What makes a "sea" song, I suppose, is the question. Why should people sing any differently depending on where they happen to be standing/sitting?

I don't mean to sound unreasonable. I think it IS reasonable to suppose a set of people—such as "professional sailors" might come to favor a certain set of material in a way that can be perceived. And maybe that is what we want to call "sea" songs. The question then is whether who they are, precisely, has much to do with having determined the material (and therefore, serving as a way to try to guess what material "might" be favored by them or should be ascribed to them) or whether it is rather more haphazard: a group of people interact socially, a set of repertoire becomes common among them, but who they are is more circumstantial.

In other words, how do we get from "People sing whatever they want (according, of course, to their values, aesthetics, and cultural exposure)" to "What these people sing are sea songs"?

I think "forecastle songs" are *nearly* an invention. I say "nearly" because, again, I think it's reasonable to observe, that at a certain time there were songs that were particularly popular for seamen to sing during leisure and that *certain* of those songs (grouped by style or lyrical content) were perceived by someone to be rather characteristic of those men and their activity.

But I say "invention" because commentators after the historical period really harp on this category of "forecastle songs" far more than the contemporary historical record (that I've seen) leads me to see was a category. The invention, then, is more of an overemphasis on the category. On one level, the category serves as a foil to "chanties," to produce a contrast between working and non-working songs. On another level, the desire to seek and highlight songs that seem particular to the Sea (and particular to the people imagined to inhabit the Sea) leads to (probably) a far narrower set of repertoire than the actual material sung in the forecastle (or at the forebits).

Unfortunately, the convenient database tools like Google Books don't offer a quick answer. The books and magazines of the 1920s are cluttered with discussions of sailors' songs of an (often undefined) earlier time rather than of that decade. The topic of chanties had caught on so strongly in that decade that it seems if "sailors" and "songs" popped into people's minds, they felt it almost obligatory to exposit on chanties or older non-working songs and ignore the current reality. Performing such a (quick) search, I find it interesting how often the phrasing turns up that says sailors sang "THEIR chanties" rather than just saying sailors sang chanties. It's as if the idea of (unique) possession of chanties by sailors is so strong that, again, when the two thoughts of "sailors" and "singing" came together in the mind, the writers feel compelled to give lip service to chanties.

We can suppose the 1920s sailors sang "Bell Bottom Trousers," but verifying the extent of that-- and thus speaking to the issue of whether it was sung enough to meet the reasonable bar of a "sea song" rather than something a lot of different people liked to sing--requires quite a bit of work.