In this post Clark at Mormon Metaphysics in the context of the Quine’s criticism of analytic/synthetic distinction puts attention on what is often taken as a prototypical analytic statement – “All bachelors are not married”.
I want to put here some thoughts on the relation of theories of concepts and analyticity in relation to this issue, and further connect that to an argument that concept-talk can be avoided in lot of cases, and that instead we can talk about awareness of multiplicity (or possibility of multiplicity), and that such avoiding is natural and explains some intuitions.
Kantian analyticity and Classical Theory of Concepts
The Kantian idea of analytic judgments is connected to the idea which is usually called ‘classical theory of concepts’. For Kant the sentence ‘A is B’ is analytic if the concept B is contained (somehow) in the concept A, and in the classical theory concepts are seen as composed of other concepts as of some kind of list of necessary and sufficient features. It should be pointed though, that in order to have analytic sentences (in Kantian sense), we need just to accept that a part of what constitutes a concept are some necessary features, and not require that sufficient features can be fully specified in terms of simpler concepts.
So, in this Kantian/classical concepts picture the idea is this: if we accept that it is a part of what bachelor means that it is male and not married; obviously by virtue of that meaning “all bachelors are not married” will be true.
It is easier to attack the traditional theory of concept as a whole, then the one that is needed for the Kantian analyticity. That is because the whole theory includes the “sufficient” part, and it is pretty easy to point to cases where the features that were proposed as sufficient are not sufficient after all. For example it is easy to point to cases where X is male and not married, but where still we aren’t inclined to call him bachelor. One example is that X might be too young to be called bachelor (or too old) or that X is the Pope (or any Catholic priest).
To attack analyticity, on other hand, we need not the case where the sufficiency of the features is attacked, but their necessity. Though it might be harder to point to such cases, I think there are… For example, we might be inclined to call a Muslim with one wife – a bachelor (though not a prototypical one), so “not married” would not be necessary feature. Or… is it clear what “male” is? We have cases of individuals whose gender identity doesn’t match their body. Now, there are prototypical males, but what to do with not-so-prototypical ones? Maybe someone will have problems identifying female-to-male transsexual as a male (instead e.g. one might insist that they are separate category), but also the same person might not see an issue in saying that such transsexual in a certain social position – will be a bachelor. Again “being male” wouldn’t be a necessary characteristic for someone to be a bachelor.
There are other examples of the sentences which seem analytical on first look, but where we can find issues with what was supposed to be necessary features. Putnam for example points that “cats are animals” is not analytic as cats can turn out to be robots (or say… group hallucination.).
(Now, while I take it that it is clear that analyticity is hard to be defended of those examples, in my opinion there is analyticity in the propositions of math, logic, metaphysics etc…, in the sense that they are true in virtue of their meaning, and not because of matter of fact. To point to example… we don’t measure sides of right triangles in the world to confirm that Pythagorean theorem is right. It just doesn’t make sense to do that.)
Concepts or Phenomena
So, we don’t have problems with the possibility of cats turning to be automatons (at least we don’t have in principle, I don’t think anybody sane believes that there is actual possibility that they will). But in a previous post (Are hedgehogs small spiny animals?) I pointed that we might have problems with something that doesn’t seem to be as essential to being a cat as being an animal is. I have on mind “possibilities” that e.g. cats turn out to be enormous, or that cats turn out to be spiny. I won’t claim that quantified propositions (“no cats are enormous” and “no cats are spiny”) of those sentences to be analytical, as for sure, a cat might turn to be enormous, and some cats might turn out spiny; but still, saying that cats are enormous, or that cats are spiny (so without quantification) doesn’t sound right to me.
What I want to propose here, is that this points to the fact that usage of common nouns (like “bachelor”, “cat”, “chair”) is grounded in people becoming aware of a phenomenon which includes multiplicity of things which are seen as similar. Now, this is vague, especially the term “similar”, but I think there are good points to be made.
1. It is hard to see why a common noun would appear if there is no multiplicity (at least assumed). If we have just one thing, we don’t need common nouns. We will use a proper name instead. But who needs common noun “cat”, if there was just one cat in history, or common noun “bachelor” if there wasn’t lots of people who come into time for marriage (in certain social conditions)?
2. If we don’t become aware of that multiplicity, nobody will need a word for it. So becoming aware of it is needed. And because of 1, becoming aware not just that there is a thing that has some features, but becoming aware of a multiplicity.
3. If there is no similarities (or some base of grouping) of this multiplicity, again we wouldn’t need a common noun. When I see a cat, and another cat after that, it is the fact that this second one reminds me on the first one, that will produce a thought “there are more of those things (multiplicity)”.
If we avoid “concepts” as some structures, and instead view things in this way, we can see why we might problems with “cats are enormous”, but not with “a cat can be enormous”. It is because the ground of “cat” is in becoming aware of multiplicity that shares the property of not being enormous. So to say the ground of common noun “cat”, is not one cat, but it us becoming aware of there being lot of cats in the world – becoming aware of the phenomenon of cats.
A separate question is where this kind of view falls in the realist/nominalist gap. This view doesn’t say that there is a list of sufficient and necessary features that all cats must share, but while the base of “cat” in this view isn’t some such thing it is still in the reality where there is a phenomenon of animals which appear similar, and the refering is to reality even if one doesn’t accept such ideas as natural-kinds. In such way, it seems to me, one can avoid Platonic forms, while still being able to avoid “concepts” as some kind of nominalist particulars. (Though, I don’t see this contradicting the possibility of talk about some kind of structures in the brain analyzed on some other level, and called “concepts”. But that there is correspondence to some such information structures in the brain and our awareness of something in the world, doesn’t mean that we should confuse our semantic and speak of our thinking in e.g. terms of concept of CAT, when we are really thinking of cats, i.e. of the real phenomenon.)
Let me just say that the difficulties of the classical theory of concepts are not shared by other theories. And this post wasn’t as much argument against any theory of concepts in particular or them in general, as much explaining why I think concepts as mental particulars are not needed (as some mental particulars). For a nice overview of different theories of concepts, you can check this older set of posts at Mixing Memory (1, 2, 3 and 4).
I would like to discuss also the issue of non-existence and historical-intentional account of names in the relation of this kind approach to concepts (or against them), but I guess I will leave it for a next post.