Joho the Blog
An Entry from the Archives

I've started re-reading Aristotle's Metaphysics (although if "re-reading" implies any serious memory of the first time, then we should remove the "re-") with an eye towards the role of categories and tree-like organizational structures. So far, that's turning out to be an oddly useful (= consistently distorting) lens.

In Book Beta, Aristotle is still examining how others have approached the problem of what things are. Previous philosophers have made the mistake, he argues, of thinking that principles and classes have to be independent things; in fact, they believe principles are the most real and the most eternal. But then you end up with the "greatest absurdity" [997b.5, p. 46], he says, for you've divided beings into an individual and the principle that makes the individual into what it is, and you have no way of getting them back together.

Worse, it's the individuals that are most real (despite Plato), but if all we have are individuals, there's no possibility of knowledge:

If there is nothing apart from concrete individuals, and if these are infinite, how is a science of infinite individuals possible? For whatever things we come to know, we come to know in so far as they are one or the same, and in so far as some general attributes belongs to them all. [998a, 23, p. 51]

But it isn't possible that the condition for knowing is incompatible with the nature of being. Can we get to something general enough to count as knowledge without denying that to be is to be an individual?

Aristotle says predecessors failed at this. Those that say principles are real beings are needlessly multiplying entities and can't explain what it means for an individual to "participate" in a principle, And, he argues, it doesn't help to point to the constituent elements of things because if that's all you have, you can't understand what makes a bed into a bed; for that you have to see how the parts are put together and understand that a bed is for sleeping...a purpose not contained in the sum of constituent elements.

But, even if we agree that knowledge of a thing is knowledge of what kind of thing it is, how general should the kind be? Do you look at its most immediate genus, or do you look at the root of the tree to see the primary genus? [998b.15, p. 49]

...what many beings have in common cannot itself be a this-somnething, but is a "what"; whereas a primary being is a "this." If we were permitted to supposed that what is predicated in common is itself a this-something, then Socrates would be many animals: himself and "man" and "animals"...These are the consequences of supposing principles to be general. But if they are not general, but individuals, they are not knowable; for knowledge of anything is general. Hence, if there is to be knowledge of such principles, there must be other principles prior to them, namely, what is general to them. [1003a.8, p. 60]

These are the problems that arise from assuming that being and knowability are one (an assumption we moderns don't make; we assume that things are apart from how we know them), and not yet seeing how categories can be different in kind from the things they categorize. We don't have a problem saying that Socrates is a human and an animal because we see that — looking upward — categories can be nested and inherit properties from their containers, and — looking downward — how categories can emerge from what the contain, just as beds emerge from constituent elements. Through the odd lens I'm using, reading this chapter feels like watching the birth of the level of abstraction required to make nested categories work.