June 4, 2002


As both the concept of culture and individual cultures have become increasingly objectified-whether through strategic self-display or through circulation in a global academic, touristic, and media-based economy of representation-the complex forms and processes of cultural presentation have been more closely scrutinized. Recent theorizing in diverse fields-including anthropology, performance studies, philosophy, cultural studies, queer theory, gender studies, and literary criticism-makes liberal use of the terms "performance' and "performativity' to describe both general processes and specific forms of social action. Both terms encapsulate broader concern with social identity, agency, and intentionality (the politics of production), and with variable relations between performers and different kinds of audiences and venues (the politics of reception). Despite-or perhaps because of-this theoretical productivity, there is often terminological confusion, contradiction, and/or conflation in the deployment of these notions. This has been exacerbated by the rise of interdisciplinary scholarship since the 1970s, with the humanities and social sciences drawing on similar models and vocabularies but often to divergent ends (Geertz 1983). This raises the important role of conceptual genealogies to help disentangle theoretical debates and tensions. In an effort to clarify and specify these terms, I will trace the development of performance theory as it is relevant to anthropological study. There have been two main branches to this development, one dramaturgical and one linguistic. Since the former is very well documented, I wilt only summarize it here as it informs the evolution (and my survey) of the latter.

Dramaturgical models. Theatrical models have provided the most directly influential vocabulary for the analysis of social life and cultural performance (see Geertz 1983; Beeman 1993; Palmer and Jankowiak 1996). From the perspective of phenomenology and sxial constructivism (Bateson 1972; Goffman 1959, 1974) came a vision of social life as a "serious game" with structuring rules, individual strategies, and a sense of ludic expressivity. Here "performance" was used to characterize how social subjects consciously


take various roles as "actors" engaging in certain temporally bounded stagings of social encounter. Purveyors of ritual theory and symbolic analysis (Turner 1969,1974; Geertz 1973,1980) approached social events as "dramas11 or'texts' with repeated, conventionalized, and structured stages,-as well as stylized rhetorical or poetic features (Femandez 1986). There was a great emphasis on the efficacy of events (whether described as religious or secular; Moore and Myerhoff 1977), and sensitivity to the social relations between performer and audience or community (whether cooperative or contentious; Fabian 1990). Here "performance" is used generally to suggest public practice and engagement in social activity (in contrast to more cognitivist theories of culture as knowledge), and specifically to refer to rituals or staged cultural events (Singer 1972; Tambiah 1985). Other scholars deal more directly with the theater itself, and explore relations between ritual and avant-garde cultural practice (Turner 1982; Schechner 1985,1993). Complicating earlier models with strict category boundaries, these studies often posit a continuum between entertainment and efficacy, performer and audience, script and improvisation. Here "performance" is used in a more restricted way to describe highly marked genres of contemporary art in their structural and formal relation to other social events. Dramaturgical language is rarely directly contested (but see Wilshire 1982; Bell 1992), and has been thoroughly absorbed into social theory's vocabulary, whether used with more or less precision. Less acknowledged is the extent to which linguistic genealogies inform performance studies, and diverge in important theoretical ways from each other and from dramaturgical models.

"Structuralism and generative grammar. Saussure (1986) transformed the study of linguistics by drawing an analytical distinction between language as an abstract sign system (la langue) and language as a spoken phenomena (la parole). The former characterized language as a constative code which is evaluated according to classic truth-functional criteria of correspondence with the world, while the latter described language as a performative or enacted aspect of social communication (Frege drew a similar distinction between reference and sense; see Lee 1999:7). Later Chomsky (1965) translated Saussure's terminology into linguistic "competence" and "performance" in terms of an individual's engagement with


language. For both Saussure and Chomsky, competence refers to (knowledge of) the abstracted system of syntactic rules which structure the generation of linguistic utterances, while performance refers to the specific production of speech by individual speakers. Chomsky was more interested in the mechanisms of the former, and he bracketed spoken language as idiosyncratic, imperfect, and largely incidental to an understanding of language as a faculty. It is important to note that in this formulation, "performance1' refers simply to the production of spoken language regardless of intent or context.

Natural Language Philosophy. A second usage of the term emerged from the philosophy of language, notably in Austin's (1962) theory of performatives, revised by Searte (1965) as Speech Act Theory. Austin, critical of logical-positivism and truth-functional analysis, approached speech as action, enunciation as social practice (Lee 1997:5). This was part of his larger effort to show how language goes beyond referentiality (signifying already existing concepts or situations) to become constitutive of social reality itself. He identified the "performative' as a specific type of verbal action which functions to bring a state of being into existence through the articulation of speech. Austin made a distinction between three kinds of speech: locutionary (standard language use for sense and reference); illocutionary (which brings a state of being into existence-among these are performatives proper); and periocutionary (which initiate a sequence of activity resulting in a change of being). Instead of being evaluated by truth-function (reference to the "real world"), performatives are subject to felicity conditions" based on their resemblance to past instantiations of similar acts and proper listener reception. Searie articulated in greater detail the "necessary and sufficient conditions" for the proper performance of speech acts. A successful performative speech act relied upon the speaker's intention to convey a certain message and a hearer's accurate reception of that message. Searie specifically wanted to account for the link between the particular intention of the speaker and the formulaic conventions of any linguistic system; that is, how people put their formalized language system into use. Here performance was recast from a type of event to a feature of linguistic efficacy. But as philosophers, they were more interested in abstract relations and logical rules


than in the messy dynamics of social fife. Though Austin and Searle were careful to avoid such strong claims, they were read as presupposing the agency or intentionality of the speaker to effect change through the performative or illocutionary force of language, while ignoring broader social contexts, specific cultural codes, and audience participation (see Finnegan 1969; Rosaldo 1982; Foster 1989).

Sociolinciuistics and the ethnography of speaking. Partially in response to such critiques and in an attempt to go beyond the limitations of structural linguistics and natural language philosophy-especially in the realm of performance-many linguistic anthropologists began attending to the social details which surround and inform the production of speech. Sociolinguists including Hymes (1962,1975), Gumperz (1964), and Sherzer (1975) carefully considered the sociocultural contexts in which language (as a code for communication) is used, seeking to formulate more detailed accounts of practice in both production and reception. Early sociolinguistics was consonant with other classificatory projects of a scientists anthropology in the 1950s and 60s, as scholars set out taxonomies of terms, conditions, and formal relations betweens types of speech event.

Hymes (1975), in his seminal article "Breakthrough into performance", signaled the larger theoretical shift away from the Chomskian view-that spoken or performed speech was an impoverished and residual type of language-to focus on performance as the emergent quality of language. He also introduced a new classification for a particular type of speech event which was neither the general structuralist "speech" nor the specific Austinian "illocutionary" act. Influenced by folklore studies, Hymes used "performance" as an honorific, a specific and highly marked speech act which makes conscious reference to its traditional predecessors in a given social or cultural context (e.g. ritual speech, political oratory, storytelling). While acknowledging traditional sources, oral narratives as performances are creative (valued for being innovative and expressive), achieved (as agentive action with intentional goals), and public (relying on listeners and witnesses for efficacy). Moreover, Hymes (1975:18) defines performance as "cultural behavior for which a person assumes responsibility to an audience." Drawing on Labov (1967), he


set out three criteria for a successful performance: interpretability, ^portability, and repeatability. He was also concerned to characterize "authentic' or "authoritative" performance as faithful reproduction of past formula, "when the standards intrinsic to the tradition are realized" (ibid). Diverging from Austin and Searle's presumption of an a priori tradition," Hymes argued that enactment helps constitute tradition itself by providing contemporary instantiations which set the standard against which future performances will be judged. Performance thus became a specific mode of language use which entails the speaker's agentive relation to an audience for the production of speech, usually within highly marked genres resonant with existing traditions or conventions. That is, performance became not merely a way of enacting the syntactic rules of language but of reproducing the cultural forms and contents encoded in particular modes of speaking.

Folkloric synthesis. As opposed to traditional linguistics, folklore studies-as informed by a fusion of Jacobsonian dynamic structuralism and Russian dialogic formalism-has specifically focused on creative, expressive language as a form of cultural production. Bauman and Briggs were especially rigorous in elaborating Hymes1 theory of performance, arguing that performance marked the production of uniquely "artistic" and creative language (poetics) within stylized cultural traditions (ethnoaesthetics). They moved away from his taxonomic project and toward a communication model-of players, events, contexts, and meanings-where performance was an individual action (a verb) and a social event (a noun). While speech act theory and sociolinguistics informed their attention to the contextual and efficacious use of language, and to the agency or intention of the speaker, they also elaborated on the larger dramatic contexts in which such language is produced by drawing on Bateson and Goffman's theatrical theory of performing selves and social roles. Exemplars of the German (Herder and Boas) influenced school of American humanistic and cultural relativism, Bauman and Briggs adopted a political project to elevate marginalized communities (e.g. rural Hispanics and urban blacks) by recognizing their "verbal art" and indeed "culture" (Bauman 1986a:7-8).

Bauman (1977,1986a) set out to study artistic language in sxial life, making the transition from literal speech to literary speech, that is, epic poems and stories, narratives, and oratory. He drew attention to genre and style as unique forms of patterning which mark performance and which supply criteria for its evaluation by audiences. While still invested in authenticity and the agentive intention of the speaker to display his cultural competence in traditional styles (1977:11), Bauman also recognized the extent to which audience reception plays a role in constituting the successful performance (see Duranti 1986). Because performances are highly conventional and contextual, he proposed detailed ethnographic analysis to uncover the dynamics of place and event specificity (1977:27; 1986a:3). He also discussed the importance of metalinguistic contextual cues-"keys" in the dramaturgical language of Bateson and Goffman-which signal to the audience the appropriate mode in which to interpret a performance. In this rubric, there are both informal and mundane modes of performance (story telling, gossip, etc) and highly marked, official cultural performances whose purpose is to reiterate a community's grand narratives and cultural values (see Singer 1972). Because they appeal to traditional or hegemonic structures, performances are means of both imposing and negotiating power, and are thus potential site for struggle in the terms of the Manchester School (Bauman 1977:44; see also Brenneis and Myers 1984).

Briggs (1988) returns such a perspective to Chomskian terms to show how competence (in cultural traditions) and performance (as creative productbn) are mutually constitutive and not categorically opposed aspects of language. Performance emerges as the "dynamic interplay of individual competence in traditional forms, stylistic resources, and a unique interactional environment* (1988:12). Briggs deals with performers as embodying and displaying cultural knowledge about traditional codes, appropriate styles, and authentic meanings. Ethnopoetics, as the study of performance, fuses formal analysis of genre and style with situated analysis of their enactment in specific social contexts. Briggs (ibid:8) invokes Austin's use of performativity as an intention to communicate, recognizing that meaning comes from the form of a message as well as its content. He also draws on Goffman's "frames" as meta-communicative devices


marking and pointing to events as performances. Thus 'art' as a highly marked objectification of culture (seen as traditional styles, forms, codes, meanings, knowledge) becomes available both to scholarly study (and potential reification) and to communities for use in reproducing themselves over-time (Bauman and Briggs 1990:73; Briggs and Bauman 1992:600; 1999).

In sum, this line of development has used "performance" at three basic levels of generality (Bauman 1986b): as linguistic or social practice in general (Chomsky); as cultural performance, ritual, and drama (Goffman, Turner, Singer); and as the poetics of verbal art and oral narrative as a specific genre (Hymes, Bauman and Briggs). When they use "performativity", Bauman and Briggs (1990:62) refer to Austin's limited sense of an individual's intentional speech acts which bring a particular effect into being. They want to expand this notion to a broader and more dialectical mode of constituting social reality and cultural (reproduction itself. Though they became more dialogical, intertextual, processual, and political about their terms and modes of analysis, and though they recognize (as does Fabian 1990) that both indigenous aesthetic presentations and academic re-presentations are equally "performative" (objectifying) and constitutive of "culture", most of these authors never questioned the inviolability of the social subject him/herself.

Post-structuralism. In sharp contrast to these developments, feminist philosopher Judith Butler (1990b, 1993,1997) drew on Austin's theory of the performative to develop a very different argument. Drawing on post-structural critics of logocentrism and the autonomy of the Western subject, she suggests that performatives not only constitute the social situation they describe, they also constitute the subjectivity of the speaker in the act of discourse production (1990a:278-9). While retaining Austin's emphasis on the emergent and constitutive nature of performtives, she denies both the theatrical metaphor and poetic genres (which privilege a highly marked kind of language use), and the a priori existence of an intentional subject (who merely draws on pre-existing discursive conventions). Instead, she claims that use of performatives brings both the speaking subject and the conventional norms of language into being.


Butler's primary influences are post-structural and psychoanalytic philosophies of language, and post-Marxist theories of ideology. She draws on contemporary, Continental anti-humanism in its rejection of the pure, autonomous, free, Bourgeois, rational, agentive, individualistic subject (a movement exemplified by Barthes' "death of the author", Foucault's "disappearance of Man*, Bakhtin's "dialogic imagination", Bourdieu's "habitus", Gramsci's "hegemony", and Lacan's "imaginary"). Specifically, Butler invokes Derrida's (1972) attack on Austin and Searle for being logocentric in their bracketing of non-serious or expressive language (thereby circumscribing "ordinary language" as normative) and their privileging of speaker intentionality. For Derrida, the role of indeterminacy in communication shifts attention away from the conscious intention of speakers toward the restricted cultural codes and the communicative contexts (within asymmetrical relations of power) in which speaking (and the potential for misrecognifon) occur. Any linguistic representations-written ones primarily for Derrida, but even verbal ones when mediated by technologies-float free from their producers (writers or speakers) and become signs severed from their signifieds. This break between intention and both convention and reception introduces the role of absences, failures, disjunctures, and significant silences, and is unexplored if not denied by the other theories of performativity (see Lee 1997:8).

Butler also calls attention to the ability of ideology and power to limit and determine how people's agency will be expressed. One of her primary inspirations, Althusser (1971) discussed how ideology works through institutions and practices (from law enforcement to education) which in effect allow the state to create social subjects in its own terms (thereby reproducing the conditions and relations of production). Through its own practices, ideology produces (and then is unconsciously reproduced by) social subjects through a process of "interpolation" (Althusser 1971:161). Subjectivity (the recognition of self as a subject) is constituted in moments when ideology (through its agents-police or teachers) "hails" you, when you recognize yourself as the object of its address. Drawing on Lacan and the mirror-structure of identity formation, Althusser described how the subject is formed in relation to the other's gaze, and how what


seems a freely entered relation is in reality one of dependence and subjection (ibid: 165-8). (In the realm of symbolic anthropology, Bloch 1974 argued much the same for religious performance, that though it seems artistic or expressive, and thus creative, it is really a mystification of the ideological structures which restrict and determine it.)

Butler rejects the notion that performance is something that the subject does, and instead posits performativrty as the discursive process of constituting the subject in the first place (see Kulick 2002:4,7). She thus contrasts "expressivity* and "performativity", whereas Goffman and Bauman & Briggs argue that the performative is expressive and that it is the self that acts (Butler 1990b:180). Butler's larger project is to deconstruct and denaturalize the category of gender itself, not revalue and elevate it by applying the honorific language of cultural humanism (which distinguishes her from some other feminists; see Aicoff 1994). As a linguistic philosopher, she is interested in the ways that language structures experience and the formation of subjectivities, and she approaches the body as a text or a discursive field which is constituted through the performance of conventional (that is, constraining, limiting, regulatory) speech 'genres* such as gender (1990b) or even biological sex (1993). Drawing on radical phenomenology, Butler suggests that gendered identities are produced through unconscious, everyday acts rather than highly marked or theatrical performances; unlike Bauman and Brigs, she in interested in the 'mundane way in which social agents constitute sodal reality through language, gesture, and all manner of symbolic signs'... using a "discourse of 'acts' that maintains associative semantic meanings with theories of performance and acting* (1990a:270-71, my emphasis). In other words, gender is "performed* in the sense of displayed and constituted publicly in the presence of (powerful) others, but it is not consciously acted. It is vital to recognize that for Butler, 1) the social subject is constituted through this action; there is no a priori gendered agent; and 2) the category of gender itself does not exist prior to its being performed; it is brought into being through performance (1990b:173). This contradicts Bauman and Briggs' assumption of both acting subjects and pre-existing traditions.


Whereas many theorists of identity and performance concentrate on normative codes and rules for behavior production, Butler also emphasizes that which can not be said, the unspeakable (see Kuick 2002). Regulatory cultural norms are seen to limit production and "performance" and thus the terms by which subjectivites can be articulated. Identity is thus defined partially through its discursive limit (Butler 1993:xi) and partially by what gets said to you (Butler 1997:24). Like Austin or Hymes, Butler would agree that what can be said is meaningful only to the extent that it reproduces past acts within normative conditions, but for her this necessarily limits the absolute intentionality of the acton pure agency, for Butler, is a denial of these regulatory restrictions, an illusion of freedom (1993:12-13). Gender (like any sxial subjectivity) is citational, it needs to be repeated and re-iterated to be meaningful; this is the ritualized aspect of performance. Like Bourdieu, she recognizes the habitus as an unconscious intemalization of ideological structures and limits on action, but unlike him she does not assume the a priori status of the cultural power formations which structure the individual habitus. For Butler, these formation emerge from discourse, which may be disjunct or ruptured from conventions (1997:134,141; cf. McNay 2000:chapter 2).

There is an obvious and acknowledged terminological tension in Butler's theory between linguistic and theatrical uses of'performance11 (1999:xxv), and a deeper political tension between humanism and deconstructivism (1993:x). Lee (1997:10) suggests that where an earlier generation of philosophers, linguists, and literary scholars used the same vocabulary, there has been a radical divergence in the use of terms, especially surrounding the notion of subjectivity. "Performance" and "performativity" are the fulcrum on which debates teeter (Sedgwick and Parker 1995). It is thes contradictions-between an assertion or denial of agency, and a general or specific account of what type of events constitute performance-that often get overlooked by contemporary anthropological accounts of performance.

Conclusion: confusion and conflation. Nowadays, when you come across the term performance in book or article titles, it is often far from clear which of these traditions are being invoked. At one point, linguists distinguished performativity as a function of speech acts rather than as performed cultural events.


Then Butler adapted performativity to refer to the consitution of subjectivities in opposition to aspects of Austin's original use. Some anthropologists use performativity to suggest specifically agerttive (as opposed to passive) cultural production (see Hughes Freeland and McCrain 1998; Hughes-Freeland 1998). Developments in other disciplines-such as performance studies and queer theory-introduce further derivations. For instance, Phelan (1993) adopts the Lacanian notion of the imaginary to suggest that transient performance art is less "representable* (and thus less objectifiable) than other art forms and is hence a prime venue for the discursive constitution of oppositional identities. Auslander (1999) replies that no performance is beyond representation, and that privileging "liveness" as if it were unmediated ignores the complicated reality of contemporary performance art and its use/subjection to technology. Some proponents of queer theory want to retain the best of both performative worlds: the deconstruction and denaturalization of sex and gender, and the appeal to agency in controlling oppositional identity formation and display (Munoz 1999). What results is a shift from identity (as a state) to identification (as a process), a seeming parallel to that between performance and performativity. However, in literalizing performance to the production of drag shows and subversive theatrical acts, there is a danger of conflating the performance of identity (as a conscious claim staking of social position) and subjectification (as the unconscious operations which discursively and practically constitute the subject in the first place) (Jagose 1996:77; Kulick 2002:15). This is not to say that the two processes are mutually exclusive, it is only to suggest that our analytical language should make the distinction more clearly.

Herein lies the underlying schism in most uses of the term performance. On the one hand, there is a legacy of using the humanizing language of performance as expressive culture, and on the other, the anti-humanist deconstruction of the autonomous subject. Yet the question of agency is orthogonal to that of theoretical lineage: dramaturgical models encompass both agentive (Goffman) and non-agentive (Btoch) approaches, just as linguistic models include agentive (Austin, Hymes) and non-agentive (Derrida, Butler)


varieties. It is necessary to be clear, in anthropological studies of performance, which intellectual genealogy one is aligned with, and what implications of the vocabulary are intended.