21 October 1996
Thank you for your letter of 18 October. I agree that the questions I set for units 7 — 9 look too tough. Question 1 would be daunting even for a graduate student! Perhaps I should have stuck with the first half of the question. Question 5 suffered from a last minute 'edit'. I had meant to ask for the metaphysical significance of the Theory of Descriptions — in other words the argument linking it to Russell's Theory of Logical Atomism. What philosophers made of it afterwards is altogether another question. My mistake. On the whole, in the course units I am dealing with difficult questions and using Kantian style 'transcendental' arguments to answer them. The intention is not to make things deliberately difficult but simply that this is the only approach in my view that ultimately works.
Now to your response to my letter:
'But what is a sense?' — (a) makes a quite different point to (b) and following. Let's tackle that first. In your sense of 'sense', a grunt or an eye flicker could have a sense. A single word, or sound, or bodily movement can, in the appropriate circumstances, convey a complete proposition with a Fregean sense. Now a theorist approaching language is seeking some kind of articulation that reduces the mass of data to a manageable form. One way to do this is present language use as guided by two kinds of principle: those that fix the meanings of words independently of context, and those which derive from the context of utterance. But as you yourself demonstrate, contextual principles can be translated into a propositional form that does not require that the listener be present when the sentence — or word, or sound, or gesture — is produced. Exploring the subtle rules of context is a rich field for linguistics (and sociology and psychology). That does not diminish the role for an investigation into the core principles of language, a theory of meaning in the spirit of Frege. (Though, as you will see, I have my own doubts about the scope for 'theory' here.)
(b) through to (e) all give examples of the word 'sense' used in a non-Fregean sense. Now it is one of the characteristics of natural language that words are used in more than one sense. This is not a defect, as Frege seemed to think, but a consequence of the all-important role of metaphor in forming new concepts from a given stock of words. By accident, there are occasionally words of the same syntactic and phonetic form with quite independent etymologies. More usually, one can trace the process whereby a word originally used in one sense acquires a different sense.
In the case of the word 'sense', however, an additional point has to be made: we are dealing with a theoretical term, a term of art. The theoretician is not obliged to respect a word's given meaning: just as the chemist John Dalton used the term 'atom' without implying that his theory was derived in any way (as Democritus' and Epicurus' theory was) from a critique of Parmenidean metaphysics. (On second thoughts, that is perhaps not the best example. Perhaps you can think of one where a word in common use is taken up by a theorist and given a new, much narrower, meaning.)
What sort of thing is a sense anyway? — I meant 'thing' in the sense of entity. I was talking colloquially, rather than using 'thing' in the more precise sense of a spatio-temporal particular. There is a metaphysical issue here, which is quite separate from the question which you then put in relation to the incident with the butler, of what gives sense — or rather the senses of particular utterances in particular contexts — their 'substance'. The metaphysical question traces back to the 'argument' between Plato and Aristotle over the ontological status of the Forms. The point I make is simply that a necessary, though by no means sufficient condition for senses to be 'entities' is that there be clear conditions for their individution. Failure to meet this necessary condition rules out senses as candidates for metaphysical 'entities', without the need to get involve in questions of ontology.
'Language is a stock of words etc. which was there before I was born...' — I like this paragraph very much, and totally agree with what you say! But do I have the right to agree, given what I say about 'sense'? I think I do. To reiterate: my approach is based on the understanding that the core principles of semantics — the philosophical and metaphysical issues raised — can be separated off from wider questions of linguisitics, psychology, sociology. Recognising the importance, in principle, of 'Forms of life' in Wittgenstein's sense does not invalidate the approach that I am taking which attempts to separate these questions out.
When we come to discuss the clash between 'realism' and 'anti- realism' it will become apparent that I am far from being sanguine about the notion that statements possess determinate truth values. Apart from vagueness, which has come up for discussion already, there is a special problem which the mathematical Intuitionists such as Brouwer raised of the idea of a proposition's 'possessing' a truth value independently of our capacity to investigate its truth or falsity. (Wittgenstein claimed that he was prompted to return to philosophy 1929 after hearing a lecture by Brouwer. The question mark raised against the determinacy of truth value is one of the main features that separates off the work of the 'early' and 'late' Wittgenstein.)
However, when I said 'we may be unclear about the senses of our words...', I had in mind a more homely point. In using language, we sometimes defer to an understanding that we ourselves do not possess, as when, in response to a query from my wife, I tell her that the car had to go back to the garage because...'.
'The sense of our discourse and our way of life are mutually sustaining but are also perishable.' — I agree with all you say here: and, equally importantly, I would argue, I have the right to agree.
Good luck with your essay!