Almost everyone assumes that there is no originalist case for constitutionalizing same-sex marriage. Same-sex sexual activity was generally illegal at the time the Fourteenth Amendment was adopted and remained so thereafter, so it seems obvious that no one then thought such activity to be constitutionally protected.
In this post I will explore whether there is nonetheless an originalist case to be made for constitutional protection. My aim is not to be definitive, but to show that the question is more difficult than commonly assumed.
To begin, I assume that the Fourteenth Amendment’s equal protection clause establishes an anti-discrimination or equal treatment rule, as apparently understood by the Supreme Court in Strauder v. West Virginia (1879) and Yick Wo v. Hopkins (1886), and by Justice Harlan’s dissent in Plessy (and as applied in modern law). I further assume that, as an original matter, the rule of equal treatment applies to marriage in the way indicated by the result in Loving v. Virginia -- that is, it is unequal treatment to say "x" can marry "y" but "z" cannot when the only relevant difference between "x" and "z" is a racial difference. (I recognize that many originalists would not accept these threshhold assumptions, but I don't think they are radically out of step with an originalist approach).
The focus, then, is on the original public meaning of “equal” in the equal protection clause. I will further assume the word had an original public meaning of (at minimum) non-discrimination on the basis of characteristics such as race (but not only race, or the clause would have proscribed only race discrimination). Thus the question is whether sexual orientation is a characteristic like race. Clearly people at the time the clause was adopted didn’t think so, because apparently no one thought the clause protected same-sex sexual conduct. Though this is sometimes called an “original expected application” it is (especially in this context) highly probative of original meaning: the best explanation of the expected application is that whatever “equal protection” meant to people in 1868, it did not mean constitutional protection of same-sex relationships.
That, though, may not be the end of the argument. It is possible that understandings of the facts underlying sexual orientation might change in a way that would affect sexual orientation’s status under the fixed meaning of “equal.” While original public meaning methodology must apply the fixed meaning of the constitutional term, changes in knowledge might change the application of that fixed meaning to particular facts.
Suppose, for example, a federal statute is enacted that provides benefits for producers of “healthful” foods. At the time of enactment, everyone assumes that butter is a “healthful” food within the meaning of the statute (although butter isn’t the main point of the statute); “healthful” means good for the human body, and everyone thinks butter is good for the human body. Many years later, however, suppose a different scientific understanding emerges: contrary to what was previously believed, butter has no nutritional value and has numerous adverse effects on human health. An originalist could now say, I think, that butter does not qualify as a “healthful” food under the statute. This conclusion does not depart from the original public meaning of “healthful.” It arises only from discovery of new facts that change butter’s status under that original meaning. In other words, in applying the statute originalism requires a fixed meaning of “healthful” but it does not (necessarily) require a fixed understanding of butter. (The situation might be different if the statute had been specifically enacted to aid butter producers, but I assume that was not the case).
A similar argument might be made for constitutional protection of sexual orientation rights. Our understanding of sexual orientation today may be factually very different from the nineteenth century understanding. Suppose that in the nineteenth century people regarded same-sex sexual activity as merely an inappropriate and arguably harmful behavior like, say, adultery. On this view, laws disfavoring same-sex sexual activity (like laws disfavoring adultery) might not raise equal protection concerns because they are not akin to discrimination on the basis of a characteristic such as race. Further suppose, however, that today scientific and sociological understandings have come to regard same-sex sexual orientation as constituting a characteristic like race. We might then be justified in thinking that discrimination on the basis of sexual orientation (or same-sex sexual conduct) raises similar issues as discrimination on the basis of race.
In this way, we might conclude that sexual orientation discrimination violates “equal” treatment in the sense that “equal” was understood by the drafters and ratifiers of the Fourteenth Amendment. As in the butter hypothetical, the meaning of “equal” has not changed; rather, the understanding of the facts surrounding sexual orientation have changed. Nothing in this analysis violates the fixed meaning canon at the core of originalist methodology, just as the changing treatment of butter in the hypothetical does not involve a change in the fixed meaning of “healthful.” The meaning of the relevant text has not changed; facts relevant to the text’s application to a particular circumstance have changed (or, rather, are now understood differently).
To be sure, this argument rests on several contestable assumptions. But though the assumptions may be contested, I don't think they are unreasonable. My goal is only to show that an originalist constitutional case against sexual orientation discrimination is plausible, not that it is conclusive.