Verificationism

The standard criticism of verificationism is that the verificationist criterion of meaning is self-defeating, fails to work for unprovable mathematical truths, and so on. This criticism shows that the criterion is not necessary for meaningfulness. And it's the necessity that the verificationists cared most about. But is verifiability sufficient for meaningfulness?

Well, first say that p is positively (negatively) verifiable iff there is a possible world where p is strongly (i.e., not just incrementally) empirically confirmed (disconfirmed). Say that x is an rmor iff x is a round mimsy or x is red. Then the claim, p, that George is a rmor admits of both positive and negative verification. In some worlds where George is an observed green triangle, p is strongly disconfirmed (because x is neither red nor round), while in some worlds where George is an observed red triangle, p is strongly confirmed (because x is red).

But of course, p is meaningless because "mimsy" is meaningless.

Maybe, though, the verifiability criterion is counterfactual instead of modal. So, p is verifiable in w iff it is true in w that were requisite observations of the items involved in p made, p would be positively or negatively verified. The above example fails then because in a world where George is a round green peg, it is false that p would be either positively or negatively verified were appropriate observations made.

But just vary the case. Stipulate that an ortum is an observed red thing or an unobserved mimsy. Then in any world that contains an ortum, the claim that George is an ortum is verifiable. For were George observed, he wouldn't be an unobserved mimsy, and one could see whether he's an observed red thing or not. But, again, "George is an ortum" is meaningless because "mimsy" is.

Total nuclear war

Assuming non-theism, a global nuclear war in the 20th century would have been a reasonable prediction to make (and indeed some made it)—but a prediction that was falsified. On the other hand, given theism, such a war would be unlikely. I think there is an argument for theism here. Let me say a little more.

Given naturalistic assumptions about human behavior, there were several approximately independent occasions (whether a short crisis or a longer stretch) during the Cold War at which it was moderately likely—say, probability 1/2—that a global nuclear war would result. Let's say there were four such occasions. Then, given naturalistic assumptions about human behavior, the probability that global nuclear war would occur would be about 15/16. On the other hand, given theism (and especially given Christianity or Judaism—specifically, the promise of the non-repetition of the vast destruction of human life in the Flood), the probability that global nuclear war would occur is much smaller, let's say 1/8. (It would be an order of magnitude harder to come up with a theodicy for a global nuclear war than for such horrors as are already hard to come to come up with a theodicy for.) Well, these two conditional probabilities show that non-occurrence of a global nuclear war incrementally and significantly confirms theism over naturalism.

If initially theism and naturalism are taken to have equal probability 1/2, and we plug the above numbers into Bayes' theorem, we get a posterior probability of 0.93 for theism given the no-global-nuclear-war evidence. If initially theism is taken to have probability 0.1 and naturalism 0.9, we get a posterior probability of 0.61 for theism.

Of course, I've made up the numbers about the probabilities of global nuclear war, though they don't seem crazy to me. A historian would be needed to give us better numbers.

Seeing versus looking

This discussion thread on an amateur astronomy site seems to me to be rather relevant to philosophical accounts of perception.

M33

Saturday night, at the Central Texas Astronomical Society's dark site, M 33 looked to me both like a spiral galaxy and like it was created by God. Therefore, probably, it was a spiral galaxy and created by God. I think this is a perfectly good argument, and it would be a nightmare to try to discredit the "spiral galaxy" part of the argument without discrediting the "created by God" part, or vice versa.

That it might not look to everyone like it was created by God is no more impressive than the fact that it might not look to everyone like a spiral galaxy (it is well known that some observers see more than others, under the same conditions).

Context

It is usual to distinguish between a piece of language and its context. This distinction is, I think, bogus. Whether "flies" in "Fruit flies like butter" is a verb or a noun is determined by context (a "typical" context is still a context, just as a flat delivery is an intonation). Consider some ways one might try to make the distinction:

1. "Context is everything over and beyond the sentence." But this is useless unless we have an account of what "the sentence" as distinguished from the "context" contains.
2. "Context is everything over and beyond what is written or pronounced." But this is false—since a part of the context may be what was written or pronounced earlier. And if one amends to say "what is written or pronounced in the sentence", the the problem in (1) comes back. Besides, this definition would make everything in sign language count as context. There is an infinite variety of ways that linguistic expressions could be realized, and there is no hope of listing them all to make a definition like this work out.
3. "Everything that is linguistically relevant but which the speaker might not know about while responsibly uttering the sentence is context." But a speaker might not know what words she is using. For instance, the discovery that a linguistic sequence can be broken up into words is a genuine empirical discovery, and a speaker might not have it. It is normal, further, for the speaker of a language to fail to know exactly the sounds she is making. (In Russian and Polish, final voiced stops get devoiced; thus, a final written "d" is pronounced "t"; however, native speakers often do not know this, and falsely believe they are pronouncing a "d"!) Moreover, there will be aspects of what is traditionally called "context" which a speaker had better know about, because it determines the parsing of the sentence (e.g., "Fruit flies like butter").
4. "The sentence is what is intentionally produced by the speaker; everything else is context." This is dubious, because one can utter a written sentence simply by cutting out an apposite sentence from a newspaper story and mailing it to someone, we may suppose with a signature appended. In that case, one has asserted whatever was in that sentence, but one has not intentionally produced that sentence, unless "produced" includes exhibiting, taking up, making relevant, etc. But if "produced" is understood in such a wide way, then what are traditionally thought of as parts of the context may well be "produced", in that the time at which one speaks, the way one's face looks, etc. can often indicate which parts of the context are relevant.
5. "The sentence is what is chosen by the speaker's intention; everything else is context." But just a speaker can choose the precise wording, she can choose the precise context, either by producing it or by waiting for it. Of course, sometimes a speaker does not choose the context and may be even unaware of it. But likewise, sometimes a speaker does not choose the words and be even unaware of them (e.g., consider cases of misspeakings or cases where habit rather than choice determines the wording).
6. "Context is that which is not governed by conventional rules of grammar." This is either trivially true, if we take "grammar" to include only the rules that govern things other than context, or else is simply false—for there are rules that involve the interaction of wording and context.

Agent causation

It seems very plausible that if event B is caused by event (perhaps conjunctive or disjunctive) A, where A contains the full set of relevant causal influences that you are praiseworthy or blameworthy for, then x is no more praiseworthy or blameworthy for B than for A. Thus, if you get drunk and this causes you, with no other causal inputs that you are praiseworthy or blameworthy for, insult or kill someone, you are no more blameworthy than you were after you got drunk. If, for instance, you were praiseworthy for getting drunk (e.g., because you falsely believed that the alcohol was a disgusting medication that you had to drink down, and so you did, with great fortitude), you are not blameworthy at all. If your blameworthiness for getting drunk was minor (for instance, because your friends forced you to drink, but you were still somewhat blameworthy for not taking enough care in the choice of your friends), your blameworthiness for the insult or killing is no greater.

The principle here holds whether or not the causation between A and B is deterministic or not. Mere event causation does not increase one's total praiseworthiness or one's total blameworthiness. (Could it decrease it? I don't know. Maybe you think that after total amnesia one is no longer praiseworthy or blameworthy for anything that happened before the amnesia. If you think this—I don't—then event causation can decrease praiseworthiness and blameworthiness.)

Since (unlike God, though it's a hard question why the case of God differs!) we all start out in an initial state we are neither praiseworthy nor blameworthy for, it follows that if event causation is all that happens, we are not praiseworthy or blameworthy for anything, which would be absurd. Hence, there is some other kind of causation, and the only plausible story about it is that it is agent causation.

Must linguistic tokens be physical?

In case you've ever wondered about the question in the title of this post, you can stop wondering: the answer is negative. Here is a simple case. Suppose I want to communicate with you. There exist a million non-physical angels, numbered in some natural way (maybe from least to most wise). Both you and I are empaths of a special sort: we can both feel the angels' emotions and impress emotions on the angels. So now I communicate with you as follows. I take what I want to say to you, and translate it into a binary bit sequence (say, using ASCII). It had better have no more than a million bits. I then take this binary bit string and encode it in the angels' emotions: if the nth bit is one, I make the nth angel happy, and if the nth bit is zero, I sadden the nth angel. You then read off the angels' emotions using your emphathic skills, and so you know what bit string I was communicating to you, and you can decode it. Eventually, the binary encoding and decoding becomes second nature, and we're talking by making angels sadder or happier. The linguistic expression—utterance or inscription—then is something entirely non-physical, a sequence of angelic emotions.

This example shows that there is very little in the way of restrictions on what could count as a linguistic token. Linguistic tokens do not, for instance, need physical realization. What is needed for communication is the ability on the speaker's part to bring about that there is a linguistic token and an ability on the listener's part to identify some of the linguistically relevant features of the token.

It need not even be the case that something positive has to be caused for communication. Maybe you and I could adopt a convention that whenever I am silent in your company, I am asserting: "I am now happy." So, if I'm not happy, I need to chatter all the time.

Sincerity conditions

The following seem plausible necessary conditions on sincerity:

• Assertion: If I sincerely asserted that p, I intended (at least) that I not be asserting something not true.
• Promise: If I sincerely promised you that p, I intended (at least) that I not be promising something I wouldn't do.
• Command: If I sincerely commanded you that p, I intended (at least) that I not be commanding something you wouldn't do.
• Performative declaration: If I sincerely performatively declared that p, I intended (at least) that I not be performatively declaring something that doesn't come off.

(I intend at least p iff I intend q where q is either p or a strengthening of p. Below I shall for simplicity just talk of "intending p" when I mean "intending at least p".)

These may not be the standard sincerity conditions for these illocutionary acts. More standard conditions would be something like this: if I sincerely commanded you that p, I intended that p or I desired that p, etc. However, these more standard sincerity conditions are incorrect. In earlier posts I've shown this for assertions and promises. The examples adapt to commands, questions and performative declarations. For instance, suppose I send you a command by mail. I may not care at all whether you get the command, but intend that if you get it, you fulfill it (imagine a case of an action which is only an exercise in obedience—it is pointless unless you actually get the command). Interestingly, the sincerity condition for commands rules out some interesting cases. It is, on this view, insincere to command something with the intention that the commandee should fail to fulfill the command and thus earn a punishment. (This rules out certain readings of Scripture, assuming that God is always sincere.) Likewise, if I name a ship "the Queen Mary", I am being insincere if the ship already has been named something else (what if it's already been named "the Queen Mary"?) and I have no authority to change the name. But I need not intend that the ship should have the name "the Queen Mary". I may have reluctantly agreed to try to name it thus, but hope that something will interrupt my naming.

What is striking about the above sincerity conditions is that they all involve truth. Granted, promises are restricted to what I will do and commands to what you will do, but all of these illocutionary acts involve a proposition, and in all of them sincerity requires that I intend not to make the illocutionary act with respect to a false proposition. Curiously, thus, in all these cases, sincerity involves an intention to avoid falsehood. There is thus a deep similarity between asserting, promising, commanding and performatively declaring.

Is this common necessary condition on sincerity also sufficient? No, for if it were, then if p reports a future action of one's own, one could sincerely promise that p under exactly the same conditions under which one could sincerely assert that p. And that isn't so. For instance, I can sincerely promise that I will quit smoking, even though I expect I won't, but I cannot sincerely assert that I will quit smoking when I expect I won't. So the sincerity conditions of some of the above four illocutionary acts must add something to the common condition. I do not know what the appropriate addenda are.

Is what I said above applicable to all illocutionary acts? Well, not directly. Certainly it is not the case that sincerely denying p requires that I intend not to deny something false! However, a surprisingly large number of illocutionary acts can be rephrased so that the above rule should apply. For instance:

• "I deny that p" → "I assert that not p", and this is sincere only if I intend not to be asserting something that isn't true.
• "I congratulate you that p" → "I congratulate you that the good G has befallen you", and this is sincere only if I intend not to be congratulating you on something that isn't true (i.e., in a case where G either isn't good or didn't befall you).
• "I thank you that p" → "I thank you that you provided me with good G", and this is sincere only if I intend not to be thanking you for something that isn't true.
• "I protest that p" → "I protest that you are doing the bad thing B", and this is sincere only if I intend not to be protesting something that isn't true.

In these cases, there is a deep propositional content that differs from the surface propositional content. "I thank you that you gave me a cookie" has a shallow content that you gave me a cookie but for purposes of analysis should be seen as having the deep content that you gave me a cookie which it was good for me to get. (That is why there is a pragmatic contradiction in saying "I thank you that you gave me a cookie that was bad for me to get." And of course "I thank you that 2+2=4" is malformed, unless addressed to Descartes' God.)

If the above moves work, then a large class of illocutionary acts have a common necessary sincerity condition that involves the truth of the proposition forming the deep propositional content of the act. Is this true of all illocutionary acts? I don't know. Is joking or asserting-on-stage an illocutionary act? If so, it would be hard to defend the generality of the claim (though maybe not impossible).

Deontology

A colleague asked me what I would say about a choice between a really minor wrongdoing and an action that has really bad consequences. I had a hard time thinking up an interesting case, because most of the minor wrongdoings I could think of are violations of ceteris paribus (c.p.) moral laws. (Lying is an exception—some cases of lying are only minor wrongdoings, but lying is not merely c.p. wrong. But I didn't want to talk about the case of lying because my colleague and I disagree on the morality of lying.) For instance, promise-keeping (which my colleague suggested as an example) is sometimes only a minor obligation, but either we will say that keeping the promise is only a c.p. duty, or we will say that a promise becomes null and void when fulfilling it would lead to great evils (this is perhaps related to the fact that promises to do something immoral are invalid). And so in neither case could one have a choice between the obligation to keep a minor promise and tolerating or producing some great evil, because given the choice, the promise would not generate an ultima facie obligation (or maybe even a prima facie one, if one takes the null and void view). So I was hard-pressed for a case.

However, there is a quite interesting family of minor moral evils which have no ceterisparibusness to them. If E is morally wrong, then to morosely delectate in E is likewise wrong, and to a degree proportional to the wrongfulness of E. It is clearly vicious to delight in an immorality, and what is vicious is also wrong. Moreover, interestingly, even if E is only c.p. wrong, delectating in a wrongful case of E is not merely c.p. wrong, but wrong simpliciter and ultima facie. For instance, it is typically a minor evil to deliberately cause a minor embarrassment to a friend, and the prohibition against causing such embarrassment only holds c.p. (there are even times when it is one's duty to embarrass a friend). However, to freely mentally delight in an actual or hypothetical unjustified causing of a minor embarrassment (i.e., one in which the c.p. clause is not triggered) to a friend is vicious, and this viciousness is not merely c.p. wrong, even though the causing of embarrassment to one's friend is merely c.p. wrong. But if the embarrassment is minor, and the delectation is not of great intensity or extended over a great amount of time, the amount of wickedness in the delectation will be merely small.

So, this gives us a nice tool for generating examples of small (as small as we like, in fact) wrongdoings that are not merely c.p. wrong. And so we get the slightly paradoxical conclusion that while it would be acceptable to embarrass a friend in a minor way to save a life, it would not be acceptable to delight in unjustifiedly causing an embarrassment to a friend, howsoever minor, even to save a life. (Imagine a mind-reading villain who will kill someone—your friend, if you want to make the case harder—unless you morosely delectate in embarrassing a friend.) I think some of the paradox is only apparent, because it is trivially also wrong to unjustifiedly cause an embarrassment to a friend even to save a life (if it were not wrong to cause it, it would not be an unjustified causing), and so, too, it is wrong to delight in such an unjustified action.

I do not find this counterintuitive. But of course some deontologists may say that it's not wrong to morosely delectate on evils, or that it's only c.p. wrong. However, I do not think the Christian emphasis on purity of heart would allow that.

Beauty: an ontological problem

Suppose, with Aristotelians, that the good is perfection. Thus, the good is always the good of some entity, and it consists in that entity's being in some way perfected—its nature being in some way realized. Now consider this problem: many cases of beauty are not cases of an entity being beautiful. A hole could be beautiful in shape, and the wave-like arrangement of air molecules could be beautiful in sound, yet it does not appear that the right ontology should include holes and wave-like arrangements of air molecules.

Now, it is always good that there be a case of beauty. We've seen that a case of beauty is not always a case of a beautiful entity, so we cannot simply say that this good is a good of that entity which has the beauty. What is it a good of? Not the perceiver—for an unperceived beauty is still good (if this is denied, the following either falls apart or needs to be redone almost completely).

Consider, for instance, a sand mandala. Suppose for simplicity (and contrary to fact) that each grain of sand is an entity, but we do not burden our ontology with bunches of grains. What entity's nature is in some way being realized in the beauty of the mandala? Is it that grains of sand, in addition to their more ordinary ways of being perfected, have in their nature a special way of being perfected, by being beautifully united with other grains? That is not a completely crazy idea, but I think only a theist is likely to think it. It requires a certain pre-established harmony. For just as the grain of sand has a nature that specifies the possibility of its being perfected by being put in a beautiful arrangement, so does a tiny colorful snail—and exactly the same arrangements can be made from tiny snails as from grains of sand, and they will still be beautiful and valuable. Thus, all physical entities have in common natures that are perfected by being put in certain beautiful arrangements. This is only plausible if the natures of things have a common origin, and a common origin to which value matters.

Alternately, we might suppose that the value of the beauty of the mandala is a perfection not of the grains of sand, but of the people who made the mandala. On this view, which is also plausible, the value of beauty is grounded in the nature of the artist. But if this is right, then unless there is a creator, natural beauty—the beauty of sunsets and nebulae—will be in trouble.

So it may well be that the idea that beauty has a value independent of observation requires theism.

More on lying

As per my previous post, the two main accounts of the wrongfulness of lying are that what is wrong is insincerity—saying something that one doesn't believe—and that what is wrong is false-telling. Here is another argument against the insincerity account, and hence in favor of the false-telling account. However, I will end by offering yet another account of the wrong in lying, which has significant attractiveness.

Case 1: You are the manager of a motel and you hate the owner. Consequently, you always put up a "no vacancy" sign when you have a vacancy and a "vacancy" sign when you don't, in order to ensure the motel not only does not get much custom, but gains ill-will among travelers. Plainly, you are lying by putting up the sign, and both the false-telling and insincerity accounts condemn your action. You are wronging not only your employer, but also wronging readers of the sign and not just by inconvenience. The action is made wrong by that which makes lying wrong.

Case 2: You are employed by a motel, and you install a computerized sign, linked to the motel's occupancy computer system. You program the computer system to post "no vacancy" when there is a vacancy and otherwise to post "vacancy." You do this with the same motives as the manager in case 1.

I don't think there is a significant moral difference between cases 1 and 2. But in case 2, we cannot analyze the deceitful sign in terms of insincerity. Granted, when the sign says "vacancy", you don't have a belief that there is a vacancy. But even had you programmed the sign to be correct, it might well be the case that you or anyone else would have a belief that there is a vacancy when the sign says "vacancy"—the computer has the information, and maybe no human does. So the fact taht you don't have the belief that there is a vacancy when the sign says "vacancy" is not the reason you've done wrong. Since in case 1, the action is, inter alia, made wrong by that which makes lying wrong, we have to say the same about case 2.

The false-telling account handles case 2 fairly easily. While it is not quite the case that you're asserting that there is or is not a vacancy, you are the person responsible for the correctness of the communication, the one who is being trusted by the reader. (That "vacancy" is not a sentence in standard English is clearly irrelevant, as it means something like that tThere is a vacancy at this establishment.) The false-telling account can be extended to say that one should not make oneself responsible for the correctness of a false communication. The precise sense of "responsible for its correctness" needs some more clarification—an interesting project—but it is plausible that there is a univocal sense of responsibility for correctness present in cases 1 and 2.

I suppose the best bet of the defender of the insincerity account is to modify the account to be conditional: One should not make oneself responsible for a communication's correctness (say, by causing that communication to be produced) when one does not believe that the communication would in fact be correct.

However, it now occurs to me that there is a third account of the wrong in lying, differing from the insincerity and false-telling accounts. This is the intention account. On this account, one is only permitted to make oneself responsible for a communication's correctness when intending that communication to be correct. What is nice about this account is that it handles the counterexamples to the insincerity account, both from this and the previous post. For instance, when one says a German sentence which one believes to express a truth that one does not believe, one is intending to make a correct communication. Likewise, when one makes a prediction that is only going to be true if one speaks (say, the statement: "I am saying the difficult word 'supercilious'"), one is intending to make a correct communication. And finally in the crooked computer programmer case, one is not intending to make oneself responsible for a correct communication.

In fact, the intention account is superior to the false-telling account and to the conditionally-modified insincerity account in the computer programmer case. Suppose the programmer is not crooked, and does things correctly, but knows that sometimes the computer memory will malfunction and misinform the code that controls the sign as to the hotel's vacancy status. On the false-telling account, the programmer has done wrong when such malfunction happens—she has caused a false communication in a way that makes her responsible for its correctness (computers aren't responsible, so all the responsibility flows back to the programmer). On the conditional insincerity account, the programmer knew that sometimes there would be a malfunction, and so she caused a communication that she believed would be incorrect—not specifically, though, but as part of a set of communications. So both of these accounts seem to make the innocent programmer who knows about the likeliness of screwups a wrongdoer.

One might bite the bullet and say that she's inculpably doing wrong, but the intention account is much neater: While she foresees that she will be correctness-responsible for a communication that is false, she does not intend that. She intends the communications to be correct, and regrets the times when they won't be.

My previous post's analogy with adultery then suggests the following moral rule: A married person is only permitted to have sexual relations when the intention is to have sexual relations with the spouse. This has the interesting, and I believe importantly correct, consequence that it is wrong for a married person to have sexual relations in which the fact that the other person is their spouse is deliberationally irrelevant, and wrong in the way adultery is. It would be equivalent to adultery for a married woman to intend to have sex with the nearest attractive person, even if that nearest attractive person were the spouse (unless "attractive" is understood in a way that makes the spousal status be a part of the attraction). This is closely related to John Paul II's remark that it is wrong to lust after one's spouse.

This may, in turn, suggest an argument against pre-marital sex. If pre-marital sex is permissible, then it seems plausible that the couple who are married could intend to have sex with one another in the same way in which they had intended it when they were not married. In other words, it would be permitted for them to intend to have sex without intending marital sex. But on the above account of adultery, it is required that the married couple intend marital sex, and hence it is very plausible that sex without that intention is wrong simpliciter.

Lying

There are two ways of seeing the morality of lying: The insincerity account and the false-telling approach. The insincerity account is that what is prohibited (at least prima facie) is saying something that one does not believe. The false-telling approach is that what is prohibited is saying something false (or maybe something not true, if one doesn't accept bivalence). The false-telling approach is naive: it is the approach of small children whose theory of mind is not sufficiently developed as well as the approach of uncharitable adults (it is sadly common to hear a politician be accused of lying just because the politician said something false).

For a while I've been suspecting that the naive false-telling approach is actually right, but my main reason was based on analogies with other moral issues. Thus, adultery is defined by a married person's sexual relations with someone other than one's spouse, rather than by sexual relations with someone one does not believe to be one's spouse.

However, my post on "sincere assertion" has given me a much better reason to accept the false-telling approach, namely that the insincerity approach is simply extensionally incorrect, for there are cases where it is not even prima facie wrong to say something one believes to be false.

The cases in that post could, I think, be handled by a messier insincerity account on which what is prohibited is saying S when one does not believe the conditional that were one to say S, one would be saying something true. However, this "conditional insincerity account" is implausible. What made the insincerity account plausible in the first place was seeing assertions as the sort of thing that should be an expression of belief. But the examples in the post make that untenable. The conditional insincerity view is thus not motivated as well as the standard insincerity view was. Moreover, it is far from clear that we typically have such conditional beliefs when we speak sincerity.

The main problem with the false-telling account is that we think that (a) in typical cases (ones that don't exhibit the sort of things involved in the cases in my post) it's wrong to say something one doesn't believe and (b) it's not wrong to say something false if one believes it or at least believes it with good reason.

I think (a) is very easy to handle. It is always wrong to do something that one believes to be wrong—it is a violation of the duty to obey conscience. When (except in the weird cases) one believes p to be false, and one has normal moral beliefs, one also believes that it's wrong to affirm p. In that case, whether p is in fact true or false, to affirm p is to disobey conscience, and hence wrong.

Problem (b) is harder to handle. The simplest approach is to say that, yes, it's always wrong to say something false, but one is not always culpable. This is what one should say about adultery—it's always wrong for a married person to engage in marital relations with someone one is not married to, but if the partner in adultery is believed with good reason to be one's spouse, one is not culpable.

However, there is something odd about the idea that my calculus students continually acted immorally when answering questions on my exams. We do not think mistakes on exams to be immoral. I think this is a bullet that one can bite, and then one should bring in the culpability stuff.

Alternately, one might argument that the illocutionary force of an exam answer is different from the illocutionary force of ordinary assertion, or that exam answers are within the scope of some truth-canceling operator like "I believe that according to the material of the course".

Sincere assertion

This is a point exactly parallel to my last point about promising. It is widely thought that x can only sincerely say that p if x believes that p. That claim is subject to a simple counterexample. An expert entomologist, German speaker and friend of mine gives me a sentence s of German, and he assures me that the proposition p that s expresses (a) is about insects, (b) is true, and (c) is not believed by me. I know no German. I then utter s to some German speakers. In so doing, it seems, I say that p, though I do not believe p. But I am sincere—at least, I violate no requirements of integrity in speech, of which sincerity is one.

Maybe this example does not impress. Perhaps you're worried that by uttering s I do not manage to say that p, because speaker meaning is essential for truthtelling. Or perhaps you modify the sincerity condition to say that one only sincerely utters s if one either believes the proposition expressed by s or believes that s is true. So let's move on to a second counterexample.

The following is sufficient for sincerely saying that p: one knows that were one to say that p, then p would be true. Antecedent belief that p is not required for sincerely saying that p. And consequent belief that p is not required, either, because sincerity in speech is not affected by what happens after one has spoken—and we can imagine that one is killed right after saying p. Here is a fun case. George, whom I trust, has promised me that he'll dance a jig whenever I tell anybody in his presence that he'll dance a jig. So I say to you: "George will dance a jig." I need not have any antecedent belief that George will dance a jig, because unless I succeed in saying the sentence, he probably won't dance it, and I am not sure I'll manage to say the sentence, because I have lately had trouble enunciating the word "jig". Nonetheless, I do not offend against the aspect of the virtue of integrity that has to do with sincerity.

It likewise follows that sincere assertion of p need not be the expression of a belief that p. One might think that perhaps it is the expression of a conditional belief: if I were to make this assertion, then p would hold. But it is very implausible to suppose that our assertions express this belief. Surely when I say that 2+2=4, I am not speaking about speakings.

So, it does not appear that sincere assertion need be the expression of any belief at all. However, a sincere assertion presupposes a belief, though not necessarily a belief that is expressed in the assertion.

The above leaves open the question whether believing that p, while not necessary for sincerely saying that p, might not be sufficient. Here, I do not know. The following case is one to think about. I falsely believe that I cannot utter the word "frog" due to a deep trauma. Let s be the sentence "I cannot utter the word 'frog'" (I count uttering "'frog'" as an uttering of "frog"—make this stipulative). I consider uttering s. I believe I can't succeed. But I can still try. And so I try, and succeed. Was I sincere? I am not sure. I knew that I was trying to say something which, if I succeeded, would be false. However, it is also not obvious to me that I am insincere. It's a tough question.

Promises and intentions

It seems to be standard to say that a sincere promise requires an intention to do the promised action. However, I doubt this is true. Here's the argument:

1. It is sufficient for the sincerity of one's promise to do A that one knows that one will fulfill one's promise by doing A and that one does not intend not to do A or not to fulfill one's promise.
2. It is possible to make a promise to do A while knowing that one will fulfill one's promise by doing A, without intending to do A or to fulfill one's promise or to not do A or to not fulfill one's promise.
3. Therefore, it is possible to make a sincere promise without intending to do what one has promised.

If this is right, then we need an explanation for why it is usually thought that a sincere promise involves, or even expresses (thus Searle), an intention to do the promised action. One explanation is that typically the intention to do the promised action helps to fulfill one's duty to do the promised action, and sincerity requires that one already be doing (within reason) whatever contributes to the fulfillment of the promise.

In premise (1), I deliberately phrased the condition strongly: one knows not just that one will do A but that one will fulfill one's promise by doing A. It is not enough for the sincerity of a promise that one knows one will do A. For instance, I would be insincere if I promised to have breakfast tomorrow if I also knew that an alien in five minutes would wipe all memory of the promise from my mind, even if I nonetheless knew that I would eat breakfast tomorrow.

One way to argue for (1) is that, except perhaps in really exceptional circumstances, it is the virtue of integrity requires that one only make sincere promises. But the virtue of integrity as exercised in the making of a promise is fully satisfied in respect of sincerity when one knows that one will fulfill the promise, and has no contrary intention.

I think (2) is easy to argue for, too. Just suppose that I am the sort of person who knows himself to always keep his promises. I then promise to do something for you, but I don't care at all about the action promised. Perhaps you agreed to do B if and only if I promised to do A. In that case, my intention is to promise to do A, but not to actually do A. Nonetheless, I know that I will fulfill the promise. But foreknowledge is not intention.

Here is another argument. The sincerity of a promise does not require an antecedent intention to do the promised action. At most what I antecedently need is an intention to do the promised action if I succeed in making the promise, and of course to make the promise, I need an intention to make the promise. (If I accidentally say "I promise", I haven't promised.) Again, take a case where I have no desire to do A, but in exchange for my sincere promise to do A tomorrow you offer to do B today. I thus promise to do A with an antecedent intention to do A if and only if I succeed in promising. For instance, I might have no intention to do A should I start coughing horrible mid-promise and be unable to finish promising. For in that case you're not going to do B for me, as far as I know. So even if an intention is needed, all it needs to be is a conditional intention.

Perhaps, though, we can still say that the promise expresses this conditional intention. But why? An honest person might very well always have a background conditional intention to do whatever she promises. Taking for granted the speaker's honesty, there is no need to express the intention to do A should the promise to do A come off. And if the speaker's honesty is not to be taken for granted, the expression of the conditional is pointless, since saying "I am honest" is of little value.

Promises

It is tempting to define promising in a way that centrally involves some condition like "S intends that the utterance of T will place him under an obligation to do A", as Searle does. Of course, there will be other conditions, but this is the one that most clearly distinguishes promises from other illocutory acts.

However, there are many ways that an utterance that expresses that one will do A could place one under an obligation to do A, without the utterance having the illocutory force of a promise. For instance, when I assert: "I will not deny having made this assertion", I have thereby put myself under an obligation not to deny having made that assertion, not by making a promise (I was asserting, not promising, hence my use of "will" in place of "shall"), but simply because it's wrong for me to lie. Searle's other conditions will rule this out as a counterexample, but it is easy to imagine a social institution like promising but whose moral oomph is not that of the promise, but of something else, like obedience to an order.

For instance, Fred might command me that whenever I say: "Pursuant to Fred's command, p", I must bring it about that p. (I think that I can make a case like this fit all the other Searlean conditions.) If Fred is a legitimate authority, then by making an utterance like that, I gain an obligation to bring it about that p, but I do not gain promissory obligation but an obediential one. An easy way to see that the obligation is not promissorily grounded is to note that if I say: "Pursuant to Fred's command, I will take you out to lunch", the obligation is not to my interlocutor, as it would be in the case of a promise, but to Fred. Of course, one might modify the central condition by saying that the utterance will place S under an obligation to the interlocutor. But, nonetheless, I can suppose that I say "Pursuant to Fred's command, I will take you out to lunch" to Fred. In that case, I may owe the obligation to Fred, if Fred is the sort of authority obedience to whom is owed to him (God is that sort of authority), but it is still an obediential and not a promissory obligation.

I wonder whether promising can be defined in a vocabulary that uses normative and even moral vocabulary but that does not distinguish promissory normativity from other normativity.

This is a general worry about "normative" accounts of various phenomena. Consider, for instance, some normative account of assertion, where a central condition for T to be an assertion that p is that T is only permissiible if the speaker is justified in believing that p. Unless one specifies the kind of permissibility at issue, we open ourselves up to weird counterexamples, such as when a superior commands us to write down a list of yes/no questions to which one justifiably believes the answer is affirmative—it is, then, permitted to write down the question whether p only if one is justified in believing p. Here, the normativity is of the wrong sort, and a correct definition of assertion needs to specify what the relevant kind of normativity is. And I don't know that one can do so without saying that we're dealing with assertoric normativity.

(In the end, I think all normativity is moral. But there are different kinds of moral normativity, as there are different kinds of moral reasons.)