Varieties of Social Necessity

2.1 Social kinds, social structures, and social modality

The first reason to think that there are real social necessities is that certain claims about the social are counterfactually stable. More generally, one reason to think that a claim is necessary is that we can reasonably assume that claim remains true under a non-trivial range of subjunctive suppositions or hypothetical scenarios (e.g., see Kment 2014; Lange 2009). Plausibly, there are certain claims about social reality that are counterfactually stable in this way. To borrow a type of example commonly used in the philosophy of science regarding natural laws, consider the following:

Assuming that (1) is true, it only seems to be accidentally true in that if it had been any other day of the week or time of day, it could have turned out that very few people socially classed as women took the train to Providence. However, assuming (2) is true, it seems to be more counterfactually stable and so supports many counterfactual judgments, e.g., if more people socially classed as women had entered the tech industry in the 1980s, they would have been paid less than their equally qualified counterparts socially classed as men.

Relatedly, and second, certain claims about the social allow us to make predictions and offer explanations. In this sense, many claims about social kinds and properties seem to be similar to claims about natural kinds and properties (see Khalidi 2015; Mason 2016: 842–843). Arguably, insofar as the predictive and explanatory power of a claim about the natural world indicates it is a natural necessity, the predictive and explanatory power of a claim about social reality seems to indicate it is a social necessity. Below I will argue that this is not accidental. Jenkins (2023: 83–88) argues that social kinds are partly constituted by the way social practices constrain and enable people in patterned ways within a given social structure. In turn, I will argue, these patterns of constraints and enablements also determine the space of what is socially possible.

Third, I take the most important reason to think there are social necessities to be that certain claims about the social indicate the limits of what is socially possible. Kment (2014: 20) notes that one way of thinking about real necessity is as something that indicates a restriction on reality’s freedom of variation (see also Kment 2006). So, we have reason to think that a claim is necessary insofar as it shows a limitation to how much the world can change. Arguably, there are claims about the social history and social structure of the actual world that seem to indicate objective blocks, constraints, and resistances that exist in social reality and limit what alternative arrangements of the social are possible.⁹ I think Barnes (2017: 2423) provides an instructive metaphor:

…think of wheel ruts (the deep indentations in a dirt road made by wagon wheels). Ruts are caused by repeated patterns of travel… [O]nce there are ruts in a road, those ruts explain why wagons continue to travel the way they do, because once there are ruts in a road it’s really hard to drive a wagon anywhere but the ruts.¹⁰

Similarly, following Rosen (2006), there are arguably certain truths operating in a social context that provide the ‘combinatorial essence’ of the social within that context, i.e., what the basic elements of the social are and the limits on how those elements can be combined. For example, the 2023 movie Barbie depicts Barbieland where women hold all of the social and political power. However, I suggest that insofar as (2) reflects deeply entrenched facts about how social relationships are structured on the basis of unjust gender kinds, it indicates that the social and political arrangements portrayed in Barbieland are either not a nearby social possibility or not a social possibility at all with respect to actual social contexts. Below, building on recent work in Burman (2023), Jenkins (2023), and Ross (2024a, 2024b), I will further argue that one way to understand the space of social possibility as something distinct from the space of metaphysical possibility is in terms of social systems and the structural constraints that follow.

2.2 Initial worry and replies

Before moving on, I would like to consider one objection to the existence of social necessities based on the debate over whether there are laws in the social sciences. Some argue that there are no laws in the social sciences as follows. Laws are supposed to be exceptionless – e.g., no two naturally possible bodies attract one another according to an inverse cube law. However, most of the generalizations made by social scientists seem to admit of exceptions. Thus, some argue there are no laws in the social sciences (e.g., see Roberts 2004). Similar issues might arise when we ask about whether any truths in social inquiry broadly construed are necessary. One might argue that necessary truths are exceptionless, yet it seems that many explanatory regularities in social inquiry, which would be good candidates for social necessities, are not exceptionless. Thus, the worry could go, there are no necessary truths in social inquiry.

For example, consider:

On the one hand, (3) seems to indicate certain constraints on what is economically possible – e.g., it seems that a retailer will not be able to successfully increase the price of some commodity if demand remains constant but manufacturers have recently been able to produce more of that commodity without an increase in overhead. On the other hand, it is well noted that claims such as (3) readily admit of exceptions, e.g., the existence of federal regulations that implement a price floor on the commodity. So, to borrow the metaphor of possible worlds, one could argue that there is not a shortage of possible worlds that show (3) is false and so it makes no sense to claim that (3) is necessary.¹¹

I cannot fully address this worry here, and I am not primarily concerned with whether supposed laws in the social sciences are metaphysically necessary or necessary in some more restricted sense. However, I think there are a number of promising responses to this kind of worry. First, Kincaid (2004) suggests that laws in the social sciences are generalizations over forces. Note we might further argue that, insofar as the powers of supply and demand to influence price are essential to the property of price, and since the property of price figures into the law of supply and demand, (3) is a social necessity, if not also a metaphysical necessity (cf. Ellis and Lierse 1994; Shoemaker 1998). But, for now, Kincaid responds to worries that the laws in the social sciences are not exceptionless by arguing that this is not a real issue since proposed laws such as (3) are strictly about the influence of supply and demand on price and does not address other potential influences or forces, e.g., political interventions or mass hallucinations.

Second, even if claims like (3) are not exceptionless, there are ways to hedge them, e.g., increases in supply will typically or will likely cause a decrease in price, and this hedged claim will hold for a non-trivial range of possible worlds. Necessity, like inviolability, arguably comes in degrees (see Kincaid 2004: 173; Kment 2014). Presumably, there is some set of possible worlds where social regularities like (3) hold without exception. These seem to be the explanatorily relevant possible worlds in social inquiry. So, we could argue that this space of possible worlds indicates the limits of economic possibility. Relatedly, even if probabilities are involved, we might think that the objective frequencies or probabilities involved in certain claims themselves hold of social necessity. Furthermore, in this case, to say that a social necessity is metaphysically contingent would be to say that there are metaphysically possible worlds, not where there are improbable exceptions, but where certain objective frequencies in social reality do not converge with the actual probabilities contrary to what is socially possible (cf. Sidelle 2002: 312–313).

Finally, it is important to note that many potential social necessities are often claims about the existence and nature of social structures, or systems, rather than generalizations over individuals.¹² For example, taken as a universal generalization, someone can try to argue that (2) is false by pointing to examples of individual people socially classed as women in heteronormative societies who occupy dominant social positions, e.g., former Prime Minister Angela Merkel or Oprah Winfrey. But arguably, such examples miss the point in that social inquirers are not looking for exceptionless definitions of social kinds or properties but ontic structures that explain the reproduction of certain social regularities. For example, following Cohen (1983), a potential social necessity is given by

I suggest that, if true, Cohen’s claim reflects a constraint that limits the space of social possibility. Furthermore, the entrenched nature of (4) reinforces various forms of oppression such as social and economic marginalization and exploitation. Cohen critically addresses arguments against (4) that point to the existence of individual proletarians who manage to build enough wealth to ‘escape’ the proletariat and so are not forced to sell their labor. His response is that the oppression of the proletariat is collective – while individual proletarians might be free, the proletariat as a whole is unfree. That is, class shouldn’t be thought of in terms of universal generalizations over individuals since it is a structural phenomenon that constrains potential outcomes of an economic system. For our purposes, the point is that many social necessities are not best interpreted as exceptionless generalizations but as claims about social structures and social kinds.

4.1 The etic theorizing argument

I want to first look at a formal argument sketched at the end of Raven (2024) that also seems to show that social necessitarianism is false. Raven’s main objective is to show we can explain the application of actual social notions to non-actual scenarios – i.e., etic theorizing – without appealing to the notion of anchoring. Raven is noncommittal about etic theorizing but argues that using actual social notions as ‘universal tools,’ as a tools that can be applied to non-actual scenarios, means that certain claims involving those notions, namely social rules, must be necessary in some sense, e.g., that acts of atrocity against humanity are war crimes. However, Raven suggests that whatever necessity is involved in social rules it cannot be the same as metaphysical necessity. This is because metaphysical necessity adheres to Axiom 4 of S5 – that is, □φ → □□φ. But the grounds for social rules are themselves contingent. For example, the Geneva Conventions may have never been enacted, and so it seems relative to that possibility it would not be necessary that committing an act of atrocity against humanity would be a war crime. Raven goes on to say

Social rules are often associated with conventions. But they may be especially alike in sharing a distinctive kind of modality. If so, the metaphysical import of etic theorizing is its highlighting of this distinctive kind of modality at work in social reality (98)

It follows that social necessitarianism is false.

I am sympathetic to this argument, but I do not think it is quite enough to show that social modality is distinct from metaphysical modality. Importantly, many who give an account of metaphysical necessity in terms of some form of conventionalism still take metaphysical necessities to be necessary, i.e., to adhere to S5. As is well-noted in the literature on modal conventionalism, we have to be careful and distinguish between the modality of a claim and the modality of the grounds of that claim (see Einheuser 2006; Finocchiaro 2021; Kocurek et. al 2020; Sidelle 2002; Thomasson 2020b). While the grounds of a necessity – e.g., certain conventions that ground social rules – might be contingent, it doesn’t follow that the necessity itself is not necessary. In other words, we must distinguish between counterconventionals and counterfactuals. In the latter case, we consider changes in the actual facts. In the former case, we consider changes in conventions, which themselves are certainly not necessary. But in etic theorizing, even if we consider a possible world with different conventions, e.g., a world where committing an act of atrocity is not considered a war crime within that world, we can still evaluate what is true at that world based on our social rules, which Raven claims the etic theorist must assume is necessary in some sense. Since the etic theorist is assuming the necessity of (5), i.e., that committing acts of atrocity against humanity is necessarily a war crime, there is no reason to assume that the necessity itself is not necessary. This is because (5) is about the factual relations between war, atrocities, and humanity – it is not about the enactment of the Geneva Conventions, which is contingent. On the other hand, if we wanted to consider how social reality might be differently carved from the non-social, e.g., the legal consequences of changing the definition of a war crime, we would engage in counterconventional reasoning rather than counterfactual reasoning. So, more needs to be said about what distinguishes social modality from metaphysical modality.

4.2 The conceivability argument

I will now provide a conceivability argument against (N1). However, it is important to keep two kinds of modal claims, and thus two kinds of conceivability arguments, distinct. One kind of social necessity will primarily reflect a socially necessary correlation or structural connection between various things, properties, or kinds. Another kind of social necessity will purport to reflect something about the nature, identity, or essence of some social thing, property, or kind, e.g., to be a φ is to be a ψ. For example, recall the following from above:

Plausibly, this claim can be true without being read as stating the nature or identity of what it is to be a woman. Yet this need not diminish the real necessity of (2), which, if true, not only points to a social category or kind of interest to social inquirers, but also points to deeply entrenched social patterns that obstruct social changes (and so what recombinations of the social are possible) as well as supports a non-trivial range of counterfactual judgments. I will address the second kind of social necessity, potentially dealing with essences, in the next section.

With this distinction in place, one might argue it is relatively easy to show that some social necessities are not metaphysically necessary since it is relatively easy to imagine metaphysically possible scenarios where an alleged social necessity fails. First, suppose that (2) is a social necessity. We can arguably imagine a heteronormative society where most people socially classed as women occupy privileged social positions – presumably, Barbieland is an example of this. Thus, (2) is not metaphysically necessary.¹⁸ To consider another example, suppose that there is a socially necessary connection between being socially classed under a certain gender category and having a certain gender identity (cf. Jenkins 2016). However, there doesn’t seem to be anything incoherent in imagining a metaphysically possible world where people have no personal attitudes towards gender expectations and yet people are still socially classed as women, men, or gender non-conforming at that world. Setting aside the issue about whether social position or identification with a social position gives us the nature or identity of gender kinds, if either, there doesn’t seem to be anything metaphysically incoherent in supposing that the above correlations fail to hold even if they are socially necessary.

To give one more example, recall (4), which states that the proletariat are forced to sell their labor while being marginalized from other productive forces. Arguably, this reveals something about the essence, identity, or nature of being a proletarian – in which case, there might be something incoherent in conceiving of proletariat who are not collectively forced to sell their labor while being excluded from other productive forces. In other words, (4) is arguably a thin modal claim like (5). But consider the following claim:

Suppose that (7) is not only true but is also a social necessity in that it reflects, among other things, deeply entrenched forms of socio-economic oppression that limit what arrangements of the social are (ontically) possible. It seems that (7) is a thick modal claim. One might still argue that it is possible to conceive of a metaphysically possible world, perhaps even a utopia, where all productive forces are privately owned and exchanged in a free market and yet there is little to no socioeconomic oppression. Presumably, Nozick (1974) assumes this is metaphysically possible. Of course, in this possible world, the proletariat do not exist insofar as (4) tells us what the nature of being a proletarian is, i.e., it is metaphysically impossible for a proletarian to exist and not be oppressed. But unless we take oppressive relations of production to be metaphysically constitutive of a capitalist economic system, (7) is a metaphysically contingent social necessity.¹⁹

Thus, it seems that (N1) is false – not all social necessities are metaphysically necessary. Before moving on to objections, I want to utilize Fine (2002: 261–265) to develop an argument against the even stronger claim that social modality can still be explained in terms of metaphysical modality, i.e., develop an argument against (N2). The basic idea behind (N2) is that – even if not all social necessities are metaphysically necessary – what makes a claim socially necessary is that the claim metaphysically supervenes on, or is metaphysically entailed by, basic social claims about how social kinds and properties are distributed across possible worlds. The main worry with (N2) comes with trying to explain the social necessity of the basic social claims themselves. The problem here is that the modality of these basic claims will be left unexplained since every basic social claim will either metaphysically entail, or supervene on, itself and so we will be left without a non-trivial explanation of the social necessity of these claims. Moreover, and I will return to this argument below, metaphysical necessity is typically taken to be universal and invariant in a way that, given what I argued above, social necessity is not. Unlike the space of metaphysical possibility, the space of social possibility that is relevant for one project may not be theoretically adequate or useful with respect to another. This means understanding the necessity of a ‘basic’ social claim, especially modally thick claims, will typically require inquiry into, and potentially direct involvement with, the historical, empirical, and causal specificities of whatever social system is relevant in the context of social inquiry.

5.1 Objection from critical theorizing

The objection from critical theorizing starts with the idea that what we can conceive and what appears to us to be possible partly depends on our conceptual competence. But our apparent concepts about the social do not reliably reflect objective truths about social reality or even the operative concepts that actually structure our thought about the social.²⁰ So, there is always a risk that the concepts driving our modal intuitions about what is possible with respect to the social in fact reflect social unrealities (imaginaries) based on socially dominant perspectives that fail to accurately reflect social reality.²¹ Thus, social inquiry requires that we take a critical perspective on dominant concepts that shape our understanding of the social world, which means we often need to be skeptical of our judgments towards counterfactual scenarios involving the social.

I think all of these considerations are true, but they do not show that conceivability fails to be a guide to what is metaphysically possible about social reality. Suppose that it is a social necessity that gender relations are structured in a way that is unjust. This means that there is no socially possible world where most people socially classed as women occupy dominant social positions. One interesting consequence is that the space of social possibility, i.e., what is socially possible relative to a given social system of the actual world, is unjust. That is, similar to Jenkins’ (2020, 2023) notion of ontic injustice, I am suggesting there is a modal injustice here such that what is socially possible for certain individuals is determined by a set of social constraints and enablements that is wrongful to them (cf. Jenkins 2023: 70–71).²² Yet once we assume this critical perspective, we can easily conceive of metaphysically possible worlds that are more just, e.g., worlds where genders exist and yet sexed and gendered bodies are treated equitably, worlds where trans inclusivity is a factual default, or even worlds where all gender expectations have been completely destabilized.²³ Further, it follows that social and political activism is not only aimed at changing how social reality is but also what is socially possible. So, the above objection to the conceivability argument does not save the claim that all social necessities are metaphysically necessary. What the objection shows is that understanding both social and metaphysical modal spaces cannot rely on unreflective modal intuitions based on ordinary conceptual competence. Instead, we must adopt a critical perspective when trying to understand the space of what is socially possible.²⁴

5.2 Objections from a posteriori social necessities and real essences

One could still raise worries for conceivability arguments by appealing to the notions of a posteriori necessities or real essences. Of course, many claims about real essences will also be a posteriori, but I want to first talk about a posteriori necessities that do not necessarily reflect the real essence of anything.²⁵ For example, returning to (2), one could argue that the necessary connection between gender kinds and properties and hierarchical social, political, and economic relationships is an a posteriori metaphysical necessity. In turn, this is supposed to completely undermine any appeal to modal intuitions based on conceivability in the way that scientific discoveries supposedly show that conceivability cannot be a guide to what is metaphysically possible even if we think we can imagine situations where water is XYZ or where objects with mass exert a force on one another inversely proportional to the cube of their distance. These situations might seem to involve violations of natural laws and necessities, but really they just involve deviant kinds or properties such as thwater or schmass.²⁶ Likewise, when we conceive of scenarios like Barbieland where most people socially classed as women occupy dominant social positions, we are not conceiving of people with actual gender properties but people with deviant schmender properties. Thus, these are not scenarios where social necessities fail to hold.

But this kind of argument faces a dilemma. Suppose that, as a matter of social necessity, there is a connection between gender classes and hierarchical social, political, and economic relationships. It follows, then, that it is socially necessary that schmender properties do not exist. To see this clearly, it is helpful to compare with natural necessities – the actual laws of nature preclude the nomological possibility of the existence of objects governed by an inverse cube law, i.e., the existence of schmass. That is, if the inverse square law is a universal law that governs the relationship between gravitational force and mass for all objects across all nomological possibilities, then there cannot be any objects that follow an inverse cube law across those possibilities. Similarly, the existence of a socially necessary connection between gender and unjust social, political and economic relationships precludes the social possibility of the existence of schmender properties, i.e., there is no socially possible world where most people socially classed as women occupy dominant social, economic, and political positions (e.g., Barbieland is not a socially possible world).²⁷

However, either worlds where people socially classed as women occupy dominant social positions are metaphysically possible or they are not. On the one hand, if worlds where people socially classed as women occupy dominant social positions are metaphysically possible, then (2) is a metaphysically contingent social necessity, and so social necessitarianism is false. On the other hand, if worlds where people socially classed as women occupy dominant social positions are not metaphysically possible, then we seem to be committed to a problematic form of modal determinism about the social, in this case, about the connection between gender classes and unjust hierarchies. Suppose that, because social necessities are also metaphysically necessary, there is no metaphysically possible world where people socially classed as women occupy dominant social positions. On the face of it, this seems to endorse a problematic form of social inevitability such that no changes in the actual underlying features of socially constructed kinds seem to be metaphysically possible. This result seems to be contrary to the goals and values of many social inquirers who are not only engaged in a descriptive project of social reality but also a critical project where a major goal of social inquiry is to uncover pernicious aspects of social reality so that we may intervene on them, and by doing so not only change what is true about the social world but also change the space of what is socially possible. Alternatively, by taking social necessities to be metaphysically contingent, we allow for the possibility that certain social kinds, like race and gender, can exist without oppressive social systems and the structural constraints that follow (cf. Jenkins 2023: 187). Hence, it seems better to accept that there are metaphysically possible worlds where people socially classed as women occupy dominant social positions, in which case (2) is a metaphysically contingent social necessity.

This leaves open the question as to whether the properties instantiated in a scenario where people socially classed as women occupy dominant social positions can be considered gender properties of some kind or other. Again, some social necessities will primarily reflect necessary connections between various social entities without reflecting the nature, identity, or real essence of those entities. However, we might think that the latter is what really matters to understanding the nature of social reality (see Passinsky 2021: 943–947). In that case, one might try to argue that a claim such as (2) reflects something about the real essence of gender, e.g., gender is essentially a social position such that to be a woman or gender nonconforming is to be socially, economically, or politically subordinated according to the dominant ideology of the relevant society (see Haslanger 2012b; Barnes 2020).²⁸ Furthermore, because real essences are metaphysically necessary, there is no metaphysically possible world where gender is not fundamentally an unjust social position. Thus, (2) – taken as providing a real definition of being a woman – is not only socially necessary but is a metaphysical necessity, and so worlds with real gender properties that are like Barbieland are metaphysically impossible.

Again, if we want to allow for the possibility that certain social kinds, like race and gender, can exist without oppressive social systems and the structural constraints that follow, it seems that we should not take those oppressive features to constitute the real essence of those kinds (cf. Pagano 2024: 91–92). Of course, many disagree that (2), and related claims, provides an account of the real essence of gender properties or that there even is a singular account to begin with (Jenkins 2022). So, to give a less contentious example, consider the following:

We can elaborate on the metaphysical necessity of (8) by distinguishing the real essence of a social entity from its nominal essence. To start, it seems that the nominal essence only reflects features of the social world that are salient to us as members of the social world, while the real essence reflects the unique identity of social entities as well as (unobvious) features of the social world that are explanatorily significant. For example, one might argue that correlations between gender presentation and presumed biological sex are salient features of different genders, but these correlations only reflect the nominal essences of different genders, which in turn show up in manifest concepts of gender that structure dominant thought about gender (cf. Stoljar 2024: 390–391). However, the real essences of different genders are given by more explanatorily significant features such as social position or gender identity (or both). In turn, whatever the real essence turns out to be, this essence is what explains why there are apparent correlations between gender presentation and presumed biological sex and furthermore helps reveal what concepts are actually operative in structuring thought about gender (cf. Haslanger 2012a). Thus, while the nominal essences of different genders might enable someone to conceive of a possible world similar to the actual world where apparent genders are, say, biological properties, these cannot be actual genders. This is because the real essence of gender is social, and since essential properties are metaphysically necessary, there is no metaphysically possible world where gender is a biological property. If this is true, then at least some social necessities are metaphysically necessary, e.g., gender is necessarily socially constructed. I do not disagree with the conclusion of this argument, though I will next argue that we do not need to endorse metaphysically substantive accounts of essence or necessity in order to get this result.

6.1 Social modality and essentialism

First, metaphysical necessities are generally thought to be mind-independent and invariant in the sense illustrated by the idea that they are true across all possible worlds. However, on the account of social modality I have defended, the kinds of social necessities useful for understanding and intervening on the social are highly embedded, i.e., they are centered on historical, empirical, and causal specificities of the actual world, which themselves seem to be metaphysically contingent. The embeddedness of social modality reflects the ways in which social systems and the resulting constraints and enablements shape the space of what is socially possible. For example, Haslanger (2016) and Ross (2024a, 2024b) argue that social structures are explanatory insofar as they reflect constraints on what is possible. Haslanger (2016: 123–124) gives the example of Lisa and Larry who are equally intelligent, educated, skilled, have no prejudices about gender roles and are equally capable and willing to fulfill domestic and child rearing tasks. However, Lisa and Larry live in a community without affordable childcare, in a sexist job market with a significant wage gap between genders, and in a society where domestic work is unpaid. When Lisa and Larry have a child, Lisa ends up quitting her job to stay at home and assume childcare duties while Larry keeps his better paying job, all of which in turn puts Lisa in a more economically and socially vulnerable position. Haslanger argues that the lack of affordable childcare, wage gaps, and socioeconomic factors surrounding gender, childcare and domestic work are ‘structural constraints’ that ‘limit the possibility space’ for Lisa and Larry. Additionally, Ross (2024b: 165) discusses Haslanger’s example of Jason, who cannot afford a car and must take public transit to work. When the city makes massive budget cuts to public transit, Jason is forced to quit his job (see also Jenkins 2023: 103–105). Ross (2024b: 174) argues that

structure can be so limiting that it affords agency no choice, as it solely determines the outcome. In these cases, social structure operates as an extreme constraint making it impossible for an agent to exhibit any other outcome. Social structure is the main explanatory factor because it explains why it is impossible for the outcome to be any other way.

Importantly, these examples are not just about practical limitations on what is possible, which are arguably too diverse to be systematized into a theory of ‘practical’ modality (i.e., whether it is possible to travel to Paris from Beijing in less than four hours). Instead, these and similar examples reflect explanatorily significant limits on what is possible and are unified by facts about social systems and the resulting constraints and enablements. So, rather than explaining social modality in terms of context invariant metaphysical necessities, a better explanation is that the space of social possibility reflects facts about materially and historically situated social systems.

Moreover, this arguably puts limits on what kinds of essentialism will be useful for giving an account of social kinds. Some claims involving stronger forms of essentialism will be metaphysically robust in that they entail metaphysical necessities and involve fixed and mind-independent properties. Insofar as social kinds give rise to social necessities that are metaphysically contingent, these stronger forms of essentialism will be too strong to explain social necessities (or at least the social necessities that are modally thick). I argued above that social inquirers should not only be concerned with what is in fact socially possible but also concerned with a form of modal justice, i.e., with making the space of what is socially possible more just. This requires changing the space of what is socially possible, which requires understanding and intervening on social systems. That is, the space of what is socially possible will change with changes in material, technological, political, and historical circumstances. However, insofar as essentialist claims entail metaphysical necessities, those claims will need to reflect mind-independent and invariant aspects of the world, e.g., the property of being prime or having an electric charge. Furthermore, while stronger forms of essentialism might make sense for certain natural phenomena, e.g., natural kinds, one might worry that appealing to stronger forms of essentialism to account for social kinds involves a kind of category mistake (cf. Ryle 2000). For example, one might argue that (some) social kinds are juridical, not ‘scientific,’ in that they reflect evolving systems of social power rather than the stable kinds sought after in the natural sciences (cf. Burman 2023).²⁹ Nonetheless, there is still room for weaker forms of essentialism that are compatible with the historical and relational features of social kinds and social modality.³⁰

That said, nothing I have argued so far conclusively rules out the existence of some social necessities that are also metaphysically necessary. First, consider the case of natural necessities. Even if there is a kind of natural modality that is not fully explained by real essences but is better explained by something like the dispositions of natural properties, one could argue that these dispositions are themselves constrained by certain fact about real essences (see Vetter 2021). Similarly, even if there is a kind of social modality that is best explained in terms of social systems and the resulting constraints, the space of social possibility might be constrained by certain metaphysical necessities, e.g., that gender is socially constructed. However, even if this is the case, I will now argue that this is best explained by a kind of conventionalism (see Sidelle 1989; Sidelle 2002; Sidelle 2023) or normativism (see Thomasson 2020b) about metaphysical modality, i.e., about metaphysical essences, identity, and necessity.³¹

6.2 Social modality and modal normativism

First, there are important methodological questions that come with introducing technical notions of metaphysical necessity and metaphysical essence into social inquiry. When we introduce these technical notions into our theorizing, we need to understand what exactly we are adding to our understanding of social reality, how we are improving our ability to debunk unjust understandings of social categories, and how we are putting ourselves in a position to make progress towards social justice. If introducing notions of metaphysical necessity and essence is useful in these respects, I propose that it is not because social inquirers are trying to describe modally ‘deep’ features of social reality, which themselves cannot be known through conceptual analysis or methods of inquiry that are already available to social inquirers. Instead, if it is useful to introduce talk of essences and metaphysical necessities into social inquiry, it is because it allows social inquirers to express and make explicit the methodological and political significance of certain properties or distinctions over others.

Modal conventionalism and normativism are motivated primarily by (a) giving a practical account of the methodological function of metaphysical notions such as essence, necessity, or even fundamentality and (b) the epistemological need to explain how reliable knowledge of claims involving these notions is possible. Conventionalism and normativism take the function of these notions to be pragmatic rather than descriptive. Claims about metaphysical necessity do not function to describe worldly facts about what is possible but function to express the conventions or expectations surrounding the role of certain terms and concepts that are deemed important to inquiry.

For example, Barnes (2020: 714–716) suggests that there is a distinction between being feminized and being masculinized that tracks whether a person is subordinated or privileged on the basis of presumptions about bodily features taken to be evidence of a person’s role in biological reproduction. On the one hand, one could try to argue that this distinction reflects something about real, fundamental differences in social reality. For example, one could argue that, unlike the underlying mechanisms, the distinction between feminine behavior and masculine behavior itself is not socially fundamental, or essential to gender kinds, since the norms guiding these behaviors are highly susceptible to changes in cultural mood and time. However, this raises important questions about how exactly social inquirers know whether they have a reliable grasp on what is metaphysically necessary, fundamental, or essential about social reality. Suppose these notions are meant to reflect features of social reality that cannot be known through more straightforward methodological resources, e.g., conceptual inquiry or statistical methods, case studies, ethnographies, textual analysis, etc. In that case, social inquirers might worry whether they can ever come to know the real essence of a social kind, or which social phenomena are really more fundamental than others.

On the other hand, the conventionalist and normativist can agree that it is useful to say that Barnes’ distinction between being feminized and being masculinized reflects something more ‘essential’ or ‘fundamental’ about the social world than the distinction between, say, people who are disposed to wear makeup and those who are not. But the primary philosophical explanation for this is that this kind of language allows people to express a methodological expectation that Barnes’ distinction be used within their community of inquiry. This is because Barnes’ distinction will be more conducive to realizing the goals of social inquiry, e.g., to having a more unified understanding of the social and being better equipped to intervene on the social. For example, on a normativist account of modality, claims that gender is essentially social, that supply and demand are necessarily inversely related, that the productive forces are more fundamental than relations of production, etc. do not describe metaphysically ‘deep’ features of modal reality. Instead, these claims express rules for how to use or prioritize the explanatory concepts used in social inquiry, i.e., that social factors, the inverse relationship between supply and demand, and productive forces are all explanatorily significant (e.g., enable us to make better predictions about the social or make reliable inferences about a wider range of social phenomena that are pervasive and stable).³² Moreover, knowledge of what is socially fundamental, necessary, or essential requires nothing more than using the methods of social inquiry mentioned above, e.g., statistical methods or textual analysis, as well as doing important conceptual work (e.g., conceptual analysis or engineering).

However, the view is not that modal claims about the social world are strictly conceptual claims. For example, the normativist can agree that some social necessities are necessary a posteriori. But, on the normativist view, modal claims such as (8) – that gender is essentially socially constructed – are best explained by (a) the conceptual practices of social inquirers and (b) the empirical work done by social inquirers. For example, critical conceptual work is done to debunk common understandings of gender to show that gender concepts are in fact social concepts often hidden under the guise of natural concepts (see Haslanger 2012a; Thomasson 2020a). Then we rely on the work of activists, economists, political scientists, sociologists, etc. to fill in the empirical details behind those concepts, e.g., regarding social position or individual identifications with social norms and expectations. In addition, following Jenkins (2023: 176–179), when disagreements over the modal facts do arise, they will be resolvable through questions such as whether each party accepts the others’ explanatory purposes as legitimate (e.g., as normatively appropriate), whether parties agree about the empirical facts, or whether the parties agree about the best theoretical interpretation of the empirical facts. Importantly, there is no appeal to some discovery that only specialized knowledge of metaphysics can provide. Instead, questions about social modality are best raised and resolved through the methods and considerations already familiar to social inquirers.