The politics of purity

Introduction

In Purity and Danger, Douglas proposes that impurity occurs when anomalies present themselves within classificatory social and conceptual systems. Due to an intrinsic drive for mental and social order, ‘matter out of place’ is designated as impure, whereas that which remains within a determinate location is pure. Douglas’ account of themes of purity and impurity in Purity and Danger suggests that the overarching social or symbolic structure lies prior to any anomalies: purity/impurity classifications depend upon ‘a set of ordered relations and a contravention of that order’: ‘our pollution behaviour is the reaction which condemns any object or idea likely to confuse or contradict cherished classifications’ (2002 [1966a]: 44–5). It is this pre-existing structure which, ‘as a spontaneous by-product’, designates phenomena as impure (1975 [1968]: 58), as ‘a residual category, rejected from our normal scheme of classification’ which ‘must not be included if a pattern is to be maintained’ (2002 [1966a]: 45, 50).

Douglas asserted that it is an axiom of her work that ‘cultural categories are public matters’: the expression of the social structure for the ‘advantage of society at large’, rather than the interests of a particular group (2002 [1966a]: 48, 168); this claim was elaborated on the basis of Durkheimian theories of classification, as I have explored elsewhere. One important elaboration of Durkheimian theory in Douglas’ work was its integration with ideas from cognitive science. In the final chapter of Purity and Danger, Douglas (2002 [1966a]: 200) proposed that ‘purity is the enemy of change’, since it is the result of an innate drive within every human being for ‘hard lines and clear concepts’. She argued that such representations tend to affirm society as a whole by coercing social or mental anomalies back within the internal or external boundaries that structure society – or by mobilizing their very difference to renew these boundaries through collective rituals. Without discourses of purity and impurity, and the order they give to our reality, human life risks becoming existentially meaningless (Douglas 1966b).

When the Times Literary Supplement (1995: 39) printed a list of the ‘hundred books which have most influenced Western public discourse since the Second World War’, side by side with texts by Freud, Wittgenstein, Orwell and Churchill stood Douglas’ Purity and Danger. Moreover, the ‘anomaly theory’ of purity and impurity proposed by the text continues to serve as the major paradigm for addressing such themes today (Patterson 2011; Forde et al. 2011). In The Powers of Horror (1982 [1980]), Kristeva has affirmed Douglas’ (1966a: 44) argument that ‘dirt is matter out of place’. She asserts that it is ‘not lack of cleanliness or health that causes abjection but what disturbs identity, system, order. What does not respect borders, positions, rules. The in-between, the ambiguous, the composite’. In The Sense and Non-Sense of Revolt (2000 [1996]: 21), Kristeva describes her account as in full agreement with Douglas’ ‘first rule’ of impurity: ‘the impure is that which does not respect boundaries’; I have considered elsewhere in detail the extent to which Kristeva’s ideas on purity and impurity exceed this rule (see Duschinsky 2013a).

Urging for work beyond the anomaly theory, O’Brien (2006: 125, 143) has argued that ‘for forty years’ scholars have been led astray in theorizing the topic by the ‘formidable sway’ held by this paradigm ‘over the sociological imagination of unclean things’: the ‘thesis does not service the analytical needs of contemporary social inquiry and must be overturned’. Campkin (2007: 79) has concluded that ‘the topic as a whole has been under theorized’. Kristeva (2004: 155) herself acknowledges this: ‘My investigation into abjection … picks up on a certain vacuum’. Labrie notes that purity and impurity are just as, if not more, important in contemporary Western societies as in the societies studied by classical anthropologists – though their operation is less recognized and studied by scholars: our

desire for purity is manifested not only in nationalism and racism but also in our virtual obsession with hygiene and the pursuit of a pure art or a pure science. These manifestations obviously reflect considerable differences, which – especially given their ethical implications – are far from coincidental. On the other hand, they are parallel movements that invoke purity as an ideal’. (Labrie 2002: 261)

Against Douglas’ account of purity and impurity classifications, I shall suggest that if ‘dirt’ is indeed ‘matter out of place’ in some instances, this is a regularity rather than an explanation. When ‘upstairs things [are] downstairs’, an example Douglas gives to illustrate her theory, matter is out of place but impurity is not generally mobilized as the appropriate discursive framing. I shall argue that across various domains of discourse within (and mutatis mutandis beyond) Western societies, themes of purity and impurity are best understood as an assessment of particular phenomena in terms of what is imputed as their relative ‘self-identity’: their qualitative homogeneity and correspondence with an imputed ‘essence’ as the substratum of existence. A phenomenon is chemically pure if it contains only one element, as the basic building-blocks of matter; likewise, a citizen is taken by nationalist discourses to be pure if they are in some sense the same as other members of the nation, taken to be an underpinning essence instantiated in each of its members. The desire to exorcise impurity and maintain purity is not the result of an innate drive within every human being for order, as both Douglas and Kristeva at points suggest. It is caused by the discursive assessment and construction of subjects and other phenomena as by degrees distant or proximal to a qualitative homogeneity that is – rightly or quite wrongly – taken to be their essential truth.

Flaws in the anomaly theory

A variety of scholars across the social sciences have criticized the anomaly theory in the work of both Douglas and Kristeva for making unmediated explanatory links between categorical systems and the social structure of society as a totality (e.g. Butler 1990; Fraser 1992; Navaro-Yashin 2009; Zhong and House 2013). As Valeri (2000: 71) has posited, bypassing the realm of social practice leads the anomaly theories of Douglas and Kristeva ‘to speak of “the system”, of “form”, or of “order” as if they were one monolithic thing’. In fact, however, ‘there are many coexisting orders of classification; what is residual to one may be central to another’. Hetherington (2004: 163) has offered criticisms of the anomaly theory which agree with Valeri’s. He argues that the anomaly theory takes the classificatory system ‘as a stable and representable thing’ and as prior to the anomalies that it designates. Other scholars, such as Beidelman (1993: 1066) and Asad (2007: 77), have added that the anomaly theory presents few adequate tools for discerning processes of social oppression as a result of its functionalist commitment to social consensus.

Alley (2002: 45) and Anttonen (2004: 113) have discerned that fundamental to the problems with the anomaly theory is a lack of clarity regarding ‘purity’ as a construct. In Purity and Danger, Douglas contests the work of prior theorists who have assimilated the sacred to the pure or made appeal to the ‘cryptic’ notion of the ‘ambiguity of the sacred’. However, her own solution to this matter is unclear, since ‘purity’ is a residual term in the explanatory framework of Purity and Danger despite appearing in the title of the text. On the one hand, Purity and Danger (2002 [1966a]: 68–9) claims that things that ‘conform fully to their class’ are sacred and are symbols of the ‘divine order’ that align the world to its true categories. This perhaps slips back into an alignment of the sacred with the pure. On the other hand, the anomaly theory suggests that anomalous objects or individuals are usually treated with revulsion, but sometimes can also become sacred sources of the renewal and affirmation of society: ‘we therefore find corruption enshrined in sacred places and times’ (2002 [1966a]: 220). Both claims have been taken up independently in the secondary literature, with major scholars defending these two opposed interpretations (e.g. Moore 2000: 17; Eagleton 2003: 289). This confusion was an issue that Douglas (1975 [1972]: 285, 307; 1998: vii–viii) came to recognize in the practical application of her ideas to particular cases; however, she did not return to further address the relationship with a new social theory.

In her later work, Douglas (1980, 1997; 2004: 160) acknowledged that the anomaly theory is not an adequate account of purity and impurity. Attempting to explain this inadequacy, Douglas (2005: 95; 2002 [1966a]: xvii) acknowledges that her conservative social commitments and her ‘kindly feeling for hierarchy’ had led her work towards ‘praising structure and control’ in the operation of themes of purity and impurity. Douglas (1996 [1990]; 1992) suggests that the various problems that critics have successfully identified with the anomaly theory are caused by two false assumptions: that society is a unitary whole; and that there exists a unitary ordering mechanism within the human mind. In an informal seminar in 1997, Douglas states that she has come to recognize that there is in fact no universal desire for either cognitive or social order at the base of purity/impurity designations. Against a core assumption of the anomaly theory, Douglas further admits that there is no ‘intrinsic value to purity’ for the individual or for society. It is striking that Douglas herself proposes that there is a need for further social theoretical reflections on the topic, which address the specificity of purity/impurity outside of accounts of ‘intrinsic value’ such as the idea of sacredness: ‘the only thing universalistic about purity is the tendency to use it as a weapon or tool’ (Douglas, 1997).

By the time of Missing Persons (Douglas and Ney 1998), Douglas is willing to fully retract the claim that there is a universal human urge for a complete and ordered cognitive system. She asserts that the most common responses to social or cognitive anomalies tend to be simply to ‘tolerate ambiguity’ (Douglas and Ney 1998: 15). Douglas’ work thus bifurcates. On the one hand, when she subsequently discusses the way that anomalies may or may not be treated as dangers in contemporary Western societies, this discussion does not make even passing reference to purity and impurity (Lianos and Douglas 2000); and when her followers have wished to address the themes of purity and impurity that are present beyond integrated and hierarchical groups, they have therefore had to return to the anomaly theory of Purity and Danger (e.g. Hendriks 2011). On the other hand, writing of purity and impurity in the Hebrew Bible, Douglas maintains the functionalist conclusions of the anomaly theory, shorn of its explanatory framework. Douglas (1993: 41, 55–9) writes of ‘the magnanimity of Numbers’ political message’, which operates ‘above all by a benign doctrine of defilement’. She discerned ‘no sign’ of conflict ‘in Numbers and Leviticus’, since society is ‘controlled quite simply by rules against defilement’ (see also Douglas 1994: 112–13; 1998: 97–101; 2004: 194–5).

Purity as self-identity

The term ‘essentialism’ has been used in the social sciences to refer to a cluster of processes: the stabilization of a phenomenon within discursive practices such that it appears to be solely the expression of its own essence; or such that it appears to be necessary rather than contingent; or such that it is associated with nature rather than culture (Narayan 1998; Heyes 2000). Hacking (1999: 17) has pointed out that the term has become a ‘slur word’, making its meaning unclear. By contrast, I intend a quite specific sense of the first signification listed above. The essences that I wish to identify as particularly associated with the operation of purity and impurity discourses do not contain any heterogeneous, foreign or inferior elements; all of their elements are ‘the same’ in some relevant sense. Furthermore, the essences in question are understood to be situated at the conceptual or ontological ground of their instantiations. In Book II the Science of Logic (2010 [1832]: §11.351), Hegel expends great effort to pick out these two aspects of ‘essence’. He presents the technical term ‘self-identity’ to specifically refer to essences that: 1) contain only elements of the same kind; and 2) serve as the ontological ground for their instantiations.

Self-identical phenomena are those that are ‘identical’ in all of their constitutive elements, and which are ‘identical’ with their ontological ground. These are properties which scholars sometimes attribute to ‘essence’ without further discussion (e.g. Grosz 1994: 47; Hird 2006: 174), but which must be drawn out explicitly in order to extract the term ‘essence’ from its position as an amorphous ‘slur word’. In fact, they are particularly characteristic of – though certainly not exclusive to – certain strands within the Western cultural tradition (Moore 2000; Foucault 2013 [1971]), which invoke purity and impurity on the assumption that

truth is purity; for a thing to be true is for it to be completely and purely itself. True wine is that which is wine only, unmixed with any foreign element. Thus, the truth of a thing consists in its simplicity or lack of composition; a thing is true because it is one. It is good for the same reason. It has well-being when it is united with itself and its principle; for a thing to be good is for it to be itself, pure and unadulterated. (Collins 1974: 16)

Or again, Caillois (1959 [1950]: 33–4) observes that: ‘that is called pure, substance or line, whose essence is not mixed with anything that may alter or debase it’.

Not all cultures, or all discourses within any culture, conceptualize purity and impurity in terms of self-identity (Duschinsky 2011b). However, within such a frame – which is facilitated by the cultural heritage of Western societies but is not limited to them – qualitative homogeneity is taken to simply be correspondence with essence, and vice versa: such purity/impurity discourses map a distinction between homogeneity and heterogeneity onto a distinction between ontologically primary and secondary. Difference from the purported origin is identified with the intrusion of heterogeneous, foreign or inferior elements into a pristine and prior essence – or as the recognition of such elements which must have been in the phenomena that had been taken to be pure all along. Figurations of ‘childhood innocence’ neatly illustrate this identification of difference from a purported origin with contamination within discourses invoking purity and impurity (Duschinsky 2011a, 2013b); Foucault (2013 [1971]: 24–5) offers another illustration in discussing how purity and impurity discourses in Western societies operate as guarantors that the truth of the world and of human relations is knowable without exteriority, thereby situating as a contamination of this truth ‘the struggle of instincts, partial selves, violence, and desires’ that necessarily organize the will to knowledge.

The contingency of the construction of ‘essence’ appealed to within Western discourses of purity and impurity is illustrated by the fact that the truth of existence can and has been characterized very differently (Coole and Frost 2010; Parlati 2011). Quine (1976: 499), for instance, described ‘the Aristotelian’ account of essences as utterly ‘tenuous’ in the light of the discoveries of quantum theory. He allowed that people may legitimately speak of ‘essence’ within particular contexts, since ‘some predicates may play a more basic role than others’ (1981 [1977]: 120–1). However, he believed that the more general notion of ‘essence’, as a homogeneous and originary foundation for existent phenomena, numbered firmly among the ‘scientifically undigested terms of metaphysics or religion’ (1986: 157). Quine’s work illustrates that the idea of essence as a qualitatively homogeneous ground for all real phenomena is only one, contingent, way of characterizing reality in contemporary Western societies.

A conflation of biblical, non-Western and Western purity/impurity discourses can be situated as the cause of the fluctuating account of the relationship between ‘purity’ and ‘sacredness’ which may be observed in Purity and Danger. When addressing the scriptural prescriptions for forming the Hebrew nation (e.g. Douglas 2002 [1966a]: 68–9), the anomaly theory predicts that pure phenomena are ‘holy’ when they remain within categorical boundaries; when drawing on ethnographic findings from Africa (e.g. 2002 [1966a]: 220), it predicts that ‘sacredness’ is the property of anomalous phenomena when they are used to affirm society as a whole. Moreover, as Douglas later admits, the anomaly theory is not applicable to numerous other areas of purity/impurity discourse within the Hebrew bible, such as discussions of sacrifices (Douglas 1998: vii–viii, 2004: 111, 160; Klawans 2005).

In contemporary Western societies, politics, religion, race, law, sexuality, embodiment, class, food, psychology, geography, health, and science all make appeal to purity and impurity. Thinking with Bourdieu, it can be noted that these domains each have in common that they are fields of practice within which relatively contestable claims can be made about the essence of reality or human life, building upon the particular Western cultural heritage which separates essence from existence – ‘a form of fetishism which is expressed in almost identical forms in the various fields’ (Bourdieu 2000 [1997]: 114). Discourses of purity and impurity can be conceptualized as mobilized by different relational subjects in fields in which competing visions of ‘the origin, the spirit, the authentic essence’ are precisely at stake in their interaction and struggle with one another, such that ‘“Essentialist thought” is at work in all social universes and especially in the fields of cultural production – the religious field, the scientific field, the literary field, the artistic field, the legal field, etc. – where games which have the universal at stake are played out. But it is quite clear in that case that “essences” are norms’ (Bourdieu 1993 [1976]: 74; 1996 [1992]: 298–9; see Duschinsky and Lampitt 2012).

Even forms of physical dirt, which the anomaly theory treats as synonymous with impurity, are only likely to become coded as impure and bad when, by degrees, they are constructed as deposing or decomposing a phenomenon taken to be underpinned by a homogeneous, originary and valued essence by actors within a field able to make such claims. In the laboratory, for example, ‘not only is what is to count as polluting variable, but the desirability of purification is itself contestable. Sometimes it is best not to remove contaminants, nor to transform a polluted thing into an unpolluted one’ (Mody 2001: 29); one might also consider the role of dissonance in music. Household dirt can indeed be impure, as Douglas observes. But this is not due to the presence of a single classificatory system, but rather due to the interaction of quite different essentializing discourses, such as on human dignity, health and sanitation, domesticity, femininity and race (see Heneghan 2003; Berthold 2010). It is this discursive, material and affective organization, rather than a desire for ‘order’ built into human nature, that makes the bathroom, the kitchen and the bedroom – as sites where heterogeneity and embodiment are negotiated – such crucial spaces associated with contemporary Western purity/impurity discourses.

Douglas (2002 [1966a]: 44) uses dirt as a synonym for impurity, and Kristeva (1982 [1980]: 96) presumes that blood is impure because it connotes the menstrual blood of female ‘fecundation’. Yet, for instance, Klemperer (2000 [1957]) has documented the role that ‘blood and soil’ may play as quintessential metaphors of purity, with the very materiality of the substances serving as semantic support for their deployment as symbols of national self-identity: a qualitatively homogeneity (whether of culture, biology, skin colour, faith, language, etc.) as the seat of belonging, and the position of every subject as manifesting the nation as a fundamental ontological principle (see also de la Cadena 2005; Mavroudi 2010; Butterworth 2010). Thus even forms of bodily waste are only likely to become coded as impure and bad when, by degrees, they are constructed as deposing (or decomposing) self-identity. Within the frame provided by such discourses, physical congress risks the entry of heterogeneity and thus a ‘falling away’ from an ideal of phenomena or forms of subjectivity as standing solely upon its own essence.

From regularity to explanation

Popper (1963: 326; 1981 [1975]: 94) argues that new theoretical work ‘must always be able to explain fully the success of its predecessor’; whilst not wishing to take such a strong position, I do think it is helpful to consider the reasons why the anomaly theory has been successful. The anomaly theory is often correct when it suggests that exceptions to social and cognitive norms tend to be classed as impure. It hits upon a regularity, but more precision is needed in considering what kind of classification is in play as they do not all lump and split phenomena, or operate socially or epistemologically, in the same way (see Zerubavel 1996). Discourses of purity/impurity do not attend the breach of any classificatory norm as a ‘spontaneous byproduct’, as Douglas predicts. Rather, they are best conceptualized as discourses that adjudicate the relative correspondence of phenomena and forms of subjectivity with an imputed self-identity. For instance, a blob of ink might be an irritating disruption but it is not perceived to disturb the ‘purity’ of a sheet already covered in polka dots, even though categorical boundaries have been disturbed; this blob might well, however, be considered an impurity if it disturbs a sheet of a qualitatively-even colour such as red or green, or – especially – white (see Warren 2003; Duschinsky and Brown 2014).

What distinguishes Douglas’ examples of ‘dirt as matter out of place’ that are in line with her theory (e.g. animal species, geographical spaces, types of human being) from those that disobey her own theory (e.g. ‘upstairs things downstairs’, ‘under-clothing appearing where overclothing should be’) is that in the former cases phenomena are assessed in terms of their self-identity: their qualitative homogeneity and correspondence with essence. These properties are sometimes presumed but sometimes lost in Douglas’ concept of the ‘anomalous’; the contemporary sense of the concept is more expansive, but perhaps retains a flavour of its original Greek meaning of a qualitative unevenness (see e.g. Plato, Laws 1.625d referring to terrain). Douglas’ theory that ‘dirt is matter out of place’, illustrated with the position of animal species as food in relation to the nation, addresses a trio of sites precisely in which reference to self-identity (qualitative homogeneity and correspondence with essence) is particularly salient. The examples which best illustrate Douglas’ anomaly theory in Purity and Danger are those in which the purity/impurity classification in question is one or more of:

Firstly: relating to the nation, community or family, when these are constructed as discrete and essential identities.

Secondly: taxonomic in form. Douglas does not theorize the different epistemology of various classificatory processes, and thus deploys the terms ‘taxonomy’, ‘classification’ and ‘identity’ interchangeably (see e.g. Douglas 2002 [1966a]: 197, 206). Yet taxonomies are a special form of classification in which there is a systematic arrangement of self-identical phenomena within exclusive categories (cf. Foucault 1970 [1966]: 164–74; Valeri 2000; Hacking 2007). The point can be illustrated through the example of territory: phenomena in the wrong section of land are not automatically impure, but will be designated as such if they breach territorial ‘borders’ which, like those of the nation, are taken to demarcate a discrete domain that manifests an essence within space. In this light, it is no coincidence that Douglas’ theory of impurity as a breach in a categorical system was first developed in the 1950s (see Douglas 1965 [1955]) with reference to species of animals, where these species were taken to be each the same as others in the category and where the boundaries of the categories were taken to reflect ontological divisions in the nature of the world.

Thirdly: relating to phenomena which are considered to be natural substances within the particular discourses of the society in question. Though chemistry may show their complexity, ‘natural substances’ are those which everyday ontologies discursively situate as elemental, such that they stand in a particular (even perhaps synecdotal) intimacy with a form of essence (e.g. human flesh, milk, blood and soil). Carsten (2004: 111), in analysing substances involved in cross-cultural discourses on kinship, defines natural substances as types of matter, which are taken themselves as distinct forms of essence, and out of which other entities may be composed. Like ‘essence’, ‘substance’ is a culturally contingent term originating in the Medieval interpretation of Aristotle’s concept of ouσíα (Marcuse 1968 [1938]), and means a phenomenon which manifests its own essence in its materiality. When Douglas uses the term ‘matter’ to mean any phenomenon, she is in fact making use of the fact that matter is a substance, and thus ‘matter’ in the place where other ‘matter’ should stand necessarily means a contamination of essences.

Leaving all production behind

In Western cultures in particular, impurity is thus not best theorized as ‘matter out of place’ according to a pre-existing classificatory system, but is better conceptualized as a representation of elements that are discursively situated as drawing phenomena away from their self-identity: their qualitative homogeneity and correspondence with their imputed essence. The ascription of purity or impurity is therefore a covert form of ‘Western metaphysics’; indeed Plotnitsky (1991: 1210) has argued – drawing on Derrida – that, in general, ‘Metaphysics, including the metaphysics of justice, is forgetting or blinding oneself to the violence of the pure.’

Within discourses where purity marks the degree of self-identity, pure phenomena and forms of subjectivity are constructed as dependent on or marked by nothing beyond themselves: they appear to depend upon no more than their own discrete essence. As a result, pure phenomena or forms of subjectivity seem to have essential and discrete boundaries, and no meaningful debt to anything or anyone. Appearing to depend exclusively on their own essence for their presence in social practice, pure phenomena may thereby be situated as outside of history and devoid of relations of power. Pure processes, things, or people appear to be simple expressions of essence, with no dependence on anything outside of themselves: for instance, in their analysis of the social production of the division between science and technology, Bauchspies, Croissant and Restivo (2006: 12) have documented that ‘once something is labelled “pure”, it symbolizes that it is free of entanglement and relationship. Labelling something “science” in essence tends to declare its purity and ignore all the antecedent work that went into its construction.’

Pure phenomena are also figured as devoid of and prior to complexity and the dynamics that organize social and material inequalities. As Bourdieu has shown in his analysis of cultural consumption (see Duschinsky and Lampitt 2012), it is precisely this feature which makes purity and impurity a discursive, material and affective resource peculiarly adapted to facilitating social consensus, compelling a shared practical demand to protect or attain purity through the deployment of mechanisms of social exclusion, and social and self-regulation. Douglas’ functionalist perspective intuits this process, but does not trace it and so is unable to do other than valorize any existing purity/impurity classifications as necessary for the maintenance of society as a whole.

Heidegger (1973 [1964]: 6) suggests more generally that discourses that depict phenomena as expressions of essence may hide what are in fact their social and political conditions of possibility. It ‘leaves all production behind and is thus immediate, pure: being in presence’. Classifications of purity/impurity can therefore be used to paper over the social or epistemological problems of essentialization. I believe that Douglas (2002 [1966a]: 202) is attempting to discern this when she states that ‘whenever a strict pattern of purity is imposed on our lives it is either highly uncomfortable or it leads into contradiction if closely followed; and if not observed, hypocrisy. That which is negated is not thereby removed.’ In treating purity as a representation of any classificatory system, however, Douglas does not identify the kind of discursive operation enacted by purity and impurity discourses, which impute an essence to reality in the course of merely assessing the self-identity of phenomena. Discourses of purity/impurity reduce the moral and epistemological significance of complexity, and in doing so permit the assemblage of powerful narratives. Their tacit appeal to essence helps make them apt tools for justifying and operating processes through which actors achieve grapple on other members of a particular field of practice or on the population more generally. This semantic organization means that appeals to purity and impurity are peculiarly adapted to the strategic smuggling of assumptions into debates about the true nature of people, processes or the world. I would like to term this process ‘the politics of purity’, and regard it as an explanation for the elective affinity between purity discourses and black-and-white worldviews.

In order to achieve a representation of correspondence with a natural and essential origin, the practical staging of the representation of pure phenomena and forms of subjectivity may need to be both disavowed and regulated. This is not to say that phenomena in contemporary Western societies that are assessed with respect to their purity simply do not exist or are ‘mere’ social constructions. My interpretation is that the processes that permit a phenomenon to seem a direct expression of an essence are organized to varying degrees by elements that must be sidelined as merely supplementary (see Duschinsky 2013c). Scientific truth, childhood, femininity, rural spaces, whiteness, intentionality, water, artistic beauty, diamonds, mathematics, moral values, religious belief, virginity, or organic food – each is, to differing extents, dependent for its existence on more than its expression of an essence. To take one example, in order for a new child to instantiate the nation as a timeless essence, a whole messy and contingent biopolitical process is necessary (Lenoir and Duschinsky 2012).

Purity and impurity discourses can therefore be considered as unusually well adapted to activating a powerful epistemological move, masking relations of dependence. For instance, Haraway (1991: 204) has described how scientific study of ‘the immune system’ has increasingly revealed the contingency of what are generally taken to be natural boundaries, such as the alignment between inner/outer, mine/yours, pure/impure. Antigens, for instance, can potentially be classed within either or both of these poles. Yet, at the same time, metaphors associated with the immune system are deployed in social and political discourses to situate the ‘inside’ as homogeneous and originary compared to an outside, producing a narrative that frames a pure self in danger from or the victim of an impure invader.

At the heart of purity and impurity discourses is the strategic construction of phenomena as in a state of relative fidelity or infidelity, integrity or perversion, through their assessment in terms of self-identity: qualitative homogeneity and correspondence with essence. This is significant because many important phenomena and forms of subjectivity can only be discursively positioned as simply expressions of essence if the heterogeneous and temporalized processes upon which they depend are sidelined as merely contaminants, rather than always already constitutive of the phenomena. Pure ideals may be used to justify the authority of the actor in the political field, for example, and to make his or her person and programme natural and neutral, a ‘delegate’ of an essential ideal outside and above sectional interests (Bourdieu 1991 [1975]). This political mobilization occasions the operation of discourses of purity as symbols of society as a whole, a process which Douglas does identify but which she reifies and depoliticizes as the necessary ‘function’ of purity images in discourse.

Conclusion

Classifications of purity/impurity in modern society should not be understood, following Douglas, as a homeostatic system, driven by a universal human desire for ‘order’. It is proposed here, instead, that purity/impurity discourses will primarily be mobilized to characterize a breach in a norm or classificatory system when phenomena are being adjudicated for their relative ‘self-identity’: whether they are qualitatively homogeneous and in correspondence with their imputed essence. Within such a discursive frame – which has been facilitated by the cultural heritage of Western societies, but which may certainly occur elsewhere and elsewhen – pure phenomena are those which have been depicted as devoid of heterogeneous, foreign or inferior elements. By virtue of their imputed close and transparent association with essence, such phenomena are seen as dependent on or marked by nothing beyond themselves. Sometimes – as in the case of chemical purity – phenomena require little or no sleight-of-hand to be understood as pure, without anything disrupting this representation. Other cases may occur, however, in which representations of purity occur regarding phenomena which are thereby depicted as the simple expression of a self-identical essence, rather than in part the product of contingent historical and material processes. In such cases, discourses of purity and impurity can be deployed to smuggle unwarranted, and often political, assumptions into debates about the true nature of people, processes or the world. I have termed this operation the politics of purity.

In offering this analysis, I have not intended a totalizing characterization or evaluation of all purity and impurity discourses. My investigation has sought to examine when, actually, dirt is matter out of place – discerning the reason for the regularity observed by Douglas, and in doing so presenting a more critical account than Douglas of the politics of purity. Two qualifications are fully needed in closing, however. First is the issue of cultural variability: the way in which ‘essence’, full presence, has been understood as the foundation of existence in Western societies. Marcuse (1968 [1936]) has facilitated the forms of purity and impurity discourse under discussion here, which assess qualitative homogeneity and correspondence with essence as equivalent to one another. I do not agree with Douglas (2002 [1966a]: 43) that ‘the difference between pollution behaviour in one part of the world and another is only a matter of detail’. Yet, as Birnbaum (2003) has argued, an analysis of discourses characteristic of the cultural heritage of Western societies may have applicability, mutatis mutandis, for those examining particular contexts around the world which are selectively incorporating and reconstituting Western purity/impurity discourses. For instance, surveying 20th-century genocides in comparative perspective, Savage (2007: 405, 417) has stated that appeals to purity and impurity ‘premised on, or deriving from, the discourses of Western imperial modernity can shape action in many non-Western countries, whether through the legacy of colonialism or through global influence’. He has concluded that ‘such rhetoric is commonly found in the words of the perpetrators of genocidal episodes’. Addressing a quite different case, Mankekar (1997) has also observed such a spread of purity/impurity discourses:

In post-colonial India, coexisting with theories that posit that children bring with them memories of a previous life, are modernist discourses that conceive of childhood in terms of a linear progression into sovereign adult/subjecthood. These modernist constructions of childhood reveal a tendency to think of children as ‘uncontaminated’ by the temporal world in which they live and, therefore, as a space of purity. (Mankekar 1997: 58)

A second point to make in closing relates to the ethical valence of purity and impurity discourses. Though our age has been ‘scarred by the actions of regimes in pursuit of purity’, Mackenzie (2004: x) has urged that we recognize that purity and impurity can help advance the work of critical thought. He states that ‘purity is an ideal that secures many of our most deeply felt attachments to our sense of self, our relations with others and the ebb and flow of life’, and that the categorical rejection of purity and impurity discourses by critical theorists has left few tools for assessing the ethical potential of different discourses. The purity of art, of utopia, of love – can these hold emancipatory potential? The politics of purity has here been identified as a form of essentialism, facilitated by the way that we generally understand purity and impurity as assessments of self-identity. And, indeed, Spivak (1988: 205) has cautiously valorized the ‘strategic use of a positivist essentialism’ as a political strategy; without appeal to such collective categories anti-oppressive projects may become difficult to defend. My analysis suggests that purity and impurity discourses used to assess self-identity are not simply ‘bad’, but that they are perhaps more ‘dangerous’ than possible alternatives in the way in which they appear to merely judge but in fact also construct the difference between an ideal and reality. Yet purity and impurity discourses to be found in the past, in other cultures, or particular domains of Western society which are less oriented by the division between essence and existence may, as I and others (e.g. Foucault 2013 [1971]; De Nooy 1998; Moore 2000; Duschinsky 2011b, 2013a) have explored elsewhere, offer a means of further specifying what the ethical possibilities of these discourses might be.