To: Malcolm P.
From: Geoffrey Klempner
Subject: Wittgenstein on sensation 'S'
Date: 5 February 2003 12:03
Thank you for your e-mail of 26 January with your first essay for the Philosophy of Language program, on the question of 'S'.
This is an excellent first essay, a real treat for me.
It is easiest for me to respond section by section. Just a warning: I don't think that the 'S' that you are talking about is the 'S' that Wittgenstein intended to attack - but more about that later.
Your eye example can easily be turned round. Developmental biology has much to say why the eye has the particular physical structure that it has. However, what an eye does, its function has been successfully reproduced by human technology. On some planet somewhere in the universe, there are beings who never evolved eyes, who have nevertheless managed to construct video cameras which reproduce frequencies of (what are to us) 'visible' electromagnetic radiation as two dimensional maps of sounds, or smells or in some other way.
Similarly, in the latter part of the 19th century Gottlob Frege invented a language -- first-order predicate calculus -- which it could be argued owes nothing to the evolutionary past of human natural language. Some philosophers, notably Quine and Davidson, would claim that first order predicate calculus could, in principle, be the language of all intelligent beings in the universe.
Against this, Chomsky famously argued that natural language incorporates an essentially human grammar, wired in the brain by evolution. (This is on the basis of extensive studies with children. The conclusions from these studies have been a bone of contention amongst rival 'schools' of philosophy of language.)
In raising the question about the 'normativity' of language, it seems that I am looking for something that could, in principle, be called a 'test' for whether a putative 'language' counts as a language by 'our' standards. The 'language' of the bee dance, for example, is a natural phenomenon, a way in which mother nature has arranged for information to be transmitted from one place to another. With human language, on the other hand, one can evaluate language use as 'wrong' or 'right' in a normative sense, which is tied to other concepts such as reason, justification, knowledge. A bee with a parasite infection might dance the 'wrong' dance, but this is not a matter of linguistic understanding.
It looks as though here we are proposing a sharp test which separates to sorts of 'language'. What about the cases in the middle? I don't see any inconsistency here with the thought that the two classes have a fuzzy borderline. Sometimes the question whether a killing is 'murder' or not lies on the borderline. That does not mean that we do not possess a coherent concept of 'murder'.
1) The Wittgenstein of the Tractatus saw the problem of language purely in terms of defining the 'proposition'. It is difficult to evaluate how far the later Wittgenstein came to agree with your claim that 'any theory assuming language exists to form propositions will misfire', because there is an unclarity here about the notion of 'what language exists for' which you need to think about. (As the nature of communication changed, so did what language was 'for'.)
You should look at the work of Paul Grice for an approach which seeks to define linguistic meaning in terms of 'communicative intentions'. For Grice, the main distinction is between making noises or gestures which are aimed simply at producing an effect, and making noises or gestures which are aimed at producing an effect *through the hearer's recognition of the speaker's linguistic intentions*.
2) I like the stuff about 'I'.
3) - 4) Saul Kripke and Hilary Putnam have argued that concepts for 'natural kinds' such as horse or gravity cannot be understood on the model of a nominal description, of the kind one might give of 'unicorn'. This has now become the accepted view in Anglo-American philosophy. However, it seems that Locke was saying something very similar in his 'Essay on Human Understanding' back in the sixteenth century (see J.L. Mackie 'Problems from Locke' OUP).
5) In talking of the 'perfect language for describing a horse' someone might understand you to be alluding to the language of atomic propositions in Wittgenstein's Tractatus. On logical grounds, W. argued, there must exist a translation of statements about a 'horse' into conjunctions and disjunctions of elementary propositions, even though we can never give the translation. This is different from the thought that the horse and its doings might be described by fundamental physics, which is an empirical claim. (The accepted view in the philosophy of science is that facts about horses are non-reductively 'realized' in facts about distributions of fundamental particles. The laws of biology cannot be reduced to laws of physics, but there is ultimately nothing in the universe apart from fundamental particles.)
We will never know all there is to know about horses. When we use the concept 'horse', however, our language is fully 'adequate'. This is because 'horse', like other natural kind terms, is understood as a theoretical term, which points to what horses are 'in reality' (whatever that might ultimately be).
6) What you say about 'S' here would be fine if we were talking about a feeling or sensation with a physical cause. The object of Wittgenstein's attack, however, is the Cartesian idea that 'I know what I feel' irrespective of what story can be told about the physical world. Hence for the term 'S' understood as a term in a 'logically' private language, 'a definition of the sign cannot be formulated'. (In this context, "'Giddy' refers to what you feel when you have too many goes on a roundabout" is a perfectly sound definition.)
7) - 8) Wittgenstein would agree with all you say about the nature of emotion.
There is more to say here about the nature of 'subjective objects'. You might find it useful to look at my essay, 'Truth and Subjective Knowledge':
All the best,