Thanks for the reply.  I'll try to reply comprehensively, sorry if I miss anything.  To start with - Aristotle.

What Aristotle Taught

Was Aristotle not an originator of a school of syllogistic logic that treated concepts somewhat similarly to the logical positivists?  

I'm going to break this into two parts - the part about logic, and the part about concepts.  Logic first.  Aristotle indeed wrote six works on logic and reasoning, which are most often collectively called the Organon.  Most of it is developing a valid system of syllogistic logic.  The really nice part about syllogistic logic is that correct syllogisms are indisputably valid (but not indisputably sound).  Aristotle is totally clear about this.  He showed - correctly - that logic, correctly applied, makes your conclusions as true as your premises (i.e. logic is valid); but that alone still doesn't entitle you to certainty about your conclusions, as you can't trust your premises any more than you could from the start (i.e., validity is not soundness). 

In The Parable of Hemlock, ctrl+F "the Greeks."  Eliezer's issue isn't with syllogism.  It's with something different: the assertion that "all men are mortal" by definition.  Aristotle says nothing of the sort, least of all in the Organon; he just uses the statement as a hypothetical premise to demonstrate the form of valid syllogism, the same way you might use a sample like "all frogs are green, Harold is a frog, Harold is green" as a lesson of validity in a logic class, regardless of whether purple dart frogs exist.  The text that most clearly shows this is the Topics, where Aristotle characterizes good arguments as constructed by using syllogism (as characterized in the earlier works of the Organon) or enthymematic syllogism, especially when the syllogism begins from established beliefs (endoxa) as premises.  Explicitly, these endoxa like "all men are mortal" are not certain or guaranteed to be true; but they are better than wild speculation, especially if you are trying to persuade someone.  So Eliezer's attack on the Greeks is off base, mistaking the assertion of validity for the assertion of soundness.

There's nothing wrong with syllogistic logic, as long as you don't make too much of it.  Eliezer's top-line conclusion is that "logic never dictates any empirical question [with certainty]"; I think you would be extremely hard-pressed to find a sentence in Aristotle which disagrees, and Eliezer's clear imputation that they did disagree is ignorant and uncharitable.  Logic is a useful tool for reasoning from premises you are reasonably confident in, to conclusions you can be similarly confident in.  

It's no straw man to say that Aristotle liked logic.  The straw-manning comes when Eliezer asserts that "the Greeks" thought you could derive certain empirical truths from logic alone.  (Parmenides, Spinoza, and Kant attempted this, but not Plato, Aristotle, or most philosophers.)  Rather, Aristotle's logic is all about taking established pretty-good beliefs (which are not called certain, but are generally acknowledged and are the best we have to work with) and having a sure way to arrive at exactly equally good beliefs.  Putting this in writing was an incredibly valuable contribution to philosophy.

Now for the part about concepts.  Did Aristotle treat concepts similarly to the logical positivists?  Honestly, I think not; my impression is that the average positivist was a nominalist about the question of universals, while the best summary of Aristotle's view on the topic probably heavily uses the word hylomorphism.  It's kinda his own deal, like how Plato was Platonist.  I don't love Aristotle's metaphysics, and I think there are powerful skeptical/nominalist critiques of hylomorphism, which is after all a formalist view of one kind or another.  But I don't think Eliezer really advanced them, or understood Aristotle's (or any Greek's) phenomenology of concepts at all.  For a little taste of how nuanced Aristotle's thoughts on words and concepts actually were, here's another bit from the last book of the Organon:

It is impossible in a discussion to bring in the actual things discussed: we use their names as symbols instead of them; and therefore we suppose that what follows in the names, follows in the things as well, just as people who calculate suppose in regard to their counters. But the two cases (names and things) are not alike. For names are finite and so is the sum-total of formulae, while things are infinite in number. Inevitably, then, the same formulae, and a single name, have a number of meanings.  [emphasis added]

Relevant Reading (By philosopher)

specific pre-moderns you think identify and discuss this problem.

If we're discussing the problem of "gee whiz, in what sense do concepts exist and truthfully inhere in an ever-changing world?"  Virtually all of them!  Here's a short rogue's gallery, take your pick if you're intrigued by one in particular.

Plato: Plato's answer is formalism.  But even (or especially) if you think that's absurd, his treatment of the question is incredibly valuable.  Plato is deeply aware and deeply disturbed by the fact that the world around him is changeable, that appearances and naively-constructed concepts deceive, and that nothing certain can be found in them.  And the core of many of his dialogues are devoted to proving exactly that.  Take the Theaetetus, where he talks about certain knowledge.  Can we get it by sense perception?  Not quite, appearances can deceive.  What about judgment?  Fallibility would indicate no.  Is it justified true belief?  Perhaps, but "justification" demands prior knowledge of the thing itself, so this is invalid by circularity!  Plato strongly hints at his solution of formalism, but to pave the way to it, he demolishes more standard accounts first by trying to prove the slipperiness of ordinary concepts and the inaccessibility of certainty.  Skeptical accounts can find a great deal to like.  (Ever wonder why J.L. Mackie's skeptical "argument from queerness" begins as a steadfast defense of Platonism as the only way to objective morality?  For generations, skeptics have made hay by starting with Plato's objections to others, then attacking Plato's rehabilitative view as the final step of a deflationary account.)  Parmenides is also recommended reading, as most of it is criticism of the theory of forms.  But it's not for the faint of heart, you'll need some really good secondary lit - or far better, a supportive professor to read through it with.  Trying to read and understand it by yourself is an aneurysm risk.  

Aristotle: Often denser than Plato.  But he's far more methodical and much easier to interpret, since he's not writing dialogues with Straussian readings or citing myths which he didn't believe or any of that artistic jazz.  The Nicomachean Ethics may be a good place to see him apply his method of discourse about the natural world, but the writings of his that are most relevant to this conversation are definitely Physics and Metaphysics. (fun fact: the field was named for the book; "meta" is just Greek for "after", so "Metaphysics" just means "after physics", "more physics," or maybe "physics 2".) 

Stoics:  Chrysippus is your boy here.  He is taken to be one of the first nominalists (a general term for one of the most popular non-realist views, i.e., that universal properties are words alone and not things in their own right).  https://iep.utm.edu/chrysipp/#H5 has a summary you might like, and it may be the best we can do, since virtually all of Chrysippus' actual writings are not extant (his views were passed to us by way of others' summaries of them), and most other Stoics (like Epictetus or Aurelius) spent more time talking about ethics, with physics receiving more of a passing mention.

Epicureans:  Really just Epicurus, as his teachings were passed down by Lucretius in De Rerum Natura.  Virtually nothing else from this school is extant, but their influence is very significant.  Steadfast materialists, atomists, atheists, and hedonists.  This community would like their teachings a lot.  I'll take this opportunity to point out a trend which is commonplace throughout ancient philosophy; Epicureanism is atheist, but the text sings paeans to gods, using them as stand-ins for abstract concepts.  This is weird, but not at all rare in ancient philosophy.  Anytime you see someone invoke a god or a myth, before dismissing it as superstition, see if it's useful to treat it as metaphor or conjecture instead.  Remember that, for all the talk of gods and myths he engaged in, one of Socrates' two crimes that he was killed for was impiety.  

Skeptics:  You will agree with these people less than their names imply you will.  They thought some weird stuff; Academic or Pyrrhonian, either way it sometimes comes off as worshiping ignorance.  In any case, formalists they were not, and their eponymous attitude comes across in their writings, which are very clear that if there are in fact universals, we are either unable to come to know them, or even morally forbidden to try.

Peripatetics, Cynics, Cyrenaics and more: there are so many ancient Greeks.  Many of them may not have written anything of value on this question, I can't say.  This is the part where I confess ignorance of and wonder at the true diversity of Ancient Greek thought.

Another big gap in my knowledge is Christian and medieval thought, but I had enough friends who studied it to understand that my received caricatures of it were misplaced.  Aquinas apparently contributed things to metaphysics in the vein of Aristotle.  Maybe Augustine has dope metaphysics, no idea.  God features prominently, of course, so know thyself and whether that's a turn-off.

Early Moderns: Spinozism is super weird and monist and stuff.  Maybe not that.  Kantianism is incomprehensible, even in the original German, but if you can find a good professor to walk you through it (preferably in a classroom environment), there is a reason he was so influential.  The obvious suggestion is the Critique of Pure Reason, and it is definitely the one that is relevant here. (It's where the separation of syntheticity from prioricity comes from!  I don't think it's a good separation, but you will need to understand what it means if you want to understand many metaphysicians after him, most of all Kripke.) I personally like The Critique of Judgment too.

Continentals: Another gap in my knowledge.  A friend read a lot of them and said "there's no there there", but I would guess that had as much to do with that friend as the writing itself.  Another said Hegel is apparently very fun "in the right state of mind" (I think they meant psychedelics.  This is not an endorsement of illegal drug use.)  As with other categories on this list, I will acknowledge my ignorance of whatever brilliance might be here.  For what it's worth, if you are interested in critiquing the "classical" method of counterfactual reasoning - or reasoning in general - you may find allies here, even if they are strange bedfellows.

Moderns:  Jumping right up to the 1900s.  Meinong gets a bad rep but I still like him (do square circles exist?  Maybe as much as anything else does!)  Russell and Wittgenstein, you cited already.  Tarski is also a great one, who created a modal logic ("T-schemas" is a search term you can start with) which is intended to be generalizable over different uses of language.  Almost certainly has connections to anything philosophy of language-related.  I like Carnap a whole lot, and he did a lot of philosophy of science which you may find relevant.  I dislike Kripke a lot, but there's no question that his thought is an intensely relevant to any philosophy which deals directly with the idea of meaning (he doesn't think it's a thing, or at least, wants a deflated version of it to be the norm).  He took himself to be in the tradition of Wittgenstein.  

Counterfactual Reasoning

I really like, and generally agree with, your summary of how edge cases and obtuse counterexamples have pushed people to somewhat absurd conclusions.  I'll provide some pushback, but first let me indulge myself in agreeing, and providing an example.  My undergraduate senior paper employed an unfortunately complex variant of the trolley problem (guess how many tracks were involved?) to contest an arcane ethical principle relevant to a facially absurd variant of utilitarianism.  It was truly approaching self-parody, and I was well aware, I just wasn't sure what other topic I had an idea about which would fill enough pages.  (funnily enough, I can write more than enough pages on random internet fora, though.)

For all that ethics should be able to provide us with answers, and there should be answers even for corner cases... it is extremely clear to me that academic ethics has gone over the deep end.  Ethical views are now defined based on cases which are often so ridiculous that whatever decision one would make in those situations is probably a noncentral example of ethical or unethical behavior.  It's clear enough to me how we got here, given a certain kind of steadfast realism about ethics, and it's unclear what exact countervailing view I think should prevail... but somewhere, somehow, we have gone wrong.

Is the source of the problem counterfactual reasoning itself?  Perhaps a certain too-strong form of it.  But I also think that a mature version of "conceptual engineering" would see a lot of it employed.  

I’m sure there are at least a couple other major sources of concept drift and sense accumulation, but I struggle to think of how often counterfactual arguments lead to real linguistic change. Can you provide an example?

The example, or family of examples, that I want to give you and propose as an incredibly useful analogy here, definitely one where there are lots of examples of "concept drift and sense accumulation", is law.  It's not exactly common usage, but legal language has a bunch of desirable features as an analogy here to apply "conceptual engineering" to.  The boundaries of initially-vague concepts like "probable cause" or "slander" are often decided based on past definitions and laid-out sets of necessary and sufficient conditions in case law.  But they are also subject to shift when corner cases are encountered which clearly do or don't fall into the category - previous understandings of the necessary and sufficient conditions be damned.  Ultimately, the courts converge on definitions that are useful at the very least, and they use a number of methods to do it, counterfactual reasoning and N&S conditions being some of the tools in the toolbox.  Do you think law should dispose of those tools, and do you think it would lead to better decisions if they did?  My answer is "no"; I think they're great pragmatic tools in conjunction with other tools; and that makes me think that N&S conditions and counterfactual reasoning aren't the real problem here.  They can be useful ways to engineer concepts, rather than just a destructive way to attack them with corner cases.

Legal language is also nice because it gives us a clear sense of an evaluative objective, a way to "grade" our engineering project - in a word, we might say "justice."  (Meanwhile, to engineer common language, we might grade based on "clarity" or "intersubjectivity".)  When the existing body of rules and conditions still leave room for doubt, we can employ and develop our terminology to produce results that accord with a notion of justice.  

I hope you like that proposed application of the theory.  Interested to hear your thoughts on whether it's fitting, or if not, why not. 

Thanks, I picked the name myself.  This is a new account because I haven't commented before, but I'm long familiar with this community and its thought - and its norms.  Given those norms, I probably should have cooled off a bit before posting that comment.  Let me try again.  I apologize in advance for the length of the below, but charity takes more work and therefore more words.

Fairness to the Ancients

I think we're talking past one another.  Plato was definitely a Platonist, and he definitely employed counterfactual reasoning.  Congratulations to your Ancient Phil professor on achieving tenure; I studied under others (I won't say who or where for privacy reasons), and they likewise taught that Plato believed in essences.  I was not trying to imply that I think otherwise.  I simply don't think that the thing Eliezer attacked in "A human's guide to words" was, in fact, Platonism; I think it was a straw man.  And I took you to also be putting up that straw man, and associating it with all philosophers before Wittgenstein.

You did not cite Aristotle; I brought him in because you cited "a human's guide to words" as a paradigmatic example of a good argument against "Platonic essences."  And yet, that sequence is not really arguing against Platonic essences, it's arguing against misapplying Aristotelean syllogism.  Eliezer attacks the idea that the logical validity of "Socrates is a man, all men are mortal, therefore Socrates is mortal" entitles you to conclude things with certainty in the real world.  Eliezer attributes that view to "the Greek philosophers", calling them "fond of certainty."  He ridicules this view often throughout the sequence.  I think the passage I quoted in my original comment shows this to be a straw man of (among others) Aristotle.  Aristotle acknowledges that when your premises are uncertain, your conclusions will be too; and that seeking certainty about uncertain or ill-defined concepts is a fool's errand.  For that matter, I would say every Greek philosopher I am aware of would have acknowledged this, and many wrote about the problem!

The other citation that seemed unduly dismissive of the ancients was your citation to Bishop as saying that philosophers "aprioristically" reasoned from their armchairs prior to the 1900s.  For the life of me, I can't find that in Bishop 1992 (ctrl+F "aprior" and "armchair", 0 results); if you can cite more specifically, I would appreciate it.  I would almost certainly have qualms with any assertion of his saying "[X idea] wasn't considered before [Y date]", if he did in fact say anything along those lines.  

I definitely agree that Plato was a Platonist; I'm not going against philosophical consensus on that front.  What I took you to be doing was taking the label "platonism", attaching it to Eliezer's straw man, and then saying that philosophers prior to 1900 all believed it and therefore have nothing to contribute.  

I took you to agree with Eliezer because you cited him, and I really strongly dislike his mischaracterization of Aristotle, and even further dislike the fact that he takes that view and attributes it to "the Greeks", whom he slurs together.  I took you to be reproducing that straw man, attaching the name "platonism" to it, and generalizing that view to an even wider range of philosophers who endorsed nothing like it.  I still think the article as written can create that impression, but it sounds like that wasn't your intent, and I'm sorry for jumping the gun into what amounted to an attack on your intelligence.  

I'll stand by my assertion that "a human's guide to words" straw mans the ancients.  Again, virtually none of the Greeks agreed with the view he attributed to them, and for that matter, attributing just about anything to "the Greeks" is bound to be wrong, given the vast differences between the diverse thinkers in the ancient Hellenistic world.  I took my irritation at Eliezer's ignorance about the ancients, unfairly assumed you agreed in full with his assessments and characterizations because of your citation of that sequence, and extended that irritation towards you, thinking to myself, "as a philosopher, this person should know better!"  

Points for Further Discussion

Finally, I want to thank you for taking the time to write a response to an ill-tempered crank; I hope I've acquit myself honorably enough in this follow-up to receive another.  If you'd like to continue the conversation to more productive discussion of conceptual engineering itself, rather than disputing the ancients, I'd be interested to hear your thoughts on the following propositions (which are, of course, derived directly from ancient thinking):

1. Counterfactual reasoning (/"Conceptual Analysis") is the primary tool which has been used to demonstrate the vagueness of concepts, so disposing of it is dangerous to any project which is premised on the vagueness of concepts.  It is one extremely useful tool (among others) for engineering and streamlining useful conceptual frameworks which align well with language.
2. A good account of concepts should include how concepts change.  For better or for worse, concepts change when people argue about them - often counterfactually.  This means that a project which sets out to understand concepts, but neglects to include counterfactual reasoning as an element of the project, may run into some very hard times very fast.  "Conceptual engineering," as laid out in the article above, is not (yet?) equipped with the necessary tools for this.

I wonder if anyone ever reads the ancients anymore. Because around here I frequently hear the word "Platonic", applied to a naive straw man of formalism that I'm not sure anyone ever believed, let alone Plato's character of Socrates. That caricature has its genesis in "A human's guide to words", which characterizes it by relying on (of all things) Aristotelean syllogism. It should go without saying that this is confused.  

It irks me when supposed philosophers so badly misunderstand the history of their own field, and diagnose and psychologize philosophy's past with the ailment of "aprioristically" and simplistically reasoning based on "necessary and sufficient conditions" alone.  Here is Aristotle, in the preface to the Nicomachean Ethics, the most commonly read of any of his works:

Now fine and just actions, which political science investigates, admit of much variety and fluctuation of opinion, so that they may be thought to exist only by convention, and not by nature...We must be content, then, in speaking of such subjects and with such premisses to indicate the truth roughly and in outline, and in speaking about things which are only for the most part true and with premisses of the same kind to reach conclusions that are no better. In the same spirit, therefore, should each type of statement be received; for it is the mark of an educated man to look for precision in each class of things just so far as the nature of the subject admits; it is evidently equally foolish to accept probable reasoning from a mathematician and to demand from a rhetorician scientific proofs.

The guy whose logical system is supposedly the source of all this aprioristic confusion was very content to characterize in general terms. He explicitly understood that everyday concepts are "fuzzy," or "inconsistent," or "polysemous", and explicitly cautioned against over-rationalizing them. So please understand before you criticize, and appreciate the ancients before slurring them. (No, you get no credit for hiding behind Bishop's words; you should be able to stand behind your citations, rather than credulously repeating the ignorance of another.)

Why is this a problem? There's worse exegesis of the ancients out there. And there are enough people unjustifiably worshiping Plato, that a few people uncharitably misunderstanding him on a minor forum doesn't upset the balance. This is a problem because if you don't understand Aristotle, you can't understand where he failed, and why. "All previous philosophy neglected the vagueness of ordinary concepts; we will correct this" is the exact kind of arrogance that will lead directly to a whole set of problems that have already been encountered (by Kripke, by Kant, by Epictetus, by Aristotle). And on the way, due to your dismissiveness of prior work, you'll incorporate a whole bunch of new jargon like "polysemous" that will make it harder for you to relate to the existing literature, and obscure the connections to all the good answers from the past to the questions you're investigating.

Ever wonder why there are twelve different definitions for the same idea in philosophy? It's because every so often, someone like this comes along pledging to solve philosophy anew, throwing out the old terms and confusions and starting fresh. The end result is always just to add another set of terms and confusions to the pile. To make it even more explicit: You will worsen the problem you are trying to solve.

I suppose I shouldn't be especially upset.  Welcome to the grand tradition of philosophers misunderstanding the history of philosophy; there is nothing new under the sun. But somehow, I am especially upset by this, and a part of me I'm not proud of will experience especially intense schadenfreude as it watches this nascent fad fall into the same old philosophical pitfalls that the author slept through when they were covered in Intro to Ancient Phil.