Why we keep separating the ‘inseparable’: Dialecticizing intersectionality

Abstract

Disputes about how to understand intersectional relations often pivot around the tension between separateness and inseparability, where some scholars emphasize the need to separate between different intersectional categories while others claim they are inseparable. In this article the author takes issue with the either/or thinking that underpins an unnecessary and unproductive polarization in the debate over the in/separability of intersectional categories. Drawing on Roy Bhaskar’s dialectical critical realist philosophy, the author argues that we can think of intersectional categories as well as different ontological levels as both distinct and unified and elaborates on the issue of how significance of the dialectical notion of unity-in-difference for intersectional studies. As part of the argument the author addresses the issue of what it actually means for something to be distinct or separate as opposed to inseparable or unified with something else, demonstrating that lack of clarity about this is at the heart of polarized arguments about separateness versus inseparability in intersectionality theory.

What precisely do we mean when we say that something intersects with something else? Whether it be used narrowly so as to depict the intersection between different axes of power and identity categories, or broadly so as to include an infinite range of possible ontological intersections between levels and dimensions of reality, intersectionality gives rise to the question of the exact nature of intersection itself. Disputes about the character of intersection often pivot around the tension between separateness and inseparability, an ambiguity that can arguably be seen as constitutive of intersectional analysis. When Kimberlé Crenshaw (1991), who coined the term ‘intersectionality’, argued that gender must be analysed as intersecting with for instance race, so that gendered identities be seen as intrinsically racialized, the very term ‘intersectionality’ at the same time implies that the entities intersecting are distinct from one another in some way – otherwise they could not intersect.

For the most part intersectionality scholars in action treat categories like gender, race and class as somehow both separable and inseparable. However, when the issue of separability versus inseparability is explicitly addressed, it is often played out in terms of a polarization between those emphasizing the inseparability of intersectional categories (the ‘anticategorical’ approach in Leslie McCall’s [2005] famous formulation) and those maintaining the need to separate them (the ‘intercategorical’ approach). For instance, I recently attended a seminar set up as a transversal dialogue between two intersectionality scholars seeking ways of bridging their opposed positions in this in/separability dispute (Lundberg and Strid, 2014).

In this article I take issue with the either/or thinking that underpins what I see as an unnecessary and unproductive polarization in the debate over the in/separability of intersectional categories. Drawing on Roy Bhaskar’s dialectical critical realist philosophy, I argue that we can think of intersectional categories as both separate and unified, an argument that actualizes the question of what it actually means for something to be distinct or separate as opposed to inseparable or unified with something else. I begin by introducing Bhaskar’s conceptualization of reality’s differentiations and interconnections. Giving some examples of ways of arguing in different areas of feminist thought that seem to be premised on a view of separateness and inseparability as mutually exclusive, I show how Bhaskar’s dialectic, centring round the figure of unity-in-difference, can counter unproductive dualisms between separateness and unity. I go on to examine some examples of how intersectional theorists, broadly defined, address the theme of in/separability, taking issue with accounts emphasizing either separability or inseparability and developing my own view on the in/separability dilemma. While the notion of intersection is mostly used to characterize the relations between different identity categories or power relations, I include the relation between different ontological levels, which is also paramount in intersectional debates.

Dialectical critical realism

Dialectical critical realism was developed by the British philosopher Roy Bhaskar (2008 [1993]) as part of the broader school of critical realism, which brings ontological issues to the fore of social meta-theory, challenging both positivism and postmodernism (Bhaskar, 1997 [1975], 1998 [1979]; Dy et al., 2014; Gunnarsson, 2011, 2013, 2014; McCall, 2005; Sayer, 1992, 2000; Walby et al., 2012). Bhaskar defines dialectics as ‘the art of thinking the coincidence of distinctions and connections’ (2008 [1993]: 180). What in particular distinguishes realist dialectics from other kinds of dialectical philosophies is its pronounced standpoint that the reason why we need to think dialectically is because reality is dialectically structured, via different modes of unity-in-difference whereby things have both points of identity with one another and points of divergence. Such coincidence of difference and unity may be constellated in a variety of ways, ranging from the necessary tensions built into our existence to historically accumulated splits between things that are fundamentally unified, such as humanity’s exploitative separation from the nature that sustains it (Gunnarsson, 2013, 2014).

The initial reason why I was long ago attracted to critical realism generally and dialectical critical realism specifically was precisely this non-absolutist approach to distinctions and relationality, which is anchored in the view of reality as a stratified and differentiated whole whose elements are both intrinsically connected and relatively autonomous from one another. ¹ I was frustrated with what I saw as a tendency among many feminist theorists to challenge atomistic and dualistic modes of making distinctions by altogether denying separability, since this move in fact reproduces the atomist’s basic view of reality: either things are absolutely separate and autonomous, or they cannot be separated at all. This mode of reversal can be seen in the way that some feminist theorists argue against the distinguishability of ontology from epistemology (Barad, 2003, 2007; Hekman, 2010; see Gunnarsson, 2014). It can be identified in the posthumanist scepticism towards separating the human from the nonhuman (e.g. Hird and Roberts, 2011: 109), and in many neo-materialists’ reluctance to delimit the social from the natural (e.g. Davis, 2009: 67; see Gunnarsson, 2013). Similarly, the deconstructionist approach to the subject is characterized by a tendency to deny the relative autonomy (and, hence, realness) of the subject, on the grounds that it is constituted by processes outside of itself (e.g. Butler, 1999 [1990]). While this move is often seen as a radical challenge to the atomist view of the subject as self-contained, as I argue elsewhere it in fact replicates its absolutist notion of autonomy (Gunnarsson, 2014; cf. Alcoff, 2006). Finally, the fact that gender coalesces with other power relations and identity categories has led some scholars to ‘completely reject the separability of analytical and identity categories’ (McCall, 2005: 1771; e.g. Butler, 1999 [1990]; Spelman, 1990; see Gunnarsson, 2011).

This emphasis on inseparability is an understandable reaction to rigid modes of separation, which deny co-constitution, fluidity and mutual transformation. We might indeed interpret these kinds of interventions as a matter precisely of emphasis, rather than as a complete rejection of distinctions. What supports such a reading is the fact that all the authors mentioned above, who explicitly deny the possibility of separating ontology from epistemology, the social from the natural, the subject from its constitutive context and gender from other intersectional categories, at the same time make use of such distinctions in their writing. However, even if it is just a matter of accentuation, the tendency shared by atomist and anticategorical theorists to emphasize either separateness or inseparability is problematic in itself, since it easily reproduces absolutist and undifferentiated notions of difference as well as unity.

Dialectical critical realism takes issue both with accounts that obscure the co-enfoldment of phenomena and with claims that phenomena cannot be at all separated if they have intrinsic ties to one another. If applied to the relation between epistemology and ontology, these can be seen as both separate and ‘constellationally unified’ (Bhaskar, 2008 [1993]: 114). As Alan Norrie highlights in his work on dialectical critical realism, ontology and epistemology are co-enfolded in one another in the sense that ‘knowing is … a subset of being, and the study or theory of being (onto-logy) is already epistemically committed’ (2010: 17, emphasis in original), but they must nevertheless be distinguished since one cannot be collapsed into the other; they are different things.

On the dialectical critical realist view, being is an interconnected, open-ended whole, whose different parts and dimensions are both intrinsically connected and relatively autonomous from one another, due to processes of differentiation, stratification and emergence. Reality has a structure to it, but is also a processual becoming, and neither the structuredness nor the fluidity of being are absolute but become meaningful in relation to one another. Bhaskar’s figure of dialectical totalities conveys that something can, however paradoxical it may sound, be both part of something else and separate from it, as in the case of knowledge being both part of and separate from reality, and of humanity being both part of and different from nature (Bhaskar, 2008 [1993]; Gunnarsson, 2013, 2014). What underpins all these kinds of dualities is the notion of unity-in-difference, which challenges what I see as the most basic and problematic of all dualisms, that between separateness and inseparability itself (cf. Kirby, 2008).

The unclear meaning of ‘inseparability’

I begin my examination of ways of dealing with the in/separability of intersectional categories by shortly turning to a statement by Judith Butler, which exemplifies a line of thought that is equally ambiguous and influential (cf. e.g. Brown, 1997; Geerts and van der Tuin, 2013). Although Butler is rarely labelled an intersectional theorist, she does address intersectional issues, and the anticategorical intersectionality identified by McCall owes much to the deconstructionist approach to gender and the category ‘woman’, of which she is an important representative.

In Gender Trouble Butler states:

If one ‘is’ a woman, that is surely not all one is … gender intersects with racial, class, ethnic, sexual, and regional modalities of discursively constituted identities. As a result, it becomes impossible to separate out ‘gender’ from the political and cultural intersections in which it is invariably produced and maintained. (1999 [1990]: 6, emphasis added)

What does Butler mean with this statement about the impossibility of separating out ‘gender’? At first glance the claim seems valid. As highlighted by Wendy Brown, who represents a similar line of thought, gender, ethnicity, class, etc. are not ‘discrete units’ (1997: 86) that we can disentangle in a clear-cut way. If understood in this basic sense, however, Butler’s statement appears trivial; it would arguably be difficult to find someone who did not agree. We could also interpret it as a stronger claim against any kind of separation of categories. In this case, though, Butler seems to engage in a performative contradiction, in that while stating that gender cannot be separated from other modalities, she nevertheless separates them in this very statement. If gender were not in some sense separable from the other relations she lists, there would simply be no need for Butler to call them by different names (cf. Jónasdóttir and Jones, 2009).

The ambiguous meaning of Butler’s statement raises the question of what we mean when we say two things are separate or distinct from one another, an issue that is in my view too rarely raised by theorists. Some authors arguing against separability seem to apply the implicit criterion that for things to be separable their discreteness must be tangible. For instance, as part of her argument against the category ‘women’, Elizabeth Spelman highlights the impossibility of pointing to a particular ‘woman part’ of herself that is not also a ‘white part’ (1990: 134). Although Brown’s line of reasoning is more sophisticated than Spelman’s hyper-empiricist mode of arguing, she seems to display a similar concern with the tangible, when highlighting that ‘we are not fabricated as subjects in discrete units by … various powers’ (1997: 86, emphasis added). I find it difficult to discern the sense of these kinds of statement. Of course we cannot identify discrete woman, white and middle-class (or other) ‘parts’ or ‘units’ in a person. I doubt that anyone arguing in favour of the separability of gender, race and class would be ready to base their claim on such a premise.

From a critical realist perspective, something’s existence as separate from something else is premised neither on the possibility of distinguishing it as a tangible unit, nor on the kind of absolute autonomy that precludes co-constitution and intra-connection. Instead, critical realists apply a ‘causal criterion’ for ascribing a distinct reality to something (Bhaskar, 1998 [1979]: 12). If something has impact on the world that is irreducible to the causal effects of other entities we can talk of it as a distinct reality, even though this reality is wholly premised on its relations with other things. For instance, we can talk of the human subject as a distinct entity, that has a reality outside of the relations that are also, paradoxically, inside it, in that they constitute it. This is because by means of its emergence ² from its constitutive relations, the subject has its own distinct properties and powers, which cannot be inferred from the relations through which it is constituted. When applied to intersecting categories, this perspective allows for a separation between gender, race, class, etc., despite the fact that these ongoingly co-constitute one another and are seamlessly unified in concrete subjects and social processes. They can be seen as relatively autonomous since the characteristics and impact of one cannot be derived from or explained solely in terms of the properties of the others.

The intersectional in/separability tension is sometimes dealt with by means of a distinction between the different levels of social reality. For instance, although Brown states that the ‘powers of subject formation are not separable in the subject itself’ (1997: 86), she also highlights that these powers are different in kind and, consequently, require ‘distinctive models of power’ (1997: 87). For her, this amounts to a paradox that creates problems for intersectional theorizing. By contrast, Nira Yuval-Davis (2006), who represents a more realist-materialist intersectional approach, seems to see a clearer separation between the different levels of intersectional relations as a way out of the in/separability dilemma. She shares the view that intersectional categories cannot be separated in concrete subjects and experiences, but argues they ought to be separated on the deeper structural level where ‘each social division has a different ontological basis, which is irreducible to other social divisions’ (2006: 195). At this level, she holds class to relate to ‘economic processes of production and consumption’; gender to ‘a mode of discourse that relates to groups of subjects whose social roles are defined by their sexual/biological difference’; sexuality to a discourse ‘relating to constructions of the body, sexual pleasure and sexual intercourse’; and ethnic and racial divisions to ‘discourses of collectivities constructed around exclusionary/inclusionary boundaries’ (2006: 201). Yuval-Davis’s way of solving the problem of inseparability versus separateness is thus to argue for a vertical separation of levels, so as to then argue for horizontal separations on the level of ontological basis, while claiming inseparability on the level of concrete identity and experience.

Although I elsewhere make a similar argument (Gunnarsson, 2011, 2014), as will be clear throughout the article my way of dealing with the in/separability tension differs somewhat from Yuval-Davis’s. To begin I want to complicate her claim that ‘the ontological basis of each of these divisions is autonomous’ (Yuval-Davis, 2006: 200–201), by stressing that even if analysed on the deepest level of abstraction, I doubt that we can think of economic, gendered, sexual and racialized relations as absolutely independent from one another. I am unsure how starkly Yuval-Davis’s claim about autonomy should to be interpreted. What seems certain, though, is that many will interpret such a claim as one about absolute autonomy, in turn provoking one-sided arguments about inseparability. This highlights the need for anti- and intercategorical theorists alike to be clearer about what they actually mean by ‘separate’, ‘autonomous’ and associated terms.

Intra-action and ‘processification’

In order to affirm the inter-permeation of the different dimensions of intersectionality, some theorists prefer to use the term ‘intra-action’, rather than ‘interaction’, to depict their interplay (Egeland and Gressgård, 2007; Geerts and van der Tuin, 2013; Lykke, 2010, 2011). In the field of feminist theory the term ‘intra-action’ was introduced by Karen Barad, who states that ‘in contrast to the usual “interaction”, which assumes that there are separate individual agencies that precede their interaction, the notion of intra-action recognizes that distinct agencies do not precede, but rather emerge through, their interaction’ (2007: 33). Barad takes issue with ‘[t]hingification – the turning of relations into “things” ’ (2003: 812), stressing the ontological priority of relations over the entities they produce. Nina Lykke judges the Baradian notion of intra-action to fit well with many intersectional researchers’ ‘agreement that intersectional interplays between categorizations should be analyzed as mutual and intertwined processes of transformation and not as a mere addition of gender, class, ethnicity, race, sexuality and so on’ (2010: 51).

Interestingly, Bhaskar (2008 [1993]) also uses the term ‘intra-action’ in his dialectical work. In addition, there is an interesting affinity between Barad and Bhaskar in that they both adopt realist worldviews (in Barad’s case Niels Bohr’s agential realism). There is also a crucial difference, though, associated with the theme of either/or versus both/and thinking that is central in this article. The same kind of either/or thinking that sustains the dualism between separateness and inseparability can be identified in the way that Barad emphasizes the processual roots of entities at the cost of their relative autonomy and stability. A closer look at her formulations about intra-action reveals an underlying temporal dualism between the product and agent of intra-action. For Barad, ‘distinct agencies do not precede, but rather emerge through, their interaction’ (2007: 33, emphasis added). From the Bhaskarian perspective things can instead be both a product of intra-actions and precede, be the point of departure of, new inter- and intra-actions. Even if something is a product of intra-active processes, it may have a relatively stable existence as product – or ‘thing’ if you like – from the point of view of which it then inter/intra-acts with other dimensions of reality. We do not need to replace thingification with ‘processification’. Instead, I suggest we think of the entities in the world as multifaceted in character, being simultaneously products, producer and process.

The theme of things versus processes is intimately linked to that of in/separability, inasmuch as things are commonly thought of as bounded whereas processes are not. A statement in Evelien Geerts and Iris van der Tuin’s intersectional intervention, drawing on Barad’s work, highlights this link: ‘The idea that patterns are constantly evolving, and that different patterns intra-act instead of merely interacting with one another … generating an assemblage that is holistic – i.e., not separable – is one of Barad’s onto-epistemological key points’ (2013: 176, emphasis in original). Here the emphasis on processual intra-action leads to an emphasis of inseparability. Although the formulation ‘instead of merely interacting’ suggests there might be aspects of unbounded intra-action as well as interaction between relatively delimited entities, Geerts and van der Tuin go on to state that ‘the concept of interaction stands for a traditional atomist ontology’ (2013: 176), seemingly wanting to throw it out altogether. Whereas the intersectional theorists drawing on Barad’s work (Egeland and Gressgård, 2007; Geerts and van der Tuin, 2013; Lykke, 2010, 2011) seem to suggest we use the term ‘intra-action’ to denote all kinds of intersectional relations, as I read him Bhaskar instead sees intra-action as one kind of interplay between phenomena. Sometimes, when the interplay has little internal impact on the phenomena in interplay, the term ‘inter-act’ might be more adequate. This multifaceted view of how things relate to one another is one of the strengths of the dialectical critical realist ontology, which could be drawn on in intersectional analysis.

Seeing boundedness and boundlessness as two interdependent aspects of the tissue of being works as a remedy against tendencies to get stuck in debates about whether gender, race, class, etc. are separate or not. Bearing in mind that there are many modes as well as degrees of connection and distinction, we can put our energy into more precise theorizations of how different power processes work for and against each other in different spatio-temporal locations and for different subjects. What, in various geohistorical locations and on different levels of abstraction, is the precise inter- or intra-active relation between structures like male dominance, heteronormativity, racism and capitalism? To what extent are they independent from one another and to what extent co-constitutive and intra-active, in the past and in the present? Do they support or contradict one another, or both?

This kind of analytical endeavour was commonplace at the time when feminist theorists were more structurally oriented, even though the term intersectionality was not yet invented (Lykke, 2010). For instance, an important topic in discussions among Marxian feminists was the exact relation between capitalism and patriarchy, including how they co-constitute versus conflict with one another (Hartmann, 1979). On one hand, capitalism is fundamentally entangled with patriarchy in the way, for instance, that it is premised on a gender-based division of labour based on women’s subordination to men. Also, capitalism is inherently patriarchal and misogynist in that it necessarily underprivileges the reproductive practices on which it depends, and cannot by itself provide the means needed to protect and support the agents of reproduction. On the other hand, in a historical perspective capitalism’s centring of the individual has been central for the emergence of the women’s movement and in this sense it has constituted a challenge to male dominance. In addition, capitalism’s need for an increased supply of workers has challenged the structure of male dominance in that it enabled/-s women to enter the labour market. Rather than making a case here for the exact nature of these intersections, my point is to illustrate that all these kinds of intersectional interrogations can be fruitfully informed by the guiding thread offered by the dialectical theme that unities-in-difference can be differently constellated.

Structures and subjects

I now want to turn more focused attention to the relation between the different levels of intersectional analysis, complicating further the somewhat messy picture of crisscrossing separabilities and inseparabilities. Like the relation between categories of gender, race, class, etc., this relation is characterized by the kind of unity-in-difference that is, as I see it, bound to generate theoretical confusion and unnecessary polarization unless approached from a dialectical ‘both/and perspective’.

As noted above, Yuval-Davis emphasizes the need to distinguish ontological levels that in her view are often confused in intersectional theorizing (cf. Dy et al., 2014). She states that in contemporary literature identities, ‘the individual and collective narratives that answer the question “who am/are I/we” ’, are ‘often required to “perform” analytical tasks beyond their abilities’ (2006: 197). In her view, a crucial task of intersectional studies is to examine the relationships between structural positioning, identity and political values, but this is ‘impossible if they are all reduced to the same ontological level’ (2011: 160). Here it is interesting to note that whereas some intersectional theorists see the connections between levels as an argument against separating them, for Yuval-Davis it is the fact that they relate to one another that motivates their separation.

An important strength about Yuval-Davis’s account is that she makes a case not only for an analytical separation between different aspects of reality, but highlights that this need is due to an ontological distinction. This departs from the common strategy of solving the in/separability dilemma by claiming to make only an analytical distinction between things that are really indistinguishable (e.g. Fraser, 1998: 12, 15). From a realist perspective, instead, the reason why analytical distinctions make sense and are efficient means for grasping reality is because they accord with a differentiation in being itself.

However, again, Yuval-Davis’s claim about separability is characterized by some lack of clarity, which opens up for anti-separative critique. Sumi Cho, Kimberlé Crenshaw and Leslie McCall (2013) take issue with what they see as an upsurge of intersectional work distancing itself from a preoccupation with identities and subjectivities in order to focus on structural inequalities. They are concerned about the division between structural power and identity on which this debate is premised, suggesting that ‘the opposition between identity and power is itself a rigid and nondynamic way of understanding social hierarchy’ (2013: 797). Although they do not refer to Yuval-Davis’s work, it is not far-fetched to assume that what they have in mind is the kind of intersectional writing of which Yuval-Davis is a leading figure.

The way that the respective perspectives of Yuval-Davis and Cho et al. become positioned in relation to one another here illustrates how the tendency to see inseparability and separateness as mutually exclusive creates an otiose polarization. Both Yuval-Davis, with her emphasis on separateness, and Cho et al., with their claim about inseparability, are in my view right. Communication would work smoother, though, if they could also affirm the other side of the picture. In my view, Cho et al.’s argument draws a lot of its strength from contrasting itself against a largely ‘invented target’, to borrow Andrew Sayer’s expression (2000: 68). What it challenges is ‘the opposition between identity and power’, but I believe it is hard to find a theorist arguing for the existence of such an opposition, whereas many would claim they are different things and must thus be analytically distinguished. Opposition denotes a sense of conflict and mutual exclusion; by contrast, a mere distinction between identity and structure is a premise for analysing how they enable one another and interact in mutually transforming ways.

Cho et al.’s intervention resembles Maria Carbin and Sara Edenheim’s poststructuralist argument about the relation between the different levels of intersectionality. They take issue with how some intersectional theorists frame the division between ‘structuralist’ and poststructuralist perspectives in terms of their respective focus on structures versus subjectivity/agency. Against this, they argue that poststructuralism constitutes a challenge to the very ‘binarism between structure and agency’, on which this scheme is in their view based (Carbin and Edenheim, 2013: 241, emphasis added). Similarly to Cho et al.’s charge of ‘opposition’, this formulation is problematic in that it confuses the act of identifying ‘two analytical levels’ (Carbin and Edenheim, 2013: 241) with the institution of a binarism, a term that connotes a dualistic opposition that is not necessarily entailed by a distinction. As I highlight elsewhere (Gunnarsson, 2013, 2014), there is a remarkable lack of clarity in much feminist work as to the difference between mere distinction or difference on the one hand and dualism or binary on the other, indicated by the fact that they are often used interchangeably (e.g. Hird and Roberts, 2011: 109). In their conventional usage, dualisms or binaries refer to the kind of absolute separation which ignores any interconnection and mutual constitution between the two terms in question, while distinction simply means that two things are not the same. Conflating these different kinds of difference easily makes us see distinctions as such as the problem. This is troublesome for, as highlighted by Yuval-Davis (2006), it is only by distinguishing phenomena that we can honour their respective irreducible constituencies and examine their interrelation, thereby avoiding that one is subsumed under the logic of the other. ³

Abstraction and the continued need for ‘separatist’ theory

Intersectional theorists often take issue with more conventional modes of discrimination analysis, which tend to reify categories like ‘Black’ and ‘woman’ in a way that ignores how they are (co-)constituted through historically formed processes of power. As Catharine MacKinnon highlights, any analysis which obscures the hierarchies underpinning raced and gendered categories ‘mirrors the power relations that form hierarchies that define inequalities rather than challenging and equalizing them’ (2013: 1023). How, then, are we to think of the relation between categories and the social processes underpinning them? Floya Anthias makes clear that we cannot circumvent the analytical tension implied in this relation:

Arguably one danger with the notion of intersections is found in constructing people as belonging to fixed and permanent groups. … This undermines the focus on social processes, practices and outcomes as they impact on social categories, social structures and individuals. This is further complicated by the fact that, despite the danger of seeing people as belonging to fixed groups, groups do exist at the imaginary or ideational level as well as the juridical and legal level. (Anthias, 2008: 14)

While it is common to think of categories in terms of ‘groups’, it is important to bear in mind that this is not the only meaning of categories. Intersectional theory deals with group categories such as whites, women, gays and middle-class, but also with the relational categories of race, gender and class, which are constitutive of these groups as well as supported by the categorical divisions on which groups are based. Whereas categories are an indispensible tool for representing the patterns and structuredness that is such a pervasive part of reality, as expressed in the grouping of people and the distribution of resources along these lines, they are less fit to do the job of grasping the ultimately processual-practical character of power relations, which tends to eschew categorical delimitations. As soon as we refer to these processes by means of abstract categories like gender, race and class, claiming they intersect, it is easy to lose sight of the full picture of how these relational structures are processually constituted. However, against some pessimistic accounts (e.g. Brown, 1997), I do not think that this impossibility of perfect representation is a reason for seeing intersectional categorizations as inherently problematic. This is the way language works, and we need not despair because we cannot get the full picture in one particular statement or study, as long as we are clear about our partiality and imperfection (Gunnarsson, 2011; Sayer, 1992).

For many intersectional theorists abstracting one category from concrete reality while putting others aside amounts to a violation of the complexity of reality, in which, as Anthias puts it, ‘classes are always gendered and racialised and gender is always classed and racialised and so on’ (2008: 13). Appreciating the relative ontological autonomy of intersecting categories as well as our always necessarily partial epistemological outlook, we need not come to this conclusion, though. As long as we are clear that an analysis of for instance gender on its own terms relies on an abstraction of some processual parts from an infinitely complex social whole (Jónasdóttir, 1994), it is desirable that some theorists engage in ‘separatist’ theoretical explorations of what precisely this ‘gender’ (or ‘race’, ‘class’, ‘sexuality’) is, in a certain geohistorical location. Otherwise there is a risk that we reproduce unreflected notions of their ontologies. As Sylvia Walby, Jo Armstrong and Sofia Strid state, similarly drawing on critical realism, there is a need to ‘systematically address the ontological depth of each of the inequalities’ (2012: 231), so that their dynamic can be ‘made more available for analysis’ (2012: 236; cf. Dy et al., 2014). Whether we conceive of gender as, for instance, an effect of the reiterative enactment of norms organized by a heterosexual discursive matrix (Butler, 1999 [1990]) or as a historically shaped social-organic practice involving flows of erotic and caring powers (Jónasdóttir, 1994) will significantly affect the intersectional analysis of any concrete situation.

Within the field of intersectionality studies Anthias and Yuval-Davis have begun to draw the contours of a promising dialectic between separatist and intersectional theorizing, by mapping the different ontological bases of social divisions (Anthias, 1998, 2008; Anthias and Yuval-Davis, 1992; Yuval-Davis, 2006). Separatist modes of theorizing need not be carried out within an intersectional framework, though, but theorists of gender, race, class, sexuality respectively should in my view feel free to focus on specifying the dynamics of each of these relations, thereby creating platforms for intersectional theorists to further the analysis of their interplay. One problematic consequence of one-sided claims about inseparability is that they tend to stand in the way of such in-depth separatist theorizing of the specificity of racial, gendered, classed and sexual power dynamics.

Before concluding I shall point to a final way in which my way of dealing with the in/separability tension differs from Yuval-Davis’s approach, which on the whole has crucial affinities with my own realist take on intersectionality. Whereas Yuval-Davis has figured as my prime example of an intersectional theorist highlighting the need for analytical distinctions, she takes an anti-separatist stance when it comes to concrete experiences of oppression. She rejects the classical formulation that Black women suffer from ‘triple oppression’, as Blacks, women and members of the working class, basing this on the fact that ‘in concrete experiences of oppression, being oppressed, for example, as “a Black person” is always constructed and intermeshed in other social divisions’ (2006: 195; cf. Anthias and Yuval-Davis, 1983). Although I concur that concrete experiences are complexly formed wholes that are difficult to divide into neat categories, I do not agree that talking about someone as oppressed specifically as Black, woman or working class is necessarily an ‘attempt to essentialize “Blackness” or “womanhood” or “working classness” ’ (Yuval-Davis, 2006: 195). Sometimes it does make sense to think of somebody as oppressed by virtue of her racial identity rather than her gender, although this racial identity is concretely articulated only in mediation with gender and not altogether separable from it. When Hannah Arendt said ‘If one is attacked as a Jew, one must defend oneself as a Jew’ (in Bernstein, 1996: 21), she had no difficulty abstracting Jewishness from the concrete complexity of experience and identity and such abstractions would make no sense unless they had some bearing on reality. As I argue elsewhere (Gunnarsson, 2011), we can think of for instance ‘womanhood’ in other terms than a reified attribute, namely as an abstraction that points to that dimension of a woman’s situation that is produced via her positioning in a socially produced gender structure, a position that intra-acts with other social positions but which nevertheless has its own irreducible properties (cf. Alcoff, 2006).

It seems to me that Yuval-Davis seeks to escape the ontological ambiguousness of reality by making a somewhat too clear-cut separation between the level of structural ontological basis, where it is possible to separate neatly between different categories, and the level of identity and experience, where such separations are totally impossible. I think this is making it too simple. In fact, this is an example of how too sharp a separation in one place tends to be accompanied by a neglect of differentiation on another level. It is because any concrete identity or experience has points of intrinsic interconnection with each of the ontological bases that Yuval-Davis lists (rather than being altogether separable from them) that it sometimes makes sense to distinguish between how these different ontological bases are articulated in a concrete situation.

I think we need to take a humble stance towards our analytical tools, recognizing that their way both of splitting up and holding the world together are never perfect, ultimately because of the processual character of reality. The structure, stratification and differentiation of reality make it viable as well as necessary to make the kind of distinctions in which human language excels. However, structuredness and boundedness exist only in dialectic with processes and flows, and these are difficult to grasp with analytical categories. Hence there can never be a perfect fit between knowledge and being. What also characterizes analytical language and thinking is its aversion to paradox and I think this lies behind much of the either/or thinking that I have mapped in this article. If we as intersectional theorists could accept the tensions and imperfections that are necessarily involved in theorizing, due to the clumsiness of the linguistic categories that we are bound to use, I think some of the conflicts among us might be recognized as largely semantic in kind.

Conclusion

Dialectical thinking is remarkably absent from contemporary feminist theorizing. Magnus Granberg notes that whereas ‘[d]ialectics was once seen as the major alternative to positivist approaches; perhaps presently poststructuralism and its offshoots in gender, postcolonial and intersectional studies have usurped this position’ (2013: 2). Peculiarly enough, while dialectics offers tools for transcending malestream dualisms, for many poststructuralist thinkers, in particular those influenced by Gilles Deleuze, dialectics is instead understood as part and parcel of the dominant western tradition and associated precisely with the kinds of dichotomous thinking that it challenges (Braidotti, 2011 [1994]). Against these anti-dialectical currents, in this article I have sought to demonstrate the usefulness of Bhaskar’s dialectical critical realism for intersectional purposes, focusing on how tendencies in intersectional studies to emphasize either the separability or inseparability of categories can be challenged by the dialectical insight that difference and unity are not mutually exclusive but two coexisting aspects of most relations. The strength of the dialectical critical realist perspective is that it offers analytical tools for challenging both atomist separations and conflationist accounts which overemphasize connectedness and inseparability at the cost of differentiation and stratification.

As part of my argument I highlighted that it is often unclear what authors mean when claiming that something is impossible to separate from something else. Given that all kinds of theorizing are largely based on the making of distinctions and connections of different kinds, I call for more systematic reflections about this epistemological-ontological issue. For intersectionality theory this is of particular importance, inasmuch as many conflicts within the field pivot around the issue of in/separability. Most intersectional theorists seem to work with an implicit notion of intersectional categories as somehow both separate and inseparable, while emphasizing only one pole in this duality when being more explicit about the issue of in/separability. I hope I have been able to furnish some of the implicit dialecticians with a theoretical structure that can work against this kind of either/or thinking, since it in my view produces somewhat artificial conflicts among theorists, which stand in the way for dealing with the more substantial and politically pressing issues of intersectionality.