5,000 Years of Marriage?
Leaving aside both the argument that what we have always done we must always continue to do and any of the merits or demerits of gay marriage, the basic factual claim being offered is hugely false in a number of ways. If we really were bound by the commonality of historical precedent in human history, our marriage statutes would look quite radically different than they do today, and quite different from what the Federal Marriage Amendment proposes.
For one, to make contemporary marriage continguous with the generality of historical precedent, we'd absolutely need to have a provision for polygamy. Over the broad span of human experience, it's a much more common form of marriage arrangement, and often has been legally or quasi-legally sanctioned and defined.
We'd probably need some statutes recognizing legal or contractual arrangements like concubinage, which have also been somewhat common.
We'd need some provisions for eunuchs and their relationship to married households--reasonably though not universally common.
We'd probably need some statutes regulating bridewealth and dowry options: very common in the past 5,000 years.
I don't think there would be a huge amount of precedent for dealing with exclusively homosexual marriages or contracts in that 5,000 years of history, but certainly many societies, including classical Greece, somewhat formally or quasi-legally recognized male-male relations as legitimate and even approved. Lesbians, as far as formal precedent in past human societies go, are kind of out of luck: there's far fewer structural or institutional forms relating to female-female bonds or contracts. (Though lesbian acts and relations have probably been as common or more common than male homosexual acts or relations over the past 5,000 years.)
If the supporters of the FMA want to be clear that they are defending the modern, Western innovation of the exclusively heterosexual companionate marriage, then I have no factual quibbles. If they want to assert that the general norm of human sexuality over the longue duree is heterosexual, I might be able to agree, as long as they'll agree to the proviso that homosexual acts have been a part of every society I know of and have often not been stigmatized or suppressed. But they can't possibly claim that the long-term history of marriage in every human society is one man to one woman, till death do they part. Marriage and its associated practices (divorce, inheritance, household structure) is one of the more variable social forms in human experience. The FMA has little to do with 5,000 years of human history: this is a distinctively modern debate, being fought about the terms of modern life.