Notes on:  Denzin, N. (1990) Harold and Agnes: a feminist Narrative Undoing. Sociological Theory 8 (2):198-216

Dave Harris

Garfinkel's story of Agnes claims to offer a  'pure ethnomethodological account of the social production of gender', and to exemplify ethnomethodological method. It shows how stable practical activities are produced in everyday life. It challenges positivism.

Garfinkel reveals that Agnes lied throughout the investigation, but Garfinkel never went back to reread his account. Denzin intends to offer 'a deconstructive, feminist, narrative analysis'. The title of this piece suggests a deep relationship between Garfinkel and his subject, and this needs to be opened up to a new reading [with implications for similar interpretive traditions including SI]. There is 'a masculine preoccupation' with theorising the origins of social situations and problems. The post-modern project to theorise textual reflexivity also needs to be deconstructed — at the moment, 'the leading modern and post-modern theorists are implicated in an interpretive process which entangles theory with the reality [sic] that is being described (n 2, 198). Derrida is to be used as a critical base.

Empirical interpretive sociology illustrates Derida's metaphysics of presence or logo centrism — 'subjects are present in the world, not only to themselves in their thoughts and in conversations, but to others through their speech and actions… Lives and their meanings can be grasped and located in a sociological text' (199). We require a methodology to uncover the subjective meanings, commonly autobiographical biographical or ethnographic approaches, or their variance such as life history. Garfinkel's method never clears itself from this 'documentary approach' [indeed he acknowledges it] [lots of examples of these approaches on 199]. The works 'presume a world "out there" that can be studied, captured, and brought into the sociologist's texts'. This is to be contrasted to Derrida where there is only the text, and Garfinkel where social facts have to be accomplished — so we can deconstruct and undo the social accomplishments of texts claiming to study social reality.

Garfinkel is radical in his use of the documentary method, where events are taken as indications of some underlying pattern. This is circular, he points out. For him, 'for all practical purposes there is no need to probe the subjective meanings behind an appearance, because it is what it is' (200). Members accounts are seen as literal representations of activities, 'situated accomplishments'. Garfinkel's own interpretations are based on these taken as evidence — 'he becomes a privileged interpreter of the events at hand'. However, the ethnomethodologist is not universally competent to know everything, and indeed, sometimes they have to do analysis. This provides an awkward position both inside and outside their own texts. The world out there is mapped by sociology in a distinctive way, however, which 'exists only in our texts'. This is normally glossed by flourishing '"methodological" words', like interview or documentary method, with 'favourite terms', like definition of the situation, indexicality etc. There are implications for what is meant by the subject – n 8 suggests a notion of authorship, being a subject in your own discourse, then being a subject for others like readers, then taking other texts as a subject for this text, including texts which reflect a subject's experiences. Overall, Denzin thinks that we can no longer maintain 'a division between fiction and social science data. Fiction does not mean falsehood… Good ethnographies are true fictions' (201) [citing himself and Clifford].

We can deconstruct Garfinkel's texts in four ways: examine the real and its representations; text and author; presence and lived experience and their representations; subjects and intentional meanings.

All writing 'is a narrative production' [citing Richardson!], which assumes a separation between writer, text and subject matter, but we can deconstruct this logic to show there are no firm dividing lines. Reports from the field appear as 'realism' [citing Clough]. Specifically the ethnographer becomes 'a masculine hero who confronts and make sense of the subjects life situation', through struggle to understand the other. This confers a privilege on the author of the only one who can understand the subjects history. Oedipal undertones sexualise narrative and thus reproduce gender stratification. The realism of the text is amplified through 'factual realistic accounts', based on a 'realist epistemology' which suppresses the role of the subject in the discourse [still Clough] and puts the narrator outside discourse. A circular process ensues, where the ethnographers understandings lead to records as instances of experiences, reflections of cultural order. n9 tells us that 'concerns for the accuracy, validity and reliable coding of field notes belie a commitment to this realist epistemology'. In conversation analysis, textuality affects both the actual conversation, the tape-recording of it, the transcription of the tape and the analysis of the tape and 'the speaker a subject presumably is present in each textual formation' there is a structure of supplement, whereby transcriptions supplement tape recordings which supplement conversations. Many sociological narratives also take on melodrama in the interests of happy endings, and these reproduce the 'master tales that circulate in the popular culture', such as the Oedipal background of the classic morality tale, featuring 'the three steps of seduction, corruption and redemption' commonly found in accounts of deviants and moral careers. Often the sociological text stops short of redemption as if sociological description is the end of the story. Researchers themselves are sometimes seduced or corrupted in the field, only to achieve a happy ending through method, although these stories are typically separated from the substance.

An ethnographic text is like cinema with concrete representations of subjective experiences. This is done by interpreting the subjects experience, that is 'using sociological words to describe what the subject has reported or what has been observed' (202). The subject's word can also appear directly in the text as field notes or excerpts. In this way, the text is the only way to represent concrete events.

As Derrida argues, the whole thing presumes an identity between voice and being. Usually, what a person says is what they mean and it is assumed that people know what they mean because they know what they think — language 'permits this self-knowledge'. Concrete experiences guarantee the presence of the subject and the connection with meaning, so to report words and actions is to achieve 'concrete documentation the meaning of her experience'. However, for Derrida, language does not permit self understanding because it involves constant 'deferral, delay and transformation' with no simple presences or absences, only 'differences and traces'. Speech is not a direct mirror of thoughts or intentionality. A speaker is 'never fully present to himself… Because his language never permits him to state with finality or clarity what he means; what he means is always part of something else'. There is indeterminacy, ambiguity, even with subjects reports or cinematic texts which are particularly good at providing a sense of presence. Ironically, 'an ethnographic text… Can only capture or represent that which is absent — the actual talking subject — through narrative illusion' (203). This is achieved through constructing 'ethnographic realism', through things like detailed field notes. These imply that texts capture experience, denying all the qualities of language above like deferral and delay.

Turning to Agnes, we can summarise the story. We can note immediately that there are materials from other accounts in Garfinkel's text [like the physician's notes]. Garfinkel recorded 35 hours of conversation with Agnes and transcribed them. He added later information about the lie, but manages to use that to justify ethnomethodology nonetheless, in studying practical accomplishments which affect even what looks determinate and objective. Garfinkel offered to write a subsequent study using material gathered from more detailed disclosures of the lie.

Using the lie to support the method is the problem. How was the first text produced as a plausible account of what turned out to be a lie. Garfinkel's text includes conventions from detective stories and melodrama which help make it plausible. From the beginning, 'Garfinkel saw Agnes as the person she wanted to be', and as someone 'who had accomplished passing' and was thus a suitable case study. The case also helps him address gender and develop 'a game model of interaction'[very promising stuff this — how does Denzin produce interesting stories about his own work as a crusade against capitalism? What techniques of emotional realism are deployed? How are victims of colonialism taken as examples for use in his own project? NB this subtlety contrasts with this account of the role of the media in dominant ideology. How come some people are duped but others express their own experiences in critical performance? -- HE seems to be the difference -- you need it to find your voice. Handy]. Garfinkel was 'duped' primarily by Agnes's appearance, which has to be located in a particular gendered space [NB n16, 204 refers to 'a powerful psychoanalytic subtext to the story' which Garfinkel glosses and thus misses a chance to see 'the sexual underside of Agnes's story' which Garfinkel himself hints at.

There is much more conventional commentary on gender, when Garfinkel asserts that Agnes was a typical girl, and contrasts her to stereotyped male homosexuals. He insists that '"her manner was appropriately feminine with a slight awkwardness that is typical of middle adolescence"' (205). Denzin says that this 'diminishes her standing in the female community'. This is a voyeuristic and masculine reading.

Describing his interaction with Agnes, it is clear that Garfinkel is interacting in a classic way to confirm her orthodox femininity — she likes it when he offers to hold the car door open, for example. This is 'leading Agnes into femininity. His actions call out these characteristics in her, which he then sees and reports upon'. He becomes the mirroring male other. He is the only one entitled to tell her story, even though the physician's story is also cited once or twice. He divides her life into pre-operation and postoperation, itself divided into three sections, leading to her achieving the properties of '"the natural normal female"' and then passing. This is a melodramatic realist narrative where aligned individual and victim achieves a happy ending by overcoming fate. Agnes is allowed to speak for herself now and then, but 'Garfinkel does most of the interpreting'. [Examples of the remarks at the different stages then follow, and there are minor contradictions with the physician's account]. Garfinkel 'fills in' these condensed remarks with commentary about Agnes, how she felt, what she acted like, and how she responded to his questions, for example whether she ever compared herself with homosexuals and transvestites. Garfinkel refers to this as 'her presentation of the 120% female'. (206) Here there are more substantial contradictions with the account she gave to the physician. Garfinkel is narrativising the tragedies while simultaneously 'suppressing' his own construction of his subjects through his own interpretations. Garfinkel writes her story for her, in an attempt to sidestep 'the problem of presence and the representation of the subject in the text'. Overall, she is 'talked about more than she talks herself'. Postoperation, it seems Agnes suffered quite serious side effects, says the physician, eventually to be diagnosed as an acute loss of oestrogen. Garfinkel continues to report Agnes's speech, including her feelings about her boyfriend — he refused to talk to Garfinkel himself. Garfinkel then uses a 'game model of interaction' (207) to describe the various strategies Agnes used so she could pass herself off as fully female, although there is an appendix which gives a happy ending — she is successfully passing five years later.

Garfinkel's text can be undone citing the problems of author, text, presence, realism and so on. Firstly, the text is really two texts, Garfinkel's version and the psychoanalytic subtext [apparently also developed in the physician's story]. This subtext is given a presence, but also subverted and displaced, in order to produce 'the definitive account of "passing"', which surpasses Goffman in going on to talk about a more general model of sociology, society, experience and sociology's texts, a 'transcendent claim' (207). This is to be developed by gaining access to Agnes's sexuality. Garfinkel poses as a 'gentle, fatherly interviewer and conversationalist… even matronly'. This helps him become 'the masculine hero who will make sense of Agnes's life for her'. She is presented as a strange other, but never treated as a man and thus 'as his equal': she remains instead 'the perfect female subject'. She is then located 'within the Oedipal framework'. This is sketched in by reference to her being raised in an all woman household, which 'shaped and moulded' her identity. Her boyfriend takes the place of a father, his mother becomes a proper mother teaching her to cook and how to act in male company. All this is seen as learning to pass. He never actually gets to interview Agnes's mother, however.

The text is thus more than a series of thoughts about passing and how people rationalise it. He wanted to do more than just offer '"one more authoritative version of what everyone knows"' [the danger of much collaborv autoethnog?]  (208). He had lofty aims to describe sociological phenomena and how dependent upon they are being able to give good reasons, how these reasons are produced from within situations, how matters of constancy are achieved with values and objects and how this is tied to impression and management, how members produce '"stable accountable practical activities, i.e. social structures of everyday activities" '. [Denzin and the others do a directly comparable talk up on the routine disappointments and frustrations of everyday life which have to be turned into oppression by racists and positivists]. His text has to both map Agnes's experiences and also show how accountable practices are produced. This assumes that Garfinkel's own 'common sense understandings' are the same as Agnes's; that the text he writes maps those understandings accurately so that what is heard 'becomes proximate with the text that he produces' [citing Derrida on the proximity of voice and the meaning of being] [shown equally well in the appalling literalness of language in the qualitative researchers he likes] presents guarantees authenticity both with his recording of conversations and his interpretations. By Oedipalising the subject, he converted himself into an authority [same goes with diagnoses of cultural politics the evils of capitalism and so on]. He also wrote a self reflective text showing how realities were constituted through description — but these descriptions refer to Garfinkel's version of Agnes's reality, not hers. In this way reflexivity turns back on the author [the problem of infinite regress also apparent in Denzin's work — if everything is political so is his].

Agnes is present in the text through a series of imposed oppositions like male-female, homosexual transsexual and so on. These are arranged in a hierarchy with male and female and normal and deviant. There is supplementarity, where his commentary supplements her words [all academic work does this — only experts can see the oppression in positivism]. Paradoxically, the other is only present through these interpretations which are themselves uncertain — however Garfinkel's text privileges his voice despite his methodological claims.

Overall, he fails to describe some level of reality beyond the mundane, and to penetrate the 'problematic world of experience that for Agnes was primordial' (209) [an example of his own priorities expressed in the dominant voice] 'Agnes wanted' various things, including being a conventionally sexualised woman in a female body, becoming 'the sexual object that her culture prescribed '[so she was a cultural dupe after all? Coward is cited 'on beauty and the commodification of the female body']. She wanted to be a woman [not just successfully pass as one? But this is Garfinkel's point, surely — we all pass as our stereotypes, and this indicates that commodification is by no means entirely successful?]

Garfinkel was fascinated by Agnes's strangeness and 'taken in by her femininity' he wanted 'to disprove her homosexuality' and did this by confronting her with the classic images of the homosexual or transsexual male, to clarify the ambiguous spaces that contained these types of '"aberrant"' sexuality. 'This is where he wanted her to be', to be a homosexual, transsexual or transvestite, the only categories that preserve the structural oppositions of male and female that he subscribes to. There can be no in between for her or her lover. Note 28 adds that this is because Garfinkel's project 'is ontological, not epistemological — 'his theory of being (ontology) assumes that being and appearance are the same (epistemology)' (209) [evident very strongly in the claim that political incorrectness means racism]. The notion of a level of being which lies beneath surface appearances is dismissed [by Denzin too in the piece on Goffman]. Heidegger is quoted as singling out a 'more primordial' level beneath appearances, where we can find references to 'the meaning of existence'. 'Garfinkel's project does not presuppose the second layer or level of existence; mine does [!]' So Garfinkel could not see another layer to Agnes's experiences and this is what undermines his own text from within.

The problems can be discussed with reference to the '"coroner's problem"' and the '"potter's object"' [these apparently are terms used in Garfinkel Lynch and Livingston]. The coroner works backwards from a dead body to determine the cause of death, but the potter takes an object and shapes it into a cultural object which gives meaning to it through production — '"first – time – through"'. Garfinkel pursues the coroner's problem, where Agnes's passing is 'a fact to be explained' by working backwards [hints of the asymmetry argument about CCCS work — obviously affects Denzin's cultural politics as well]. However, Agnes was not a passive 'dead body' but someone with a desire and an intention to mislead, hence her potterish account where she forms her own sexual identity through interactions with herself and others. There was no determinate history or nature 'she never was who she said she was' and 'Garfinkel was duped' [note 29 says that the text also appears to be potter-ish 'because he searches continually for clues and reasons to explain how she learned to pass — this is a common academic narrative, however where abstract research appears to deliver what you suspected all along]. As a result Garfinkel could not grasp fully the sexual subtext [and privileged a Freudian account] — Agnes simply wanted to be rid of her penis and to have a suitably deep vagina 'that was beyond suspicion'. Garfinkel missed the chance to grasp this complex femininity and sexuality. This is because he subscribes to the view that femininity and masculinity 'are socially defined ingredients which are added to biological gender' [very contemporary debate]. To go beyond that requires a feminist approach asking how people are '"formed through their sexuality"' [citing Mitchell] is not a matter of learning how to add femininity but rather of showing 'how gender exists through sexuality — gender and sexual identity interact', but this is missed and thus 'he fails to tell Agnes's strange sexual story'.

Agnes understood this better. Sexuality needs biological genitals, but goes beyond that need to involve 'a mode of sexual being'. For her this had to be fully feminine without any contamination of maleness, hence she sought castration, because she realised that the phallus was as a major way to define the difference between the sexes. She wanted to remove the phallus and did not fear castration, which contradicts Freud [note 30 refers to hints in the physician's text implying that she had had too much infantile contact with her mother's body and a psychologically absent father]. She wanted real castration not symbolic. She refused her penis and thus 'any male privilege'. 'Like Lacan' she [must have] understood that castration involves humanisation, the emergence of fully human sexual differences. But Agnes rejected her biologically given identity with its accompanying images, including the insistence by her family that she was male, or Garfinkel's suggestion she was homosexual. 'In Lacan's terms' [!] 'Agnes rejected her mother because she rejected her own phallus and hence the male identity'. This led to reject conventional family relationships with mother and brother. She rejected 'the paternal metaphor' and 'sought a sexual body that would allow her to go beyond her missing father and her displaced mother', and this would 'contain the wild sexuality that looked in her imagination — she wanted to have intercourse with men [this is very much a third level commentary of course, where Denzin is commenting on Garfinkel's comments on what Agnes said]

This required a vagina created surgically. The use of a plastic mould shaped like a penis in the operation means 'Agnes became a being who had inseminated herself with her own male seed' (211) [citing some fanciful stuff about vaginas as folded spaces, from Derrida and Spivak]. This was 'her own sexual union, a union forever deferred until she found herself in the arms of a man making love as she imagined lovemaking to be. She folded herself into herself; her castration was a necessary negation on the way to full sexual selfhood' [very confident commentary]. Her identity was still seen in relation to the phallus, both present and absent and which became her vagina. She also 'took a feminine stand within language and rejected all male signifiers that could be applied to her', including refusing to accept Garfinkel's terms. 'As if she had read Lacan, she knew' [!] that feminine sexuality belonged to masquerade, because it is always constructed with reference to male signs. She wanted to become active sexually so that she could be a woman and this required her to overcome male signifiers. As Lacan argues, she realised that 'all women… [live their ]… sexuality under the male sign of patriarchy. She agreed to become a research subject in order to achieve her own goals.

Garfinkel also learned to pass in these interviews, by offering her 'scientific findings of which he wasn't sure'. He uses the unfortunate term telling a 'cock and bull story'. Perhaps Agnes realised that she had to tell her own cock and bull story.

What we have here is a story that is 'emblematically tragic' [Denzin's preferred narrative]. It shows the 'lengths to which women must go in order to be women in this culture'. If they want to change their form they must submit to dominant males who will transform their bodies and also tell the stories about their transition. Agnes was of course reticent, and it is not surprising that she failed to give information on critical issues like whether or not she had taken hormones, how exactly she collaborated with her mother, how she had managed to male feelings, whether she'd ever used the penis sexually, how she managed to satisfy herself and her boyfriend sexually, whether or not she had homosexual feelings, and whether she saw herself as phony. These questions were refused — 'why should she answer'. She had already 'fully and completely accepted her sexuality'. She'd already experienced hostility. She was being turned again into a subject by 'male representatives of this patriarchal culture'. 'She used her femininity against science, as she got from science what she most desired… Agnes created her own text' and Garfinkel could not penetrate it.

Derrida reminds us that the text is 'always in motion, doing and undoing itself as it moves forward' (212), but it is always waiting to be violated by interpretation, 'by the [dis] dissemination of the writer's pen', a 'male act'. But Agnes 'let's the semen fall in advance of penetration and interpretation' [oh dear]. She greets Garfinkel with silence, will not be an available object for scientific penetration. She engages in 'proactive manipulation, characteristic of all clever subordinates who know how to work their superiors'. This is typically feminine but in a subversive way — 'she turned woman against man'. The account actually offers a challenge to Garfinkel's text 'and the scientific model on which it rests'. She resisted any method that would enclose her life story within a narrative [the implication is she somehow recognised conventional patriarchy and scientific dominance in ethnomethodological narratives — Denzin's ventriloquism]

We see an underlying commitment to essentialist metaphysics of presence in Garfinkel's account. This appears in texts as a subject that is present in a 'firm and incontrovertible' way, but subjects like this are there 'only because analysts were looking for them in the first place and because subjects allowed themselves to be so represented' [Gale and Wyatt are exactly these firm subjects]. We need to remember there are different levels of subject and experience — 'the worldly or flesh and blood, the empirical, the analytic, and the textual' [further shuffled into generic versions with subtypes in note 33]. Worldly subjects are located in historical moments as 'a universal singular' [for Sartre] and can express 'deeply felt emotions that may escape textual representation'. This is Agnes in her own words. The analytic subject however is a construction of the subject as a social type, or ideal type, a second-order construct. The textual subject contains both and others — a medical text, 'a first-order textual representation' where direct commentary is included, and a 'second-order empirical textual representation' when sociological commentary is added. A particular type of sociological rewriting of subjective experiences is 'analytic textuality' where a subject's experiences are rendered as those of an ideal type' [exactly as Denzin does above when Agnes becomes an ideal type feminist woman]. We see interaction between these levels 'sometimes silently or subtextually'. As one voice speaks it displaces and traces other voices. Garfinkel 'cannot be separated completely from Agnes', although we can try to do this: however if we do we are left with 'an analytic text about an analytic subject'.

Traditionally interpretive ethnography, which includes Garfinkel, privileges second-order empirical representations as analytic textuality. This 'privileges the researchers gaze'(213). Agnes was able to resist such gazes, although she helped Garfinkel and the physician develop their stories. Nevertheless, her story transcends 'gender, ethnomethodology, and psychoanalysis'.

As a result, 'the traditional divisions among author, subject, and text can no longer be maintained'. Experience cannot be represented directly, but given 'only through traces and movements of the federal, or indirectly' — through subject enunciations or through analytic textuality. This confirms the claim that there is only the text. Garfinkel's interest in how texts are produced ignores that 'what is happening is the text' [note 34 notes that texts may produce accounts of things that may not have happened]. There is no zero point in a text. When we deconstruct texts we show how 'it's very constructions maintain the illusion that real or analytic subjects have been discovered and analysed'. This produces results that will be 'dear to our post-modern hearts'  [not sure if the textual construction or the deconstruction does this]. A logic of narration sustains conventional interpretive sociological writing. [Note 35 acknowledges that his own position can be deconstructed because he has assumed a privileged position to Garfinkel's text, and we can see how he did that — 'I staked my position on Agnes's sexuality and on my construction of a level of her being that escaped Garfinkel's gaze']. Stories are located in real time and space which reduces experiences to particular points where interaction is occurred. This helps the illusion that we can capture experience. Texts presume an enacted and self-sufficient world. Temporality gets confused with causality as in a typical narrative. The narrator is required to enclose the events in the story. 'This strategy creates a teleology, ally to the effect that there are no accidents in this world. Everything is sensible and can be made sense of' [precisely Gale and Wyatt]. Post-modern novels reject this approach, but it still structures ethnography and ethnomethodology. It is an illusion of 'a knowable universe in which knowledgeable scientists as writers make sense of the commonplace and the out of the ordinary'

Instead there exists a '"wild," uncontainable, worldly, flesh and blood version of the subject who continually eludes narrative capture'. [Something IS outside the text?] This is what Heidegger meant by being, life below the surface representations, life which struggles with matters of being emotional and existence — and sexuality. 'The sexual Agnes escaped Garfinkel's net'. We learn from the example that 'one level of the subject, then, always stands outside or alongside any single text'. There can be no exhaustive system of representation. There are always unrepresentable 'presuppositions' or 'domain of application' (214). 'There will be alternative versions of the subject that are never contained within an narratives'.

'Harold and Agnes are one and the same'. Garfinkels story shows how fragile interpretive projects are. In effect, 'Agnes… is a creature of his ethnomethodological imagination'. Together they were able to produce accounts that pass for something else. But we should focus on the something else, because 'the texts that we write make a difference; these differences often have effects on the lives of "real" flesh and blood people'.

back to social theory page