siderea linked to
this article on the sex lives of White-Throated Sparrows, which inspired this post.
But first, consider how sexual reproduction works in our world, in our species. I'm only talking about biological/chromosomal sex here, not anatomical or social or legal, and for simplicity I'm assuming that every individual is of one unambiguous biological/chromosomal sex -- no XXY or XYY trisomies. A man and a woman can produce offspring of either sex; both sons and daughters are genetically close to both of their parents (although sons are
slightly closer to their fathers because that's where they got their unpaired Y chromosome). The result is basically a single gene pool for the whole species (except for mitochondrial DNA and a few Y-linked genes).
Now imagine that each human carried only
one sex chromosome, rather than a pair, and that men and women were both capable of homosexual reproduction: two women could produce a (necessarily female) child, and two men could produce a (necessarily male) child. If this happened only occasionally, it wouldn't have much effect on the population. But if homosexual reproduction were universal, there would be genetic divergence, with males and females gradually becoming two mostly-separate gene pools and then two separate species. If there were an
occasional instance of heterosexual reproduction against a mostly-homosexual background, the offspring would have genetic material from both pools, and thus (if the offspring reproduced) pull the two pools back together.
So, back to the sparrows. White-throated sparrows have two genetically-determined color morphs: brown-streaked and white-streaked, each forming roughly half of the population. If brown-streaked sparrows preferred to mate with other brown-streaked sparrows, and white-streaked with white-streaked, they would diverge genetically over time and we'd eventually have two separate species: brown-streaked and white-streaked. But in fact, females (both brown- and white-streaked) prefer brown-streaked males, while males (both brown- and white-streaked) prefer white-streaked females. As a result, white-streaked females and brown-streaked males get their pick of mates, which is each other, leaving the remaining brown-streaked females and white-streaked males to make do with one another. As a further result, essentially all mated pairs are either white-streaked-female to brown-streaked-male or brown-streaked-female to white-streaked male. Both sorts of pairs can produce any combination of male, female, brown-streaked, and white-streaked, so there's no genetic isolation and no speciation event: there are effectively four "sexes", but they're all still bound together into a single genetic population.
(Another point in the article: brown-streaked sparrows of both sexes tend to be more nurturing, while white-streaked sparrows of both sexes are more aggressive. Which has interesting implications of its own, but it's peripheral to the point I'm making here.)
By amazing coincidence, the
same day that I read that article, I read Ursula LeGuin's short story "The Wild Girls", about a human society with three castes named Crown (nobility), Root (merchant class), and Dirt (slaves). In our world, wherever a caste system has arisen, cross-caste marriage is strongly discouraged if not forbidden, which (given enough time, and strict enough enforcement) would be expected to produce genetic divergence and speciation. But in "The Wild Girls", cross-caste marriage is mandatory: Crown men are only allowed to marry Dirt women, Dirt men with Root women, and Root men with Crown women. It's not entirely clear in the story, but I think a child's caste is always the same as its father's. In any case, there are effectively six "sexes", and they're all bound together stably into a single genetic population by the cyclic marriage rules. And although you can tell someone's caste by clothing, there's no genetic difference among castes and thus probably no way to tell the caste of a naked person. (Which raises plot ideas....)
And then I thought about fingerloop braiding. Consider a simple two-loop "braid", as discussed
here. There are two interesting operations you can do: you can pass one loop through the other, and you can twist a loop on its own axis ("taking thy bowes reversed", as the middle English source says). If you just twist each loop on its own axis, never passing one through the other, you end up with two independent two-ply twisted cords, connected only at the anchor. If you
only pass one through the other, never twisting either one, you again end up with two two-ply twisted cords, connected at both ends but otherwise independent. But if you pass one through the other, then twist one, in alternation, the two operations lock one another in place and you end up with a single four-strand braid. Likewise if you have three, or four, or five loops: if you only twist each one, you get five independent two-ply cords; if you only pass them through one another, but always "taking thy bowes unreversed", you get two 3-ply, 4-ply, or 5-ply braids connected to one another only at the ends. But if you pass loops through one another
and reverse them, alternating operations reasonably often, you get a single bound-together braid of 6, 8, or 10 strands. If there's an
occasional reversal in a mostly-unreversed braid, you get a cord with large "eyes", holes where the two halves run parallel but independent between one linkage and the next.
Do with that what you will.