Skepticism About Meaning 2

Wittgensteinean Meaning Skepticism

Primary:               Wittgenstein, L. (1974) Philosophical Investigations, Oxford: Blackwell

                                Kripke, S. (1982) Wittgenstein on Rules and Private Language, Oxford: OUP

Secondary:            Miller, A. (1998) Philosophy of Language, Montreal: McGill-Queen’s University Press

                                Kenny, A. (1973) Wittgenstein, London: Harmondsworth

Wittgenstein's Claim

We really have to begin this section by emphasizing that it’s very difficult to say that any philosophical doctrine was or was not held by Wittgenstein. He claimed to take a view of philosophy as an activity that could not aim at the discovery of Truth, but could show why the ideas that we do have are nonsensical. Nevertheless, some of the arguments he uses to show that certain ideas can’t be right seem to suggest that he thought certain other ideas were less wrong. The ideas that we’ll be looking at in this lecture are those that Kripke has extracted from Wittgenstein’s writings as the sorts of things that he might be expected to prefer. It seems to be generally agreed that Wittgenstein did not actually hold the precise views that Kripke attributes to him. For these reasons you may sometimes hear this described as Kripke’s Wittgenstein, or Kripke-Wittgenstein, or Kripkenstein. Anyway the position of Kripkenstein is that:

There is no fact of the matter about what something means, and talk about meanings is not talk about facts but has quite a different purpose.

The Rule-Following Problem

a.             Structure of the Argument

K’s argument for the factlessness of talk about meaning has three parts:

i.                     First, we accept that if there are facts which are going to make sentences true or false then those facts are going to belong to some general category of facts, be it behavioural, conceptual, dispositional, or whatever.

ii.                    Second, it is claimed that if we have unrestricted epistemic access to all facts of the relevant category – i.e. we are in an ideal fact-knowing situation with respect to the important facts – then, if those facts are determinative of truth of falsity of meaning claims, we will be able to determine the truth or falsity of meaning claims.

iii.                  Third, K shows that even if we do have unrestricted epistemic access to all facts of the relevant category – and he tries several likely categories – it is quite impossible to know the truth or falsity of a meaning claim. (Note: it’s not impossible to make a good guess at it, but that ain’t knowing.)

Note that it might look like this is going to be an argument about the epistemology of meaning facts, but it should really be thought of as an argument over what is constitutive of meaning. This will become clearer later, I hope. Anyway, we won’t bother too much with the development of the argument through points one and two. Instead we’ll get to the real meaty stuff as soon as possible. This comes when K. begins his more particular argument for the factlessness of talk about meanings, which he bases on his considerations of what it is to follow a rule.

b.             The Possibility of Quaddition

He takes as his example of rule-following the case of a person who is adding two numbers together – or who thinks that they are adding two numbers. K wonders what facts there were/are that make it the case that in the past he meant addition when he used the plus sign, ‘+.’ Suppose, for example, Bob is asked to add together two numbers that he’s never added together before. Because Bob is less than infinitely old we can be sure that there are such numbers. Let’s suppose that the numbers are 68 and 57. Let’s also suppose that those two numbers are larger than any numbers I’ve ever had to add together in the whole course of my life. (Again, we know there are such numbers, let’s just pretend that they’d be a manageable size.) Doubtless, some of you have already performed the addition and arrived at the answer 125. You are happy to claim that 125 is the correct answer. But what makes you think so? K thinks that it’s possible to be sceptical about that claim.

Now suppose I encounter a bizarre sceptic. This sceptic questions my certainty about my answer, in what I just called the ‘metalinguistic’ sense. Perhaps, he suggests, as I used the term ‘plus’ in the past, the answer should have been ‘5’! Of course the sceptic’s suggestion is obviously insane. My initial response to such a suggestion might be that the challenger should go back to school and learn to add. Let the challenger, however, continue. After all, he says, if I am now so confident that, as I used the symbol ‘+’, my intention was that ’68 + 57’ should turn out to denote 125, this cannot be because I explicitly gave myself instructions that 125 is the result of performing the addition in this particular instance. By hypothesis, I did no such thing. But of course the idea is that, in this new instance, I should apply the very same function or rule that I applied so many times in the past. But who is to say what function this was? In the past I gave myself only a finite number of examples instantiating this function. All, we have supposed, involved numbers smaller than 57. So perhaps in the past I used ‘plus’ and ‘+’ to denote a function which I will call ‘quus’ and symbolize by ‘Q’. It is defined by

                x Q y = x + y if x, y  < 57

                          = 5 otherwise

Who is to say that this is not the function I previously meant by ‘+’

Naturally we find this to be a ridiculous position, but the trick is to discover just why it’s ridiculous. We are inclined to say that there is just a fact of the matter about what’s the correct meaning for ‘+’, and that we know what it is. This is the point where K allows us to pick some realm of facts and to knock ourselves out trying to prove that the facts of that realm – if fully knowable to us – will establish, or could establish in the idealising limit, that the meaning of ‘+’ is what we think that it is. His argument, you’ll recall, is that we’ll never be able to prove any such thing.

c.             Two Types of Facts for Meanings

As a matter of pedagogic convenience K proposes two likely seeming realms of facts, which we can use to try to construct a proof of the correctness of our claim about the meaning of ‘+’. He gives us the realm of facts about our behaviours and the realm of facts about our mental states. We’ll look at them in just a second. It will turn out, however, that we can’t do it and the reason that we can’t do it is the sort of reason that can be applied equally well to the attempt to derive such demonstrations using any realm of facts.

1.                   Behaviour

OK. Consider then behaviour as one sort of fact that K suggests as a possible source of meaning truths. The idea is that all our previous behaviour – including linguistic behaviour (utterances and suchlike) – is epistemically accessible to us, and that given this knowledge we are able to say that one interpretation of the rule amongst all others is the correct interpretation.

But this can’t possibly be right, because we know that all our behaviour relates only to the addition of numbers that are less than 57. That was one of the things we stipulated. And this just means that behaviours and stimuli and so on that are available to us cannot distinguish between addition and quaddition. Any behaviour, stimulus etc., that indicates that 125 is the correct answer for 68 + 57 will also indicate that 5 is the correct answer for 68 Q 57.

                        Interlude

We do rather have to wonder, though, whether we couldn’t claim that what we do when we set out to learn to add is not simply to look at a bunch of examples of things that are claimed to have been ‘added’ and try to extrapolate a correspondence of some kind from them. Isn’t it rather the case that we try to learn a rule; an operational procedure by which a function could be instantiated? And isn’t it the case that when this rule is internalised we are able to simply apply the rule by running the appropriate procedure – or something that is effectively identical to it? If we consider an example of what such an operational instruction is going to look like for addition, I’m sure you’ll agree that it is much more like a plausible story of our learning to follow a rule. Perhaps you remember how you were taught to ad things together as a child. Some parents like to say look here little Bobby, you’ve got 2 apples in this basket and you’ve got 3 apples in that basket. How many apples are there all together? Well if you put all the both apples in this basket into this bucket, and all the apples in this basket into the bucket too, how many are in the bucket? Count them 1, 2, …, 5. So 2 + 3 = 5. Now, Bobby, whenever you’re asked to add two numbers together you follow a procedure that is effectively equivalent to this process of counting the set of objects that is the union of two sets whose cardinalities are the numbers that you were asked to add.

Unfortunately this isn’t going to get us very far. K thinks that the skeptic could simply argue that

True, if ‘count’, as I used the word in the past, referred to the act of counting (and my other past words are correctly interpreted in the standard way), then ‘plus’ must have stood for addition. But I applied ‘count’ like ‘plus’, to only finitely many past cases. Thus the sceptic can question my present interpretation of my past usage of ‘count’ as he did with ‘plus’. In particular, he can claim that by ‘count’ I formerly meant Quount, where to ‘quount’ a heap is to count it in the ordinary sense, unless the heap was formed as the union of two heaps, one of which has 57 or more items, in which case one must automatically give the answer ‘5’.

And similar arguments can be pressed against any attempted clarifications of count that would try to immunise it from this sort of indeterminacy. In fact similar arguments are going to be able to be pressed against any story that requires the interpretation of instructions. So explaining how to interpret items in terms of understanding and following instructions is not going to help at all.

2.                   Mental States

OK. So behaviour won’t work. What about the other option that K proposed, that mental states are the sorts of things that provide the facts required in order to make a normative judgement about which rule is the rule that I mean to follow? Could that work? K. thinks not. He distinguishes at least three cases corresponding to three different sorts of mental item that would be supposed to be doing the work here

i.                     Normal Mental States

We think first of the sorts of mental states that people usually have in mind when they say that meaning is to be explained in terms of mental states. Things like mental images, perhaps. Well, we’ve already seen plenty of reasons not to take this suggestion too seriously, but let’s suppose that there are representational states that are significantly different from images, and let’s wonder whether these are the sorts of states that provide the facts about meanings for language. K says they can’t be, and for reasons that we didn’t bother looking at last time we bumped up against the idea theory. First off, we should note that K’s view of how mental items are described as providing meaning facts for sentences is that a sentence has a meaning because understanding it is associated with the occurrence of the mental item. In response to this (possible straw man) he simply says that ideas are neither necessary nor sufficient for providing meanings.

a.                    They don’t seem to be necessary because we don’t at all think that there are mental items that have to come before the mind at the moment of understanding. That just seems to be a fact that we know about how our minds really do work. (Naturally we’re’ thinking of conscious ideas – we can’t have any data about unconscious ideas, if that even makes sense in this context.) And even if we notice that every understanding of ‘cat’ involves an image of a cat, we wouldn’t be inclined to conclude that we couldn’t understand ‘cat’ without that image being present.

b.                   On the other hand, they aren’t sufficient because we can have all sorts of images in our heads at the time that we’re understanding ‘cat’ without their contributing at all to the understanding of ‘cat’. As I mentioned last week I could be thinking about pork chops all through this lecture. In any case, the mental image here is just another thing that needs to be interpreted. It is not special because it’s inside your head. It is just as indeterminate as the picture that you draw (supposedly) of a cat.

ii.                   Special Mental States

K also proposes that perhaps the sort of mental item that accounts for meaning is not like any other sort of mental item that we’re familiar with, but is of a special kind. Of course the only thing that we know about this new mental type is that it provides meanings. We’d have to do a bit better than that if we’re going to make it more than a simple piece of ad hockery. And there’d still be a problem with explaining how any such presently existing mental item could regulate the use of words and meanings in the future.

iii.                 Dispositions

The final possibility that K puts forward is that the mental facts that are to account for truths about meanings are dispositional facts. If, in the past I was disposed to add 68 and 57 when I thought I was trying to find 68+57 (ie. I would have found 125 if I’d done so), then that’s what makes it true that in the past ‘+’ meant addition for me. If, on the other hand, I was disposed to quadd 68 and 57 when I thought I was trying to find 68+57 (ie. I would have found 5 if I’d done so), then that’s what makes it true that in the past ‘+’ really meant quaddition for me.

But this isn’t a satisfactory story because it merely tells what I am disposed to do; it does not at all touch on what it is correct to do. It misses out the normative element altogether. I could very well be disposed to do all sorts of things when asked to find 68 + 57, and few of them would have anything to do with the truth of mathematics. I could be disposed to run away, or to talk about the weather, or to completely ignore the question. But none of these dispositions are going to be the sort of things to tell me what I ought to respond to an arithmetical query. Someone who gives a dispositional account of what I do when I am asked to add 68 and 57 is giving a merely descriptive story. We want a prescriptive story.

The Skeptical Solution

a.                    Pro

The story for ‘plus’ is the story m. m. for any word. K concludes that the evidence so far is that there is no fact of the matter about meaning. Of course, we’ve only looked at a few of the possible stories that could be told, but K suggests that the sorts of difficulties that we have seen with those stories will be repeated in different forms for any other story you might like to tell about the facts that are going to be taken to lie at the heart of meaning. Put in this way the argument at this point is a hostage to the future, because there has been no demonstration that all possible fact-giving stories about meaning are susceptible to the style of critique offered here, and thus there is always the theoretical possibility that some future story will turn out to be successful in explaining the facts about meaning.

Let’s suppose that this turns out to be impossible. What does that leave us with? It leaves us with what Kripke call the sceptical solution, which comes in two parts.

i.                     There are no truth-determining facts about meaning.

ii.                    We can talk about things usefully even when those things don’t have truth-determining facts.

The first part of this seems obvious enough, but what can possibly be the point of the second part of the Kripkensteinian thesis. Kripke makes a general claim that in the Philosophical Investigations Wittgenstein is interested in establishing that sentences can be justified or properly asserted, and that these sorts of judgements about sentences have a certain role in our linguistic lives.

I’m not going to talk any more about this here, but I may come back to it in a later lecture. At the moment I think it’d be better if we looked at a criticism of the sceptical solution. Not the only criticism, but an interesting one.

b.                    Con

An objection to the whole sceptical argument by Kripkenstein is mounted by Zalabardo, inspired perhaps by a naïve objection that we can select between alternative possible interpretations on the grounds that we should prefer the simplest interpretation. Thus in the example of addition/quaddition for ‘+’ we’d prefer addition over quaddition because the instructions that are required to describe the rule for addition is somehow ‘simpler’ than the instructions that are required to describe the rule for quaddition.

Now this doesn’t work as an objection to the K position because – quite apart from the problems with determining what’s going to count as evidence for simplicity – the objection seems to misunderstand the nature of the skeptical conclusion. The sceptical position is not that there are determinable hypotheses that can be compared, but that those hypotheses don’t have a determinate meaning.

[S]implicity considerations can help us to decide between competing hypotheses, but they can never tell us what those hypotheses are. If we do not understand what two hypotheses state, what does it mean to say that one is ‘more probable’ because it is ‘simpler’? If two competing hypotheses are not genuine hypotheses, not assertions of genuine matters of fact, no ‘simplicity’ considerations will make them so.

But Zalabardo thinks that for very much the same sort of reason the positive part of the sceptical solution also fails. That solution seems to assume that the difficulty concerns an indeterminacy about what rule is to be followed. But that is not what the argument shows. The argument shows – or is supposed to show – that there is a difficulty in the notion of rule itself. If the skeptical solution’s positive doctrine is correct in claiming that a choice can be made between hypothetical interpretations (of rules) on the grounds of the community’s behaviour with respect to assertions and acceptance of sentences, then this is just a standard solution to the rule-following problem. It constitutes a factual basis by which we can talk about meanings.  And if that were the case, then it would be necessary to determine whether this story of meaning is immune to the skeptical arguments that were made against the other meaning theories.