> > "A relationship of husband and wife involving mutual rights of sexual intercourse, life in common, and an enduring union. The last two characters distinguish marriage, respectively, from concubinage and fornication. The definition, however, is broad enough to comprehend polygamous and polyandrous unions." Pretty much anyone from any culture in history would recognize this as a marriage. It's just foolish to assume that marriage can be changed at this point--in precisely the same way that it's foolish to try to come up with a new meaning of parenthood. It's an objective description. If we tried to change it, we'd end up with something that wasn't really marriage (or parenthood) at all, which would defeat the whole point.
>
> The terms "husband and wife" seem singular to me, so the idea of it covering polygamous (which includes both polygyny and polyandry) relationships seems false to me. But that's just nit-picking.

It was correct. There are as many marriages as there are members of the more numerous sex. In a polygamous (the word is virtually always synonymous with polygynous) union of a man and two women, the women are not considered to be married to each other, but only to the man.

>But it seems to me that by removing the gender-specific part and replacing it with "two or more people", you can go from covering "pretty much anyone" to absolutely anyone. Or was there a reason why the same-sex marriages entered into in Sparta and several North-African or Polynesian cultures were ignored?

They were ignored because they are aberrances. In precisely the same way, kibbutzim raising children communally do not invalidate the meaning of the word "parent".

> I'm also not entirely sure what you mean by "mutual rights of sexual intercourse". Does that mean simply that they are allowed to have sex with each other (in which case I'm not sure how it differs from being unmarried)

Yes, and the difference was explained in the definition.

> Then, what is an "enduring union"?

When you can tell me what a heap is, then I'll consider answering this pointlessly argumentative question.
http://plato.stanford.edu/entries/sorites-paradox/

> There's little or no evidence about the types of relationships entered into before about 20,000 years ago.

Very well. The desire for exclusivity is essentially hard-wired into us and into some apes (it's sometimes one-way, though: exclusivity for my partner but not for me), and I just assumed that we would have followed through on our desires long before we could write history.

> I've been trying to work out how I could have been allowed to immigrate via such a contract. This is something that only the state can regulate ...
> There are many rights that it would be easy to arrange privately, like joint taxes or next of kin rights. These would simply require the parties to file declarations of intent with several hundred independent government agencies...

Granted, in the modern disgustingly interventionist state system, there are lots of troubles. But I have no sympathy for this argument. It's begging to be drawn into general political theory, and I'm as radical a market anarchist as they come. Generally, my positions are summed up by the Roman maxim, "Let justice be done though the heavens fall," rather than any utilitarian scheme. The extreme unlikelihood of my views being implemented is why I said that state recognition of gay and any other kind of marriage is an acceptable compromise. You and your wife should most definitely be allowed to share 'life in common.'