Intersectional what? Social divisions,intersectionality and levels of analysis

Introduction

Intersectionality approaches provide an important corrective to essentialising identity constructs that homogenise social categories. Although social divisions such as gender, ethnicity and class have been understood through the lens of intersectionality for at least two decades, and have had a profound effect on feminist theories in particular, this approach has only recently acquired a more central place in academic and political life. Moreover, intersectionality has now become part of policy initiatives, which have begun to recognise multiple intersecting inequalities (for example within the Equality Act 2010 in the UK (Equality and Human Rights Commission, 2010)).

For Knapp, it is because intersectionality has become ‘a formula merely to be mentioned, being largely stripped of the baggage of concretion, of context and history’ that it has been such a ‘fast travelling concept’ (Knapp, 2005: 255). This article focuses on the specific issue of the analytical levels that need attending to in any framing which is dedicated to the complexity and mutual interdependency of social divisions. Winker and Degele (2011) also raise the issue of levels and argue for an intersectionality approach which focuses on ‘interactions between inequality-creating social structures (i.e. of power relations), symbolic representations and identity constructions that are context-specific, topic-orientated and inextricably linked to social praxis’ (2011: 54). However, although this is a position that I concur with, I use the term ‘societal arenas of investigation’ for the types of foci that they refer to as different levels of analysis. I propose a theoretical framing which attends to different levels of analysis instead of questions about what is being referred to (social categories or concrete relations), societal arenas of investigation and historicity (processes and outcomes). I also explore the notion of intersection, which I argue is one of the pivotal concepts used (see also Bilge, 2010).

Intersectionality: A range of approaches

This article makes the assumption that ‘intersectionality’ does not refer to a unitary framework but a range of positions, and that essentially it is a heuristic device for understanding boundaries and hierarchies of social life. I will refer, therefore, to an intersectional framing rather than to intersectionality theory. This differs from some current views that it constitutes a theoretical and empirical paradigm (Hancock, 2007: 249–250), on the one hand, or the idea that it is a buzzword (Davis, 2008), on the other. However, the plurality of the ways in which it can be advanced is acknowledged.

At a very broad level, and put simply, intersectionality posits that social divisions interrelate in terms of the production of social relations and in terms of people’s lives. For example, gender is seen as inflected by other social divisions such as ‘race’ and class, and they are seen as mutually constitutive (Brah and Phoenix, 2004; Collins, 1993). This highlights divisions amongst ‘women’ (and other social categories), by pointing towards processes of racialisation and class, and the disadvantages that follow. ¹ The focus on gender, ethnicity/race and class does not imply that sexuality, age, disability, faith and so on cannot be incorporated. For example, there has been important recent writing on sexuality (Taylor et al., 2011) and disability (e.g. Meekosha and Shuttleworth, 2009), thereby broadening analysis away from the ‘big three’. As Dhamoon notes, ‘the privilege assigned to this trinity is not intrinsic to the study of categories but indicative of the choices researchers have made  ... in specific historical contexts’ (2011: 5).

It is worth noting, in this context, that the idea of the existence of interrelations in the domain of the social is not new and arguably constitutes the very foundation of classical sociological theory. ² The concern with the links between different forms of identity and hierarchy is not new either, and there has been a long-standing interest, both theoretical and political, in exploring the connections between different forms of subordination and exploitation. ³ In other words, the interconnections between social divisions were recognised without being named as ‘intersectionality’. The impetus in the formulation of intersectionality arose within a feminist politics articulated powerfully by black feminists (see, for example, Combahee River Collective, 1977/1982; hooks, 1981). The political underpinnings of intersectionality are clear in the work of pioneers such as Patricia Hill Collins and Kimberle Crenshaw amongst others.

The work of Patricia Hill Collins on gender, race and class became central to feminist theory and method before the term intersectionality was coined. She treats these as ideological (Collins, 1990, 1993, 1998, 2001) or discursive practices emerging in the process of power production, and as historically contingent (as would be suggested in the work of Foucault, 1972). ⁴ Using the notion of interlocking oppressions organised through a ‘matrix of domination’ (1990: 276) relating to various domains of power (identified as structural, disciplinary, hegemonic and interpersonal), she provides an approach which relates to the social, economic and political context ‘within which intersecting oppressions’ linked to gender, race and class are organised.

Kimberle Crenshaw, a socio-legal theorist, has been credited with coining of the neologism ‘intersectionality’. Her approach focuses on the overlapping of categories of discrimination (Crenshaw, 1994). This has been very influential, leading to an interest in the production of data or policy research and practice that recognise the specificity of the discriminations experienced by racialised women, who have suffered from intersectional invisibility (Crenshaw, 2000). Within the UK, the work of Avtar Brah as well as Floya Anthias and Nira Yuval Davis located the articulation of social divisions within the context of power relations and the state, and was central to the entry of intersectionality approaches within the European context (Anthias and Yuval Davis, 1983, 1992; Brah, 1996).

There has been a proliferation of recent writing on intersectionality, pointing to the limitations of some existing intersectionality frames and attempting to provide better theoretical and methodological underpinnings (e.g. Bilge, 2010; Brah and Phoenix, 2004; Davis, 2008; Dhamoon, 2011; Erel et al., 2011; Ferree, 2009; Gimenez, 2001; Hancock, 2007; Knapp, 2005; Levine-Rasky, 2011; Lutz et al., 2011; McCall, 2005; Taylor et al., 2011; Verloo, 2006; Walby, 2007; Winker and Degele, 2011; Yuval Davis, 2006). The positions taken range from the idea that intersectionality is the best means we have for exploring the multidimensional and complex articulation of forms of social division and identity (Brah and Phoenix, 2004), through the view that intersectionality serves as a useful buzzword but cannot aim to be a theory as such (Davis, 2008), to the view that intersectionality tends to reduce all forms of difference to a list and treats them as equivalent (Erel et al., 2011; Gimenez, 2001).

It is not my intention to review these contributions but suffice it to say that pitfalls associated with intersectionality include the ‘listing’ of differences (often reduced to identities) that intersect and the impossibility of attending analytically to this plurality, as well as potential competing claims about which are the most important of these, or how many differences should be incorporated. ⁵ There is also a potentially endless list of hybrid positions or cross-cutting groupings that can be yielded (such as black working class, lesbian, young, poor, rural, disabled and so on). Ultimately, there is the danger of a focus on individual differences (raising the question about what differences matter, when, where and how).

The idea of sections that come together is also a problem as these may be treated as given (rather than constructed in social practice). This may fail to consider the construction of social categories, identities and divisions as part of the exercise of power within particular contexts and in relation to the state, economic interests and practices, as well as in relation to discursive facets of social relations (see Ferree, 2009, for an emphasis on the political and the discursive, and Hart and McKinnon, 2010: 1048, for the view that they are part of the organisation of social order). Issues of power and social structure as well as political economy have arguably been sidelined, as have those of class. ⁶ There is also the issue of how political mobilisation can take place given either a deconstruction or an individualisation of difference. Another issue relates to whether the concept of intersectionality has universal applicability or is only applicable to marginalised groups (e.g. Nath, 2009). If it purports to understand social relations, however, it must aim to be more generally applicable (see Anthias and Yuval Davis, 1992, for such a view).

There are therefore a broad range of issues which are raised and not all can be covered by this paper. I propose that we should not focus too much on the problems of the metaphor of intersectionality but retain its early concerns found in anti-racist feminism (e.g. Combahee River Collective, 1977/1982; hooks, 1981), which challenged dominant and exclusionary academic and political frameworks and knowledge production. This requires locating social categories and divisions within a broader social framing that attends to power, hierarchy and context – both spatial and temporal. Greater analytical rigour would contribute to a more robust intersectional framing in terms of ‘what’ is being referred to as intersecting, and what that means. This does not require that the objects of reference (seen as intersecting) should be the same (in fact this opens up the possibility of different foci), nor does it mean that ‘intersecting’ has to have one meaning – merely that it must be defined in particular instances. In the following sections I will explore these issues.

What is being referred to: Different levels of analysis

Societal arenas of investigation: Embodied practices

Concrete social relations in terms of social divisions relate to positionalities and hierarchies as they are embodied and articulated within different societal arenas at particular conjunctures. Each arena acts as a context (Bourdieu, 1990) for the others and enables an exploration of how they interlink with each other.

This is very different from systems or domain-based approaches to social relations (e.g. Derek Layder, 2006) or to social divisions (e.g. Sylvia Walby, 2007). The argument here is not so much that these arenas exist as autonomous or actual societal structures; it is rather that setting them out helps to organise the types of issues that we focus on, and allows comparison across these; positing societal arenas is therefore a heuristic tool. The distinctions presented below are heuristic rather than actual, therefore, and enable different foci to be investigated:

Organisational (structural position): this focuses on how population categories are organised within institutional frameworks, e.g. family structures and networks, educational systems, political and legal systems, the state apparatus and the system of policing and surveillance.

Societal arenas are different for the purposes of analysis but are interrelated aspects of social relations and can be analysed in relation to one another. The arena of the representational, for example, can be analysed both in terms of social practices, which entail the production of texts, media, knowledge and so on, and in terms of how actors narrate, assimilate, respond and agonistically challenge these in their everyday understandings and actions.

In order to highlight the potential of focusing on societal arenas, in the ways I have defined them above, I will take the example of the experiential arena which relates to narrations by subjects. Intersectionality at this level can be illustrated with reference to recent work, particularly in the UK, that has been concerned with collecting narratives of people relating to ‘class’ identities. A particular example is found in the work of Mike Savage (Savage, 2010; Savage et al., 2005), which concerns the choices middle-class people make for being in neighbourhoods with others ‘who are like them’, using the term ‘elective belonging’. It is argued that this plays a key role in people’s definitions of who they are and also provides access to other resources, such as those of education. Accounts of gentrification, for example in south east London (Butler and Robson, 2003), also focus on the choices people make to move to areas where there are people like themselves (Watt, 2009). A major argument here is that this plays a role in middle-class formation.

However, there is the question of the role of distancing from the ‘other’, which may be more than a classed other. As Bourdieu says ‘tastes are perhaps first and foremost distastes, disgust provoked by horror or visceral intolerance (“sick-making”) of the tastes of others’ (Bourdieu, 1984: 56). As Watt (2009) argues, choices about place and attachments are racialised as well as classed: ‘Such judgements are at their most acute when the expressions of taste and distaste operate hierarchically, i.e. when they are directed against “others” with lower volumes of capital in subordinate class or racialised positions’ (Garner, 2007; Tyler, 2003). Therefore, narrations of people here have been far too narrowly placed within the class construct, as they spill over into other forms of difference and disadvantage such as gender and ethnicity as well as locality.

Class elements can also be found within narrations around ethnic resources and ethnic belonging. In research into minorities in the food sector, some subjects used strategies of locating oneself within an ethnic category where positive ethnic capital was involved, and distancing oneself where this was perceived as negative, showing how narrations about ethnicity are often imbued with class elements (Anthias, 2007; Anthias and Cederberg, 2009). Lamont (1992) showed that working-class black men use notions of masculinity derived from white men to define their own class identity, illustrating how narrations of gender are tied to class and ethnicity.

This discussion briefly shows how, at the concrete level of analysis, social categories operate within the societal arena of the experiential in intersectional ways and that at this level it is difficult to maintain that narrations of belonging are only about class, ethnicity or gender, for example.

Historicity and the notion of intersection

The issue of historicity entails looking at processes and outcomes and is crucial because the social divisions appear differently in terms of this distinction. Outcomes (synchronic, like still shots, e.g. unemployment at a particular time or as a trend at a particular time) and processes (diachronic, like a moving film, e.g. racialisation and its patterns over time) need to be unpacked. Unpacking gendered processes and gendered outcomes is dependent on their framing within a time context; the distinction actually refers not to differences between them (for example, sexism can be both a process and an outcome), but on the focus of enquiry.

There is a danger of treating outcomes (e.g. gendered or racialised violence) merely as relating to the processes that produce them in the most mechanical way, i.e. to do with sexual difference and ‘race’ difference. What this means is that you cannot invoke only gender processes to understand gendered practices or outcomes, you cannot invoke only ‘race’ and racialisation to understand racist practices and outcomes and you cannot invoke only class processes to understand classed outcomes. This is one of the important insights of intersectional frames found in both the earlier versions of intersectionality and the more recent contributions. This brings me to considering further the notion of intersection, which I see as the key axiomatic principle of the intersectionality framework.

Within an intersectionality framing, if what is new is the insistence on intersection, we need to ask what is the process of intersecting? What does it mean to intersect? Is it more than noting the co-dependency of different spheres in the realm of the social? As noted earlier, there is of course the problem of the ‘sections’, which might assume unproblematic social entities, thereby essentialising them. Intersection usually denotes some variant of the idea of ‘mutual constitution’, as noted earlier. The mutual constitution of gender, ‘race’ and so on refers essentially to the postulate that they are affected by and affect each other. I would interpret this in a dialectical way rather than a deconstructionist or reductionist way, therefore retaining the existence of categories themselves as important elements of the social landscape. I do not think that it is possible to dispense with the social categories themselves, as they are a necessary component of analysis; they must therefore be specified before any intersectional analysis can take place. This is because, if they are seen as mutually constitutive in a deconstructionist sense, then this entails that the categories themselves do not possess any autonomous or systemic features in their own right and in relation to social processes. It is therefore the salience of the distinct categories that produces the derivative although specific saliency of intersectional or hybrid categories (such as ethnicised women) hailed by some forms of intersectional framing.

One problem with the powerful metaphor of intersectionality is that it may be misleading, as it suggests that what takes place is similar to what happens at an intersection (where things collide or crash together). However, the sources of the social relations, or, more pertinently for most intersectionality paradigms, the inequalities and forms of violence experienced by people, might not be a product of the intersection at all but may be manifested in that space; in other words something happens at the junction which is not necessarily a product of the different roads that lead to it. The idea of points or instances of intersection is in fact problematic as it does not attend to processes. Therefore I concur with those writers (such as Ferree, 2009; Glenn, 2002; Walby, 2007) who are not so much concerned with pointing to the space of an intersection but focus on the dynamic and located dimensions of inequality and division in terms of relationships with each other. Therefore the intersection does not denote specific places occupied by individuals or groups (e.g. working-class black women). It is a process; for example, ‘class’ takes on racialised or gendered inflections for specific people in specific places and times within the arenas of organisation, representation, intersubjectivity and experience.

Moreover, the notion of intersection says nothing about the ways in which the production and reproduction of discrimination/subordination take place. One example here is that you need to look at the operations of inequality and violence through the state and other institutional frameworks in which power and economic interest are exercised, and not just at the categories and practices of gender, race and so on. Indeed, broader power relations within social processes and practices need to be considered (as recognised by Collins, 1993, for example). This means we should not take the metaphor too literally (it should therefore be seen as signposting complexity as well as multiplicity). It is not therefore just a question of finding a better metaphor.

My view is that forms of social distinction and inequality are produced in complex combinatories of social location in its broadest sense, forged through multiple sites. Hierarchical relations linked to social divisions are emergent and subject to historical contingencies depending on different constellations of power in different time/space frameworks. This means that social categorisations are not equally positioned or salient at all times. One or other of the divisions does not always matter in particular contexts or some may matter more than others; for example, ethnicity does not always matter either at the structural or at the identity level – it has a spectrum of intensity and identification (Pieterse, 1997). In other words, it is important to be sensitive to the relationships between social categories, rather than presuppose them.

There is also the question of dialogical and contradictory positions and positionings. In relation, for example, to issues of structural position, social divisions can be mutually reinforcing (e.g. as in the case of particular racialised migrant women), but they may also function to produce multiple and uneven social patterns and contradictory locations at particular conjunctures. For example, a man may be subordinated in class terms, but is positioned advantageously in relation to his female partner and may exercise patriarchal forms of power over her. Another example is where a woman may be subordinated as a cleaner but has a degree that gives her good life chances in some contexts. An individual may be positioned higher in one social place than another. For example, migrants who occupy a low ranking in the migration country may achieve class benefits as they display relative wealth to poorer villagers on visiting or returning to their homelands.

However, there are social locations where the dialogical articulations produce an amplification of inequality and disadvantage, and these need attending to. This involves examining both the specific combinatories that produce systemic inequalities and their transformatory potential. The approach has to be historically sensitive as there are new and emerging complex constellations of disadvantage.

Concluding remarks

This article argues that for an intersectional framing to deliver its promise there is a need to distinguish different levels of analysis in terms of questions about what is being referred to (social categories or concrete relations), arenas of investigation (organisational, representational, intersubjective and experiential) and historicity (processes and outcomes).

In this article, I have argued for the need to go beyond a focus on intersectional categories and to look at the broader social landscape of power and hierarchy. I have also argued for the need to consider the categories themselves and not only focus on their intersection, as this does not exhaust the field of investigation of these very categories. In other words, I have tried to locate this exercise in both a broader and a more defined set of questions without abandoning the intersectional project itself. Rather, the latter can be applied at specific levels in different ways but particularly relates to the level of concrete social relations. I have argued that being clear about the levels of analysis we use is one important means of delivering some of the promise of an intersectional framing.

One particular issue is the extent to which intersectionality can be incorporated within social policies that are concerned with multiple inequalities. The issues raised are difficult, as indicated by discussions of the multiple discrimination frameworks of the UK and the EU (Kantola and Nousiainen, 2009; Verloo, 2006). In terms of legal or policy initiatives, legal frameworks identify fixed legal categories and may have difficulty in dealing with complexity. This also relates to the problem of dealing with social correction on the basis of identifying people in terms of the social categories they inhabit (however intersectional) as opposed to dealing with inequalities in a more systematic structural way. There is also the difficulty of dealing with the inequality of women within their own communities and within their own families, i.e. outside what is usually thought of as the public domain. It is also difficult to identify points of intersection. Particular issues include how the equality duty (as in current UK legislation; see Equality and Human Rights Commission, 2010) can be implemented at intersectional levels and what public bodies need to do as they cannot predict which combinatories intersect in producing specific outcomes. There is also the danger of fixing some categories and making others invisible. Even if we assume that public bodies can only be responsive, there is still the problematic requirement that a subject be able to identify where the disadvantage comes from. Given these difficulties, I believe a more radical rethinking of forms of inequality and the social relations involved needs to accompany the exercise of improving public provisions to deal with discrimination and disadvantage.