and Meaning Communication
Guy Longworth [e-mail]
Meaning and Communication
Reading for this time:
P. Grice, "Utterer's Meaning and Intentions" chapter 5 in his Studies in the Way of Words. S. Neale, "Paul Grice and the Philosophy of Language" (1992), 4-6, pp. 541-556 only, Linguistics and Philosophy vol. 15.
Recap
Grice's project can be viewed as an attempt to give an account of word and sentence meaning in two steps: (Step 1) An account of speaker's meaning (what speakers mean on occasions) by appeal to their communicative intentions. (Step 2) An account of (timeless) word and sentence meaning by appeal to speaker's meaning.
Recap
Is there reason to think that this a worthwhile project?
Recap
Motivation for project 1. It seems implausible that word and sentence meaning is a primitive feature of the world. In particular, it seems implausible that, even if there were things a bit like words and sentences in the absence of speakers, those word-like and sentence-like things would have meaning.
1.
Recap
Motivation for project 3. Suppose that words and sentences could have meaning independently of speakers. Now it is difficult to see why speakers couldn't use those words and sentences in ways that have nothing to do with the meanings they have independently of speakers.
Recap
Motivation for project 3. In that case, it is plausible that we would need to appeal to word and sentence meanings that accord with the use made of those words and sentences by speakers.
Recap
Motivation for project 3. Put another way: even if all possible words and sentences with all their meanings (all possible languages) were part of the `fabric of reality', still two questions would arise: Q1. Which of those languages are our languages? Q2. What is it about speakers that determines the answer to Q1.?
Recap
Motivation for project 3. Hence, even if there could be meaning independently of speakers, still the meanings that we care most about would be those that accord with speakers use of language. So it seems that language use must play a role in an account of word and sentence meaning.
Recap
Motivation for project So, we have some motivation for expecting that a satisfactory account of word and sentence meaning will make reference to something about speakers. Plausibly, this will be something about speakers' psychologies: their beliefs, desires, intentions, and so forth.
Recap
Last time, we began looking at the first step, Grice's account of speaker's meaning, an aspect of meaningNN.
Recap
Last time, we began looking at the first step, Grice's account of speaker's meaning, an aspect of meaningNN. Grice begins by marking an intuitive distinction between natural meaning and non-natural meaning (meaningNN).
Recap
Roughly, this is the distinction between natural meaning cases like (1), where meaning is factive, and non-natural meaning cases like (2) where meaning is nonfactive: (1) Those dark clouds mean that it will rain. (1) Kim meant, by uttering "It will rain", that it will rain.
Recap
Intuitively, in cases of both natural and non-natural meaning, one feature of the world can serve to indicate to us some other feature of the world. E.g. clouds can indicate rain. E.g. Kim saying `It's raining' can indicate that it's raining.
Recap
The immediate intuitive difference is that in one case, but not the other, the indicator is under the intentional control of people--speakers. It is that feature that appears to allow for the possibility of `false indication'. Since people can decide whether or not to produce an indicator, they can produce the indicator whether or not what it indicates is in place.
Recap
This difference appears to underwrite Grice's view that we cannot account for non-natural meaning just by appeal to causal interaction with the beliefs of speakers and their audiences, for why the simple Causal Theory is wrong.
Recap
It is plausible, then, that we need an account that appeals to speakers' intentions, perhaps directed towards the beliefs of their audiences. The central question is: What precise form should the account take?
Recap
By considering, and rejecting, simpler accounts, we arrived last time at the following account: S meantNN, by x, that p iff (i) S intended x to induce in A a belief that p, and (ii) S intended A to recognize the intention in (i). (iii) S intended A to form the belief at least partly on the basis of the recognition in (ii).
Recap
S meantNN, by x, that p iff (i) S intended x to induce in A a belief that p, and (ii) S intended A to recognize the intention in (i). (iii) S intended A to form the belief at least partly basis of the recognition in (ii).
on the
Recall that the idea is that intention (iii) serves to link (i)
and (ii) together, to make recognition of intention essential to communication, so ensuring a special form of openness.
Assessment and development
This time, the aim is to assess the account to this point and to see whether it should be developed further (and, if so, how). We want to test how the account fares by considering whether there are cases in which:
(a) The candidate condition holds, but this is not a case of meaningNN. In that case, the candidate is not sufficient for meaningNN. This is a case of meaningNN, but this is not a case in which the candidate condition holds. In that case, the candidate is not necessary for meaningNN.
(b)
Is the condition sufficient?
What we are looking for are cases in which: (i) S intended x to induce in A a belief that p, and (ii) S intended A to recognize the intention in (i). (iii) S intended A to form the belief at least partly on the basis of the recognition in (ii). But in which it's not the case that S meantNN, by x, that p.
Is the condition sufficient?
Consider the following case (I hope it's not familiar to Warwick undergraduates!) (Cases of this type are originally due to Stephen Schiffer and P. F. Strawson). I think that the house you are about to move into is infested with rats. Not wanting to say this outright, I let some rats loose, realising that you are watching me and also that you do not realise that I realise that you are watching me.
Is the condition sufficient?
Here, we have that: (i) (ii) (iii) I intend you to form the belief that there are rats in the house. I intend you to recognise my intention that you form the belief that there are rats in the house. I intend you form the belief that there are rats in the house on the basis of recognising my intention that you for that belief (rather than, e.g., on the basis of seeing the rats, since you know I released them deliberately).
Is the condition sufficient?
So all the candidate conditions are met. Do we have that I meantNN, by releasing the rats, that the house is infested with rats? Plausibly, not.
Is the condition sufficient?
The problem appears to arise due to a failure of openness. My intentions are somewhat sneaky. Although I meet the conditions specified by our candidate account, I fail a more demanding condition.
Is the condition sufficient?
I do not intend you to recognise that I intend you to form your beliefs on the basis of recognising my intention that you come to believe that there are rats in the house. As far as I'm concerned, it is sufficient that you merely recognise my intention that you come to believe that there are rats in the house.
Is the condition sufficient?
How should we respond to the latest difficulty?
Is the condition sufficient?
How should we respond to the latest difficulty? There are a number of options.
Is the condition sufficient?
How should we respond to the latest difficulty? There are a number of options. The most obvious would be to add the intention whose absence appears to be responsible for the present difficulty.
Is the condition sufficient?
We might try the following account:
S meantNN, by x, that p iff (i) S intended x to induce in A a belief that p, and (ii) S intended A to recognize the intention in (i). (iii) S intended A to form the belief at least partly on the basis of the recognition in (ii). (iv) S intended A to recognise the intention in (ii).
Is the condition sufficient?
The addition blocks the `rats' counterexample. If I had released the rats intending that you recognise my intention that you recognise my intention that you form the belief that the house is rat-infested, it is plausible that I by meant, releasing the rats, that the house is rat infested.
Is the condition sufficient?
Suppose, for example, that I encourage you to attend to what I'm doing, and then deliberately release the rats. Here it is plausible that, by releasing the rats, I meant that the house is rat-infested.
Is the condition sufficient?
So, the most obvious revision to the account solves the immediate difficulty. But is it ultimately acceptable?
Is the condition sufficient?
Problems: 1. 2. 3. Epicycles Regress Psychological implausibility
Is the condition sufficient?
Problems: 1. Epicycles The most immediate problem is that, with a little ingenuity, an analogous problem can be raised for the new proposal.
Is the condition sufficient?
A version of Schiffer's version: Alf has a hideous singing voice. Alf wants Ralph to believe that Alf wants him to leave and so sings "Jerusalem" in his hideous voice.
Is the condition sufficient?
Alf: (i) intends Ralph to form the belief that Alf wants him to leave; (ii) intends Ralph to recognise that Alf intends Ralph to believe that Alf wants him to leave; (iii) intends Ralph to form his belief on the basis of his recognising Ralph's intention (i) (iv) intends Ralph to recognise the intention in (ii)
Is the condition sufficient?
So, Alf meets our candidate conditions, (i)(iv). However, Alf intends and expects Ralph to believe that Alf believes that the reason for Ralph believing that Alf wants him to leave is Alf's hideous singing. However, Alf's true intention is that Ralph should form his belief on the basis of recognising Alf's intention that he should believe that Alf wants him to leave.
Is the condition sufficient?
The problem, then, is that Alf's last intention is not intended to be overt. So we have another failure of openness, and this appears to be responsible for the intuition that this is not a case in which Alf means, by singing "Jerusalem" that Ralph should leave.
Is the condition sufficient?
At this point we could add another clause of the same sort, such that S should intend the intention in (iv) to be recognised. But it is plausible that, with sufficient ingenuity, further counterexamples could be produced.
Is the condition sufficient?
The epicycles problem is that we appear to be faced with the prospect of a cycle of fix, followed by counterexamples, followed by fix, followed by counterexample... The fact that the same sort of counterexample can apparently be produced at each stage suggests that we are just making ad hoc fixes to what is really a broken analysis.
Is the condition sufficient?
See Schiffer, Meaning, Oxford: Clarendon Press, 1972: ch.11 for further discussion of this type of counterexample.
Is the condition sufficient?
One response to the epicycles objection is that, perhaps speaker meaning is just more complicated that we first thought. Perhaps the account needs to be quite complicated, but it is on the right lines, and with enough extra intentions filled in, it will be immune to further counterexamples.
Is the condition sufficient?
However, even if we achieve a counterexampleimmune account after the next few epicycles, the account would be rather ugly and unnatural. That might be reason for rethinking, trying to find a more natural account.
Is the condition sufficient?
It might even be reason for thinking that we should view speaker meaning as primitive. After all, unnatural accounts tend to be less illuminating than straightforward accounts.
Is the condition sufficient?
However, to this point, we've been given no reason to think that there will be a stopping point, a point where our account becomes immune to further counterexamples. This brings us to the second problem arising from the Strawson/Schiffer counterexample: Regress.
Is the condition sufficient?
The regress worry is the worry that we will never reach a natural stopping point, that our analysis will have to continue indefinitely adding further and further clauses. If that were the situation, we would appear to lack a proper analysis of speaker meaning.
Is the condition sufficient?
Whether or not the analysis is regressive, so that it can never be completed, still we must face the third problem: psychological implausibility.
Is the condition sufficient?
Here there are two sorts of worries: 1. It's far from obvious that we ordinarily have all the required Gricean intentions when we seek to communicate with others. There might be speakers who are congenitally unable to have suitably rich intentions, e.g. small children.
1.
Is the condition sufficient?
There is room here for philosophy to interact with psychology in attempting to test Grice's account. On the assumption that ordinary speakers should be able to engage in speaker meaning, we might be able to test the account by seeing whether all ordinary speakers either do, or can, meet the conditions spelled out in our account. That would be a partly empirical project, not something to be pursued from the armchair.
Is the condition sufficient?
We've seen s...