On one view of philosophical method, we are concerned with words rather than things, whether we realize it or not. Call this the language-first view of philosophy. Sometimes you see philosophers push back against language-first view, and insist that they are interested in X, rather than the meaning of ‘X’. But what do they mean?

As an illustration, I will consider the metaphysics of personal identity. Derek Parfit seems to reject the language-first view when he is defending the use of thought experiments to reason about what selves are (Reasons and Persons, 1983). He notes that Wittgenstein and Quine would have been sceptical about such procedures, and quotes the latter, who writes: “The method of science fiction has its uses in philosophy, but. . . I wonder whether the limits of the method are properly heeded. To seek what is ‘logically required’ for sameness of person under unprecedented circumstances is to suggest that words have some logical force beyond what our past needs have invested them with.” But Parfit thinks that the fact that we have “strong beliefs” about such imagined cases neutralizes the worry. He states that, in teletransporter cases etc., these strong beliefs are “not about our words, but about ourselves”.

I am struggling to make sense of the distinction that seems to be assumed in such remarks between matters of language and matters of fact, as it is sometimes put. So again, I ask, what does it mean to be interested in (or to investigate) X, rather than in the meaning of ‘X’? Or to be talking about X, rather than talking about the meaning of ‘X’?

On the crudest reading, ‘being interested in X’ could amount to refusing to attempt philosophical disambiguation and precisification. As such, it appears to be an attitude at once stubborn and willfully obtuse – the attitude of the grossly caricatured scientist who lacks conceptual acuity. Some terms may be primitive, but their status as such should only be granted after thorough attempts are made at definition, precisification, disambiguation and so on. It is no good assuming from the outset that everybody knows what you mean by ‘X’, and claiming to be investigating whatever that picks out. You cannot even assume that you know what you mean by ‘X’, in some sense of ‘know’ which may be characterized as knowing that, rather than knowing how (i.e., knowing how to use ‘X’, for instance).

A more charitable variant of the first reading is the following: the cases where we are appropriately interested in x rather than the meaning of ‘x’ are cases in which the meaning of ‘x’ is clear. If we think in terms of concepts, these will be cases where we have a felicitous (if not perfect) analysis of the relevant concept. Or, if this is too strong, for the application conditions of the concept to be relatively uncontroversial (even if we cannot articulate them in terms of necessary and jointly sufficient conditions). (If we think in terms of words, these will be cases where the meaning of the word is uncontentious. But I don’t want to make a big fuss over the distinction). When the meaning of ‘x’ is clear, we can turn our interest to the extension of ‘x’, and to the properties of the things falling within its extension (at least, when we are dealing with referring expressions). For example, we know what chairs are, so we can inquire into the properties of chairs, and where in the world we find chairs (interesting I know). But if we are interested in these questions, then there must be things to learn about chairs that we didn’t already know. After all, it seems to be a necessary (but not sufficient) condition for n being interesting is that there is further information to be learned about n. So, the properties we are looking for must not be the properties that identify chairs as chairs, otherwise we would already know they had those properties. In other words, we must be inquiring into the non-essential properties of chairs. When it comes to extension, things are easier to state: we are just finding out about the world to discover, derivatively, as it were, where there are chairs (i.e., where the conditions required for the concept <chair> to apply obtain). Both of these kinds of inquiry seem potentially interesting. But what is at issue, in many philosophical inquiries, is precisely the extension of a given concept. That is, we are not interested in the accidental properties of various instances of <c>, we are interested in what counts as a <c>. This seems to be an apriori inquiry, or at least, it can be an apriori inquiry (we can think of ways to prosecute the ‘method of cases’ through ‘real-world experiments’ rather than thought experiments which do not seem apriori). Indeed, it seems to be just the sort of apriori inquiry that philosophers very often take to be distinctive of their discipline.

A third reading of this sort of remark is the following. There is a weighty conception of conceptual analysis according to which it is aimed at understanding the structure of reality, rather than the meaning of words. This is one way in which it is sometimes distinguished from mere definition. In giving a conceptual analysis, philosophers are, on this picture, attempting to articulate the necessary and jointly sufficient conditions for something to have some property or be of a certain sort. I fear this does no better in helping me understand the distinction.

I will return to the issue of personal identity to see if I can get in the head of the proponents of this distinction. After all, I think this issue falls within the category of cases which come closest to vindicating the language-first view: those where the object of inquiry really matters to us. When we are thinking about personal identity and persistence, it seems sensible to insist “I don’t care what the word ‘self’ means, I care about what the self is! I don’t care about (or don’t just care about) whether the person that steps out of the teletransporter at t2 is, according to the ordinary meaning, 'myself'. I care whether about whether I will step out of the teletransporter at t1.”

But I remain unconvinced. I think these debates can also be thought of as about what the relevant words mean, or what the concept <self> entails. We can think of what role we want to concept <self> to play. Do we want it to play a social and ethical role in tracking, for instance, who is responsible for what, who deserves what, who has which relationships and commitments to whom, and so on? Do we want it to play the role of identifying continuity in values? There is no role of ‘tracking where the self goes’, even though this might be what we want to say, since clearly the specification of this role involves the concept whose role it defines (vicious circularity; semantically empty).

At this point, my opponent might say the following. In saying that ‘I want to have experiences at t2. I want to exist and hence have experiences at t2’, we are saying that we want to be identical to the person who has those experiences. We want to be the same self as that person. We don’t just like experience. We like to have experiences. We don’t just like there to be experiences out there, in the past, present or future. We want to be subject of those experiences. (‘Boy that’s optimistic!’ you may reply. Yes, it is. The assumption here is that we aren’t horribly depressed and also that we regard our prospects as reasonably good, at least better than death).

But does it make sense to talk about a unique enduring subject of experience? The intuition that this is a sensible thing to say – that there is some fact of the matter about whether we will be identical to some future person – is behind, I think, resistance to the idea that this question is simply about words. There seems to be some deep fact of the matter which determines the truth of the statement: ‘I will/won’t be the subject of experience had by the person at the other end of the teletransporter at t2’. But until the meaning of the concept <self> is specified, I just can’t make sense of this idea. All I see is experience in different places at different times, which may be connected in various ways. I agree with Parfit, then, that there is no deep and mysterious fact about the self. It’s a kind of prejudice, and no doubt an evolutionarily useful one, to think that there is. But I wouldn’t go the next step and say that there is no self, or that it only exists for a moment of experience and then disappears. That is equally nonsensical, since it buys into the same ‘deep’ and mysterious notion of persisting subjecthood and numerical identity of time. I have been rejecting this notion precisely because it is without content (as far as I can tell). I choose not to use the word in this way. The mysterious version of the <self> concept doesn’t refer not because there is nothing in the world that satisfies its conditions, but because its conditions are unspecified! It doesn’t refer in the same way that gibberish like ‘lutrem’ doesn’t refer. But there are candidate concepts of self that do refer, because their conditions can be specified, or at least adumbrated. Our choice is simply to identify which better serve the roles we want the concept of <self> to serve – and perhaps, if there are different and important roles, to distinguish between different concepts of <self>.

What this example illustrates, I think, is that we cannot tease out metaphysical (or naturalistic, for that matter) inquiries and semantic or conceptual inquiries. So I think there is a good sense in which philosophy is about language (or the concepts underlying linguistic expressions).

But I still feel as though I'm missing something. I'm not quite grasping the alternative view of philosophical method is supposed to be.