is a sad truth that he can. Perhaps he ought not to, but that merely reminds us of what we knew already, that falsehood is one kind of objection to assertions, and does nothing to show that it is fatal. (Williams 2002, 68.)

In this passage neither ‘ought’ nor ‘can’ has the sense of that sort of modal expression to which I have been alluding, the sort which can serve to express a rule constitutive of a game or practice. Let us start with ‘ought’. It seems that Williams would interpret the statement ‘He ought not to have made that assertion’ as either (a) criticising a person’s character or rationality, i.e. as imputing some sort of defect to the person, or (b) claiming that there was, all things considered, more reason not to make the assertion than to make it. (Or perhaps both.) As for ‘can’, ‘it is a sad truth that he can’ shows that Williams has in mind only (what might be called) empirical psychological possibility/capacity. It’s very easy to carry on lying: anyone can do it and lots of people do.

It is worth taking a moment to consider Williams’ use of ‘ought’. Relevant here is a sort of ‘paradox’ sometimes alleged to arise in connection with practical normativity, illustrable by the following sort of example: I, a doctor, am summoned to an unconscious person’s side; in a case of this sort the standard procedure is to inject the person with X so as to rouse them (or to prevent terrible injury or whatever); I inject Smith with X; unknown to me, Smith has a very rare intolerance of X, such that to inject Smith with X results in his death. He dies. Question: ought I to have injected Smith with X?

Yes and no. The question is ambiguous. Given what I knew, I had best reason to inject X; but given the whole truth, I had—or there was—best reason not to. So ‘I ought to have injected X’ is true in the sense that not to do so would have impugned either my rationality or my character, but it is false in the sense that the best thing for me, or anyone, to do in that situation was to not inject X.

The two ways Williams explicitly mentions by which to block the inference to ‘He ought not to have made that assertion’, viz. ‘He may have made it in good faith, on convincing evidence…’, rely on the sense of ‘ought’ which concerns a person’s rationality or character—i.e. on the sense of ‘ought’ associated with (a), above. I added (b) since Williams’ phrase ‘and so on’ might be thought to allow in addition a different kind of inference-blocking move, the kind which relies on the ‘objective’ test for what there is best reason to do: the sort of test which yields the conclusion that the doctor ought not to have injected X. After all, many will agree that telling a falsehood can sometimes be objectively justified all things considered, if only in extremis (e.g. to put the Gestapo off the scent). But if you tell me you’ll shoot me dead if I move my king, that doesn’t render false the statement ‘You have to move your king, it’s in check’, even though I now have best reason not to move it. A game-internal modal statement is not in competition with ‘all things considered’ statements of practical reason; rather, it supplies one of the reasons which get weighed up when one has to ‘consider all things’.

Thus Williams’ statement ‘No-one can hold that if A’s assertion is recognized to be false, it follows that he ought not to have made it’ is only true on certain readings of ‘ought’. If ‘ought not to have made it’ employs the past tense form of a stopping modal, then what Williams claims doesn’t follow does in fact follow, if I am right that the norm of truth is internal to the practice of assertion in the way in which the rules of chess are internal to the game of chess.

What then is the practice of assertion, that the norm of truth should be internal to it? It is not sufficient for asserting that P that you utter ‘P’ where ‘P’ is a sentence capable of being true or false. For you could be telling a story, or a joke, or could be acting in a play, or reporting another’s speech. It is thus very natural to characterise ‘proper’ assertion by saying that assertions express, or purport to express, the beliefs of those who make them, in the way Williams does. The question is: does saying this much about assertions really tell us what they are?

Let’s say I believe that there is vegetable life outside this galaxy. How shall I convey or express my belief—using my vocal cords, say? I could make any number of noises. But in fact I have no hesitation in confidently emitting the noises I do emit. Is this confidence the (‘epistemic’) confidence that those noises are the right ones to make?—that they will indeed succeed in expressing my belief? Surely not; for how would I answer the question ‘How do you know those noises are the appropriate ones to express whatever it is you believe?’ Do I have any inductive evidence for thinking that these noises are the right ones? Maybe I’ve never heard the noises until this moment of my own uttering them. ‘Ah, but you know from experience that when you find yourself uttering noises in a certain way they do express your belief…’ So the utterance itself becomes an involuntary action which I observe taking place. And there is the question what way ‘a certain way’ is: an assertoric sort of way, perhaps—as opposed e.g. to a story-telling sort of way. (Something to do with tone of voice.) In reality, if we do say that I am confident these noises I have uttered express the belief which I wish them to express, that is because I have a ready answer to ‘And which belief of yours is it that you wish to express by means of these noises?’ My answer takes the form of repeating the noises I made. Not only is the answer a ready one, it guarantees its own correctness. The appropriate noises to produce (among English speakers) in order to express the belief that there’s vegetable life outside this galaxy are: ‘there’s vegetable life outside this galaxy’. (Moreover, your own production or imagining of certain noises when presented with the marks on paper/screen with which you’ve just been presented enjoys the same sort of confidence: not epistemic confidence (‘credence’) but confidence in action. NB here imagining is a kind of doing.)

These remarks are intended to make vivid how close is the internal connection between a belief and the assertion of it. Despite his conceding that very often the criterion for whether one believes that P is simply whether one’s prepared to assert ‘P’, Williams I think fails to do justice to this closeness. The sense in which an assertion is an ‘appropriate tool’ for expressing a belief appears in Williams’ account as akin to the sense in which a hammer is an appropriate tool for fixing a nail in a wall. (And can’t we define ‘hammer’ in terms of banging nails and similar things?) We already know what believing something is—now we can understand ‘assertion’ just as the procedure we language-users often find it handiest to employ when we want to express our beliefs.

To see what’s wrong with this approach consider the difference between an actor’s saying something and a person’s asserting it insincerely. In both cases it’s true to say the speaker doesn’t believe what she says. And no doubt behavioural and other evidence could be adduced in both cases; she doesn’t really think she owns a house in Bermuda, for look!—she never goes to Bermuda, can’t answer questions about the place, and never appears in company with a suntan. The insincere person or liar pretends to believe what she says; but isn’t that true of the actor also? Williams would suggest that the difference lies in whether the speaker intends her audience to believe that she believes what she is saying. But need an insincere person have that intention? She may know that her audience regards her as self-deceived, or as whistling in the dark or as just bullshitting—i.e. as not really believing what she’s saying. She may know this about her audience and not care. Her insincere assertions are assertions for all that.¹² So—a last attempt—is the important thing that the insincere person, unlike the actor, intends her hearers to believe what she’s saying? But you can only rationally intend to make something happen if you think your action makes it likelier to happen; and a person can assert something, sincerely or insincerely, without thinking any such thing as regards her hearers’ beliefs, e.g. if her aim is simply to be a witness to the truth or (when insincere) to pretend to be a witness to the truth. (She might be standing on the soapbox at Speaker’s Corner.)

What then is acting? The answer has to do with the wider context, the ‘surroundings’ as Wittgenstein would say: the kinds of human situation in which the utterances occur, other people’s reactions, the further consequences, and so on. But the ‘situation’, ‘reactions’ and ‘consequences’ cannot be described in brute terms to do only with noises and movements; their descriptions will invoke notions like ‘theatre’, ‘hamming it up’, ‘following the story’, ‘correcting a line’ (which a theatre prompt might do), ‘booing the villain’. Acting in a play is a certain sort of rule-governed practice, and the notions I’ve just mentioned belong within that practice—you couldn’t understand them unless you grasped (enough of) the rules of that practice.

In the same sort of way, I suggest, we should understand ‘proper assertion’ as a kind of rule-governed practice—or perhaps family of practices. The contrast between proper assertion on the one hand and joking, acting and the like on the other is a genuine contrast, in that the items being contrasted belong within a single (if fuzzy) family. Wittgenstein’s term ‘language-game’ might be the aptest name for that family. In this sense of ‘contrast’ one contrasts novels or faces or jewels. As with acting, the fact that someone asserted that P is down to the surroundings and (actual and potential) consequences of their utterance. ‘I was just joking’ can often be rebutted merely by mentioning the context in which the utterance was made—e.g. a court of law, or a diplomatic summit.¹³ It’s not that there’s an ‘expectation’ that people remain serious in such contexts—after all, the odd joke is allowed—but the rules governing what is said and done in these contexts mean that, in the absence of certain defeaters, an utterance of ‘P’ will count as, i.e. will be, an assertion. The response to ‘I was just joking’ won’t be ‘We don’t believe you!’ but simply ‘No you weren’t. You can’t have been.’

Williams does in fact allow that there exist conventions within institutional contexts which, as it were, relate strictly to ‘what was said’ rather than to the beliefs of the speaker, conceived as standing behind what was said. His particular interest is in the distinction, on which in general he wishes to cast doubt, between lying and speaking so as to mislead.

In the British Parliament there is a convention that ministers may not lie when answering questions or making statements, but they can certainly omit, select, give answers that reveal less than the whole relevant truth, and generally give a misleading impression. (Williams 2002, 108.)

It is not quite clear how this statement is compatible with Williams’ own definition of a lie (see p. 3, above). Perhaps he is merely to be taken as reporting how ‘lie’ is used by those who lay down and enforce the convention in question. Be that as it may, he is content to describe the situation in such a way that a minister’s statement will (absent defeaters) count as a proper assertion, and will be governed by a norm of truth, or at least by a norm of non-lying. Again, any excuse along the lines of ‘I was just joking’ will be met with ‘No you weren’t, for (e.g.) you were answering an Opposition question in the House of Commons’. But Williams thinks that what holds in these sorts of institutional contexts doesn’t hold in most of ordinary life. As well as parliament there are, he says, such contexts as courts of law and certain areas of commercial activity; but ‘apart from these cases, not much of life has just this structure of expectations’ (p. 109).

What is wrong with this statement, I think, is the word ‘expectations’. It may of course be true that a Minister of the Crown can be expected not to lie, just as a football referee can be expected not to take large sums of money from the manager of one of the teams. But such an expectation is nothing other than the expectation that a person will abide by the rules. The rules have logical priority. There will probably be an explanation of why people have adopted those rules, in terms of the human point or function of the game/practice/institution; but that does not turn the rules into hypothetical imperatives. The sui generis character of Anscombe’s modals remains intact.

Williams is right to say or imply that ordinary assertions won’t sensibly be regarded as always being made within the context of specifiable institutions such as parliament. But after all, telling a story or a joke isn’t sensibly called an institution either. Nevertheless what differentiates these activities from ‘straight talk’ cannot be captured simply by adducing psychological notions like belief and intention. Intention is or can be relevant insofar as what the speaker says she intended/meant can decide whether she was e.g. joking or not in cases where doubt and ambiguity are real possibilities. In a similar way, the sense of a semantically ambiguous utterance is normally decided by what sense the speaker (sincerely) gives to it. It doesn’t follow that a person’s fiat determines the meanings of all of her utterances—that way lies Humpty-Dumptyism. As to joking vs assertion, ordinary conversational interactions are often such that there is room for doubt and ambiguity as to what ‘game’ the speaker is playing, so there will sometimes be occasion to invoke the speaker’s intentions. (See footnote 13.) But the important point is that here we have different games, different kinds of procedure or practice, whose conduct is rule-governed in the sorts of ways I have sketched. And among the rules constitutive of the practice of assertion is ‘Speak the truth’.

I began this chapter by saying that Williams’ account of assertion gives a prima facie logical priority to belief over assertion. But is this fair? Williams refers in a footnote to Dummett’s statement that assertion is not ‘the expression of an interior act of judgment; judgment, rather, is the interiorization of the external act of assertion’¹⁴—and he remarks: ‘Leaving aside the word “rather,” I do not think that there is an opposition between the two views’ (p. 289). In other words, we can think of assertion as the expression of an interior act of judgment as well as thinking of judgment as the interiorization of the external act of assertion. Ignoring the difficulties faced by any picture of belief as involving interior acts of judgment, I think that if Williams intends here to give the impression that he regards belief and assertion as logically on a par then what he says is belied by what he actually writes, in ways I hope to have spelt out above.

It is Frege who is credited with instigating that line of philosophical thought concerning logic which has been dubbed ‘anti-psychologism’. In Wittgenstein’s work, both early and late, we find a continuing critique of the tendency to regard meaning, truth, reference, assertion, and so on as essentially psychological phenomena. The critique remains forceful even where the notion of the psychological has evolved in ways in which it certainly has evolved over the last century or so (roughly, becoming less Cartesian). In the later Wittgenstein the technique of criticism often consists in reminding us of certain sorts of case which we are liable to forget about when in the grip of our psychologistic generalising:

In what circumstances does one say ‘This appliance is a brake, but it doesn't work’? That surely means: it does not fulfil its purpose. What is it for it to have this purpose? It might also be said: ‘It was the intention that this should work as a brake.’ Whose intention? Here intention as a state of mind entirely disappears from view.

Might it not even be imagined that several people had carried out an intention without any one of them having it? In this way a government may have an intention that no man has. (Zettel sec. 48)

These remarks relate to the concept of intention. And in our discussion of what Williams says about assertion and belief a similar sort of reminder may be in order; consider for example signs, labels and notices. A sign with ‘Toilets’ written on it with an arrow underneath pointing left says something, something we might render as: ‘Toilets are situated in the area to your left’. This is an assertion; it can be true or false, and it is not a joke, poem, or whatever: if it is false there is a ‘fatal objection’ that can be made to it. Somebody painted it perhaps, someone else put it up there, and some person or committee ordered that it be put up there. But it would be ludicrous to take this assertion as the expression of a belief of any of these people, certainly if we are thinking of belief in the way Williams is in his account of assertion. As we might put it: here belief as a state of mind entirely disappears from view. If the sign is false I can perhaps lodge a complaint with some person or body whose job it is to deal with such complaints (e.g. if the result of my trusting the sign was an embarrassing accident), and that person or body might even issue an apology on behalf of the institution—the airport, for instance. These are the consequences attaching to the sign’s having been an assertion, and they are expressible using Anscombe’s modals—as, ‘The sign’s meant to be accurate’, ‘You can complain to Customer Services’, ‘They have to respond to your complaint within a fortnight’. It is to such consequences that we should look to find what it is that makes the sign an assertion, not to the expression of anyone’s belief.

Let me return to those pairs of rules, ‘Assert what is true’ and ‘Assert what you believe to be true’ on the one hand, and ‘Do what is right’ and ‘Do what you believe to be right’ on the other. I want to connect these and similar rule-pairs with the sort of ‘paradox’ I mentioned, exemplified by the case of the doctor who asks himself the question ‘Ought I to have injected Smith with X?’ And before going back to Williams on assertion I wish briefly to indicate how clarity on these matters would be beneficial in thinking about two philosophical topics—first, ‘subjective/objective consequentialism’, and second, Gettier cases. I broach these two topics not merely to indulge in a philosophical interlude but in the hope that what I say will help to bring some of our main themes into sharper focus.

Some consequentialists take their task to be that of giving a criterion of ‘right action’. For a ‘subjective consequentialist’ the right action is the action the agent judges would promote the good on the basis of what they know, while for an ‘objective consequentialist’ the right action is the action that would in fact promote the good. (The terminology of ‘subjective/objective’ and also the phrase ‘promote the good’ are Peter Railton’s.¹⁵) As with our previous rule-pairs, an agent who aims to do what will promote the good thereby aims to do what they think will promote the good, and vice versa. You might then wonder whether it makes sense to claim that one of the two varieties of consequentialism gives the correct criterion for being a ‘right action’.

‘Right’ seems to be connected with ‘ought’ and the sort of question we will be asking ourselves is exemplified by our earlier ‘Ought I to have injected Smith with X?’, alias ‘Was my injecting Smith with X the right action?’ It looks as if Railton is right and there are two conflicting answers available according to whether you’re subjectivist (‘Yes’) or objectivist (‘No’). But as I’ve argued, the correct answer to the question is ‘Yes and no’, for it is simply an ambiguous question. The issue isn’t, as is often made out, whether our moral theory should be guiding people’s actions as opposed to just ‘attributing the property of rightness’ to certain types of action (whatever that means); for the ambiguity I have referred to is still there if our moral theory is admitted to be of the action-guiding kind. The ambiguity relates to two separate questions, whether an agent’s action indicated a defect in the agent, and what there was best reason to do all things considered. Meanwhile if there is any difference between aiming to do what will promote the good and aiming to do what one thinks will promote the good this will come out in such phenomena as that of regret, about which Railton et al. are silent. As I have said, rational regret appears to be more easily accounted for if our guiding rule is ‘Do what is right’ than if it is ‘Do what you think is right’. (Whether ‘promoting the good’ is an appropriate way of cashing out ‘doing what is right’ is another question.)

Turning from the practical to the theoretical, let’s consider Gettier cases.¹⁶ In Gettier’s example of Smith, Jones, the job and the coins, Smith’s belief (B) that the man with ten coins in his pocket will get the job is said to be justified, having been derived by deduction from one true justified belief (‘Jones has ten coins in his pocket’) and one well-grounded but actually false belief (‘Jones will get the job’). (B) also happens to be true.

However, in what sense is Smith’s belief (B) justified? Ought he to believe that the man with ten coins in his pocket will get the job? The question is analogous to our ‘Ought I to have injected Smith with X?’ And the answer is ‘Yes and no’—the question is ambiguous. There’s reason to say yes because in adopting belief (B) Smith shows himself to be rational; and there’s reason to say no because the belief cannot be adequately defended or supported. Rather than talking of ‘defending beliefs’ however let’s talk of ‘defending assertions’, since (a) that’s what defending beliefs is in practice, and (b) the concept of adequacy I’ve invoked belongs properly within the language-game of enquiry and response, more specifically the language-game of asking for reasons for an assertion and (in response) giving some. So: can Smith adequately defend his assertion that the man with ten coins in his pocket will get the job? No. As soon as he says, ‘Jones will get the job…’ we are in a position to reject what he says—for it is false. He cannot now respond, ‘So it’s false! So what? I had it on good authority.’ In the context of this enquiry-and-response game such a move evidently fails.

In the literature the issue of whether Smith’s belief (B) is justified is typically taken as one on which two (conflicting) positions can be adopted, yes or no. But this is not so. The same goes for ‘Is (B) a justified true belief?’ What the consequences are for epistemology I leave to the reader to mull over.

As I put it above, for Williams the rule ‘Assert what you believe to be true’ has a certain priority over the rule ‘Assert what is true’. The priority in question relates to what Williams terms the value of truthfulness, as against the value of truth. Truth has (or is a) value, roughly, because—or in the sense that—(a) human beings have a general need for information-sharing and (b) ‘belief aims at truth’ (see p. 3, above). But ‘the internal role of truth in the belief-assertion-communication system gets us no further at all in delivering the values of truthfulness, once the questions arise to which truthfulness helps to provide the answer’ (p. 85). To what ‘questions’ is Williams referring? The main question he mentions is the question whether to lie, or in other words—as he sees it—the question whether, in a certain situation, to follow the rule ‘Assert what you believe to be true’.

It is at this point that the ethical import of Williams’ position begins to make itself felt. If my decision whether to speak the truth is quite independent of any internal normative connection between assertion and truth, then it looks as if it will be a case of weighing the pros and cons, much as one does when deciding whether to invite a certain cousin to one’s wedding. I will ask myself such questions as whether my interlocutor ‘deserves the truth’ (cf. ch. 5, sec. 6 of Williams 2002), what effects on our relationship my being discovered to have lied would have and how probable such discovery would be, what benefits would follow from my interlocutor’s being misled, whether I’m in danger of subconsciously prioritising my own interests, and so on. On this account the question whether to lie appears as a much more open question than it does on an account of truthfulness which ties the value of that trait to a norm of truth governing assertion as such.

If the rule ‘Assert what is true’ governs the practice of assertion in the way I have sketched, the truth of ‘P’ supplies a default reason for asserting that P rather than not-P. This is a default reason in the sense that I will need sufficient countervailing reasons to break the rule. As always with rule-governed practices, the onus of proof is on the rule-breaker, not on the rule-follower. If I speak falsely to my interlocutor he will have a prima facie complaint or objection to my doing so. Adequate excuses include: that I was in fact joking (NB this as we saw is not something about which I have first-person authority, although in cases of doubt or ambiguity my say-so can decide the matter); that I didn’t realise ‘P’ was false (and my ignorance was not culpable); or that there were sufficiently good reasons for me to deceive him. But if this last can function as an excuse, will there be any difference between the present view of the matter and Williams’? I think so, if only because for Williams the contest between reasons in favour of lying and reasons against doing so starts from a position of equality. This feature of his view is not nullified by his pointing out that in ordinary life most of us most of the time ‘come out with’ the truth pretty unreflectively. The issue is really one of how much work a person will have to do in building a case for the permissibility of lying to another.

No calculus is available for weighing the default reason against countervailing reasons. This is because there is no single currency of the sort imagined by consequentialists—pleasure, utility, or the like. The reason ‘One is meant to assert the truth’ is of a different species from such a reason as ‘If my interlocutor knew the truth he’d commit an outrage’. How then do I weigh the reasons against one another? How do I arrive at a decision? If ‘how’ is asking after a method or algorithm, no answer is available. One considers reasons pro and con, and if one has experience, knowledge, intelligence and practical wisdom one will arrive at a good or reasonable decision.¹⁷ These are facts about practical reasoning which are quite general and they should come as no surprise. Moreover I suspect Williams himself would not wish to dispute them.

‘You’re meant to assert the truth’, with its Anscombean forcing modal, supplies an external reason, in the sense of ‘external’ that is involved in Williams’ famous denial that there are any external reasons for action.¹⁸ The force of ‘You’re meant to assert the truth’ is in other words independent of the agent’s own desires or ‘motivational set’. ‘Force’ doesn’t here mean ‘efficacy’; analogously, a good reason to believe something may lack all efficacy when presented to an intellectually stubborn person. That doesn’t stop it being a good reason. The force of ‘You’re meant to assert the truth’ is ultimately a matter of its being a constitutive rule of a practice which it would be understating things to describe as ‘conducive to human flourishing’; for the practice in question—assertoric language—is one of the most basic elements of our form of life.

Williams’ view that the only reasons a person can have for doing something depend on what his personal desires, projects and values are combines with his denial of any dependency of the value of truthfulness upon an assertoric norm of truth to yield a position according to which people of good will who are blessed with practical wisdom, experience etc. will go in for a certain amount of lying. And the amount of lying in question is greater than would be predicted of such people if Williams’ views are wrong in the ways I have suggested. These remarks of mine can be regarded as conceptual in nature. At the same time they seem to show something about the tenor of Williams’ philosophy.