Conferralism and Intersectionality: A Response to Ásta’s Categories We Live By

1 Introduction

Ásta’s (2018) book Categories We Live By is a landmark text in contemporary social ontology. It investigates human social kinds, such as gender and race, in a way that seamlessly blends metaphysics, social philosophy, social ontology, and feminist philosophy. As Ásta observes, “camping out at that intersection can be a cold and lonely endeavour” (6): each group has a different conception of what philosophy should aim to do, and there is plenty of mutual suspicion. As a fellow camper at that intersection, I know this all too well – which is why I am so heartened by the way that this excellent and exciting book manages to make that intersection a warmer and more inviting place.

I find the conferralist account of social properties that Ásta develops and defends in Categories We Live By persuasive in many ways. For one thing, it resists the temptation to simplify the social world too much, instead engaging head-on with the multiplicity and variability of social kinds. For another, the focus on constraints and enablements strikes me as philosophically important and politically helpful: as Ásta puts it, these are “categories that matter to our social life” (29).

In this paper, I make an internal critique of conferralism, arguing that conferralism could do better, by its own lights, at handling the phenomenon of intersectionality. In order to help make this point, I will first suggest a friendly amendment to the schema for conferrals that Ásta offers. This in turn will help me to explain the difficulty concerning intersectionality. Finally, I shall suggest a way of developing the conferralist account that would resolve this difficulty.

2 The Role of Constraints and Enablements in the Conferral Schema

Ásta offers a schema to illustrate how conferrals work. This schema consists of five aspects that matter in a conferral:

Conferred property: what property is conferred…

Who: who the [conferring] subjects are…

What: what attitude, state, or action of the subjects matter…

When: under what conditions the conferral takes place…

Base property: what the subjects are attempting to track (consciously or not), if anything… (21)

Here is an example of a filled-in conferral schema that Ásta gives:

Conferred property: being elected president of the US

Who: the current US vice president, as president of the US Senate; this is the entity in authority

What: the declaration that someone has received the most electoral college votes for US President

When: on January 6, following a November election, starting at 1 p.m.

Base property: the majority of electoral college votes, that is, 270 or more (23)

I want to suggest that something odd is going on with the first element of the conferral schema, the conferred property. To understand my concern, it is necessary to note that, for Ásta, a social property is a social status (49), and a social status consists of constraints and enablements (2). So one might think that the conferred property element would list the constraint and enablements that constitute the status that is the property in question. However, the examples Ásta gives of the “conferred property” component instead always given the name of the conferred property, such as “being elected president of the US”, or “being a woman”, and not the constraints and enablements that constitute the status in question. But there is a problem here: the same name can refer to different properties – different, that is, in the sense of being equated with a different status and hence consisting of different constraints and enablements.

For example, consider the following two possible worlds. In w₁, the property being the president of the US is exactly as it is in the actual world. It follows the conferral schema specified by Ásta, and involves all of the powers, duties and prerogatives with which we are familiar. In w₂, the procedure for selecting the president of the US is exactly the same as in w₁, but the presidency is exclusively a ceremonial role, involving far fewer powers, duties and prerogatives. For example, in w₁ the president has the nuclear codes, and in w₂ they do not. These are two different properties, because they are constituted by different constraints and enablements. Yet this difference vanishes in Ásta’s conferral schema: if we were to fill out the conferral schema for the property of being the president in w₂, it would look exactly the same as the schema for the property of being the president in w₁.

The problem, I think, is that under the “conferred property” element of the schema, Ásta typically gives the name of the property; whereas what we actually need to specify in order to fully understand the conferred property is the constraints and enablements that constitute the social status. It is these constraints and enablements that are different with regard to the property being the president in w₁ and w₂, for instance. To my mind, the most natural way of fixing this problem is to split the “conferred property” element of the schema into two components: (1) the name of the property (“property name”), and (2) the constraints and enablements that constitute the conferred status (“status”).^[1] The resulting schema would look like this:

Property name: the name of the property that is conferred, e.g. “being the president of the US”

Status: The constraints and enablements that constitute the social status that is the conferred property

Who: who the conferring subjects are

What: what attitude, state, or action of the subjects matter

When: under what conditions the conferral takes place

Base property: what the subjects are attempting to track (consciously or not), if anything

This expanded schema will help me to explain the difficulty that conferralism faces concerning intersectionality.

3 Intersectionality: A Brief Explanation

The concept of intersectionality, roughly speaking, captures the idea that different dimensions of oppression are not additive, simply stacking one on top of the other; rather, what may seem to be different dimensions of oppression actually interact in complex ways, and may even be impossible to fully distinguish from one another. The term was coined by Kimberlé Crenshaw, although the concept has a much longer history in Black feminist thought (Combahee River Collective 1986; Crenshaw 1989, 1991). A note about language: I will refer to kinds that require a multi-part label, such as Black women, as “intersectional kinds”, and I will refer to kinds that have a single-part label, such as women, as “single-moniker kinds”.^[2] Following Ann Garry (2011), I take intersectionality to be a “framework checker”. On this view, intersectionality as such does not constitute an account of social categories, power, agency, identity, and so on; rather, it provides a standard that theories purporting to illuminate these phenomena should meet (see also Grillo 1995).

There are at least three different insights that are included in the concept of intersectionality. First, there is the thought that oppression is non-additive, which is to say that we cannot gain knowledge about Black women’s oppression, for instance, simply by adding together some general claims about race-based oppression and some general claims about gender-based oppression. For example, suppose that, in a certain social context, we know that women are less likely to be in paid employment than men, and Black people are less likely to be in paid employment than women. An additive approach would license the assumption that Black women are especially unlikely to be in paid employment – they are unlikely “twice over”, as it were. But this may well be wrong. For example, it might be that Black women are employed at higher rates than White women and higher rates than Black men, because Black woman, unlike White women, are subject to economic pressure to take on paid work, and because there are many domestic service jobs available for which Black women, unlike Black men, are considered suitable employees.

The second, closely related thought is that oppression is non-separable: the oppression experienced by woman of colour, for example, cannot be neatly separated into oppression experienced “on the basis of race” and oppression experienced “on the basis of gender”. For example, suppose that a Black woman is presumed by the police to be a sex worker, and is arrested. Did this happen “because she is a woman”, or “because she is Black”? Non-separability highlights the misguided nature of questions such as these, which assume that the effects of race and gender categorization are neatly separable. Rejecting the fragmenting of identity represented by these questions (Grillo 1995), non-separability directs us to treat race and gender as thoroughly intermeshed and impossible to separate.

There is an intimate connection between non-additiveness and non-separability, but they are slightly different: non-additiveness cautions us against thinking that we can make predictions about intersectional kinds based on what we know about single-moniker kinds, whereas non-separability cautions us against thinking that we can use single-moniker kinds to explain the experiences of members of intersectional kinds.

The third thought, which is a rather stronger claim than the first two, is that oppressive categories involve cross-constitution or mutual construction. It is not just that race and gender interact to underpin forms of oppression that are non-additive and non-separable; rather, race plays a crucial role in constructing the reality of gender, and vice versa. In other words, gender could not be what it is if race did not exist, and race could not be what it is if gender did not exist (Lugones 2007; Garry 2011; Bernstein 2020).

I take it that cross-constitution is necessarily a metaphysical claim, whereas non-additiveness and non-separability can be interpreted either as metaphysical claims – claims about the way that oppression is – or as epistemic claims – claims about what we can know about oppression. Of these two options, the metaphysical version is stronger and places more demands on a theory. In order to have the broadest possible appeal, then, an account of social kinds should be compatible with the metaphysical versions of non-additiveness and non-separability, as well as with cross-constitution.

4 Ásta on Intersectionality

Ásta clearly intends for conferralism to be sensitive to intersectionality (6; 81; 125). In regard to the intersectional nature of social properties, she writes:

The general story is that we have features, and some of those features have social significance in a context, and the status we enjoy in a particular context is the result of the constraints and enablements that the presence of each and every one of our socially significant features brings, where the presence of some features can trump others. (125, italics mine)

The acknowledgement that “the presence of some features can trump others” conforms with the idea that oppression is non-additive, which is the first component of intersectionality. I take it that the thought here is that although a person’s overall social status is the result of all the different constraints and enablements that they are under as a result of each of their various socially significant features, this result is not a straightforward matter of adding up all the different constraints and enablements, but something much more complicated, in which different features can interact.

However, the way that Ásta talks about overall, intersectional status as the “result of the constraints and enablements that each and every one of our socially significant features brings” (125) seems to imply that there is a set of constraints and enablements that a given feature brings. For example, Ásta gives separate conferral schemas for race and gender properties. That is, she gives conferral schemas for properties such as being a woman (74–75), and being Black (99), and these conferral schemes do not make reference to one another. My worry is this: in giving separate conferral schemas for single-moniker properties in this way, even ones that are admitted to interact, Ásta appears to commit herself to the view that race and gender kinds can be neatly distinguished from one another, thereby going against non-separability, and that they exist independently of one another, thereby going against cross-constitution.

Now, it is possible that non-additiveness is the only aspect of intersectionality that Ásta accepts, and that she would be content with an account that is not compatible with non-separability or with cross-constitution. So she could in theory respond to this worry by simply rejecting the demand to accord with non-separability and cross-constitution. However, there are two reasons why I think this response would be a mistake.

Firstly, it seems to me that these further components of intersectionality are widely accepted among feminists, such that this move would seriously narrow the appeal of conferralism and should, for that reason, be avoided if possible. Secondly, the picture of separate single-moniker properties that interact (perhaps non-additively) seems to render the constraints and enablements that constitute each such property somewhat mysterious, such that there is a reason to probe further even if we were satisfied with just meeting non-additiveness. To put the point simply, it is not clear to me where the constraints and enablements that constitute a single-moniker property are meant to come from.

This point can be put more precisely by drawing on the amendment to the conferral schema that I suggested above: I am puzzled as to how we are supposed to spell out the status element of the schema for a single-moniker race or gender property. For example, if the ways in which a Black woman is constrained/enabled in a certain context are significantly different from the ways in which a White woman is constrained/enabled and the ways in which a Black man are constrained/enabled, what determines the constraints and enablements that constitute the social status Black or the social status woman? I have not been able to find a clear answer to this question in Ásta’s account of conferralism. My aim in the rest of this section, then, is to expand on Ásta’s account of constraints and enablements in order to answer this question in a way that renders conferralism compatible with all three components of intersectionality.

5 Understanding Constraints and Enablements

A good place to begin is by asking what, exactly, it means for someone to be subject to a constraint on or enablement to their behaviour in a context. Ásta states that “social categories get formed and maintained through the individual actions of classifying and placing” (125). She also makes it very clear that what matters for the creation of constraints and enablements is the actual behaviour of individual agents in social contexts, not the mental associations that the agents have concerning the members of the social status kind of the feature that conferrals are tracking (48). This means that communal properties can be conferred without the conscious awareness of those doing the conferring. I take it that if there are regularities in the actions and responses of the conferrers that are sensitive to certain properties, and if these regularities make a difference to what is possible for the relevant individuals, then we can say that there is a conferral taking place that involves the tracking of a base property (or properties). If the conferring subject is not aware of the regularities in her responses or of the role played by the base property, we can say that the conferral is implicit and the tracking is unconscious, but it is still a conferral. The exact psychological mechanisms that underpin these regularities need not be specified, nor do they need to be consistent – there may well be many different mechanisms underpinning different instances of behaviour. What ultimately matters is not what the agents think about the people or properties in question, but how they behave, because it is their behaviour that generates the constraints and enablements that constitute social status.

Once such constraints and enablements are understood in this way, it seems to me very natural to interpret them as probabilistic. For example, even in contexts where very sexist attitudes and implicit bias abound and women are routinely talked over, talked down to, and so on, it is not uncommon for a particular woman to manage not to be treated in this way on a particular occasion. But it does not follow that there is after all no constraint concerning women’s interpersonal interactions in such contexts. Rather, the conferralist can think about the constraints and enablements as probabilistic: to be under a constraint concerning behaviour B does not necessarily mean that one can never manage to do B; it may equally mean that the chances of a given attempt to do B succeeding are lower for one than they are for members of the relevant comparison group. Brännmark (2018a) offers a similar account of constraints and enablements as probabilistic, although his focus is solely on deontic constraints and enablements (social perceptions of duties, prerogatives, and so forth).

Taken together, these two points suggest that constraints and enablements are constituted by actual incidents of behaviour that make it the case that an individual’s having certain features in that context raises and/or lowers the probability that they will be able to do or to avoid certain things. A socially significant feature, then, is one that has predictive power when it comes to describing the patterns in these incidents in a certain context. A background view that supports this picture, and which I endorse, is the idea that social kinds are explanatory kinds, and that kinds are explanatory (or not) relative to particular explanatory purposes or projects (Haslanger 2016).

The key to squaring conferralism with the second and third insights of intersectionality lies in recognising that there are different levels of generality, or granularity, at which we could specify these constraints and enablements. For example, specifying separately the constraints and enablements on women who are differently racialised in a context is more fine-grained than specifying the constraints and enablements on women simpliciter. Saying that women have a 70% chance of being interrupted when they try to talk in a certain meeting is less fine-grained than saying that White women have a 50% chance, Asian women have an 85% chance, and so on, for differently racialised groups. And this in turn is less fine-grained than if we were to also separate women by age, or by class, or by their job role in the relevant organisation, and so on. The level of granularity that is most useful to us will depend on our explanatory purposes. Here, my view accords with the approach to intersectionality taken by Brännmark (2018b, p. 15) in relation to institutional kinds (kinds characterised by deontic relations).

It is worth noting that there may be contexts in which there are no patterns that unite the different intersectional kinds that constitute a single-moniker kind. In such cases, there are no single-moniker kinds, but only intersectional kinds. Imagine a bar that is both extremely racist and extremely sexist. Black women work as waitresses. Black men work as bartenders. White men are customers who order and pay for drinks. White women are patrons, but not customers: they only frequent the bar as the guests of White men patrons, who order and pay for drinks on their behalf. If a White woman arrived at the bar, she would be assumed to be there to meet a White man, rather than being a customer in her own right. If a Black woman arrived, she would be assumed to be a waitress (or someone seeking work as a waitress) rather than a patron of any sort. In this case the constraints and enablements that characterise the status of the White women and the constraints and enablements that characterise the status of the Black women have very little, if anything, in common.

The social dynamics of the bar can be best explained by way of intersectional kinds such as Black women and White women, not single-moniker social status kinds such as women. Suppose Audre, a Black woman, enters the bar by herself and tries, unsuccessfully, to order a drink. Why can not Audre manage to order a drink? Because Audre is a Black woman, and everyone assumes she is a waitress, or perhaps is looking for work as a waitress. Now, we could try to answer the question by saying instead: because Audre is a woman, and women do not order drinks here. But this is much less explanatory: it does not tell us what people expect Audre to do instead of ordering drinks, and it obscures the difference between Audre’s situation and the situation of Betty, a White woman, who also cannot order a drink. In Betty’s case, this is because she is assumed to be the guest of a White man, who will order on her behalf; whereas in Audre’s case, this is because she is not considered to be a patron at all. We get a much more informative and accurate picture of why Audre cannot order a drink if we operate with the intersectional kind Black woman than with the single-moniker kind woman. Accordingly, in this specific social context, we might want to say that there is no such conferred kind as woman simpliciter.^[3]

6 Meeting the Demands of Intersectionality

As we saw above, the way in which Ásta talks about intersectionality seems to imply that there are constraints and enablements for single-moniker kinds that combine (potentially in non-additive ways) in order to produce the constraints and enablements for intersectional kinds. It is this picture that seems to stand in tension to the second and third aspects of intersectionality, and also to stand in need of further explanation. The understanding of constraints and enablements that I have set out here, however, reverses the picture. Ontologically speaking, it is not the case that we begin with constraints and enablements for single-moniker kinds which then combine (potentially in non-additive ways) for intersectional kinds. Instead, what comes first, ontologically speaking, is the actual incidents of behaviour, and the consequent abilities of different individuals to effect or to avoid various outcomes in a certain context. Patterns in these incidents of behaviour allow us to articulate constraints and enablements, initially for multiply intersectional kinds (e.g. disabled Black lesbian women). We can then take a more and more coarse-grained approach to describing these patterns that will give us simpler intersectional kinds (e.g. Black women) until finally we arrive at single-moniker kinds (e.g. women).

This analysis treats intersectional social status kinds as more ontologically basic than single-moniker social status kinds, but does not require us to reject the existence of single-moniker social kinds.^[4] Of course, establishing the level of description that is helpful for a particular enquiry is a substantive task, and I take it that one of the lessons of intersectionality theory is that we ought to be very careful not to assume that the most general level of description is usually going to be the most helpful. Granted, this model is somewhat messier than one on which we only acknowledge the existence of the most coarse-grained, single-moniker kinds; but the social world is messy, and we should expect some of this messiness to have to be reflected in our theories of it.

This way of understanding constraints and enablements renders conferralism compatible with non-separability and with cross-constitution. It conforms to non-separability because conferral schemas for race kinds and for gender kinds are only an approximation of the actual constraints and enablements that are at play, and are acknowledged to involve a large amount of generalisation. As a result, they should not be understood as implying that those constraints can be perfectly or truly separated into those relating to race and those relating to gender; the separation is a rough approximation of the complex truth. The understanding I have developed is also compatible with cross-constitution, because each single-moniker gender kind is a generalisation based on a number of intersectional kinds that involve race as a component, and vice versa for single-moniker race kinds and intersectional kinds involving gender as a component. Thus, race plays a role in constituting the reality of gender, and gender plays a role in constituting the reality of race.

7 Conclusion

I have advanced an internal critique of conferralism based on its handling of intersectionality, and I have suggested a modification to the account to solve this difficulty. First, I showed that Ásta’s conferral schema obscures the important difference between the name of a property, on the one hand, and the constraints and enablements of which it consists, on the other. I proposed an expanded conferral schema that includes a “property name” component and a “status” component as separate elements. Second, I used this expanded conferral schema to help argue that conferralism appears to struggle to meet some of the demands of intersectionality, specifically those relating to the non-separability of race and gender properties and their nature as cross-constituting. I proposed an understanding of conferred properties that gives ontological priority to the most fine-grained intersectional properties, and treats single-moniker properties as generalisations or abstractions that may be helpful in the context of particular explanatory projects. The result of incorporating this understanding would, I believe, be a version of conferralism that is more attractive by its own lights, with a broader appeal and greater explanatory power.