Butler’s Sophisticated Constructivism: A Critical Assessment*

Abstract

This paper aims to investigate whether and in what respects the conceptions of the body and of agency that Judith Butler develops in Bodies That Matter are useful contributions to feminist theory. The discussion focuses on the clarification and critical assessment of the arguments Butler presents to refute the charges of linguistic monism and determinism.

During the last decade, a new paradigm has emerged in feminist theory: radical constructivism. Judith Butler’s work is most closely linked to the new paradigm. On the basis of a creative appropriation of poststructuralist and psychoanalytical theory, Butler elaborates a new perspective on sex, gender, and sexuality. A well-known expression of this new perspective is Butler’s thesis, in Bodies That Matter (1993), that not only gender but also the materiality of the (sexed) body is discursively constructed.

Since the publication of Butler’s Gender Trouble in 1990, however, the emergence of radical constructivism as a new paradigm has not gone uncontested. One of the most focused critical assessments of this new paradigm takes place in the context of the discussion in Feminist Contentions (Benhabib et al. 1995) among Seyla Benhabib, Butler, Drucilla Cornell and Nancy Fraser on the merits and dangers of postmodernism. Especially Benhabib and Fraser formulate important critical objections and questions with respect to subjectivity, agency, and the political-normative force of Butler’s work. Benhabib accuses Butler of a complete debunking of any concepts of selfhood, agency, and autonomy (Benhabib et al. 1995, 21), and she doubts whether Butler’s theory of performativity can explain not only the constitution of the self but also the resistance that this very self is capable of in the face of power/discourse regimes (Benhabib et al. 1995, 111). According to Fraser, Butler’s Foucaultian framework is structurally incapable of providing satisfactory answers to the normative questions it unfailingly solicits, for instance the question of how to distinguish between better and worse practices of subjectivation (Benhabib et al. 1995, 68).

Butler, however, attributes most of the objections to a misunderstanding of radical constructivism and/or poststructuralism that tends to reduce both to linguistic monism and determinism. The claim, for instance, that language constitutes the subject does not entail that language fully determines the subject (Benhabib et al. 1995, 135). Likewise, the notion of agency as the effect of discursive conditions does not entail that these conditions control the use of agency (Benhabib et al. 1995, 137). Both in her second contribution to Feminist Contentions (1995) and, more explicitly and extensively, in Bodies That Matter (1993), Butler defends her position as a sophisticated version of radical constructivism that overcomes the charges of linguistic monism and determinism. ¹

These charges have to be countered if radical constructivism is to be validated as a feasible or useful paradigm for feminist theory. A determinist theory, that is, a theory that does not allow a viable concept of agency, obviously debilitates the feminist project. Though the practical consequences of linguistic monism are less obvious, a theory that negates materiality by reducing it to some sort of linguistic substance is not a good starting point for a feminist theory of the body. For this reason, I focus my discussion of radical constructivism almost exclusively on the ways in which Butler tries to refute the charges of linguistic monism and determinism. Her Bodies That Matter (1993) is in my opinion, up to now, the most extensive and theoretically sophisticated account of the radical constructivist position. Instead of recapitulating and commenting upon the already wide and varied discussion of Butler’s work and its poststructuralist framework, I intend to assess the viability of the radical constructivist position by scrutinizing the claims and lines of argument in this book that underpin this position. Furthermore, as Butler’s arguments in Bodies That Matter (1993) are often intricate and sometimes quite impenetrable, an in-depth discussion of several arguments serves to clarify partially the radical constructivist position. There is some truth in Butler’s suggestion that her position often is misunderstood, but this, I’m sure, is at least partly because of her writing style. Daunting in its incessant use of highly abstract jargon, not seldom confusing in its rhetorical effects, and often implicit in its argumentation, Butler’s is one of the most difficult but also one of the most provocative texts I have been reading the past few years.

I. Linguistic Monism: the Question of the Body

How does the problem of linguistic monism arise? One of the aims of Butler’s Bodies That Matter (1993) is to deconstruct the notion of the body as a natural, prelinguistic given. This is how Butler puts it: “The body posited as prior to the sign, is always posited or signified as prior. This signification produces as an effect of its own procedure the very body that it nevertheless and simultaneously claims to discover as that which precedes its own action. If the body signified as prior to signification is an effect of signification, then the mimetic or representational status of language, which claims that signs follow bodies as their necessary mirrors, is not mimetic at all. On the contrary, it is productive, constitutive, one might even argue performative, inasmuch as this signifying act delimits and contours the body that it claims to find prior to any and all signification” (1993, 30).

Butler’s deconstruction of the body as a natural given results in the claim that the body is always already linguistically constructed. Obviously, this claim evokes the charge of linguistic monism: doesn’t the claim entail a sort of linguistic metaphysics of the body? ² What needs to be examined, however, is the exact import of this claim: is it an ontological or an epistemological claim? Does the claim entail that the body is ontologically coextensive with its linguistic constructions, in other words, that the body is nothing but a collection of linguistic constructions? Or does it imply that the body is only epistemologically accessible as a linguistically constructed body? Only the former, not the latter, would justify the charge of linguistic monism.

I examine two lines of argument that in Butler’s opinion undercut the charge of linguistic monism. The first one, concerning the notion of referentiability, can be construed as a general epistemological argument about language and its relation to reality. The second argument is more complex, beginning from the claim that language is the condition of the appearance of materiality. The import of this claim is ambiguous; it can be construed as either ontological or epistemological. Though I conclude that Butler succeeds in refuting the charge of linguistic monism, the way in which she solves this problem raises new questions. On the one hand, she ends up defending an epistemological position that is not only too restrictive but also, in my opinion, has negative consequences for a feminist and queer theory of the body. On the other hand, certain passages suggest another, more phenomenological approach that, though hardly elaborated, opens an interesting and more fruitful perspective on such a theory.

Referentiability

The claim that language constructs the body does not mean that language originates, causes, or composes exhaustively that which it constructs. Rather, it means that there can be “no reference to a pure body which is not at the same time a further formation of that body” (Butler 1993, 10). In more general terms, instead of merely describing or referring to a body that is simply there, the constative or referential use of language is “always to some degree performative” (Butler 1993, 11). This argument can be interpreted in the following way. Though referential or constative language seems to offer a direct connection to extra-linguistic reality, it depends on the prior semantic definition of the words used. The statement “this body is female” for instance, can be uttered only by a speaker who has acquired some knowledge of what these words mean. Referential or constative language is not only dependent on the prior semantic definition of the words used; precisely on the basis of that semantic definition, it also effects a certain delimitation of what is taken as extra-linguistic reality. Apart from referring to extra-linguistic reality, the statement “this body is female” at the same time delimits the body it refers to as female, with all the connotations this term carries. If the words “this body is female” appear to represent or mirror an extra-linguistic reality, then it is only because, in everyday usage, we systematically forget that the possibility of referential or constative language depends on prior semantic definitions of the words used, delimiting or highlighting the reality that is referred to in certain ways. In other words, we are so accustomed to the referential use of terms such as body and female that often we are not aware of the ways in which the connotations of these terms delimit and inform our view of extra-linguistic reality.

Yet the fact that referential or constative language introduces a certain semantic construction of the reality that is referred to does not imply that there is nothing but these semantic constructions. That the body is linguistically constructed does not preclude that it has, as it were, a life of its own. Rather, Butler’s argument implies that language always already mediates our knowledge of the body, of reality in general: to have an idea of what a female body is, we need to know the meaning of the words body and female. This epistemological argument does not, by itself, entail the ontological conclusion of linguistic monism that reality consists of or is reducible to some sort of linguistic substance. What, if any, ontological conclusions are to be drawn from this argument is an open question.

Language as the Condition of Appearance of Materiality

In the context of a similar argument about referentiability, Butler proceeds with what looks like a much stronger claim: “Can language simply refer to materiality, or is language also the very condition under which materiality may be said to appear?” (1993, 31). Though put in the form of a rhetorical question, the implied claim that language conditions the appearance of materiality is confirmed subsequently in Butler’s discussion of Foucault: “‘Materiality’ appears only when its status as contingently constituted through discourse is erased, concealed, covered over. Materiality is the dissimulated effect of power” (Butler 1993, 251). The quotation marks around materiality indicate that Butler refers to the common, everyday notion of materiality as a given, extra-linguistic reality. The long-established discourse on the naturalness and givenness of material entities and bodies conceals the fact that language constructs material reality. Because of this erasure bodies and material entities are perceived as a given, extra-linguistic reality.

The import of the claim that language conditions the appearance of materiality differs from the one discussed above because of the introduction of the notion of “appearance,” an ambiguous notion with strong ontological and epistemological connotations. Depending on one’s interpretation of “appearance,” the claim either implies linguistic monism or a very strict epistemological assumption.

In the phenomenological-ontological tradition, the notion “appearance” is one of the key terms. For Martin Heidegger, being is synonymous with appearing; only in so far as it appears, something can be said to exist, to be there. ³ In this sense reality is all that appears. According to Heidegger, language enables the interpretation of what appears, it does not enable that things appear. Hence, reality is epistemologically dependent but ontologically independent of language. The ontological independence of reality is reversed into dependence if, according to the above claim, language conditions the possibility of appearance. In so far as it enables that things appear, language determines the limits of reality. As language is contingent and variable over time, so are the limits of what we call, and perceive as, reality. Though much more sophisticated and defensible than the absurd conclusion that reality consists of some sort of linguistic stuff substance, this position still amounts to linguistic monism for it precludes the possibility of an ontologically independent, extra-linguistic reality. What appears or is perceived as extra-linguistic reality is still a hidden effect of language or, in terms of Butler, the dissimulated effect of discursive power.

In the Kantian tradition “appearance” is an epistemological notion, albeit one with ontological repercussions. According to Immanuel Kant, knowledge of reality is restricted to appearances (phenomena). What things are in or by themselves (as Ding an sich) we cannot know. The upshot of this argument is that knowledge does not determine the ontological limits of reality. Hence, the ontological independence of reality is preserved even though it is epistemologically beyond our reach. Following this Kantian logic, Butler’s claim can be interpreted as denying, not the possibility as such of an ontologically independent, extra-linguistic reality, but the possibility of access to an ontologically independent, extra-linguistic reality. In this Kantian context, the claim that language conditions the possibility of appearance means that language, as the epistemological condition of accessibility, determines the way in which reality appears to us. Thus, if materiality appears as a given, extra-linguistic reality, then this extra-linguistic reality is still an effect of language, but materiality or, in general, reality, is not reducible to these linguistic effects, that is, to the semantic constructions of language.

May we conclude from the Kantian interpretation of Butler’s claim that reality, in so far as it is not reducible to language, is ontologically independent of language, in other words, that language does not determine the ontological limits of reality? And, hence, that the charge of linguistic monism is refuted? I think so. Butler, in a recent Dutch interview, dispels the ambiguity that clings to many passages on this topic in her Bodies That Matter (1993) by explicating her position as a post linguistic turn Kantian position: “the ontological claim can never fully capture its object, and this view makes me somewhat different from Foucault and aligns me temporarily with the Kantian tradition as it has been taken up by Derrida. The ‘there is’ gestures toward a referent it cannot capture, because the referent is not fully built up in language, is not the same as the linguistic effect. There is no access to it outside of the linguistic effect, but the linguistic effect is not the same as the referent that it fails to capture” (Butler 1997, 26). ⁴

In whatever way language constructs reality, no construction nor all constructions together can ever fully capture it. With a linguistic turn of Kant, Butler says that the reach of language and, hence, the reach of knowledge is limited: the signifying processes of language always leave, as it were, an ontological remainder. But if language does not determine the ontological limits of reality, it does determine the epistemological limits of access to reality. According to Butler, there is no access to reality without language. This epistemological assumption is more strict than the one that follows from the claim I discussed in the previous section for it implies that language conditions not only the intelligibility of reality but also its accessibility. I have my doubts about this assumption. To posit language in both cases as condition implicitly equates intelligibility and accessibility. Yet isn’t the reach of the latter wider than the reach of the former? Can’t we have access to phenomena we do not understand?

Intelligibility and Accessibility

Whereas everything that is intelligible to us is also accessible to us, the reverse is not true. Phenomena that are intelligible to us are phenomena we do understand in some way or other. At the most basic level, to understand something means to be able to name or refer to it. As understanding involves the capacity to name, to refer, or to articulate that which is understood, it is always mediated by language. To equate intelligibility and accessibility would mean that we cannot have access to phenomena we do not understand, that is, phenomena we cannot articulate. That does not seem plausible. By following the hermeneutic model of understanding, I try to show that we can have access to phenomena we do not understand, that is, cannot articulate, though this access is not completely independent of linguistically mediated understanding.

In daily life, our behavior and actions are guided by a mostly implicit understanding of the world we inhabit, an understanding that is based upon the ways in which this world is semantically constructed. Even so, our daily routines are on occasion slightly, and sometimes profoundly, disrupted because we are confronted with people, situations, actions, images, texts, things, bodily sensations etcetera that defy our understanding. The context of habitual understanding enables these confrontations or encounters. ⁵ To become aware of something we do not understand, we need a context of what we do understand. ⁶ By giving us access to what we do not understand, the context of habitual understanding does, as it were, indicate its own limits. We register these limits not simply as a lack of understanding but, more precisely, as a lack of our capacity to articulate. The nagging feeling or awareness of something we cannot put in words is nothing unusual. This fact of everyday life implies that the range of accessibility is wider than, though not independent of, the range of intelligibility. Whereas the latter more or less coincides with our linguistic capacities, the former indicates that these capacities do not (fully) determine our awareness of and contact with reality.

The Body as “Demand in and for Language”

Mostly, in Bodies That Matter (1993), Butler implicitly equates intelligibility and accessibility. However, some passages suggest a wider, more phenomenological notion of accessibility. The most intriguing one is to be found in the context of a subtle reinterpretation of Freud’s notion of the bodily ego and Lacan’s mirror stage. In this context, Butler examines the indissolubility of the psychic and the corporeal. The psyche is to be understood “as that which constitutes the mode by which [the] body is given, the condition and contour of that givenness. [But] the materiality of the body ought not to be conceptualized as a unilateral or causal effect of the psyche in any sense that would reduce that materiality to the psyche or make of the psyche the monistic stuff out of which that materiality is produced and/or derived” (Butler 1993, 66).

Butler’s use of the phenomenological terms given and givenness reveals how we are to understand the indissolubility of the psychic and the corporeal. Givenness is always transitive, it is always givenness to. Hence, by itself, givenness implies the interdependence of two instances—something that gives itself and something to which this givenness is given. In more concrete terms, the body can only give itself if there is something, that is, the psyche, to which it gives itself. But if the givenness of the body is dependent on the psyche, the converse is also true. The psyche cannot constitute the mode and contour of this givenness without the body giving itself. It would, as it were, have nothing to work upon. Thus, the indissolubility of the psychic and the corpereal refers to the mutual dependence, not simply of body and psyche but of a body that gives itself, that is, that somehow signals its presence, and a psyche that receives, translates, and transforms these signals.

The phenomenological conception of the body as that which gives itself or signals its presence denotes what might be called the persisting dynamic materiality of the body. With respect to the notion of the body as persisting materiality, Butler says something interesting:

We might want to claim that what persists . . . is the “materiality” of the body. But perhaps we will have fulfilled the same function, and opened up some others, if we claim that what persists here is a demand in and for language, a “that which” which prompts and occasions, say, within the domain of science, calls to be explained, described, diagnosed, altered or within the cultural fabric of lived experience, fed, exercised, mobilized, put to sleep, a site of enactments and passions of various kinds. To insist upon this demand, this site, as the “that without which” no psychic operation can proceed, but also as that on which and through which the psyche also operates, is to begin to circumscribe that which is invariably and persistently the psyche’s site of operation; not the blank slate or the passive medium upon which the psyche acts, but, rather, the constitutive demand that mobilizes psychic action from the start. . . .

(1993, 67)

Persisting materiality or, in my terms, the body that signals its presence, is interpreted here as a demand in and for language. Though maybe strange at first sight, a plausible explanation—one that has some interesting practical consequences—can be given. The body that signals its presence is the accessible body, but not necessarily the intelligible body. In so far as they are not intelligible, these signals might be registered as a lack of the capacity to articulate. However, the urge to fulfill the lack often accompanies the feeling of lack; the lack of language turns into a demand for language. In language, that is, in our linguistic existence, a demand for language may manifest itself. This demand for language is the urge to articulate, hence to understand, the unintelligible.

Thus, the body that makes itself felt as a demand in and for language is, in my explanation, the—as yet—unintelligible body. The body is unintelligible insofar as it exceeds not only, in general, the limits of the linguistically constructed body but, more specifically, the limits of the sexed, gendered, and sexualized body. This specification is important, for sex, gender, and sexuality are anything but neutral categories. The construction of the sexed, gendered, and sexualized body is, at least partly, regulated by oppressive norms among them, as Butler rightly stresses, the norm of heterosexuality. If the intelligible body is a body that is sexed, gendered, and sexualized according to (oppressive) norms, then these norms will permeate not only our understanding but also our lived experience of the body. What such norms prescribe often will be understood and lived as natural and hence inevitable facts of human life. But lived experience is not restricted to the intelligible body. If it were, it would be hard to explain why and how, for instance, a teenager who has grown up in a community that considers heterosexuality as a natural fact feels desires that don’t fit this fact. If accessibility were restricted to the intelligible body, it would be hard indeed to explain such phenomena. Because we have access to the unintelligible body with its sometimes unintelligible desires we not only may come to feel ill at ease with the intelligible body but also may come to conceive of, for instance, heterosexuality as an oppressive norm.

The Unintelligible Body as Critical Force and Creative Resource

It is important to distinguish between intelligibility and accessibility for two reasons. First, we need this distinction to explain why aspects of the intelligible body may be experienced as not fitting and even oppressive. Experiences of this kind presuppose the awareness of an alternative. If accessibility were restricted to the intelligible body, access to an alternative experience of the body would simply be impossible. The second reason has to do with motives and resources for criticizing and changing oppressive aspects of the intelligible body. Access to what exceeds the intelligible body might be a powerful motive and resource for criticism and change. In other words, the unintelligible body may come to function as a creative resource and critical force. Manifesting itself as a demand in and for language, the unintelligible body may mobilize us to articulate new meanings and new discursive practices with respect to the body. The signals of the unintelligible body and their demand may, if we are responsive to them, initiate an effort of articulation that is not only creative but also critical insofar as it challenges oppressive norms that regulate the intelligibility of the body.

Though inspired by Butler’s characterization of the body as a demand in and for language, this interpretation of the unintelligible body is incompatible with the epistemological position Butler defends, in general. If language determines the limits of accessibility, then access is restricted to what we can name, articulate, and in a basic sense, understand. This epistemological position precludes the possibility of (pre)conscious experience of the unintelligible body. The psychoanalytical interpretational scheme which dominates Bodies That Matter (Butler 1993) confirms this conclusion. According to this scheme, the normative standards of sex, gender, and sexuality that regulate the intelligibility of the body initiate the psychic operation of repudiating and abjecting the unintelligible body, thereby excluding it from the realm of (pre)conscious experience. Though this psychoanalytical interpretation makes sense in specific cases, Butler applies it as a generalized interpretational scheme to Western culture as a whole. Applied as such, it is both empirically implausible and politically self-defeating. ⁷ The force of normative standards is not always and everywhere the same nor are the circumstances and backgrounds of individuals the same. Depending on these two variable factors, psychic reaction with respect to the unintelligible body may vary from repudiation and abjection to acceptance and even celebration. The political consequences of the generalized scheme are more serious. Though Butler stresses the subversive potential of the abjected, unintelligible body, this kind of subversion is not likely to change anything for the better. What is excluded from consciousness does not disappear, on the contrary, it haunts conscious experience and it does, indeed, exert a subversive pressure on intelligible reality. But to what good? Unless we can take control of it, its very subversiveness will lead to either defensive reactions, hence it will consolidate the normative standards that regulate the intelligible body, or mental and/or emotional breakdown, hence to suffering instead of change for the better. Only if the unintelligible body is accessible to (pre)conscious experience and conscious efforts of articulation, its subversive potential may turn into the positive potential of critical force and creative resource.

To sum up, it is not only the resolution of the problem of linguistic monism but also the phenomenological notion of accessibility that validates Butler’s theory of the body as a useful contribution to feminist and queer theory. However, the tendency, reinforced by psychoanalytical interpretation, to equate accessibility and intelligibility or, in other words, to restrict (pre)conscious experience to the intelligible, is debilitating for such a theory. It leaves us with a body the intelligibility of which can neither be contested nor transformed by our own experiences of the body. That, to me, is both implausible and depressing.

II. Determinism: the Question of Agency and Power

It is clear from certain passages in Bodies That Matter (1993) that Butler assumes a viable and defensible conception of agency, but the theoretical framework on the basis of which this conception of agency can be accounted for is deficient in at least one respect and obscure in other respects. What is the conception of agency Butler assumes? She says that “drag is subversive to the extent that it reflects on the imitative structure by which hegemonic gender is itself produced and disputes heterosexuality’s claim on naturalness and originality” (1993, 125). She describes queer politics as a laying claim to terms “through which we insist on politicizing identity and desire” and as “a self-critical dimension within activism” (1993, 227).

“Reflection,” “dispute,” “laying claim to terms,” and “self-criticism” require an intentional and reflective subject, capable of deliberate and purposive action. An explicitly elaborated conception of such a subject, however, is not to be found in Bodies That Matter (Butler 1993). Butler apparently fears that this conception of the subject will reintroduce the ghost of the humanist subject, fully capable of self-determination and self-creation. The latter however, does not necessarily follow from the former. On the contrary, the conception of an intentional, reflective subject, capable of deliberate and purposive action, can be fully compatible with the poststructuralist notion of a subject implicated in chains of signification and relations of power that it cannot (fully) control—the notion Butler works with. The general problem with Butler’s theoretical framework is that its conceptualization of the relation between language and subject is too restrictive. The relation typically is described as unilateral, with language forming and constructing the subject. Consequently, it is difficult to see how this subject can be capable of the kind of action Butler typically attributes to it, namely resignifying the significations that have formed this subject.

Apart from poststructuralist linguistic theory, another important theoretical source of Butler’s notion of agency is Foucaultian theory of power. Inspired by this theory, Butler characterizes agency as a practice of rearticulation or resignification that is “immanent to power” and not in “a relation of external opposition to power” (1993, 15). Agency is “implicated in the very relations of power it seeks to rival,” it is a “turning of power against itself to produce alternative modalities of power, to establish a kind of political contestation that is not a ‘pure’ opposition, a ‘transcendence’ of contemporary relations of power, but a difficult labor of forging a future from resources inevitably impure” (Butler 1993, 241). Apart from the question of how to interpret Butler’s conception of agency, the reader cannot help being struck by the relentless “antihumanist” idiom (Fraser 1995, 67) in which it is couched. Despite its opposite rhetorical effect, however, the antihumanist idiom should not be taken as a symptom of the unwitting reduction or evacuation of agency. Though the Foucaultian framework tends to obscure the role of agency, it is possible to distill a politically engaged account of agency from Butler’s text.

Language and Subject

Butler conceives of language as a process of reiteration carried forward by the (re)citations of subjects. The speaking of subjects is recast as citing. There is nothing bizarre about this for every word we utter is, quite literally, a citation. As nobody, no single subject, invents by itself the language it speaks, its speaking is rather like a borrowing, a citing, from an already existing vocabulary. To be able to speak, however, does not simply mean that one is able to cite words but rather that one is able to apply the rules or conventions that regulate the use, that is, the meaning, of words. Thus, while citing words, we at the same time cite the conventions that regulate the use of those words.

In Bodies That Matter (Butler 1993), the possibility of agency is consistently, and correctly, located in the process of reiteration: recitations can be resignifications. Yet, if we look more closely into the way the process of reiteration is conceptualized, the possibility of agency seems to evaporate. The process of reiteration, as Butler time and again insists, can be both stabilizing and destabilizing, in the latter case opening up possibilities of agency. The idea of destabilization may look promising, but what exactly does it mean? The process of reiteration is not a process of simple repetition. Jacques Derrida, from whom Butler borrows her notion of iterability, is very clear on this. Every citation implies, by itself, a shift: the words I am citing here and now are cited in a temporal and spatial setting that is necessarily different from former or later citations elsewhere of the same words. These inevitable temporal and spatial shifts, in short, contextual shifts, are at the same time shifts of meaning. ⁸ The citation of the word woman in one context can signify something completely different than, slightly different than, or anything in between those two, but never exactly the same as the citation of the same word in another context. Because of the continuous shift of meaning, the process of reiteration is a process of change that can be either stabilizing or destabilizing, that is, reinforcing or undermining signifying conventions. Yet, the point is that the continuous (de)stabilizing shift of meaning is completely independent of any acts or intentions on the part of the citing subject. It inheres in the very movement of language itself, its “temporalization” and “spatialization” as Derrida calls it, a movement that the citations of subjects carry forward but do not control. ⁹

Though the question of agency is, at best, still open, this conception of language achieves the deconstruction of certain determinist consequences of the structuralist conception of language. The latter typically identifies certain structures, for instance, the structural interplay of castration and phallus in Lacan, as invariable, unchanging structures that determine to a certain extent the shifting signifying chains of language. As the movement of the signifying chains of language is always completely contingent, Butler’s Derridean conception of language effectively undermines the determinist result of the structuralist conception. ¹⁰ But deconstruction of structuralist determinism does not, by itself, imply that room is made for a viable notion of agency. This requires at least that the subject is not only the carrier of the process of reiteration but also a possible participant in this process. ¹¹ The trouble with Butler’s account of the relation between language and subject is, as I said above, that almost invariably the subject is cast as being constructed and reconstructed by the signifying chains of language. The subject is not completely passive, for its activity of citing is a necessary condition of its (re)construction. Therefore, Butler legitimately can claim that the subject does not simply undergo the process of (re)construction (1993, 3). Yet again, this does not add up to a viable notion of agency.

Butler’s Derridean conception of language does not simply exclude the possibility of agency. Though often it is understood as precluding the possibility of intentionality, ¹² her conception does not imply that we do not intend certain meanings when we speak or act. The point is that our intentions do not effectively control or decide the meaning of words and (speech)acts. The individual subject does not, and cannot, decide the meaning of words and (speech)acts, including its own, for two reasons. On the one hand, the subject always already finds itself in a language with more or less established signifying conventions. When we learn to speak we learn to apply those conventions, and when we have become proficient speakers our speech still has to be more or less compatible with those conventions to be intelligible to others and to ourselves. Yet, there is room for initiative here: we can, intentionally, try to deflect or guide the signifying conventions or chains we are citing in a certain direction, aiming at the resignification, the shift of meaning, that is our goal. This initiative, on the other hand, does not imply that our intentions can control the future course of these signifying chains, can decide the future meaning of the words that are being cited. But neither does it imply that the initiative or intervention on the part of the subject never leaves a trace. It does leave a trace, and it may be effective in the long run, when other subjects pick up and continue the deflecting course of the signifying chain.

Explicated thus, the Butlerean conception of language is fully compatible with a realistic notion of agency. While doing away with the misguided idea that the subject is capable of complete self-determination and full control over language, this conception retains what I take as the quintessence of agency: the possibility of initiative or intervention, a possibility that is both dependent on the past, that is, established conventions, and at the mercy, as it were, of the future, but nevertheless a possibility that may effect desired changes if—and this “if” indicates a necessary condition—the initiative is picked up by others. ¹³

Power and Agency

Agency, Butler says, is implicated in the very relations of power it seeks to rival. Though it is not easy to get a clearer picture of the ambivalent relation of agency and power, a discussion of the main characteristics of power in Butler’s text may shed some light both on her concept of power and its relation to agency.

By schematizing, one can discern four related characteristics: power is discursive, reiterative, productive, and exclusionary. Roughly speaking, the first two characteristics locate the domain of power and specify the way it works, whereas the last two explicate the effects of power. Power is located in discourses and discursive practices, more specifically in the conventions that constitute and regulate these discourses and practices. In terms of agency this implies the following. To speak and act intelligibly, I have to comply, to some extent, with the conventions that regulate the use, that is, the meaning, of words and that constitute the particular practices in which I participate. I cannot, for instance, act intelligibly as a teacher without, to some degree, following the conventions that constitute the practice of teaching. Precisely because one has to comply with them—in Butler’s terms recite them—to speak and act intelligibly, conventions are binding or normative. As such, conventions are both constraining and enabling: they enable us to speak and act intelligibly exactly in so far as they constrain us. ¹⁴ Thus, in so far as it is enabled and constrained by the conventions that constitute and regulate the domain of intelligible discourse and practice, agency is enabled and constrained by discursive power.

The Derridean notion of iterability highlights the temporal and contingent aspects of discursive power. Iterability is a necessary condition of conventions: conventions are only conventions in so far as they are not unique but, on the contrary, are reiterated or recited over and over again. Hence, if conventions are reiterable, so is discursive power. More precisely, power is located not so much in the conventions themselves, that is, in their content, but rather in their reiteration. We have seen already that reiteration is a process not of repetition but of perpetual and contingent shifts. These shifts are contingent in so far as they inhere in the reiterative process itself and are not a result of purposive, intentional agency. Because of (the accumulation of) contingent shifts in their reiterative course, conventions may either get consolidated or extenuated, stabilized or destabilized. Thus, whatever the content of a specific convention, its power increases or decreases as its reiteration consolidates or extenuates the convention in question. In terms of agency, the reiterative aspect of power implies: first, that we will influence the power of conventions only in so far as we participate—instead of merely being implicated—in its reiterative course, for only then we have a chance to intervene; and, second, that the success of our interventions always will remain dependent on, in principle, the incalculable future course of the reiterative chains or on, in general, the incalculable effects of our interventions.

Whereas in liberal politics and also in everyday usage the term power is associated with the effects of prohibition and repression, Butler stresses the productive and exclusionary effects of power. Power is productive in the sense that the reiteration of conventions brings into being or materializes what these conventions specify: we perceive reality through the sedimented grid of signifying conventions. For instance, the reiterative power of the convention or norm of binary sex/gender differentiation produces or materializes bodies that are either male or female. ¹⁵ Yet, in so far as it is productive, power is also and at the same time exclusionary. If the reiterative power of the sex/gender norm produces bodies that are either male or female, then, simulteanuously, it excludes any other type of body from the domain of social and symbolic intelligibility. Generally speaking, the reiteration of conventions both produces the intelligible and, at the same time, excludes the unintelligible. Hence, reiterative discursive power constitutes both the intelligible and the unintelligible. Explicated in terms of agency, the import of the productive and exclusionary effect of power is more or less tautological. If reiterative discursive power enables and constrains speech and action, that is, agency, then another way of saying this is that it produces intelligible speech and actions while excluding unintelligible actions and utterances.

The explication of the role of agency I have given thus far is meant to demonstrate that Butler’s reliance on poststructuralist Foucaultian theory and its antihumanist idiom does not lead to the reduction or evacuation of agency. What is not yet clear, however, is why we would seek to rival the power in which we are implicated. Why would we try “to turn power against itself”? For what reasons or motives would we contest (the effects of) discursive power?

Political Contestability

On first sight, the above explication of power does not seem to give any clear indications as to whether or why power may be politically contestable. That discursive power is constraining and exclusionary cannot be held against it, ¹⁶ for if it were not it would not be enabling and productive either. Constraint and enablement, exclusion and production are two sides of the same coin. Discursive power enables intelligible speech and action only in so far as it constrains agency, that is, excludes the unintelligible. Unless the distinction between the intelligible and the unintelligible is given up altogether, both sides belong together. To criticize this distinction simply because of its constraining and exclusionary implications does not make sense for it is constitutive of agency. I do not see how any realistic account of what it means to be able to speak and act can do without this distinction.

Yet, even though the distinction between the intelligible and the unintelligible is itself irreducible, every specific instance of it is contingent and, hence, contestable. Any specific demarcation of the intelligible from the unintelligible is contingent because it cannot but rely on conventions that are contingent or arbitrary themselves. Apart from the contingency inherent in the process of reiteration, from the viewpoint of reflective agency as well conventions are, in principle, arbitrary: we cannot provide them with a conclusive foundation or justification. No matter how well established or how well argued, no foundation or justification can ever succeed in turning conventions into necessary rules without alternatives. It is always possible, in principle, to conceive of an alternative to any specific convention, and consequently to contest its constraints and exclusions and to reinterpret its demarcation of the intelligible and the unintelligible. For example, despite the long tradition and scientific justification of the convention of binary sex/gender differentiation, it is not difficult at all to think up several alternatives, such as plural sex/gender differentiations or the conception of one sex or gender. These alternative conceptions involve a reinterpretation of the body as we know it. What is unintelligible now might become intelligible and vice versa. Heterosexuality, for instance, would lose its self-evident and perhaps its intelligible character whereas bodies neither male nor female would become intelligible. ¹⁷

Thus, because of the irreducible contingency of any specific convention discursive power is, in principle, always contestable. The question, however, of why we would contest the power of any specific convention still remains open. Merely the contingency of conventions, that is, the mere fact that it is possible to contest conventions, is no reason to contest them. Butler’s discussion of the so-called “law of sex” suggests that it is the hegemony of a convention that provides a reason to contest it. She uses the expression “law of sex” to indicate the hegemonic status of the conventions or norms that make up this law, that is, heterosexuality and binary sex/gender differentiation. The introduction of the term hegemonic implies a certain hierarchy with respect to the relative force of conventions, namely hegemonic or dominant conventions as the most forceful and minor or subordinate conventions as the weakest. The relative force of a convention affects its power. Whereas the power of hegemonic conventions tends to be compelling, the power of minor conventions leaves more room for choice.

From Butler’s discussion of the law of sex throughout Bodies That Matter (1993), I infer that the power of hegemonic conventions is compelling in so far as it determines subject status. To qualify for and maintain the status of subject, one has to comply with, that is, recite, the law of sex. In other words, compliance with the law of sex is a necessary condition for subject status. ¹⁸ Whereas noncompliance with minor conventions—for instance the conventions which regulate the practice of teaching—will not disqualify me as a subject but, in this case, as a teacher, noncompliance with the law of sex results in deprivation of subject status. Those who do not comply are degraded to the status of “abjects”: “those who do not enjoy the status of the subject” (Butler 1993, 3), who do not qualify as “fully human” (1993, 16), and whose bodies do not matter.

Whereas hegemonic conventions specify the necessary conditions of subject status, nonhegemonic conventions specify the many variations and forms of subjecthood. Within the confines hegemonic conventions circumscribe, the wide variety of nonhegemonic conventions allows many different ways to shape our identity as a subject or actor. As long as we comply with hegemonic conventions our words and deeds will be acknowledged, even if they are unintelligible. That every now and then my words and actions are unintelligible does not disqualify me as a subject, but the persistent failure to embody and realize those features of subjecthood or humanness which are deemed to be “essential” does. To be disqualified, to be an abject, means that one’s words and deeds will be ignored or dismissed, not because they are in or by themselves unintelligible but because they emanate from an unintelligible, unthinkable, and even threatening being, one whose claim to intelligibility and subject status I cannot acknowledge without jeopardizing my own secure status as a subject.

As they result in deprivation of subject status, the exclusions effected by hegemonic conventions are dehumanizing and violent. The violence of their exclusionary power provides a good reason to contest and oppose hegemonic conventions. But what exactly can we achieve, given Butler’s concept of power and agency, if we contest and oppose the power of hegemonic conventions, in this case the norms that make up the law of sex? Agency is a “turning of power against itself to produce other modalities of power,” that is, it cannot annihilate power. We cannot simply abolish these norms, we can only undermine their hegemony by resignifying them in such a way that neither heterosexuality nor binary sex/gender differentiation designate natural or essential humanness. What we may achieve through resignifications of this kind is a less exclusive definition of subject status, a definition that includes the abjects. Never, however, can we achieve a totally inclusive definition for “the ideal of a radical inclusivity is impossible” though Butler adds that “this very impossibility nevertheless governs the political field as an idealization of the future that motivates the expansion, linking, and perpetual production of political subject-positions and signifiers” (1993, 193).

The Political Ideal of Radical Inclusivity

Butler’s argument concerning the political ideal of radical inclusivity is somewhat paradoxical: we have to strive for something we will never achieve. Why is radical inclusivity impossible? Why would we be motivated by an ideal we cannot realize? In what sense does radical inclusivity function as a positive political ideal?

Total or radical inclusivity is impossible because no category can be all-inclusive. Unless they are completely empty or meaningless, categories cannot but have exclusionary effects. To be intelligible at all, the category of “the subject” or “the human” has to be interpreted, that is, assigned some meaning, and it is this interpretation that produces exclusionary effects. Even when it is very broad or abstract, no interpretation can ever be all-inclusive. It always foregrounds one or several signification possibilities at the (implicit) expense of other possibilities. However, “even if every discursive formation is produced through exclusion, that is not to claim that all exclusions are equivalent” (Butler 1993, 207). Only the violent ones are problematic, that is, the exclusions produced by hegemonic conventions and interpretations. If we cannot simply abolish discursive power and its exclusionary effects, we can, Butler suggests, change its modality or force. In the process of resignification we may undermine the compelling force of hegemonic conventions or interpretations, and consequently, reduce the violence, if not the exclusionarity, of discursive power.

Moreover, it would be politically undesirable to reduce the exclusionary effects of noncompulsive discursive power. As I indicated above, it is the nonhegemonic conventions that enable the differentiation of specific subject-positions or identities. These differentiations are in a neutral, almost tautological way exclusionary: being a teacher, philosopher, and feminist, for instance, excludes numerous other possible positions or trajectories I could, in principle, have chosen or in which I could have happened to end up. Actually, because of the negative connotations of the term exclusion(ary), it would be more helpful and adequate to speak of the differentiating, instead of exclusionary, effects of noncompulsive discursive power. It is obvious that radical inclusivity in the sense of a complete reduction of all differentiations would make for a dubious political ideal. The political condition that comes closest to realizing radical inclusivity in this sense is totalitarianism. In the definition of Hannah Arendt totalitarianism refers to the destruction of plurality, that is, the differences constituted by speech and action. I suppose it is this totalitarian reduction that Butler has in mind when she says “that there can be no final or complete inclusivity . . . and that, for democratic reasons, ought never be” (1993, 221).

Butler nevertheless offers radical inclusivity as a positive political ideal. But she does this in the context of a critical political diagnosis of the world we live in: a world in which hegemonic conventions determine who is to count as a subject and who is not. In this context, to strive for total inclusivity means to strive for a radical democratic world, that is, a world no longer regulated by hegemonic conventions, in which everyone is entitled to the status of subject. Yet, even though hers is a positive political ideal, Butler is wary of presenting radical democracy simply as an achievable goal. To become an achievable goal, the ideal of radical democracy has to be translated in a political program or movement: the abstract, as it were, empty universality of total inclusivity is translated into a concrete, meaningful universality, such as the concrete universality of human rights. Though inevitable if the ideal is to be effective, the translation has its price for it cannot fail to be exclusionary and contestable. If radical democracy is simply taken as an achievable goal, ideal and translation get conflated. What gets lost in the conflation is the irreducible and critical distance between abstract and concrete universality, between the idea(l) of total inclusivity and its inevitably exclusionary and contestable translation. Because of the lack of critical distance the translation becomes self-evident, uncontestable, in short, hegemonic, and the violence of its exclusions will be ignored, be excused, or most likely, remain invisible. In other words, without the critical distance between ideal and translation, the ideal of radical democracy, like any other ideal of a perfect world, may turn into its opposite. ¹⁹ To prevent the conflation of ideal and translation, radical democracy should be taken not as an achievable goal but as a regulatory idea in the Kantian sense. As such it enjoins us both to act as if total inclusivity is realizable in the future and, at the same time, to contest here and now any (violent) exclusions effected in its name.

Feminist politics can draw a valuable lesson from the general argument concerning radical inclusivity. In so far as feminist politics invokes the category of “women,” it seems to face a dilemma. On the one hand, it cannot but rely on this category if it is to have a basis for solidarity and empowerment. On the other hand, not even the most politically correct, multiculturalist specification of identities can fulfill the promise of inclusivity this category holds out. On the contrary, if anything, politically correct identity politics evokes rather than assuages recriminations of exclusion and lack of recognition. Instead of empowerment and solidarity, a “politicing of identity” (Butler 1993, 117) is the result. We are confronted with the dilemma of either giving up the category of “women” and hence the basis of feminist politics or resigning ourselves to proto-totalitarian identity politics. However, if we take the category of “woman” as a regulative idea(l) rather than an actual representation of all women, the dilemma is resolved. This means that we have “to learn a double movement: to invoke the category and, hence, provisionally to institute an identity and at the same time to open the category as a site of permanent political contest” (Butler 1993, 222). Only if we do not forget that the idea of total inclusivity is not realizable as such and hence that the category of “women” is permanently open to different interpretations, can the invocation of this category enable solidarity and empower feminist politics.