You seem to be under the impression that Aristotle was a nominalist about definitions. To my knowledge, that is false. Most of the Posterior Analytics is devoted to showing how definitions can be discovered. He does not believe they are simply arbitrary tags. From where in his corpus are you deriving this interpretation?

I am not talking about nominalism at all, actually; nor Aristotle's notion of horismos which is often translated as "definition" but better translated as "essence".

Rather, I am speaking about the Aristotle-influenced view (still held by many Traditional Rationalists today) of what we would call "categories" or "definitions", in terms of individually necessary and together sufficient properties for membership; and of what may be inferred from these by way of what we would call "syllogisms". (Aristotle's sullogismos being more properly translated as "deduction".)

In particular, it is the idea of categorization-based inference as a matter of logically valid deduction, that has given rise to the notion of being able to define a term "any way you like"; this is an Aristotelian notion but not necessarily Aristotle's notion.

I should note that, being unwilling to put up with Aristotle's writing style, my understanding of his work is derived from secondary sources such as the Stanford Encyclopedia of Philosophy. Sorry, but seriously, bleah.

Anyone who did not understand the above comment, my advice is not to bother.

Here the illusion of inference comes from the labels, which conceal the premises, and pretend to novelty in the conclusion.

Surely you aren't suggesting that Aristotelian categorization is useless? Assigning arbitrary labels to premises is the only way that humans can make sense of large formal systems - such as software programs or axiomatic deductive systems. OTOH, trying to reason about real-world things and properties in a formally rigorous way will run into trouble whether or not you use Aristotelian labels.

I have to agree with anonymous. Having read your discussions of "true-by-definition" and arguments about labels for the past couple of weeks, I wonder what ax you are grinding against Aristotle. Who is making the claim that logical inference yields empirically significant inferences? Why do you see the lack of empirically significant inferences as some kind of point against Aristotelian syllogism? Aristotle was one of the first, if not the first, to attempt to formalize reasoning. Sometimes when I read these posts, I feel like your are failing to distinguish between an inference and an induction. As Hume argues (forcefully, in my opinion), induction based on empirical observations can never be certain. I don't take this as a point against induction, but rather as a caution against those who use it thoughtlessly. Finally, the fact that logical inference can never yield an empirically significant result may not be equivalent to saying that logical inference is pointless. Unlike the classic proof of socrates' mortality, there are many tautologies that are not obviously tautological. The most famous of these may be "If A is a formal system that allows the development of arithmetic , then there is no set of axioms, B, such that all true statements in A are provable from B." This is a hasty statement of Goedel's incompleteness theorem. This statement is tautological, but does that make it an unimpressive inference? This is a tautology that has been extremely empirically helpful, if only insofar as it freed up the time of those struggling to prove the completeness of arithmetic.

I don't think mortal is included in the definition of human.

Shouldn't the syllogism be rendered:

All [~feathers, bipedal] are mortal. Socrates is a [~feathers, bipedal]. Therefore, Socrates is mortal.

Which is at least a little bit more interesting than you've indicated.

Compare also Mill's discussion of finding out that diamonds are combustible from "A System of Logic"