Against “the norm of assertion”
avatar

Most of the literature on the “norm of assertion” assumes the following two controversial theses:

  1. Speech acts fall into various different natural kinds, one of which is assertion.
  2. Each of these natural kinds of speech act is subject to a unique “central” or “fundamental” norm, which is especially intimately bound up with the nature of that kind of speech act.

Typically, these controversial theses are assumed, rather than defended, in this literature. However, it is far from obvious that these theses are true. Indeed, I am strongly tempted to think that they are false.

1. According to one picture of discourse, understanding an utterance fundamentally involves recognizing the communicative intentions that lie behind the utterance. So every intelligible contribution to a conversation presents itself as being done with a certain intention (in the sense that the utterance allows the audience to see which intention the speaker means them to recognize as lying behind the utterance).

On this picture, the features of an utterance that are most relevant to the question of which type of speech act it belongs to are the intentions that the utterance presents itself as done with.

But there are many different possible intentions that an utterance might present itself as being done with. It is not clear which kind of intention is essential to the speech acts that we call “assertions”.

Certainly, the word ‘assertion’ in English can refer to acts of many different kinds. (E.g. one old meaning of the word was to defend or vindicate a cause against any hostile attack; it is in this sense that Milton announces at the beginning of Paradise Lost that he will “assert eternal Providence/ and justify the ways of God to men”.)

In some contexts, the word ‘assertion’ seems to be used for speech acts that present themselves as done with the intention of expressing a true proposition (by means of the very utterance in question). For example, the answers given by candidates in oral examinations, or by contestants on quiz shows, normally present themselves as done with this sort of intention. An early Christian martyr being interrogated by a Roman magistrate might in this sense “assert” the proposition that Jesus Christ came to save all human souls.

One very special subset of these speech acts are those that present themselves as done with the intention of informing the audience of the proposition that is expressed. To “inform” one’s audience of a proposition p is to bring it about that one’s audience comes to know this proposition p. But clearly not all acts that present themselves as done with the intention of expressing a truth also present themselves as done with the intention of informing one’s audience. The answers given by candidates in oral examinations do not present themselves as done with the intention of informing the examiner; and the Christian martyrs’ declaration of their faith need not present itself as done with the intention (as opposed to a mere hope) of informing the Roman magistrate of the proposition that is thereby expressed.

In other cases, a speaker – such as a trial lawyer or a politician – might make an utterance that presents itself as being done with the intention of winning an argument, but does not clearly present itself as done with the intention of expressing a truth, let alone with the intention of informing the audience.

So there are many different categories of speech acts here; and presumably, different theoretical projects will find it useful to focus on different categories. It seems clear that no word in ordinary English stably picks out just one of these categories; and so the theorist will just have to stipulate a semi-technical sense of the term ‘assertion’ to pick out one of these categories in the way that seems most useful to the theorist’s purposes. But there is surely no good question about what the true nature of “the speech act of assertion” is.

2. Speech acts are acts, and acts are subject to lots of different norms. (E.g. there are moral norms, legal norms, prudential norms, technical norms, etc.) It is radically unclear what is meant by saying that a certain norm is the “central” or “fundamental” norm that applies to acts of a certain type. What is the central or fundamental norm that applies to acts of killing a human being, or to acts of scratching one’s head?

The best that I can do to make sense of the idea of the “central” norm of assertion is the following. In some cases, we identify an act-type by a certain purpose, and we talk about the “correct” or “proper” way to do an act that has that purpose. Thus, there is a correct way to tie a reef knot, or to play Bach’s first French Suite, etc.

If we can identify a kind of speech act by a certain purpose, then we might say that that kind of speech act is done correctly or properly on a given occasion if and only if it successfully accomplishes the relevant purpose on that occasion.

So, e.g. on this approach, one might say:

  • An utterance is the correct performance of the speech act-type whose purpose is to express a truth if and only if the utterance does indeed express a truth.
  • An utterance is the correct performance of the speech act-type whose purpose is to inform the audience if and only if the utterance succeeds in informing the audience.
  • An utterance is the correct performance of the speech act-type whose purpose is to win an argument if and only if the utterance does indeed lead to the speaker’s winning the argument.

Here, however, ‘correct’ effectively just means ‘achieves the relevant purpose’. So it is not clear what the introduction of normative terminology (like ‘correct’) adds to the basic element of this theory, which is just the identification of the relevant purpose.

So, it seems to me, the literature on the norm of assertion is in bad shape until it can make it clearer what exactly (if anything) it is supposed to be about….


Comments

Against “the norm of assertion” — 12 Comments

  1. Ralph, I agree that it’s not entirely clear (1) what kind of speech act ‘assertion’ picks out; or (2) what kind of normative property ‘the norm of assertion’ picks out. But I think there are ways of clarifying these issues in a way that makes sense of the much of the literature. This is how I set things up in a paper I wrote a few years ago now on “The Normative Role of Knowledge” which is linked here:

    http://onlinelibrary.wiley.com/doi/10.1111/j.1468-0068.2010.00787.x/abstract

    (1) Assertion can be defined as the speech act whose function is to express the attitude of belief (i.e. outright belief, rather than partial belief). There may be an element of stipulation involved here, but that needn’t matter – if belief is an important attitude, then presumably the speech act that functions to express belief is an important speech act and one we might want to ask normative questions about.

    (2) Now we ask the following kind of question: what kind of epistemic position does one need with respect to a proposition p in order for one’s assertion that p to have a certain kind of normative status? If assertion is the speech act that functions to express belief, then it’s plausible to suppose that these questions about the norms on assertion reduce to questions about the norms on belief. Agreed, there is no *unique* norm that governs belief or assertion, but then we can ask questions about what it takes for one’s assertion to be justified, correct, and so on, which reduce to questions about what it takes for one’s belief to be justified, correct, and so on.

    I’ll be interested to hear whether you think this helps with the problems you were raising, since I’m drawing heavily on ideas developed in your work, e.g. the distinction between partial vs outright belief and the distinction between norms of justification vs. correctness for belief.

  2. Thanks Declan!

    Well, yes, within the context of the sort of normative theory of intentional mental states that I have defended elsewhere, we certainly can make sense of the idea of the “central norm” that applies to belief, and of the central norm that applies to each of many other types of mental state as well.

    But these norms really do apply to mental states. The only way of applying them to speech acts is by sheer stipulation, and there would then be a real question about whether this stipulation reveals anything interesting or illuminating about speech acts at all.

    What can we learn about the nature of speech acts by studying the various norms that apply to them? Williamson in a strangely uncritical way seems to assume Dummett’s old proposal, that speech acts are like games — and uttering a truth counts as “winning” the assertion game. But I do not believe that this model is compatible with the most plausible view of games (such as the view that Bernard Suits defended in The Grasshopper).

    In general, I’m tempted to think that we get a much better understanding of speech acts if we focus on the idea of the communication of intentions, rather than the various kinds of normative status that speech acts can have.

  3. Hi Ralph, I’m quite sympathetic to your criticisms of the Williamson/Dummett model of assertion. But I do think it can be quite illuminating to think of speech acts in terms of the attitudes they function to express and that there are important descriptive and normative distinctions to be drawn between speech acts that function to express outright belief and speech acts that function to express partial belief.

    Sure, there’s going to be some element of stipulation in deciding whether to use the term ‘assertion’ in a narrow sense that includes only speech acts that express outright belief or a broader sense that includes speech acts that express partial belief. But there’s also an element of stipulation in deciding whether to use the term ‘belief’ in a narrow sense that includes only outright attitudes or a broad sense that includes partial attitudes. The choice of terminology doesn’t matter very much, but it seems to me that we need some kind of terminological stipulation in order to mark theoretically important distinctions that are left muddy in ordinary language.

  4. It might be illuminating to study speech acts in terms of the attitudes that they “function to express”. However, I doubt it. It has proved quite elusive to give a good account of the relevant notion of “expression”. After all, a lying assertion is still an assertion, but it doesn’t express the speaker’s belief; and the intention that speakers usually present themselves as acting on is usually just an intention to speak the truth — this intention does not typically itself focus on the speaker’s beliefs. So, I’m not a big fan of the view that we get an illuminating account of speech acts by focusing on the attitudes that utterances “function to express”.

    Moreover, even if this were a good account of speech acts, it still wouldn’t be true that we could derive any additional illumination by taking the norms that apply to the attitudes, and then saying that these norms can in a sense be used to evaluate the speech acts that “function to express” those attitudes. On the contrary, it seems more illuminating to say that the norms fundamentally apply to the attitudes, and that the nature of these types of speech act is given in non-normative terms, in terms of the attitudes that are involved in those speech acts.

  5. Hi Ralph,

    Thanks for the excellent post. I think I started to have similar worries after reading Thomson’s discussion of Williamson in _Normativity_. It struck me that this might be a decent way of making sense of some of the literature. If we focus on those cases where a speaker has undertaken the responsibility to share some intelligence with an audience (something that doesn’t happen whenever someone asserts that something is so), we can ask whether there’s any epistemic standard in play that determines whether the speaker has met her responsibilities. It seems clear to me that there are epistemic standards in play in these kinds of exchanges. If the intelligence you share with me was provided by monkeys banging away at typewriters, I think you’ve failed in your responsibilities even if what the monkeys typed was true. You had to have some reason to think that the intelligence was good and now the question is whether you can properly share intelligence having undertaken the responsibility to do so when you do not know if the information provided is correct. I imagine some (e.g., Kvanvig) would say that you’ve met your responsibilities provided that your belief that the information provided was correct met the requirements of some orthodox account of justification. Others might have more demanding standards. The idea is, I think, one of Alston’s. I believe (but can’t find the source) that he once said that we can think of epistemic evaluation much in the way that we think about product liability. Some will say that there is a due care requirement and nothing further (of epistemic interest). Others will go further (e.g., just as I think you can exercise due care without meeting your obligations in sending products to market, I think you can exercise due care and fail to meet your obligations if what you tell your audience is false). If we think about this as an issue of the ethics of product liability and responsibility to an audience, does this allow us to bracket difficult questions about the true nature of assertion? (Also, if anyone has any idea where this Alston suggestion comes from (article? book? my imagination?) that would be very much appreciated).

  6. Thanks Clayton!

    As you rightly point out, the range of cases in which “a speaker has undertaken the responsibility to share some intelligence with an audience” are really significantly narrower from what in any common sense of the term would be called “assertions”.

    The norm that you describe in your comment seems clearly to be broadly speaking moral norms: like the norms that we should refrain from foul dealing and free riding, this norm seems to correspond to a moral requirement that we should exercise due cognitive care and attention in playing our part in cooperative social activities.

    If that is right, then the underlying norm not just only fails to include all assertions, but also encompasses a far wider field of social behaviour — including at least all social cooperation that requires some sort of mutual understanding or the like. So yet again, these “norms” or “standards” really have very little to do with assertion as such.

  7. “So yet again, these “norms” or “standards” really have very little to do with assertion as such.”

    Agreed, but I think that might be a good thing (if, that is, there’s still a dimension of epistemic assessment so that we might sensibly argue about whether there’s anything beyond one having to do with due cognitive care).

  8. Ralph: I sympathize (a bit) with some of your complaints here. But I don’t see how the avenues you point us to help with them.

    First, you complain that 1) is a controversial thesis; but then you begin discussion of 1) with a controversial account:

    According to one picture of discourse, understanding an utterance fundamentally involves recognizing the communicative intentions that lie behind the utterance. So every intelligible contribution to a conversation presents itself as being done with a certain intention (in the sense that the utterance allows the audience to see which intention the speaker means them to recognize as lying behind the utterance).

    I’m not sure what to make of this in the case of a lying: one who lies to me has presented herself as having a communicative intention, but (unless I’m wise to it) I surely haven’t recognized the communicative intention that lies behind it; yet I presumably have understood the utterance without difficulty.

    You later say On this picture, the features of an utterance that are most relevant to the question of which type of speech act it belongs to are the intentions that the utterance presents itself as done with. Okay, but now we’re not dealing with what it takes to “understand” an utterance, but with what a speech act presents of itself, regardless of the intentions of the speaker; and this seems just like what Moore, Unger, and others say when they say that asserting/declaring/stating “represents its speaker as knowing.” Moreover it looks as though the knowledge norm advocate (e.g.) could just reply that an assertion is one which standardly presents the speaker as intending to convey a proposition known by the speaker.

    In giving diverse examples of the intentions behind an assertion, you focus on the potentially different intentions that may be present in an assertion; but this doesn’t yet show that there are many different kinds of speech acts. In each of your cases (Milton, martyr, examination candidate, lawyer or politician) the speech act is used to affirm a proposition as true (and is typically presented as one believed)… sometimes this can also involve trying to inform, win an argument, or defend/vindicate a view, etc. But in each case the speech act presents a proposition as true; that would form the core of what you’re after, given that you’re worried about “which kind of intention is essential to the speech acts that we call “assertions”.

    As for your 2), I share some of your worries, though I think Williamson handled them (perhaps not perfectly) in his discussion of the norm being “constitutive,” yet such that other norms (e.g. moral, or prudential) take priority depending on the context.

    Nevertheless, I’d like to quibble with your opening line: “Speech acts are acts, and acts are subject to lots of different norms.” Is the idea supposed to be that because an act type is subject to lots of different norms, it cannot be subject to a central or fundamental one?

    Lying is an act, and it is subject to lots of different evaluative norms. Does that tend to show that lying isn’t fundamentally subject to a moral norm?

  9. Hello, Charles Myro here,

    Well, I agree with the above writer. It seems clear, to me at least, that to assert is to state that something is the case.
    Seems very basic that all statements are such assertions, and that all statements are of the general form–x is a y, and that the purposes for their utterance must vary greatly.
    Whatever the statement , it asserts that something is the state of affairs.
    But the hypothetical is a complicated odd bird then.
    If a speaker speaks hypothetically and yet does not label his utterance hypothetical– does not reveal explicitly his intention to speak hypothetically —can he be said to have made an assertion by his utterance?
    It seems that here perhaps, consideration of his intention is important to whether he is asserting that something is the case or not.
    One could argue that if he was speaking hypothetically then there is no intention that his utterance be an assertion that something is the case. So do we call what he said a statement or pseudo statement or in form only or some such?
    One may argue that every hypothetical statement (e.g., “supposing x is a y”) is a statement or assertion in form only–a saying that something is the case but without the actual assertion behind it. One might say, that the real assertion in the hypothetical statement, the actual state of affairs being described, is that what is being stated is hypothetical. One then may ask if this hypotheticality is the actual assertion even if it is not explicitly expressed in the utterance (as in “speaking hypothetically gentlemen”).
    Presumably the speaker knows if he is speaking hypothetically, but must the hearers know also before the statement can be considered hypothetical? And if the speaker reveals only after he has spoken that his utterance was hypothetical, may his so revealing be considered a separate statement of affairs or part of his previous utterance? After all, if a statement is hypothetical it must carry with it the fact of its hypotheticality (or how may it be called hypothetical?) and so does that fact not make up part of the hypothetical statement itself–and was that fact not being asserted in the utterance?
    In any case, what is the speaker’s intention apparently has a bearing on whether he is making an assertion that something is the case though the form of his utterance is that of such an assertion.
    Are all statements then hypothetical until we know the intention?
    And does someone make an assertion that something is the case when he deliberately prevaricates and knows he does so? Is a lie a pseudo assertion? Or is it a real assertion but with part of the assertion hidden–namely that part that asserts that it is not a real assertion?
    I’m getting dizzy with this, so I’ll stop. Interesting question though.

  10. Hi Ralph,

    I don’t hold any particular brief for the Dummett/Williamson view of assertion myself, but I was wondering if you could say a bit more about why you think it’s mistaken. I think I see why our practice of making assertions (on any plausible construal of it) wouldn’t satisfy Suits’s definition of a game – but it’s not obvious to me that it’s wrong to think of it as something quite like a game in Suits’s sense – i.e. as a rule-governed practice that aims at producing certain results, even though following the rules in question will always not be the most efficient means of achieving these results in every particular case. (The last clause, of course, is why we need to talk about rules and not just purposes, on this view.) Could you say a bit about why you think this picture of assertion (or of the various assertion-like speech acts, once we distinguish them) is mistaken?

    thanks,
    Karl

  11. Hi Ralph,

    On 1:
    I’m tempted to modus tollens your modus ponens. Intuitively, the speech-acts of the examinee, quiz-show contestant, martyr, and trial lawyer, are assertions. To the extent that a theory of speech-acts makes it mysterious what they have in common, to that extent the theory is counter-intuitive.

    On 2:
    Maybe something can be made of the idea that the knowledge norm is ‘constitutive’ of the nature of assertion. But put that to one side. There’s still the question as to whether all assertions can be epistemically evaluated, and if so, whether expressing knowledge is the epistemic standard for assertions to meet. Is all well epistemically speaking with an assertion iff it expresses knowledge? That seems like a legitimate and interesting question (addressed by Jessica Brown and others).
    (This knowledge norm strikes me as compatible with your thought that the normal reason to meet the epistemic standard for assertion is an instance of a general reason to cooperate with others. The reasons to meet (or not meet) the epistemic standard are moral/prudential.)

  12. Ralph, I have another consideration here. Lets consider that Knowledge is not – as implied by you – the constitutive norm of assertion. But it occurs to me that this could pose a problem for our general understanding of the Moore’s paradox. Consider the following proposition:

    1)P,but I don’t know that P.

    It is quite accepted in the literature that (1) can be a known proposition to S, despite the fact that S himself cannot assert it. The standard explanation for this phenomena makes use of the knowledge norm of assertion, so that when S asserts (1) he is, in fact, contradicting himself.

    If knowledge is not the norm of assertion as you are suggesting, then the Moore’s paradox is vanished. In this sense, It would be ok for S to say that ‘P,but I don’t know that P’, since by asserting the first conjunct P, he would be just saying, lets say, a justified belief. Thus, assuming that assertion is equivalent to justified belief, there would be no paradox anymore.

    But this just seems wrong.

Leave a Reply

Your email address will not be published. Required fields are marked *