Skip to content
BY 4.0 license Open Access Published by De Gruyter October 29, 2021

How to Project a Socially Constructed Sexual Orientation

Peter Finocchiaro ORCID logo EMAIL logo


Was bisexuality a widespread feature of ancient Greek society? This question is an instance of cross-cultural projection -- of taking the means through which people are categorized in one culture and applying it to members of another. It’s widely held by those who think that sexual orientation is socially constructed that its projection poses a problem. In this paper, I offer a more careful analysis of this alleged problem. To analyze projection, I adapt Iris Einheuser’s substratum-carving model of conventionalism to fit the specific needs of social construction (and social metaphysics more broadly). Using this model, I show that projection is conceptually coherent, and so does not for that reason pose any problem. Along the way, I identify some of the epistemic difficulties facing projection. While these difficulties are formidable, they are not substantially affected by the constructivist claim. I therefore conclude that there is no unique problem facing the projection of a socially constructed sexual orientation.

1 Introduction

In more than one dialogue, Plato portrays a relationship between Socrates and Alcibiades that was sometimes playful and often erotic. Within an English church there is a 15th-century brass memorial that depicts two maidens, Elizabeth Etchingham and Agnes Oxenbridge, intimately gazing into each other’s eyes. And, according to the Book of Han, the Emperor Ai once awoke from a nap and decided to cut off the sleeve of his robes so as to not disturb the slumber of his favored official, Dong Xian, underneath whose head the sleeve was stuck.

More than a few people have used these -- and similar -- examples to combat contemporary heteronormativity. On the whole, our society is one where people are presumed to be straight and those who are not are seen as unusual. But the history of sexuality seems to tell a different story. In some cultures, relations between men or between women were seen as normal. So why should such relations be, in the here and now, so conspicuous? Why should we presume that there is something unusual about those who are bi, gay, or lesbian when the historical records suggest otherwise?

The richness of queer history is seen by many as a useful tool in the fight for political equality. It is perhaps understandable, then, that it has become entangled in a dense knot of issues regarding sexual orientation. I can’t hope to adequately address each of those issues here. Many of them are better addressed by historians, sociologists, and biologists, anyway. But within this knot is an issue that, in fact, is best addressed by philosophers. I think I can separate it from the others. And I hope to adequately address it here.

Some people think that our contemporary categories of sexual orientation reflect real divisions in nature. For example, the neuroscientist Simon LeVay says:

Sexual orientation is an aspect of gender that emerges from the prenatal sexual differentiation of the brain. Whether a person ends up gay, straight, or bisexual depends in large part on how this process of biological differentiation goes forward, the lead actors being genes, sex hormones, and the brain systems that they influence (LeVay (2017): 163).

They consequently take our contemporary categories of sexual orientation as “essential”, and often look for historical expressions of this supposed essence. To that end, it is extremely common for people to reference the myth that Aristophanes shares in Plato’s Symposium. According to the myth, humans used to be large, circular, and doubled -- double-armed, double-faced, and double-sexed. Zeus split these primeval humans down the middle. Ever since, humans long for their “other half”, be it the other male from a male-male double, the other female from a female-female pair, or the other sex from a mixed pair.

But other people think that our contemporary categories of sexual orientation are mere constructions of our society. For example, Michel Foucault says:

We must not forget that the psychological, psychiatric, medical category of homosexuality was constituted from the moment it was characterized […] less by a type of sexual relations than by a certain quality of sexual sensibility, a certain way of inverting the masculine and the feminine in oneself. Homosexuality appeared as one of the forms of sexuality when it was transposed from the practice of sodomy onto a kind of interior androgyny, a hermaphrodism of the soul. The sodomite had been a temporary aberration; the homosexual was now a species (Foucault (1978): 43).

According to Foucault, “the homosexual”’ did not exist until doctors began to categorize patients as homosexual (Shortly thereafter, they were understandably compelled to categorize some “non-homosexuals” as “heterosexual”). He would deny LeVay’s assertion that sexual orientation is a biological category, and he would deny that the historically recent division of people into gay, straight, or bisexual is a mark of scientific progress in our understanding of sexuality. According to Foucault, there is no unadulterated “essence” for science to discover. Foucault and others often marvel at the diversity of sexual expression in humans across space and time. As a result, they pay careful attention to how other cultures conceive of their sexuality and organize its expression. They emphasize taking others “on their own terms” and not imposing onto them our contemporary categories of sexual orientation. Thus, despite their differences, the ancient Greek erastes and the medieval sodomite were no less real in their own times than “the homosexual” or “the gay man” is in ours.

Entangled in this disagreement is a sort of reasoning that I will call cross-cultural projection (or projection, for short). We regularly sort ourselves into social categories -- black or white, man or woman, gay or straight, and so on. We project when we extend this activity to those in other social contexts. So, when in their fight against heteronormativity people claim that bisexuality was a widespread feature of ancient Greek society, they are projecting our contemporary understanding of sexual orientation onto people like Socrates and Alcibiades.

What is at stake is whether or not the projection of sexual orientation is problematic. The most prominent characterization of this alleged problem relies on two seemingly connected claims. The first claim is a metaphysical claim: sexual orientation is socially constructed; our concepts regarding who is lesbian, gay, bi, or straight do not track natural categories; if they track anything, they track something that we, as a society, have created. The second claim is a methodological claim: it is problematic to sort culturally distant figures into socially constructed categories that would be alien to them. From these two claims it follows that the projection of sexual orientation is, despite its political significance, problematic. As the historian of Christianity and medieval Europe John Boswell so plainly puts it: “If the categories ‘homosexual/heterosexual’ and ‘gay/straight’ are the inventions of particular societies rather than real aspects of the human psyche, there is no gay history’’ (Boswell (1989): 20).

In this paper, I argue that this characterization fails to show that projection is problematic. It fails because the second claim is either (a) prima facie plausible but in fact false, or (b) not impacted by the truth or falsity of the first claim. I begin in Section 2 by more clearly articulating what it means to say that something is socially constructed. I then develop a model for articulating such claims about reality. Next, in Section 3, I use that model to develop an account of our socially constructed concept of sexual orientation. In Section 4, I analyze the projection of sexual orientation and distinguish between some ways in which it may be problematic. I argue that, so far as the conceptual coherency of projection is concerned, it is not problematic. I end by tentatively discussing some of projection’s epistemic difficulties. These difficulties are legitimate, but their difficulty would remain even if our concepts regarding who is lesbian, gay, bi, or straight did track natural categories. Thus, there is no unique problem facing the projection of a socially constructed sexual orientation.

2 Modeling Social Reality

Let’s assume that sexual orientation is socially constructed. Yet to say that something is socially constructed is to leave a lot of the details unspecified. To give the assumption greater specificity, I will first say a bit more about what “social construction” means in the context of the essentialism/social construction debate about sexual orientation. Then, I will present a model for characterizing this social construction. The model is an adaptation of a neglected model of metaphysical conventionalism developed by Einheuser (2006).

2.1 Socially Constructed, But Not Socially Fictitious

There is no way to uncontroversially explain what social construction is. But for my purposes here, I will focus on the question “What is it to say that x is socially constructed?” Thus, I set aside (for now) the important motivations and methodologies of those engaged in constructionist projects.

Generally speaking, to say that x is socially constructed is to say that society plays a role in determining what x is. There are at least two importantly different roles a society may play, though. First, a society may play a causal role in determining what x is. Think of stereotypes about social categories -- for example, the stereotype that girls are not as good at math as boys. Quite plausibly, this stereotype has a causal effect on who excels in math classes, who decides to pursue math-related careers, and so on.[1] Things that are shaped by society in this way are causally constructed. Second, a society (or an aspect of a society) may be a constitutive element of x’s metaphysical definition. For example, any plausible account of what it is to be an Associate Justice of the Supreme Court of the United States of America involves a reference to -- you guessed it! -- the United States of America. Thus, the judgeship is constitutively constructed. While causal determination and metaphysical determination may come together, they often do not.[2]

Sexual behavior is obviously causally constructed. Just look at how much it varies across societies! The only plausible explanations for why pederasty was more prominent in ancient Greece than in medieval India or for why same-sex sexual behavior is more common among female prisoners than among the general female population are explanations that mention the causal role society has on behavior. But the fact that sexual behavior is causally constructed does not automatically entail that sexual orientation is causally constructed. Plausibly, an individual’s sexual orientation is dispositional: it only “manifests” in the appropriate “stimulus conditions”. These conditions are often not met in observable sexual behavior. The sexual behavior of a celibate priest (i.e. his celibacy) says very little about his underlying sexual orientation; so, too, for the sexual behavior of those who are forced into survival sex. Thus, the variation in sexual behavior across societies shows that sexual orientation is causally constructed only if the relevant conditions for its manifestation are met across these variations. But what these conditions are and whether they are met is a matter of dispute.

Nevertheless, these causal issues are not the focus of those who think that sexual orientation is socially constructed. They are more interested in the seemingly arbitrary nature of the categories we employ and our blindness to that arbitrariness. As the sociologist Mary McIntosh (who is arguably the first to acknowledge this issue) puts it:

Recent advances in the sociology of deviant behaviour have not yet affected the study of homosexuality, which is still commonly seen as a condition characterizing certain persons in the way that birthplace or deformity might characterize them […] This conception and the behaviour it supports operate as a form of social control in a society in which homosexuality is condemned […] It is proposed that the homosexual should be seen as playing a social role rather than as having a condition (McIntosh (1968): 182–184).

McIntosh goes on to compare sexual orientation to other social labels, like juvenile delinquency and race. These labels are unlike the labels that track naturally significant categories (like those for blood type or diabetes) and so it is no surprise when their scientific study leads to error and confusion: “One might as well try to trace the aetiology of ‘committee chairmanship’ or ‘Seventh Day Adventism’ as of ‘homosexuality’” (183).

Continuing this theme, the classicist and queer theory scholar David Halperin compares the study of homosexuality to the study of dietary preferences. Why should we assume that someone’s preference for men over women reveals their “personal identity” or their “innate, characterological disposition” when we do not assume the same about their preference for white meat over dark meat? (Halperin (1990): 26–27). Similarly, the philosopher and legal scholar Edward Stein compares our situation to that of an imaginary country named Zomnia. In Zomnian society, people are categorized on the basis of their sleeping habits. Those who predominantly sleep on their stomachs are called “fronters”, and those who predominantly sleep on their back are called “backers”. According to Stein, we ought to think that “[the Zomnian] practice of grouping people into backers and fronters is laughable at best, their practice of discriminating against people on the basis of their sleep habits is morally wrong, and their scientific theories that concern the ‘etiology’ of ‘backerhood’ are pseudoscientific” (Stein (1999): 73).

What these comparisons have in common are two theses about sexual orientation, one positive and one negative. The positive thesis recognizes that our categories of sexual orientation track genuine differences in human sexuality and that some claims made using these categories (e.g. “Ellen DeGeneres is gay.”) are non-trivially true or false. The negative thesis, though, denies that these differences are mind-independently significant in the same sense that differences between species are (arguably) mind-independently significant. Insofar as differences in sexual orientation matter at all, it is only because we, as a society, have decided to make them matter. Contrary to what we may have thought, the social significance of sexual orientation is an ineliminable part of the story about what it is. Thus, sexual orientation is constitutively constructed.

Along those lines, those who think that sexual orientation is socially constructed in this way do not go on to say that sexual orientation is “not real”. Some people really do prefer white meat over dark meat. Admittedly, it may be peculiar, even laughable, to continue placing such a great emphasis on something that lacks mind-independent significance. But that is not by itself a mark against its reality. Some further argument must be given to move from mind-independent insignificance to elimination. Such an argument may be hard to sustain, though, since there seem to be all manner of social features that still exist despite their mind-independent insignificance -- political appointments, gym memberships, and modern art, to name just a few examples.

2.2 The Substratum-Carving Model

To more precisely characterize claims about socially constructed features, I will now develop a model. The model is an adaptation of a neglected model of metaphysical conventionalism developed by Iris Einheuser.[3] She uses her original model to characterize prominent anti-realist positions like ontological conventionalism and moral projectivism. But with some small modifications the model can be made to work for theories of social construction and social reality more broadly.

The primary distinction upon which Einheuser’s model is built is that between S-features and C-features. S-features are the features of the world whose presence is independent of our conceptual practices. In contrast, C-features are the features whose presence constitutively depends on our conceptual practices. Take the mind-independent portion of reality, the substratum. This substratum on its own gives rise to certain features of the world. Those features are the S-features. Facts about which S-features are present and what they are like depend only on what the substratum is like. Everything else is a C-feature. Facts about which C-features are present and what they are like depend on both what the substratum is like and the conventions we impose on that mind-independent substratum.

Any example here would be controversial. But I hope the following examples are controversial in an illustrative sort of way. The class of S-features may include things like physical particles, space and time, God, and perfectly natural properties. These features are “objective” -- if they exist, they exist whether or not we recognize them. The class of C-features may include things like spatially disjointed objects, abstracta, statues, and moral properties. According to some philosophers, these things exist only as constructions of our minds. Of course, other philosophers think that these things are just as “objective” as physical particles. Their disagreement is best characterized on this model as a disagreement as to whether the thing in question is an S-feature or a “mere” C-feature.

C-features are derived from conceptual practices. Loosely speaking, conceptual practices are systematic ways of conceiving. Conceptual practices are varied and complex phenomena that raise many tough questions in their own right (Consider, for example, just how much has been said throughout the history of western philosophy about the nature of abstraction). But Einheuser distances her model from the particularities of conceptual practices by introducing the notion of a carving. A carving is a function that takes a substratum and gives a set of C-features. Intuitively, then, a carving models the ways C-features can emerge from a given conceptualization of substrata. For example, a carving that models abstraction might take the particulars existing in a substratum and give the general ideas or universals that an individual could mentally generate from those particulars. Similarly, a moral carving might take a substratum, including its natural properties, and supply the moral properties that like-minded individuals would project onto that substratum. Because carvings are functions, they do not represent “free-for-all” readings of abstraction or morality. They uniformly produce the same outputs from the same inputs, only generating C-features where the relevant S-features are present in the substratum to which they are applied.

Following Einheuser, we can formally represent a world as an ordered pair of a substratum and a carving. Consequently, for any one substratum -- for example, the one we actually occupy -- there are many worlds constructed from that very same substratum and various carvings of it. Similarly, a single carving can be paired with different substrata to generate different worlds.

Now, to adapt Einheuser’s model to work as a model of social reality, I need to introduce three modifications.

First modification. Einheuser remains fully neutral with respect to the comparative “reality” of S-features and C-features. But, arguably, such neutrality cannot be maintained by a model of social reality. As I said above, the claim that something is socially constructed is distinct from the claim that it is “unreal”. My model should therefore be interpreted as one where C-features are genuinely real and claims about their presence or absence are genuinely true or false, depending on what the world’s substratum is really like. What this means, exactly, is hard to say. On my preferred interpretation, the relationship between social C-features and non-social S-features is like the relationship between non-fundamental properties and fundamental properties. The claim that the world contains grue emeralds is genuinely true, even if it is in some way less metaphysically perspicuous than the claim that the world contains green emeralds. So, too, for the claim that my wallet contains some US dollars and the claim that Ellen is gay. Each is a true claim about a particular C-feature, and each might be the most accurate way of describing that C-feature. Nevertheless, C-features are not fundamental and so claims about them cannot be as metaphysically perspicuous as claims about S-features can be.[4]

Second modification. Einheuser originally characterizes the conceptual practices from which C-features derive as being constituted by how people think, speak, and behave. So, while Einheuser does not characterize conceptual practices as being entirely mentalistic, she does seem to suggest that they are the direct product of intentional activity. Yet such an understanding might exclude indirectly produced social phenomena, indirect in the sense that it is not necessary for people to think about them specifically. Consider economic recessions. Economic recessions are socially constructed phenomena – without minds, there are no recessions. And yet economic recessions do not seem to be the direct product of intentional activity, but rather the collateral creation of economic systems. Economists nowadays might be able to identify when an economy is in a recession. But the first recessions predate their ability, and in fact predate the entire concept of a recession. Thus, a recession is in this respect importantly different from a C-feature like money, which is (arguably) directly constructed in the sense that money needs to be conceived of by people as money. John Searle characterizes this difference as the difference between the “ground-floor” facts that are directly represented in people’s thought, speech, or behavior and the “fallout” facts that are constituted by those ground-floor facts, but need not themselves be directly represented. To adapt Einheuser’s model to work as a model of social reality, her (now potentially misnomered) notion of “conceptual” practices should be interpreted broadly so as to include the “ground floor” and the possibility of “fallout”.[5]

Third modification. In her original model, Einheuser intentionally leaves the relationship between conceptual practices and carvings underspecified so as to accommodate as many conventionalist positions as possible. But not all carvings are social carvings. For a carving to count as a social carving, it must satisfy some minimal constraints. Most importantly, social carvings must be social. A carving that models the enigmatic conceptual practice of a single individual living in relative isolation is not a social carving. Thus, for a carving to count as social it must take into consideration the conceptual practices of multiple people. This third modification leads to a formidable challenge. It takes some work to get a sense of what one single person thinks. It takes that much more work to get a sense of what one hundred, one thousand, or one million people think. But the challenge runs deeper than that. Members of a single society almost always have conflicting understandings of their world (try asking people if a hot dog is a sandwich and see what happens!). For a carving to be genuinely social, it must also model an appropriate means of reconciling these conflicts by formalizing that reconciliation into a mathematically determinate function. That’s no mean feat.

I don’t have anything new to say that would help meet this challenge. Following Searle (1995), we could say that a carving is social only if it models a constitutive rule that the members of some society collectively accept. Or, following Haslanger (2012), we could say that a carving is social only if it models the schemas that partially constitute some society’s social structure.[6] But either approach would still leave plenty of work in determining how, for example, conceptual practices combine to produce collective acceptances.

For now, I will simply assume that there is some way to reconcile the conflicting conceptual practices of a society’s members and that members of that society have more-or-less reliable access to the results. I don’t think this is an unreasonable assumption. Consider the determination of the meaning of expressions in a natural language. Linguistic meaning is ultimately constituted by the linguistic practices of members of a linguistic community and the interactions between them. How, precisely, that happens is of course a matter of great dispute. Hopefully, this issue about linguistic meaning is somehow resolved. But even if I don’t know how this issue is resolved, it’s reasonable for me to assume that it is, somehow, and that the members of the linguistic community in question have more-or-less reliable access to the results. To say otherwise is to commit to a pervasive form of skepticism about meaning. The determination of linguistic meaning is just a special case of social reconciliation; in the absence of an argument to the contrary it seems reasonable for me to assume the same holds for other special cases. That being said, it is not okay for me to assume that every member of every society knows exactly what is manifested by that society’s collective social practice. We are often lost, confused, and deceived about how, precisely, a society operates. This is especially true when it comes to features that are purported to be natural but are in fact socially constructed.

3 Carving Sexual Orientation

Now I will show how the adapted substratum-carving model can be applied to the metaphysics of sexual orientation. I will focus on characterizing the dominant conceptualization of sexual orientation in the social context that I know best: the United States of America in the early 21st century.[7]

I think that dominant conceptualization is accurately represented by the carving I am about to provide.[8] To that end, I will offer some motivations for the view and argue against a few alternatives. But my primary interest is in analyzing the problem of projection, not in defending any specific carving that we project. I encourage whoever is unpersuaded by what I have to say to substitute my carving with their preferred carving in the subsequent analysis of projection.

3.1 The Proposed Carving

Consider the fact that Ellen DeGeneres is gay. I suggest that this fact depends on society’s conceptualization of seemingly natural facts about Ellen’s dispositions. More specifically, Ellen is gay because Ellen is a woman who is disposed to feel sexual attraction toward and engage sexually with primarily or exclusively women. But there is nothing special about Ellen with respect to the relevance of these dispositions. To model this conceptualization of sexuality, I offer a carving that targets two portions of the world’s substratum.[9] First, it targets every person’s dispositions to feel sexual attraction toward and engage sexually with others and groups these dispositions on the basis of the sexual object’s gender. Second, it targets the gender of the person with these dispositions. The carving builds schemas from these two elements to carve sexual orientations.[10]

I claim that the United States’ dominant conceptualization of sexual orientation is accurately represented by the popular LGB initialism. This concept can be characterized by a carving that satisfies the following constraints:

Lesbian: For all x, if (i) x is a woman and (ii) x is disposed to feel sexually attracted toward and engage sexually with primarily or exclusively women, then x is lesbian.

Gay: For all x, if either (i) x is a woman and (ii) x is disposed to feel sexually attracted toward and engage sexually with primarily or exclusively women, or (i’) x is a man and (ii’) x is disposed to feel sexually attracted toward and engage sexually with primarily or exclusively men, then x is gay.

Bisexual: For all x, if (i) x is disposed to feel sexually attracted toward and engage sexually with women, (ii) x is disposed to feel sexually attracted toward and engage sexually with men, and (iii) the strengths of these dispositions are comparable, then x is bisexual.[11]

Straight: For all x, if either (i) x is a woman and (ii) x is disposed to feel sexually attracted toward and engage sexually with primarily or exclusively men, or (i’) x is a man and (ii’) x is disposed to feel sexually attracted toward and engage sexually with primarily or exclusively women, then x is straight.

These four constraints exhaust the features carved by US social practices regarding sexual orientation.

3.2 Motivating the Proposed Carving

I don’t claim that there is anything especially novel in the characterization I’ve provided. Indeed, the fact that it is not novel is precisely my aim -- I am trying to accurately characterize the carving that US society in fact employs. To that end, I should say more to motivate this descriptive claim. First, this account is in some important ways compatible with scientific approaches to sexual orientation that are prominent in the United States, like LeVay’s. Those who adopt such an approach often think that sexual orientation is a kind of natural feature (like blood type) connected to dispositions to feel sexual attraction and dispositions to engage in sex (sexual dispositions, for short). They then go on to offer scientific explanations for sexual dispositions.[12] But such explanations can be endorsed even by those who think that sexual orientation is socially constructed. The two parties can agree that there is an underlying naturalistic explanation for why people have the sexual dispositions that they have. What the two parties disagree about is whether or not there is a socially independent reason to care about these sexual dispositions.[13]

That being said, determining what an individual’s sexual dispositions are is not always a straightforward task. We never have direct access to a disposition. We can only infer the existence of a disposition by observing its manifestation in an individual’s sexual behavior. But, as I mentioned above, in some circumstances an individual’s sexual behavior is not a reliable indication of their underlying orientation.

So when is sexual behavior a reliable indication? I don’t know, and I don’t think the United States of America knows. That is, the dominant US conceptualization of sexual orientation is underdeveloped enough that there is no precise answer to this question. There are some conditions that seem uncontroversially mandatory. The individual must be acting consensually. They must be presently willing and able to act sexually. They must have a reasonable diversity of options (e.g. being stranded and lonely on a desert island doesn’t count). But beyond those uncontroversial conditions it’s unclear what more can be said. On the one hand, people often allude to the content of fantasy, which suggests that they think the best indications of sexual orientation involves what an individual would want to do in ideal circumstances. On the other hand, people also very clearly make judgments about an individual’s sexual orientation on the basis of ordinary circumstances -- circumstances which are presumably not ideal. There are at least two ways of reconciling these facts. First, judgments made on the basis of fantasy may just be a particularly expedient way of introspecting about what an individual would do in ordinary circumstances. Second, judgments made on the basis of what an individual does in actual, ordinary circumstances may involve a sort of abstracted inference as to what they would do in ideal circumstances. I’m not sure how to choose between these two readings.[14] I don’t think I need to, though, and so I will leave the issue undecided for now.

The dominant conceptualization of sexual orientation in the United States is also underdeveloped when it comes to the distinction between sex and gender. Though the terms are contested, many scholars think sex terms like ‘male’ and ‘female’ target biological features (like genitalia) and gender terms like ‘man’ and ‘woman’ target social features (like personal appearance or social situatedness). People in the United States regularly run roughshod over this distinction. Consequently, the dominant US conceptualization of sexual orientation is equivocal. In some prominent cases, sexual orientation is said to be a matter of gender. This is best exemplified in connection to the sexual orientation of individuals who are trans. For example, Grindr says that a trans man who is attracted to other men is gay and a trans man who is attracted to women is straight.[15] But there are equally prominent cases where sexual orientation is said to be a matter of sex. Recently, the Supreme Court ruled that Title VII of the Civil Rights Act protects individuals against discrimination on the basis of their being gay or trans. In the majority opinion, Justice Gorsuch argues that “homosexuality and transgender status are inextricably bound up with sex” and explicitly specifies that ‘sex’ indicates the biological distinction between male and female (Bostock v. Clayton County (2020), Opinion of the Court, 10).

My carving should be read in a way that preserves this equivocality in usage.[16] To capture that equivocality, I could use a neologism like ‘sex-gender’. But I’d rather not. The terms ‘man’ and ‘woman’ are already used equivocally; sometimes they are used to refer to a gender, sometimes to a sex, and other times to neither in particular. Thus, when I define being a lesbian in terms of being a woman, I do not intend to take a stand on whether people in the United States consider trans women to be “genuine” lesbians.

That being said, I do intend to take a stand on whether people consider bisexuality to be gender-inclusive. They do not. Admittedly (and thankfully), there is a growing trend of linguistic authorities defining ‘bisexual’ in a way that accommodates attraction to those who are neither men nor women. Merriam-Webster, for instance, recently revised its definition to include “sexual or romantic attraction to people of one’s own gender identity and of other gender identities”. Yet my impression is that the predominant usage of ‘bisexual’ in America is still binary. Indeed, despite its recent revision, Merriam-Webster retains the binary definition and adds this usage note:

While educational and advocacy groups tend to define bisexual broadly as applying to sexual or romantic attraction to members of one’s own gender identity as well as to members of other gender identities, the older, narrower application describing attraction to male and female people persists among English speakers, aided, no doubt, by the word’s morphology: the prefix bi- means ‘two.’ Note that while the broader meaning of bisexual can be understood as occupying the same semantic territory as pansexual, there are people who identify as one but not the other, as well as people who identify as both (

I think that as a matter of empirical fact the LGB carving I provided best characterizes the dominant conceptualization of sexual orientation in the United States of America. First, this carving has widespread official recognition. Returning to the Supreme Court Title VII ruling, even Justice Alito says in his dissenting opinion that sexual orientation follows the LGB carving, citing similar language in a psychiatry textbook, the American Heritage Dictionary, and Webster’s New College Dictionary (Bostock v. Clayton County (2020), Dissenting Opinion, 7–8). But the most compelling reason to think that the LGB carving is the dominant carving is the comparative rarity of alternatives. Consider carvings that offer more fine-grained categories. The Kinsey scale, for example, assigns to each individual a number between 0 (exclusively heterosexual) and 6 (exclusively homosexual). Though the Kinsey scale might better capture the popular idea that sexuality is a spectrum, it isn’t commonly used by people in the United States.[17] Similarly, while the use of ‘asexual’ is growing, the term has not yet reached the level of prominence possessed by ‘lesbian’, ‘gay’, ‘bi’, and ‘straight’.

I could be wrong, of course. I have not offered quantitative evidence for the dominance of the LGB carving. Those who think that another carving (perhaps one of the ones I just mentioned) more accurately characterizes the dominant conception of sexual orientation should use it in the subsequent analysis of projection.

4 Addressing the Problem of Projection

Now, with an account of a socially constructed sexual orientation in hand, we can turn more specifically to the problem of cross-cultural projection.

4.1 Stating the Problem

Cross-cultural projection is the application of our social categories onto those for whom those categories are alien. Many people think that the legitimacy of this application depends on the nature of the categories. So, if some given category is “essential”, that category reflects a real division in nature that is socially independent, and therefore the projection of that category is unproblematic. But if that category is “constructed”, it is in some way contingent and tied to the society in question, and therefore it is problematic to project onto a different social context.

To be clear: there is no canonical sense in which projection is allegedly problematic. The problem was first stated, somewhat inchoately, almost as soon as social constructionism began to rise in prominence in the late 1970s. It took time for scholars to articulate more precise statements of the problem. Despite their increasing prominence, even these more precise statements are underdeveloped and undermotivated. And, unfortunately, they diverge in philosophically significant ways. This lack of precision and consensus is, I think, understandable. The study of sexuality spans across many disciplines. So, while many scholars have wrestled with this problem, most of them have done so within their own disciplinary bubbles. It’s no surprise that sociologists, biologists, anthropologists, and historians have said things that philosophers would find, by the standards of philosophy, unsatisfying.

Nevertheless, I think that the predominant sense in which projection is allegedly problematic is conceptual. Here’s a representative statement. Recall Stein’s fictional society of Zomnia, a society where people are divided into sleep orientations. The historians of this Zomnian society project their categories of fronters and backers onto their historical figures. Stein insists that they are making a mistake. So, too, when we project our sexual categories onto our historical figures:

[I]t does not make sense to try to apply the Zomnian terms “backer” and “fronter” to people in Attic Greece or contemporary North America because the Zomnian categories of sleep orientation are (or at least seem to be) merely social kinds. Similarly, constructionists say that it does not make sense to try to apply the terms “heterosexual,” “homosexual,” or “bisexual” to other cultures such as Attic Greece. Constructionists admit that there were people in Attic Greece who had sex with people of the same sex-gender---they even admit that there were people who had sex primarily with people of the same sex-gender---but they deny that this entails that there were homosexuals in our sense of the term in Attic Greece. To apply our sexual-orientation terms to another culture, we need to have evidence that people in that culture had sexual orientations in roughly our sense of the term (Stein (1999): 97).

According to Stein, “it does not make sense” for someone to apply terms for the socially constructed categories of her own culture to those to whom these categories would be alien. Admittedly, “it does not make sense” could be interpreted in several importantly different ways. Stein does not provide much elaboration. But I think “it does not make sense” is best interpreted as an accusation of conceptual error. According to the reasoning provided by Stein, it seems to be an analytic truth that membership in “merely” social kinds is not universally attributable. Thus, given the relevant historical facts, everyone should know that the claim “Alcibiades was bisexual” cannot be true. Perhaps it is an analytic falsehood, like “There exists a round square”. Or perhaps it is in some other way infelicitous, like the nonsensical claim “Green ideas sleep furiously” or the presuppositional claim “China has stopped shipping tea to its bunker base on the far side of Jupiter”.

This conceptual problem of projection should be distinguished from an adjacent problem with which it is sometimes conflated. Theoretically, some C-features might be interactive, in the sense that the presence of that feature at least partially depends on attributions of that feature. For example, some people claim that an individual’s sexual orientation is determined by their sexual identity: when someone identifies as bisexual, their self-identification does not merely provide strong epistemic evidence that they are bisexual but in fact metaphysically determines that they are. But self-identification is not the only potentially relevant attribution. Other people claim that an individual’s sexual orientation is determined by their social position: to put it simply, someone is bisexual when society treats them as bisexual. Interactive accounts of sexual orientation create a distinct conceptual problem for projection. If bisexuality is an interactive C-feature, then someone cannot be bisexual unless they or their society has a concept of that feature. But the absence of such a concept is the characteristic feature of cross-cultural projections. Thus, if sexual orientation is an interactive C-feature, cross-cultural projections are false, and obviously so.

Call this second problem the specific conceptual problem of projection, specific in the sense that it depends on specific claims about which S-features determine an individual’s sexual orientation. The general conceptual problem that Stein articulates does not depend on these specific claims. Rather, the general problem depends on the claim that sexual orientation is a C-feature and C-features cannot be cross-culturally projected.

Some people have raised the specific problem of projection. The legal scholar Ortiz (1993), for instance, says “The answer to the constructivist question […] depends upon the content we give to gay identity” (Ortiz (1993): 1843–1844). But it is not the problem that Stein is concerned with. And Stein is not alone. The Talk pages of Wikipedia articles about LGBT topics are rife with debates over whether or not it is a mistake to even use the phrase “LGBT”, especially when discussing historical topics.[18] On the more academic side of things, John Corvino gestures toward the general problem when he says, “constructionists do not deny that there were same-sex desire or same-sex sexual acts. Rather, they claim that those desires and acts did not constitute a ‘sexual orientation’---because that mode of identity simply didn’t exist until fairly recently” (Corvino (2013): 113–114). Similarly, the anthropologist Carole Vance asks “if sexuality is constructed differently at each time and place, can we use the term in a comparatively meaningful way?” and recommends describing some acts as “same-sex” rather than “homosexual” (Vance (1989): 164–165).

An analogous issue occurs in the philosophy of race. While supporting the claim that race is socially constructed, Michael Root says “Race does not travel. Some men who are black in New Orleans now would have been octoroons there some years ago or would be white in Brazil today. Socrates had no race in ancient Athens, though he would be a white man in Minnesota” (Root (2000): 631–632). This “no traveling” constraint is widely shared by scholars who think that race is a social construction, and is often used as a desideratum for favoring one theory over another. Now, strictly speaking, the “no traveling” constraint could be substantiated by the claim that race is an interactive C-feature. But some scholars specifically reject such a claim. For them, their “no traveling” constraint is just another way to state Stein’s claim that cross-cultural projection “does not make sense”.[19]

4.2 Two Axes of Social Possibility

In Section 2.2, I defined a possible world as a substratum-carving pair. Consequently, social possibility is a two-fold notion. One axis of possibility is possibility with respect to the space of possible substrata. The other axis of possibility is possibility with respect to the space of possible carvings. Worlds vary along both axes: two worlds may have the same substratum but different carvings, different substrata but the same carving, or even different substrata and different carvings.

This two-fold notion of possibility introduces subtle challenges of interpretation. Consider, for instance, the claim “Ellen could be a regal blue tang fish”. This claim is a paradigmatic claim about what is metaphysically possible, the type of possibility that is about how things could have been different from the way things actually are. On a standard analysis of this type of possibility, “Ellen could be a regal blue tang fish” is true if and only if there is a possible world at which “Ellen is a regal blue tang fish” is true. But how should this claim be analyzed when the standard notion of a possible world is replaced with my notion of a substratum-carving pair?

Well, that may depend on the context in which the claim is made.[20] That being said, I think most invocations of what is metaphysically possible should be interpreted as implicit restrictions to the first axis of possibility. Interpreted in this way, we fix to a constant carving and survey what is true at the worlds constituted by that carving. More formally:

Substratum Possibility:♢ s φ is true at a world w1 = <s1, c1> if and only if there is a substratum s2 such that φ is true at w2 = <s2, c1>

Understandably, when people discuss what is metaphysically possible they are usually concerned with what is possible relative to their own world. Such claims should be interpreted as applying Substratum Possibility to their own carving. For us, here and now, that is the actual world’s actual carving.

What is the actual world’s actual carving?[21] Relative to any context of utterance, there is a social structure in which the speaker is embedded. That social structure’s dominant social carving is the actual carving for any utterance made in that context.[22] What carving is actual is therefore a location- and time-sensitive matter: different parts of the world have different social structures, and even the same location will witness changes in social structure over time.

We can also evaluate what is possible along the second axis of possibility. To do so, we fix to a constant substratum and survey what is made true by different carvings of that substratum:

Carving Possibility:♢ c φ is true at a world w1 = <s1, c1> if and only if there is a carving c2 such that φ is true at w2 = <s1, c2>

This axis of social possibility explores the ways that the very same underlying natural facts can give rise to social facts. Together, Carving Possibility and Substratum Possibility can exhaustively characterize social reality’s modal space. For example, “The whiskey in my glass is bourbon” is true in the actual world because (i) it was made from a mash that was 70% corn, and (ii) on the legal standards actually set in the United States, a whiskey is a bourbon only if it was made from a mash that was at least 51% corn. But “The whiskey in my glass is bourbon” is possibly false in two different senses. It is possibly substratum-false because there is a possible world consisting of the actual carving and a different substratum where the whiskey in my glass was made from a mash that was 50% corn. In addition, “The whiskey in my glass is bourbon” is possibly carving-false because there is a possible world consisting of the actual substratum, including my 70% corn-based whiskey, and a different carving according to which a whiskey is a bourbon only it was made from a mash that was at least 75% corn.

I now need to complicate the model.

Which carving is the actual carving is a contextually sensitive matter. Consider the planet Earth. What carving is actual will depend on what region of Earth you’re in and what time you’re in it. Let’s say that a substratum, s, supports a social carving, c, just in case the conceptual practices present in s are accurately modeled by c. Earth supports many, many carvings. Methodologically speaking, though, it would be better if a substratum supported only one carving. For that reason, I define a substratum as a “centered” substratum -- as a slice of the all-encompassing universe taken at a particular place and at a particular time. Thus, by definition, each centered substratum supports exactly one social carving (even if that carving is the “null” carving).[23]

This complication increases the complexity of modal space. Substrata can now agree or disagree with respect to their universe, their region of focus, or their time of consideration, and as a result the definitions I gave above must be refined and supplemented. I won’t go into all the details here; most of them make no difference. Conveniently, though, one of them will help address the conceptual concern about projection. To that difference I now turn.

4.3 Temporal Possibility and Diagonal Confusion

Sometimes we make claims about what used to be the case. In metaphysics, it is standard practice to formalize such claims using the Priorian tense operators: P, F, H, and G. But on the current substratum-carving model tense operators can be eschewed in favor of quantification over worlds with temporally-connected substrata. To make this interpretation work, a (centered) substratum is structured as an ordered triple of a universe, u, a region of that universe, r, and a time of that universe, t. A claim of the form “” can then be defined as:

Past Possibility:♢ p φ is true at a world w1 = <<u1, r1, t1>, c1> if and only if there is a substratum s2 = <u1, r2, t2> such that (i) t2 < t1 and (ii) φ is true at w2 = <s2, c1>

where “t2 < t1” means that t2 temporally precedes t1. On this interpretation, claims about the past are a subset of claims about possible centered substrata. As I’ve already suggested, substratum possibility is analyzed by holding fixed to the carving supported in the context of utterance. Thus, past possibility is also analyzed by holding fixed to that carving.

Let’s return to projection. Consider statements like the following:

  1. Alcibiades was bisexual.

I maintain that, in most contexts, utterances of (1) are simply claims about the past. They should therefore be analyzed using Past Possibility. Thus, Alcibiades was bisexual just in case it is true there was some time in the past (i.e. a time of the universe that is prior to the time of utterance) where, according to the dominant carving at the time of utterance, Alcibiades is bisexual. The dominant carving in the United States, I’ve argued, is one according to which Alcibiades is bisexual if and only if (i) Alcibiades is disposed to feel sexually attracted toward and engage sexually with women, (ii) Alcibiades is disposed to feel sexually attracted toward and engage sexually with men, and (iii) these two sets of dispositions are comparable. This was likely true of Alcibiades -- though the historical details are irrelevant to the larger conceptual point at hand.[24] Thus, in most contexts where (1) is uttered in the United States, that utterance is true.

Importantly, such utterances of (1) are not modal utterances that attribute a modal property (like would-be-bisexual) to Alcibiades. According to the analysis I am proposing, the default evaluation of claims about the past does not discriminate between attributions of C-features and attributions of S-features. So most utterances of (1) are no more modal than are utterances of an “ordinary” statement like “Alcibiades was a member of the species homo sapien”. Both are evaluated by holding fixed to the dominant carving in the context of utterance, even if that carving bears little resemblance to the carving that was dominant in Alcibiades’s social context.

This analysis of projection holds generally. In most contexts, an utterance of a claim that employs a socially contingent feature should be evaluated according to the social practices present in the context of that utterance. The time, place, or universe under discussion is irrelevant. Thus, we can speak intelligibly about bisexuality in ancient Greece, in ancient China, or in medieval England. We can even speak intelligibly about bisexuality in the completely fictional land of Zomnia, though that will require us to evaluate worlds that combine the actual carving with merely possible substrata.

Earlier, I said many people simply assume that if a category is socially constructed then it cannot be applied to individuals in a society that does not recognize the category. That claim rules out the analysis that I’ve developed in this section. Interestingly, I think the two-fold notion of possibility can be used to construct a substantive argument for their assumption. While I think the argument ultimately fails, it is initially plausible and it does arguably capture what people who present the conceptual version of the problem of projection may have had in mind.

A diagonal world is a world constructed from a substratum and the carving that it supports. The actual world is among the diagonal worlds, since it is constructed from the actual substratum and the actual substratum -- by definition -- supports the actual carving. Also among the diagonal worlds is Stein’s land of Zomnia. Its substratum contains a society of people that divide themselves into those who sleep on their back and those who sleep on their front; its carving is the carving that accurately models that society’s concept of sleep orientation. Diagonal possibility is a restriction of social possibility that considers what is true at these diagonal worlds. In general:

Diagonal Possibility:♢ d φ is true at a world w1 = <<u1, r1, t1>, c1> if and only if there is a substratum s2 = <u2, r2, t2> and carving c2 such that (i) s2 supports c2 and (ii) φ is true at w2 = <s2, c2>

Many substrata are devoid of conceptual practices and therefore support no carvings. For such substrata, what they make diagonally possible is nothing beyond what is possible simply with respect to their S-features.

Methodologically speaking, any substratum can be paired with any carving. But we are naturally interested in worlds that have substantive connections between their substrata and their carvings. In some historical investigations, we might be especially interested in the carvings supported by the historical society in question. Those who object to projection seem to turn this interest into an obsession. Thus, for them, any claim made about the past must be interpreted as a claim about the diagonal past:

Diagonal Past Possibility:♢ dp φ is true at a world w1 = <<u1, r1, t1>, c1> if and only if there is a substratum s2 = <u1, r2, t2> and carving c2 such that (i) t2 < t1, (ii) s2 supports c2 and (ii) φ is true at w2 = <s2, c2>

So here’s the argument. In every context, an utterance about the past must be analyzed according to Diagonal Past Possibility. For instance, the relevant substratum of (1) is ancient Greece, and so the relevant carving must be the carving that was supported by ancient Greek society. When social possibility is so restricted, it follows that we cannot intelligibly talk about socially contingent features that were not supported by that society. Because projections are by definition about socially contingent features that were not supported by the society in question, we can conclude that all projections are false by definition. And, surely, if projections are false by definition then they are conceptually problematic.

But the real problem lies with those who restrict social possibility in this way. I concede that some utterances of statements like (1) should be interpreted as claims about the diagonal past. Historians, in particular, may be interested in these kinds of claims. But I deny that all utterances should be so interpreted. We regularly entertain claims that deviate from diagonal space. In fact, I think that the interpretation provided by Past Possibility is the default interpretation for utterances of statements like (1). It is only in special contexts that they should be interpreted according to Diagonal Past Possibility.

To fully substantiate my interpretive hypothesis would require extensive empirical work. But the following three examples suggest that the burden of proof is on those who would limit all interpretations of statements like (1) to Diagonal Past Possibility. First, consider the following series of counterfactual claims about social reality:

  1. If Ellen were to move to Taiwan, she would not be a resident of Los Angeles County.

  2. If Ellen were not disposed to feel sexually attracted toward and engage sexually with women, she would not be lesbian.

  3. If we were to change our laws, conventions, and practices such that the conditions for being a resident of Los Angeles County required living in the city of Los Angeles, Ellen would not be a resident of Los Angeles County.

  4. If we were to change how we conceived of sexual orientation, such that the condition for being a lesbian required having never dated a man, Ellen would not be lesbian.

Each of these claims considers what would be true if things had been different along only a single axis of social possibility. (2) and (3) are countersubstratums. Intuitively, each is evaluated by holding fixed the carving supported in the context of utterance and determining if the consequent is true at the closest worlds where its antecedent is true. (4) and (5) are counterconventionals. Intuitively, each is evaluated by holding fixed the substratum of the context of utterance and determining if the consequent is true at the closest worlds whose carvings make its antecedent true. All four of these intuitive interpretations deviate from diagonal space. Thus, if the interpretive hypothesis that I reject were correct, these interpretations would be as unacceptable as my interpretation of projections. But it seems to me that these interpretations are perfectly ordinary and perfectly intelligible. At any rate, these interpretations seem more ordinary than those demanded by the restriction to diagonal possibility. Yes, in some contexts, we may consider how people would react to changes in the law. Thus, it may be true that a change in residency law would motivate Ellen to move. But in just as many contexts we consider how different laws would apply to the same material circumstances. What we say in these contexts would therefore be true because Ellen lives in Beverly Hills, not the city of Los Angeles. Similarly, in many contexts an utterance of (5) would be true because Ellen, like many lesbians, had dated a man early in her life.

Second, the diagonal restriction on modal space fits poorly with a wider range of claims about C-features. Consider the claim

  1. At some time in the past, no one spoke English.

In most contexts, someone who utters (6) would be saying something that seems obviously true. On my proposed interpretation, it’s clear how. English is a C-feature; it is the product of social practices and is not a feature of the mind-independent substratum. I claim that my utterance of (6) should be interpreted as applying the carving supported in my context of utterance. This carving establishes the conditions used for evaluating when someone speaks English. Of course, there is some time in the past where no one meets those conditions. So, on my interpretation, my utterance of (6) is true. In contrast, suppose that all utterances of (6) must be evaluated with respect to diagonal space. Then (6) would be evaluated with respect to substrata in the past and the carvings that they support. Presumably, though, the substrata in the past where no one speaks English are precisely those substrata whose carvings lack English as a C-feature. No one in ancient Greece spoke English and ancient Greek society lacked a concept of English. So it seems that the restriction to diagonal modal space makes (6) just as unintelligible as the supposedly unintelligible claims about sexual orientation. But that’s absurd. We regularly talk about the absence of some specific socially contingent feature.

Third, if social possibility were restricted to diagonal possibility in every context of utterance, then negative claims about social features that are alien to the society in question would be no better than their positive counterparts. Consider my utterance of a statement like:

  1. In ancient Greek society, no one was gay, straight, lesbian, or bisexual.

On my interpretation, my utterances of (7) is false. It is false because it is evaluated with respect to the carving supported in my context of utterance, the LGB carving I articulated in Section 3.1. According to that carving, there were such people in ancient Greek society. But, following the diagonal interpretation, my utterance should be evaluated according to the carving supported by ancient Greek society. But that carving does not recognize the LGB sexual orientations. Thus, on the diagonal interpretation, utterances of (7) are also unintelligible. Perhaps the diagonal interpretation could be modified in a way that makes utterances of (7) intelligible. Perhaps, for example, negative claims about the past are evaluated according to the carving supported in the context of utterance even though positive claims are evaluated according to the carving supported in the past. But that bifurcated interpretation strikes me as unmotivated.

The supposed problem of projection rests on an implicit restriction on social possibility. But that restriction is not generally adhered to. Because social possibility involves much more than diagonal possibility, the projection of a socially constructed sexual orientation is not conceptually problematic.

4.4 Epistemic Problems with Projection

There is no conceptual problem with projection. But that doesn’t mean projection is problem free. In fact, projection raises a host of epistemic issues.[25] These issues are especially challenging when it comes to sexual orientation.

One such issue concerns counterfactuals regarding our behavior across social contexts. Even assuming that we know an individual’s sexual orientation in their own social context, how can we know what their orientation would be in another?[26] If Alcibiades lived in the United States of America today, would he still be bisexual? If Ellen lived in ancient Greece, would she still be gay? It’s hard to say. Sexual behavior is not modally robust. While who someone sleeps with is likely in part determined by their socially-independent desires, who they sleep with is undeniably shaped by their social context. Thus, it is difficult to determine the extent to which an individual’s desires are their own rather than a reflection of their society. Insofar as these desires indicate their underlying sexual dispositions, there is a similar difficulty for sexual orientation.

Actually, though, this is a difficulty even for the essentialist. Let’s suppose that sexual orientation is not socially constructed. That doesn’t change the underlying facts that give rise to this difficulty. Who someone sleeps with will still be shaped by their social context. Everyone agrees on that. Of course, these variations in sexual behavior may not indicate a corresponding change in sexual orientation. Perhaps the relevant stimulus conditions for their sexual dispositions were not met in one or either social context. The variation in sexual behavior marked by an oath of celibacy does not indicate a corresponding change in sexual orientation because the stimulus conditions for sexual dispositions are ordinarily not met within oaths of celibacy. The same may be said for social contexts where, for example, there are strong social sanctions against some kinds of sexual behavior. But these considerations are a matter of the causal construction of sexual orientation. They are not a matter of its constitutive construction.

Another epistemic issue concerns the ways that projection may be a harmful act in an epistemically instrumental sense. When we apply our socially constructed categories onto people to whom those categories would be alien, we distort our understanding of their lives.[27] Suppose that the claim that Alcibiades was bisexual is meaningful and suppose we even know that it is true. Considered on its own, then, the claim is epistemically valuable. But it may nevertheless lead us astray. When we know the sexual orientation of someone living in the United States today, we are likely to make inferences on the basis of that knowledge. Those inferences may be reliable in a contemporary context. But they are not universally reliable.[28]

Here are two examples. First, we may infer from someone’s sexual orientation that they have certain romantic desires. Many gay people in the United States are not only disposed to feel sexual attraction toward people of the same gender, but are also disposed to feel romantic attraction toward them, with some social expectation that this will lead to the pursuit of long-term partnerships like marriage. This connection need not hold universally. In fact, in many societies it did not. Thus, by coming to believe that a historical figure was gay, we may incorrectly infer that they were unhappy in their “traditional” marriage. Second, our reliance on using our own contemporary carving of sexual orientation may result in our failure to recognize the alternatives. The ways in which other societies have institutionalized their sexual lives are extremely diverse, at times incorporating features like age, social status, sexual acts, and roles. Such features may seem to many people in the United States to be entirely irrelevant -- or, at least, explanatorily subservient to gender. But that says more about a society’s inability to recognize carvings that differ from its own. Again, though, this is not a difficulty unique to the constructionist view. Our indiscriminate tendency to categorize historical figures as gay or straight may blind us to the ways that they categorized themselves. But that remains true even if our categories are mind-independently “correct”. Suppose that our contemporary understanding of certain psychological disorders “carves at the joints”. The projection of our contemporary understanding to the behavior of ancient Greeks could nevertheless lead to a distorted understanding, especially if we were ignorant of how they conceptualized these issues. Furthermore, the essentialist can justify the value of our projections only on the supposition that our projections succeed in employing categories that “carve at the joints”. If it turned out that they did not, then the essentialist would be no better off than the constructionist with respect to the distorting effects of projection.

As I’ve suggested, there are a multitude of reasons that we may consider the sexual orientation of historical figures. Perhaps when it comes to some historical projects, we have good reason to restrict our attention to diagonal worlds. But there are other, equally legitimate, projects that should not be so restricted. Consider, for example, the historical analysis of class. Surely such an analysis need not limit itself to the conceptual practices of whatever society happens to be the target of analysis. Rather, there is genuine value in using our own conceptions of class to better understand the world. So, too, for the historical study of sexual orientation.[29]

4.5 Alternative “Two-Factor” Models

I would like to end by briefly discussing the role that my substratum-carving model plays in solving the general conceptual problem of projection.

I think that the model is particularly well-suited for the analysis of modal claims about socially constructed features. I also think that the model has advantages over other models with respect to metaphysical issues that are orthogonal to the issues I’ve discussed in this paper.

That being said, I think the problem of projection can be solved using other models. Abstractly speaking, any model that enables a “two-factor” account of social features should be able to develop an analysis of projection analogous to the one I develop in this paper. A “two-factor” account is an account that distinguishes between (a) the social conditions that determine when a particular social feature is present and (b) the social conditions that determine the rules for determining when a particular social feature is present. By distinguishing between these two factors, the model can be used to construct a two-dimensional modal space. That two-dimensional modal space is what ultimately salvages the conceptual coherency of projection.

In the substratum-carving model, this distinction comes from the difference between a social carving and the substratum to which it is applied. I suspect that many other models of social reality can also make the distinction in their own ways, though I cannot hope to give a comprehensive survey of how. Still, here is an example to illustrate what I have in mind. The model of social reality developed in Epstein (2015) distinguishes between the grounding relations that hold between particular social facts and their grounds and the anchoring relations that hold between rules -- frame principles, in Epstein’s terminology -- and the social conditions that establish those rules. Epstein uses this grounding-anchoring model to construct a two-dimensional modal space that recombines possible grounds of social facts with possible frame principles (82–84). While Epstein does not directly address the problem of projection as it relates to sexual orientation, he clearly rejects the more general claim that, as I put it, Diagonal Possibility is the only way to evaluate modal claims about social reality. Thus, though his dialectical focus is elsewhere, Epstein says that “we can look back at ancient societies, and evaluate whether there are classes or castes, aristocrats or serfs [… and] we can sensibly ask whether Caligula was a war criminal” (123–124).[30]

I think the fact that the substratum-carving model is able to provide an adequate analysis for the conceptual coherence of cross-cultural projection serves as a reason to think that the model is a good model for social reality. I would think the same of any other two-factor model that can be shown to do something analogous. But I admit that this conclusion is shaped by my pre-theoretic intuition that cross-cultural projection seems conceptually coherent. I suspect that adamant opponents of projection will take the analysis I develop as an indirect argument against the model – or, at any rate, against the parts of the model that are used in the analysis. Such is life. My main aim in this paper was to provide a defense of projection in the same way that someone may provide a defense of knowledge against the arguments of the epistemological skeptic. Such a defense need not satisfy the skeptic; but it should provide a rational justification for what is being defended.[31]

5 Conclusions

The constructionism/essentialism debate does not matter when it comes to the problem of projection. The one way in which it might -- with respect to its conceptual intelligibility -- relies on an unmotivated restriction on social possibility. That being said, there are legitimate epistemic difficulties regarding our ability to understand the sexual lives of culturally distant people. But these difficulties would remain even if essentialism were true and sexual orientation were a natural, objectively significant, characteristic of every human of every culture.

Corresponding author: Peter Finocchiaro, School of Philosophy, Wuhan University, Wuhan, P. R. China, E-mail:


I’m grateful to the many people who helped in the development of this paper from its inchoate beginnings to now: Andrew Brenner, Rebecca Chan, Ben Cross, Duen-Min Deng, Kate Finley, Liz Jackson, Mari Mikkola, Ron Mallon, Sam Murray, Timothy Perrine, Katherine Ritchie, Jessica Wilson, the audience at my session of the Meaning and Reality in Social Context conference, and the Institute of European and American Studies at Academia Sinica. Thanks to the two anonymous reviewers and the editors at the Journal of Social Ontology for their patience and feedback. Thanks, especially, to Sally Haslanger for useful feedback and for bringing Einheuser’s brilliant work to my attention. Finally, I owe an intellectual debt to Robin Dembroff, who years ago opened my eyes to what Metaphysics could be.


Ásta. 2018. Categories We Live By: The Construction of Sex, Gender, Race, and Other Social Categories. Oxford: Oxford University Press.10.1093/oso/9780190256791.001.0001Search in Google Scholar

Bostock v. Clayton County. 2020. 590 U.S. Supreme Court.Search in Google Scholar

Boswell, J. 1989. “Revolutions, Universals, and Sexual Categories.” In Hidden from History: Reclaiming the Gay and Lesbian Past, edited by M. B. Duberman, M. Vicinus, and G. Chauncey Jr., 17–36. London: Penguin Books.Search in Google Scholar

Corvino, J. 2013. What’s Wrong with Homosexuality? Oxford: Oxford University Press.10.1002/9781444367072.wbiee570Search in Google Scholar

Dembroff, R. 2016. “What is Sexual Orientation?” Philosophers’ Imprint 16: 1–27.Search in Google Scholar

Dembroff, R. 2018. “Real Talk on the Metaphysics of Gender.” Philosophical Topics 46: 21–50, in Google Scholar

Díaz-León, E. “Sexual Orientations: The Desire View.” In Feminist Philosophy of Mind, edited by K. Maitra, and J. McWeeny. Oxford: Oxford University Press (forthcoming).Search in Google Scholar

Díaz-León, E. 2020. “Relativism and Race.” In The Routledge Handbook of Philosophy of Relativism, 265–71. New York: Routledge.10.4324/9781351052306-29Search in Google Scholar

Einheuser, I. 2006. “Counterconventional Conditionals.” Philosophical Studies 127: 459–82, in Google Scholar

Einheuser, I. 2009. “Some Remarks on ‘Language-Created Entities’.” Acta Analytica 24: 185–92, in Google Scholar

Einheuser, I. 2011. “Toward a Conceptualist Solution of the Grounding Problem.” Noûs 45: 300–14, in Google Scholar

Epstein, B. 2014. “How Many Kinds of Glue Hold the Social World Together?” In Perspectives on Social Ontology and Social Cognition, Volume 4 of Studies in the Philosophy of Sociality, edited by M. Gallotti, and J. Michael, 41–56. Netherlands: Springer.10.1007/978-94-017-9147-2_4Search in Google Scholar

Epstein, S. 2015. The Ant Trap: Rebuilding the Foundations of the Social Sciences. Oxford: Oxford University Press.10.1093/acprof:oso/9780199381104.001.0001Search in Google Scholar

Finocchiaro, P. 2021. “High-Fidelity Metaphysics: Ideological Parsimony in Theory Choice.” Pacific Philosophical Quarterly: Online Early View.10.1111/papq.12349Search in Google Scholar

Foucault, M. 1978. The History of Sexuality, Volume I. New York: Random House.Search in Google Scholar

Glasgow, J. 2007. “Three Things Realist Constructionism About Race---or Anything Else---Can Do.” Journal of Social Philosophy 38: 554–68.10.1111/j.1467-9833.2007.00398.xSearch in Google Scholar

Good, C., J. Arsonson, and J. A. Harder. 2008. “Problems in the Pipeline: Stereotype Threat and Women’s Achievement in High-Level Math Courses.” Journal of Applied Developmental Psychology 29: 17–28, in Google Scholar

Hacking, I. 1999. The Social Construction of What? Cambridge: Harvard University Press.10.2307/j.ctv1bzfp1zSearch in Google Scholar

Halperin, D. M. 1990. One Hundred Years of Homosexuality: And Other Essays on Greek Love. New York: Routledge.Search in Google Scholar

Haslanger, S. 2012. Resisting Reality: Social Construction and Social Critique. Oxford: Oxford University Press.10.1093/acprof:oso/9780199892631.001.0001Search in Google Scholar

Hinsch, B. 1992. Passions of the Cut Sleeve: The Male Homosexual Tradition in China. Berkeley: University of California Press.Search in Google Scholar

Katz, J. N. 1990. “The Invention of Heterosexuality.” Socialist Review 20: 7–34.Search in Google Scholar

Khalidi, M. A. 2015. “Three Kinds of Social Kinds.” Philosophy and Phenomenological Research 90: 96–112, in Google Scholar

LeVay, S. 2017. Gay, Straight, and the Reason Why, 2nd ed. Oxford: Oxford University Press.Search in Google Scholar

Lewin, E., and W. L. Leap, eds. 2002. Out in Theory: The Emergence of Lesbian and Gay Anthropology. Urbana and Chicago: University of Illinois Press.10.5406/j.ctvvng2rSearch in Google Scholar

Lewis, D. 1979. “Attitudes de dicto and de se.” The Philosophical Review 88: 513–43, in Google Scholar

Littman, R. J. 1970. “The Loves of Alcibiades.” Transactions and Proceedings of the American Philological Association 101: 263–76, in Google Scholar

Mallon, R. 2004. “Passing, Traveling, and Reality: Social Constructionism and the Metaphysics of Race.” Noûs 38: 644–73, in Google Scholar

Mallon, R. 2016. The Construction of Human Kinds. Oxford: Oxford University Press.10.1093/acprof:oso/9780198755678.001.0001Search in Google Scholar

McIntosh, M. 1968. “The Homosexual Role.” Social Problems 16: 182–92, in Google Scholar

Miller, D. I., A. H. Eagly, and M. C. Linn. 2015. “Women’s Representation in Science Predicts National Gender-Science Stereotypes: Evidence from 66 Nations.” Journal of Educational Psychology 107: 631–44, in Google Scholar

Millikan, R. G. 2014. “Deflating Socially Constructed Objects: What Thoughts Do to the World.” In Perspectives on Social Ontology and Social Cognition, Vol. 4, edited by M. Gallott, and J. Michael, 27–40. Netherlands: Springer.10.1007/978-94-017-9147-2_3Search in Google Scholar

Ortiz, D. R. 1993. “Creating Controversy: Essentialism and Constructivism and the Politics of Gay Identity.” Virginia Law Review 79: 1833–57, in Google Scholar

Ritchie, K. 2018. “Social Creationism and Social Groups.” In Collectivity: Ontology, Ethics, and Social Justice, edited by K. Hess, T. Isaacs, and V. Igneski, 13–34. London: Rowman and Littlefield Publishers.Search in Google Scholar

Robinson, M. 2017. “Two-Spirit and Bisexual People: Different Umbrella, Same Rain.” Journal of Bisexuality 17: 7–29, in Google Scholar

Root, M. 2000. “How We Divide the World.” Philosophy of Science 67: 628–39, in Google Scholar

Searle, J. 1995. The Construction of Social Reality. New York: Free Press.Search in Google Scholar

Searle, J. 2010. Making the Social World: The Structure of Human Civilization. Oxford: Oxford University Press.10.1093/acprof:osobl/9780195396171.001.0001Search in Google Scholar

Spencer, S. J., C. M. Steele, and D. M. Quinn. 1999. “Stereotype Threat and Women’s Math Performance.” Journal of Experimental Social Psychology 35: 4–28, in Google Scholar

Stein, E. 1992. Forms of Desire: Sexual Orientation and the Social Constructionist Controversy. New York: Routledge.Search in Google Scholar

Stein, E. 1999. The Mismeasure of Desire: The Science, Theory, and Ethics of Sexual Orientation. Oxford: Oxford University Press.Search in Google Scholar

Stock, K. 2019. “Sexual Orientation: What is it?” Proceedings of the Aristotelian Society 119: 295–319, in Google Scholar

Thomasson, A. L. 2003. “Foundation for a Social Ontology.” ProtoSociology 18: 269–90, in Google Scholar

Vance, C. S. 1989. “Social Construction Theory: Problems in the History of Sexuality.” In Homosexuality, Which Homosexuality?, 13–34. Amsterdam: An Dekker.Search in Google Scholar

Published Online: 2021-10-29

© 2021 Peter Finocchiaro, published by De Gruyter, Berlin/Boston

This work is licensed under the Creative Commons Attribution 4.0 International License.

Downloaded on 9.12.2022 from
Scroll Up Arrow