To: James D.
From: Geoffrey Klempner
Subject: Russell's theory of descriptions
Date: 21 March 2002 10:40

Dear Jim,

Thank you for your e-mail of 22 February, with your third essay for the Philosophy of Language program, on Russell's theory of descriptions.

I am please that you have gone to the original text ('On Denoting') to research this essay. Let me first pick out the salient points which you have noted in your essay:

1. Definite descriptions like 'the present King of England' are 'denoting phrases' which contrast with indefinite descriptions such as 'a man' or 'all men'.

2. Definite descriptions play a key role in accounting for the distinction between knowledge by acquaintance and knowledge by description.

3. The context principle: A phrase only has meaning in the context of the statement in which it is embedded. I like the way you put this: 'phrases in themselves have no meanings they are merely cogs in a wheel'. (The picture that comes to mind is a gear mechanism, where several cogs of different sizes are working together.)

4. The meaning of ordinary proper names are given by a set of associated descriptions, although no single description is an adequate substitute for the name.

5. Surface structure is not always a reliable guide to underlying logical form: 'the language structure we use sometimes hides beneath its exterior the real, if somewhat oblique meaning of a name or statement.'

One glaring omission from your account, however, is the *analysis* which Russell offers of definite descriptions, and the *motivation* for that analysis.

Very briefly, if we take a statement like, 'The present King of France is bald', we find that both the statement and its negation, 'the present King of France is not bald' are false. This violates the principle of non-contradiction. Russell's explanation is that the correct account of the logical form of the statement is,

'There is one and only one x such that x is a present King of France, AND x is bald.'

In other words, two claims are being made by the speaker who utters the statement: (a) there is a unique individual x who is the present King of France (b) that individual x is bald.

Now we can see how the original statement and its negation can both be false: namely if it is false that 'there is a unique individual x who is the present King of France'

Note that the occurrence of 'x' in (a) and (b) shows that these two claims are part of a single 'propositional function' as Russell calls it. The explanation of the logical structure of such statements belongs to the theory of quantification which Russell inherited from Frege.

We can see here a ready illustration of the context principle (3. above) which Russell also inherited from Frege.

That is all one needs to say so far as the 'logic' of the 'theory of definite descriptions' goes. However, Russell goes on to apply the theory in his account of knowledge by acquaintance and knowledge by description.

You need to make clear that this account of knowledge involves two distinct ideas. The first idea is that we do, in fact, succeed in using proper names of people, or places, in order to communicate even though we have never been 'acquainted' with them. Here 'acquaintance' is understood in the ordinary, common sense way, as meeting someone and being introduced to them, or visiting a place. One possible explanation of our linguistic competence is that the meaning of a name is given by a set of associated descriptions. For example, when I say 'Tony Blair is in France' the meaning of the proper name 'Tony Blair' is given by, 'the present Prime Minister of Britain', 'the husband of Cheri Blair' etc.

From this relatively common sense view, Russell then makes a strikingly bold metaphysical claim: that we are not, in fact, 'acquainted' with any of the things we take ourselves to be acquainted with in the common sense way. The only objects we really 'know', the only objects with which our minds actually make contact, are our own private sense data. So, for example, the meaning of 'the Moon' is not the object which I see in the sky, but a complex of definite descriptions referring to my own private sense data.

It is important to distinguish the different claims Russell makes, because we are interested in whether what he says about definite descriptions is true.

Someone could buy the definite descriptions theory, as an account of the 'logic' of definite descriptions, but reject Russell's 'bundle of descriptions' account of the meaning of proper names.

Someone could buy Russell's 'bundle of descriptions' account of the meaning of proper names, but reject his metaphysical theory according to which the only real 'objects' are our own sense data.

- Despite the 'glaring omission' which I mentioned above, I feel reasonably confident from what you have written, that you have got the gist of Russell's theory. The most important aspect is the idea that the logical form of a statement can be different from its surface structure, Perhaps you can appreciate how that idea is one of the most important single ideas in the history of twentieth century analytical philosophy.

All the best,

Geoffrey