Kripke’s second argument against descriptivism is much simpler than his first. Schematically it goes as follows:
◊ NN isn't the F.
That means that for any description we may have for, say, the proper name ‘Tony Blair’, it always makes sense to state the above – say, ‘It's possible that Tony Blair wasn't the first leader of New Labour’. We can now say that ‘the first leader of New Labour’ can't give us the meaning of the name ‘Tony Blair’ because we can ask ourselves how do we understand the thought that Tony Blair mightn't have been the first leader of New Labour? This amounts to the very simple possibility that in the mid-1990s Tony Blair might never have become New Labour’s first leader. Clearly this is logically possible.
Let’s just say that the only thing that most people know about Tony Blair is that he was the first leader of New Labour. We can follow this by saying that if any description constituted the meaning of the name ‘Tony Blair', the description ‘the first leader of New Labour’ would be the best option. Let’s summarise this description as ‘the F’. However, let’s suppose that I'm one of those people who only know one thing about Tony Blair: that he was the first leader of New Labour. Thus:
Tony Blair was the F.
It would still be possible to understand the thought that
◊ Tony Blair wasn't the F.
If the only thing I know about Blair is that he was the first leader of New Labour, I can't say that I don’t understand the above because, in actual fact, it's another meaning that fixes the name ‘Tony Blair’. I can't say this because no other description of Tony Blair is known to me. And neither can I say that the thought that Tony Blair wasn't in fact the first leader of New labour doesn't make any sense to me. I may think it false; though I still understand it and understand it clearly. That's because even if I only know the description ‘Tony Blair was the first leader of New Labour’, then clearly the statement
Tony Blair was not the first leader of New Labour.
would make sense. Not only that: someone could, in principle, convince me that it's a fact that Tony Blair wasn't the first leader of New Labour. Again, that's logically possible. For example, say that a political historian tells me that recent research has conclusively shown that it was Gordon Brown who was the first leader of New Labour and that Tony Blair was just a front-man who obeyed Brown’s commands. Because of these ostensibly new and conclusive facts, I could easily come to believe that Blair wasn't, in fact, the actual first leader of New Labour. However, despite this piece of historical revisionism, I may, instead, have never acquired any new historical facts which refute the source of my original description of Tony Blair. Now I couldn't possibly rely on the description ‘the first leader of New Labour’ because it would contradict my new or alternative knowledge of the Prime Minister. Clearly Blair can't both be and not be the first leader of New Labour.
The example above is supposed to show us, Kripke argues, that we can’t - or don’t - actually rely of any descriptive content to fix a proper name. Following on from that, it follows that something other than a description (or descriptions) must come into play when we name someone and also when we later understand and use that name. What all this means, again, is that even if I knew sod all about Tony Blair (if that's possible), I can still effectively use the name ‘Tony Blair’. More relevantly to Kripke’s argument, I successfully refer to the man. This is a Wittgensteinian point about the communal nature of meaning. That a reliance on descriptive content would be an example of private meaning – our private descriptions of Tony Blair which ‘fix the content’ of the proper name ‘Tony Blair’. It would therefore be both a Cartesian and therefore an ‘internalist’ take on the semantics of proper names. Alternatively, if we don’t rely on private or subjective meanings or descriptions, then we must, instead, commit ourselves to semantic conventionalism in that it's the community which fixes the meaning of our names and words and thus makes such things inter-subjective in nature. Thus reference isn't only (or partly?) de re, according to Kripke, because a name (or namer) relies not only on the referent or named object itself, but also on other namers and speakers. This again shows us that descriptivism isn't only a matter of (partly?) de dicto expressions: it's also Cartesian and therefore subjectivist. Kripke, therefore, paints descriptivism as being radically non-Wittgensteinian in nature. The reference-relation between name/namer and named object is socially constituted.
Kripke also elaborates the ostensible fact that descriptions are of little or no relevance to the reference-relation.
For example, we not only may not know anything about a named object (like our previous Tony Blair): we can become adept name-users of names even when we have little contact with any users of the name or the name itself. This is another way of saying that we don’t need the name’s content (if it has any) in such situations of naming scarcity. Whatever we do need, Kripke argues, it's not name-content of any kind.
For instance, I may be on a bus and hear the name ‘Reginald Sniff-Peters’ being used by people on the back of a bus. I have little (or even no) knowledge of Reginald Sniff Peters and have never even heard the name ‘Sniff-Peters’ before. However, when I get off the bus and go home, I can quite easily refer to Sniff-Peters and use the name ‘Reginald Sniff-Peters’ in an ordinary conversation despite my dearth of knowledge of this man and his name. However, by talking to these ‘Sniff-Peters’ novices I could easily bring on board more members of the community of ‘Sniff-Peters’ name-users and even namers.
That is an example of a Kripkean causal chain.
I first heard ‘Sniff-Peters’ spoken on the bus by a group of strangers. Then I passed this name on to yet another group of people who've never heard that name before. And they too could quite easily pass it on and increase the set of users of the name ‘Sniff-Peters’. Of course the people who originally used that name (on the back of the bus) would have had it passed on to them by others. In that case, they too might have been ‘Sniff-Peters’ novices like me.
This causal chain must end, Kripke argued, with what he calls the initial "naming baptism" – i.e. when Sniff-Peters was first named.
Importantly, despite this emphasis on Reginald Sniff-Peters’ name, these new name-users are now talking about him qua person: they aren't simply playing (as it were) with his name.
Kripke, S. (1972/1980) Naming and Necessity, Oxford: BlackwellWittgenstein, L . (1953) Philosophical Investigations, trans. G. E. M. Anscombe, Oxford: Blackwell