I've written at length about gender-neutral language with respect to pronouns (e.g. whether 'he' can be gender-neutral in some English dialects). I didn't say anything there about the gender-neutral 'man' or 'men' for humans as a whole without regard to gender, but I found a good example of how the Geoff Pullum view on 'he' can't be true of this usage of 'men'.
The Pullum view is that what we have taken to be gender-neutral language isn't really. It's genuinely sexist language in that it assumes a male person unless the context clearly indicates someone to be feminine. I was arguing that in some dialects (not mine) it might be that sometimes it's really gender-neutral, and I had to offer an alternative explanation for why certain sentences that don't clearly indicate maleness still sound wrong (e.g. 'Either the husband or the wife has perjured himself'). My explanation was that whenever anything in the context forces us to think about the gender of people, even if it doesn't say specifically what the person's gender is, then it's impossible to think of the person as gender-neutral and thus impossible in these dialects to use a gender-neutral 'he'.
Now I don't want to assume that 'man' is analogous in every way to 'he', but I stumbled across a sentence today that seems to confirm my explanation of what's going on over Pullum's at least for the gender-neutrality of 'man':
"Eph 5 intends to help all men, especially all couples, to live together in peace." (Markus Barth, Ephesians 4-6, Anchor Bible, p.753).
That sounds archaic to me, but it sounds right within the dialect that uses gender-neutrally the terms I restrict to male gender, whereas Pullum's statement above doesn't sound right in either. It's clear in the sentence that the 'all men' must refer to all of humanity, without assumption of maleness, because in context 'all couples, to live together in peace' isn't intended to refer to each couple living in peace with every other couple but with each member of a couple living in peace with the other member of that couple. If Pullum's explanation for 'he' were to apply here, this would be wrong, because there's no assumption of maleness.
However, my explanation for 'he' also doesn't work with this sentence. My explanation was that 'he' can be gender-neutral only when the context doesn't even raise the gender question. That's not so here. Couples, at least to a German theologian writing in 1974, would have been assumed to be male and female and not same-sex. (Even today, I think this is usually the case anyway, but it was undoubtedly true in Barth's situation.) So I think this is good evidence against the explanation I gave in my earlier post.
Nevertheless, I'm not sure it's good for Pullum either. 'Men' in cases like this and 'he' in the previous cases are both purportedly gender-neutral terms. I don't know what Pullum thinks about 'man' and 'men' as gender-neutral, but the fact that he thinks 'he' carries an assumption of maleness and isn't gender-neutral would lead me to expect that he should think the same of 'man' and 'men'. It would be strange for one to be gender-neutral but for the other to make sense only when the assumption is maleness. Yet it seems as if both Pullum's account and mine have trouble dealing with this on the assumption that they work in parallel.
This leaves me with a couple actions. I could recant on my explanation of 'he' in gender-neutral dialects of English (not that I ever endorsed it anyway; I simply offered it as a possible account to show that Pullum's argument doesn't succeed unless he rules such options out). I wouldn't therefore endorse Pullum's explanation either. On the other hand, I could conclude that 'he' and 'man' just don't work the same way. I think this is what I should do. After all, 'he' seems to have problems when you introduce gendered people even without saying which one 'he' is supposed to agree with, while 'man' doesn't. Regardless of what your explanation for 'he' and 'man' will be, that disanalogy has to be part of it. I conclude, therefore, that the differences between 'he' and 'man' don't require me to take this result as undermining either my account or Pullum's, and we're back to square one with that issue. I offer one account, and he offers another. Mine is a little more charitable, and his is a little more suspicious of people's motives (which I think is a moral reason to prefer mine but not one that has anything to do with whether it's more likely to be true).
With 'man', though, I think the case that it really is gender-neutral is much stronger than with 'he'. Pullum's explanation of what else might be going on can't apply here, even if my explanation of what else might be going on with 'he' can't apply either. I don't think either is needed, however, because I haven't seen a sentence like his that seek to undermine seeing 'he' as gender-neutral. If I see such a sentence, maybe I'll need an explanation to retain my view that such purportedly gender-neutral terms aren't necessarily sexist in assuming maleness (even if in my dialect they sound archaic) any more than when people use clearly gender-neutral terms like 'person' will assume a male (as I think does happen with at least a good number of people). That's a psychological factor, not a linguistic one, and building it into the semantics just seems to me to be a very bad idea.