OK, quite a lot to go through here. I don't think the racial coding (basically, using elements that point to a race, without actually saying it outright) of some D&D races is generally terribly conscious on the part of the people who write it, though I think it's looped through subconscious stereotypes of how we see society. And yes, the stereotypical indigenous-style costuming of goblins is a good example of this. The term "medievalism" is used in this context to mean "replicating medieval aesthetics or popular perceptions of the middle ages" - this was the sort of conference where not everyone does coal-face history, there were plenty of people who worked more on how the public sees the past and this is more their sort of area. I think it would've been interesting to have more discussion than we did in the session on how things linked back to the "real" middle ages, but a lot of the discussion was really on how a "medieval" aesthetic is often used to justify a much more modern set of norms.
As I said in the middle of the piece, I think I tend to agree that most people don't take parallels between fantasy and real races/race constructions away from the gaming table (and that seems to be the anecdotal impression of the one masters' thesis I found and read on the topic too) - but I'd really like to see some good social science work on that, since it's a contention that people keep making (and did make in this session) that I think it'd be worth having evidence on one way or another.
The important issue about your devil's advocate para is the word "inherent". There's been some interesting work some people have done to design D&D variants that split "species" and "culture" which I think is an interesting approach - one of the issues of the D&D race system is that it includes elements that seem to be biologically natural (dwarfs being able to see in the dark) and stuff that's definitely cultural (elves having a specific aptitude to using longbows, which is something that actually takes a lot of practice and which elves would only logically have in a world where they'd been brought up in a purely elven society that used longbows all the time). The differences you mention in human cultures fit almost entirely into the "cultural" mould, I think: they're aspects of nutrition, upbringing, etc, more than genetics. I think that's the worry people have with settings presenting inherent/from-birth racial characteristics - if you take a reading of a D&D setting that treats D&D races like human races, the whole thing starts looking like nineteenth century scientific racism where people posited that these differences were all genetic, which they're undoubtedly not, pretty quickly. I don't think that's how players generally understand the system, but I can see why it's something that people can read into it.
I don't, that said, think this should be a purely academic discussion - I think it's something that designers especially might find it worth time thinking about, for a variety of reasons. I think race/racism, and whether DMs reflect it at all in their games, is a fairly sensitive area that DM handbooks could sometimes provide more support on, and that particularly with a more recent generation of fantasy books working on "humanising" orcs, goblins, etc, we seem to be moving away from the "Goblins: Always Neutral Evil" type writeups of older books, and it's worth giving some thought to how and why we're doing that from a design aspect and where this ends up. In a hobby this international it can sometimes be worth clarifying terms like "race" in a rulebook too - people will interpret them differently in different settings.
Gender is a whole other kettle of fish, of course. It very much depends where/when you are in the medieval world - which of course spans like 1000 years depending on who you ask - how heavily gender norms were enforced, and what gender norms even were. The gender norms of modern medievalisms are mostly filtered through quite restricted/binarist cultures on gender, like the Victorian/C19th stuff in UK/Europe - and filtered through the literature of that period, which overwhelmingly focussed on elite women. I think there's a tendency for people to assume that things like gender norms, racial boundaries, etc, always used to be more rigidly policed than they are today, which is nonsense. Racial and ethnographic constructions in the medieval and ancient worlds were constructed on pretty different lines to begin with, and lots of cultures worldwide had quite a bit of variation in gender roles (which again in some cases were constructed differently, there's a fair argument that for example Byzantine eunuchs were treated as a separate gender to women or other men in terms of their social functions etc).