Wednesday, November 9, 2011

Analysis of 'Sherlock Holmes does not exist'

To: Andreas N.
From: Geoffrey Klempner
Subject: Essay on Existence
Date: 15 November 2005 12:03

Dear Andreas,

Thank you for your email of 6 November, with your University of London essay in response to the question, 'Can we intelligibly claim that Sherlock Holmes does not exist?'

Let me start by saying that I am perfectly happy with recognizing two senses of exists, let's call them exists and eggsists. 'Exists' is a second-order predicate, while 'eggsists' is the predicate which applies to everything (or, to everything within a specified domain).

Arguments over whether or not 'existence really is a predicate' presuppose that there is a single, correct answer to the question what is the best way to formalize natural language. There are plenty of reasons for thinking that there is no 'best' way. However, since first-order predicate calculus has shown itself to be rather useful in analysing statements involving generality - to say the least - we are stuck with the second-order 'exists' and the only question is whether it would serve any useful purpose to also recognize the first-order predicate 'eggsists'. I am happy about this.

We are agreed that there is no incoherence in using the predicate 'eggsists', although it does have the seemingly paradoxical consequence that it is not possible to state that Sherlock Holmes does not eggsist.

One of the serious objections to existence as a first-order predicate, which is not fully articulated in Russell's rather obscure remark about a 'robust sense of reality' is the risk of falling into the fallacy of thinking that 'things' can acquire or lose the predicate 'eggsists'. Suppose I said, 'My maroon Ford Capri no longer eggsists.' That would imply that one and the same object x, eggsists at t but does not eggsist at t+1. This is absurd. To eggsist is the one predicate that an object - whether physical, mental, fictional, mythical or whatever - can never lose, *by definition*.

My maroon Ford Capri, which long ago went to the breakers yard, eggsists timelessly as an object of reference. However, if we consider the class of physical objects that 'eggsist now' (pending a suitable analysis), then my maroon Ford Capri is not a member of that class.

The Strawson/ Evans approach does not fall into the fallacy I described because we are talking about an object identified as a member of a given domain. Suppose that I had a dream about owning a maroon Ford Capri. Then one day my dream comes true. The physical car standing outside my house is not the 'dream car' which went through the mysterious metaphysical process of 'becoming physical'. There were, are and always will be two 'maroon Ford Capris', the figment of my dreams and the car outside my house.

Your dispositional account, if I understand it correctly, doesn't fall into this fallacy, but it does skirt very close. What saves the dispositional account is that you are clearly talking about the second-level predicate 'exists', not the first-level predicate 'eggsists'.

Here is where you identify the crucial problem:
Evan’s proposal however, runs into difficulties when the account is not intentionally fictional, as for example in the case of unobservable entities relating to the scientific, the religious or the supernatural. Although we can connive with Conan Doyle’s world of fiction and quite intelligibly refer to Holmes when we say that Holmes does not really exist in the physical world, can we really unwittingly connive, as has been suggested by some, with the proponents of things like phlogiston and ether, UFOs, magic and ESP without knowing in advance whether or not such things truly exist in the physical world? And if we can how much does that help us differentiate between fantasy and reality, which is surely our primary concern?'

Let me introduce you to my invisible friend, Bertie the 6 foot rabbit. No-one else can see Bertie but me. Bertie follows me around, but sometimes disappears for days or weeks on end then reappears without warning. That's all I have to say about Bertie.

Does Bertie eggsist?

We wanted to recognize fictional characters, creatures of myth and so on. Why not Bertie? The problem is - and this is brought out in your account of the role of 'theory' in a dispositional theory of existence - that we have not yet found a suitable 'domain' for Bertie. All I have given you is a couple of sentences of description. To create a 'Bertie' who eggsists we need to create a world - of fiction or myth, or (most probably) paranoid delusion. A 'world' in this sense is characterized by a 'theory', by the things that hold true of that world and the entities which populate it.

One of the main problems with things like UFOs or ESP is that, as in the case of my friend Bertie, we do not yet know what we are talking about. The incoherent fragments of description do not cohere to form a theory. We have no idea of what kind of 'world' it would be if ESP was true, because, just for a start, we have not been informed how ESP is supposed to work, the laws which operate in that world and so on. So it is much less obvious that we would want to say that some object X 'eggsists' (is a member of domain M + N) but does not 'really eggsist' (is not a member of domain M). To create a 'domain' you need a systematic theory.

It seems to me a lot more intuitive to use the second-order analysis for these cases. Someone has put forward a description, and nothing in the domain satisfies that description. End of story.

Your essay clearly exhibited a good grasp of the issues around the question of existence as a second-order or first-order notion. I have to say that I did struggle to understand the point of your 'dispositional' account (you can tell me if I've missed anything).

If you were trying to explain how an object (like my dream maroon Ford Capri) can 'come into' existence (or, rather, eggsistence) then my response would be to say, 'It can't work'. One and the same abstract structure - e.g. the story of Sherlock Holmes - can be realized in different ways (in a novel, in my dreams etc.) but in each different way it is a different *realized structure*, a different theory, and the objects which form its domain are correspondingly different.

All the best,

Geoffrey