How Sex Difference Makes a Difference: A Reply to Ilya Somin


Courts heighten scrutiny of policies that classify people by sex and other “suspect” traits. But laws defining marriage as a male-female union are different in structure. The very form of policies based on the male-female pairing—unlike every other classification—demonstrates their connection to the common good. So courts shouldn’t subject them to the special scrutiny applied to other laws that classify by sex.

Print Friendly

I recently argued that the Supreme Court shouldn’t subject male-female marriage laws to the special scrutiny it applies to other policies classifying by sex. Professor Ilya Somin of George Mason Law has posted a thoughtful reply, for which I’m grateful. But I’m afraid it misses—it’s just silent on—my main argument. Here, I’ll show why.

Equal Protection and Discrimination

To determine whether a policy denies equal protection, judges have devised a two-step process. First, they ask about the law’s form or structure: does it classify people based on a “suspect” criterion? Suspect classifications involve traits—like race, ethnicity, or sex—that we have antecedent reason to think may be used in policies unjustly. In particular, the Court tends to deem a trait a “suspect” basis for classification to the extent that it picks out long-oppressed or powerless minorities, is immutable or central to identity, and has little to do with people’s ability to serve the common good.

Second, judges apply to the law’s substance a test determined by its form. Policies that classify by race, for example, must be “narrowly tailored” to serve a “compelling” state interest. Those that classify by sex must have a “substantial” link to an “important” state interest. And laws involving no suspect classification at all just have to have some rational basis.

I had argued that even though state marriage laws formally classify by sex, they shouldn’t be subject to higher scrutiny. Courts should recognize—at the first stage of analysis—that state marriage laws are meaningfully different from other sex classifications.

The question, of course, is why.

Professor Somin’s Misreading

Somin attributes to me the answer that male-female marriage laws are different because they don’t promote male supremacy—while, for example, interracial-marriage bans did promote white supremacy. But I didn’t make this argument myself; I cited it in my summary of common arguments on whether marriage laws involve sex discrimination.

As for my own argument, Somin described—and tried to rebut—it thus:

Girgis also argues that laws banning same-sex marriage differ from laws banning interracial marriage because, unlike race, “[m]aleness and femaleness, and a certain social purpose, are necessarily inter-defined: one can’t even fully explain either concept without reference to the other and to a certain social good.” Maleness and femaleness are indeed “inter-defined.” But the exact same thing can be said of racial classifications. Racial distinctions make little sense, except in a society which has multiple racial groups defined with “reference to the other.” And that definition almost always involved real or imagined “social goods,” such as racial purity, social solidarity, and the well-being of children.

But I had made and distinguished Somin’s point about race. His critique just misfires. Below I explain why.

The Form of Marriage and the Common Good

When a policy’s form triggers extra scrutiny, again, judges pivot to examining its justifications: the state’s reasons for thinking that the classification is sufficiently linked to a legitimate interest. It’s at this second stage that judges would have considered, for example, the Jim Crow-era assumptions that once linked racial categories to the public goals of “racial purity, social solidarity, and the well-being of children” that Somin cites.

But my point was that in the case of a law classifying by sexual composition—by male-female pairing—the law’s form itself already points to its own link to the common good, apart from mere social conventions, or merely contingent data about men and women. So the justification of this classification begins to appear at the first, “formal” stage of analysis.

That’s why I emphasized that the male-female pair’s link to a public interest can be spun out of the very concepts involved (maleness and femaleness), even before judges roll up their sleeves and dig into the state’s supporting evidence of a link or social constructs it might be relying on. Nor must this link rest on stereotypes about what constitutes proper behavior for men and women, but just on the root physical potential to procreate that is implicit, again, in the very ideas of male and female.

In short, neither (a) reproduction nor (b) its social value nor (c) its link to the male-female pairing is just a social construct. So, for a policy classifying by opposite-sex composition, some of the policy’s justification seeps into its “formal” features, without any help from our social conventions. Nothing similar can be said of any of the suspect classifications.

To press the point: we can’t even explain the male-female criterion to someone ignorant of it (say, an asexually reproducing alien) without referring to a necessary social value: reproduction. For male and female are defined with respect to the physical potential for just that: reproduction. (Indeed, a description of the sexes in terms of anatomy would be meaningless to an asexual Martian if it didn’t make reference to reproduction.) And this is due to evolution or nature, not just our social conventions.

Of course, certain races have at times also been defined in terms of alleged social benefits (or harms) and not just skin color, as I had noted. But my point was that with race, these alleged connections crucially depend on social convention or contingent empirical data.

And that, in turn, is why it makes sense, when a policy makes something hinge on race, to proceed to the second stage of equal-protection analysis: to look into the ideas or data behind it and to see if they establish a tight enough link to a serious enough public interest, without reliance on stifling stereotypes or other forms of prejudice.

The same goes for individual-sex classifications. If a policy assumes a special link between, say, men and truck-drivers, empirical data will be needed to establish the link (to say nothing of showing that shaping our policy in light of that link is justified overall). You can’t demonstrate the link (or the social benefit of laws based on it) just by a formal or conceptual analysis of “male” and “truck driving.”

Again, this is all in contrast to the male-female criterion of marriage laws. Judges still in the first stage of analysis—still focused just on these laws’ form, on the classification itself—will already have evidence of the classification’s link to a social benefit. And not just because they’re relying on mere social constructs, about the link or the benefit’s value. Evidence of the link inheres in the nature of the classification’s very categories, male and female (though these concepts are also shaped by social assumptions). And the benefit to society at stake is undeniable: indeed, it’s the society’s very survival. This structural difference is what can justify not treating these laws as suspect, by going on to subject them to a higher test.

Likewise, my argument isn’t undermined by Somin’s objection that “adoption and artificial insemination” make “same-sex marriage . . . also closely connected to reproduction.” Neither adoption nor reproductive technology weakens the conceptual link tying male-female union to reproduction. So they do not justify a judicial presumption against classifications based on the male-female pairing. Even finding ways to make children out of two ova or two sperm cells wouldn’t cut that link. It would only mean that, while male-female union has a conceptual link to reproduction, other groupings could have a contingent, technologically enabled link. The latter (unlike the male-female) link would require inquiry into empirical data; and it wouldn’t be limited to same- or opposite-sex couples, but any group of adults deciding to produce and rear children.

So the Supreme Court needn’t heighten scrutiny here after all: The male-female classification doesn’t stand in need of justification by reference to something beyond its form in the way that racial and other sex classifications do. For its very form—apart from our (potentially prejudicial) conventions—already contains the seeds of its own justification as no other does.

Sherif Girgis is a student at Yale Law School and doctoral student in philosophy at Princeton University. He is co-author, with Ryan T. Anderson and Robert P. George, of What Is Marriage? Man and Woman: A Defense.

Print Friendly



Related Reading


Web Briefings

PD logo

Want more great articles?

Sign up for daily or weekly emails!

subscribe button