If it's any consolation(?) I agree with you.

Plato's theory of Forms says: we can see around us many examples of horses. What makes them all horses? Why, the ideal form of a horse. Plato, I believe, then claims that these forms are perfect and timeless and the the actual world is inhabited by imperfect reproductions of these forms. Apologies to scholars of Plato.

Aristotle's Categories, while superficially, quite different I think is tackling the same problem. His most universal abstract categories try to answer the question, "what is it that all things regardless of their nature have in common?". And so his 10 (suspicious that, that 10) categories. His categories are very much rooted in language, as in they map readily onto parts of speech.

Both are ontological investigations. Both probe the essence of what it means to be something. And so on to the present day. Even though Russell was attempting to resolve a paradox in set theory what I think he actually did was answer the question, "what is a thing (any thing) in essence?" and I think the answer is, for our purposes, a term (instance) of a type (form).

It is not at all obvious how following these ancient investigations leads us to the door-step of type theory but all indications seem to point to it. You seem to have thought about this a good deal. I'd be interested in what you think. I've left out a _lot_ of detail but I hope what I've said is useful to somebody.

"Table" is just a word. If you want to know what the word means, you have to observe the structure of the patterns in it's usage. Words acquire meaning through usage (in context.)

There is no such thing as an "approximate table" per se, just things that can be approximately described by the word "table". It makes no sense to talk about a 'real table'. Every thing that exists is real, whether you call it a table or not.

I'm also confused why you'd think the concept of continuity is blurry on this issue. Continuity means that similar inputs give nearby outputs, and this is perfectly consistent with how computation works: if you construct two results through similar algorithms, they will have closely related types. Types become a measure of similarity of different computations, and thus you can use them to group things with similar linguistic structure, and thus similar meaning.

Either way, Aristotle and Plato were both ancient people with not even a tenth the information we have today, so I wouldn't consider them authorities on any part of this discussion.