The Epistemic Contract of Bisexual Erasure

Home , BECAUSE (conference), BiNet USA, Bisexual chic, Bisexual erasure, Bisexual politics, Bi Any Other Name

TheEpistemic Contract of Bisexual Erasure

KenjiYoshino*

In thisarticle, Professor Kenji Yoshino seeks to explainwhy the category of bisexualityhas beenerased in contemporaryAmerican political and legal discourse.He firstargues that the invisibility ofbisexuality relative to homosexualitydoes not reflect the incidences of those orientations in thepopulation. Definingbisexuality as thepossession of more than incidental desire for both sexes,Yoshino shows that the major sexuality studies demonstrate that the inci- denceof bisexuality is infact greater than or comparableto theincidence of homosexuality.Yoshino explains the erasure of bisexualityby positing that bothself-identified heterosexuals and self-identified homosexuals have overlapping interestsin theerasure of bisexualitythat lead theminto an "epistemic contract"of bisexual erasure. These interests include: (1) thestabilization of exclusivesexual orientation categories; (2) theretention of sex as an important diacriticalaxis; and (3) theprotection of normsof monogamy.Noting that suchcontracts tend to becomevisible only when they are challenged,Yoshino describeshow bisexualshave increasinglycontested their own erasure. Fi- nally,Yoshino examines the effects of bisexual invisibility and visibilityin the legal realm,focusing on thesexual harassment jurisprudence of recentdec- ades.

* AssociateProfessor, Yale Law School. I thankAkhil Amar, Ian Ayres,Jennifer Gerarda Brown,Ariela Dubler, Bill Eskridge,Oren Izenberg, Robert Post, Bill Rubenstein,Vicki Schultz, RevaSiegel, and Amanda Tyler. I am also gratefulto participants in workshops at ColumbiaLaw School,Fordham Law School,and Yale Law School,as wellas studentsin myTheorizing Sexual- ityseminar at Yale and LarryLessig's Advanced Constitutional Law seminarat Harvard.Rick Baker,Romana Mancini, Ravenna Michalsen, Zachary Potter, Rose Saxe, and EricSonnenschein suppliedexcellent research assistance.

353

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 354 STANFORDLAWREVIEW [Vol.52:353

INTRODUCTION ..... 356 I. THE ERASUREOF BISEXUALS ....363 A. BisexualInvisibility . . .364 1. Bisexualinvisibility defined . .364 2. Evidenceof bisexual invisibility . .368 B. BisexualErasure . . .370 1. Bisexualitydefined . .370 2. Thestudies . .377 a. Kinsey(1948 & 1953).380 b. Mastersand Johnson (1979) .382 c. Janusand Janus (1 993) .383 d. Wellings(1994) .383 e. Laumann(1994) .385 f. Critiquesof the studies ...... 386 C. BisexualErasure as a Cause ofBisexual Invisibility ... 388 II. THE EPISTEMICCONTRACT OF BISEXUALERASURE ...... 388 A. DifferentExplanations ...... 389 B. TheEpistemic Contract Defined ...... 391 C. Strategiesof Erasure ...... 395 1. Straightdeployments of thestrategies ...... 395 2. Gaydeployments of the strategies ...... 397 D. TheEpistemic Contract as a Cause ofBisexual Erasure ... 399 III. MONOSEXUALINVESTMENTS IN THE EPISTEMICCONTRACT ...... 399 A. Stabilizationof Sexual Orientation ...... 400 1. Sharedinvestment ...... 400 2. Straightinvestment ...... 402 3. Gay investment ...... 404 B. BisexualityDestabilizes the Primacy of Sex ...... 410 1. Sharedinvestment ...... 411 a. Destabilization...... 412 b. Thetension between public and private treatmentsof sex ...... 413 2. Straightinvestment ...... 415 3. Gay investment ...... 417

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 355

C. BisexualityDestabilizes Norms of Monogamy ...... 420 1. Sharedinvestment ...... 421 2. Straightinvestment ...... 423 3. Gay investment...... 426 D. OverlappingMonosexual Investments as a Cause ofthe Epistemic Contract ...... 428 IV. SELF-IDENTIFIEDBISEXUALS AND THE EPISTEMIC CONTRACT ...... 429 A. BisexualCapitulation to theEpistemic Contract ...... 430 B. BisexualResistance to theEpistemic Contract ...... 431 C. TheDissolution of the Epistemic Contract ...... 434 V. BISEXUALITYAND SEXUALHARASSMENT LAW ...... 434 A. A Summaryof the Development of the Sexual HarassmentJurisprudence ...... 436 B. BisexualVisibility-The Recognition and Closingof the BisexualHarassment Exemption ...... 439 1. Therecognition of the bisexual harassment exemption- bisexualvisibility ...... 440 2. (Incoherently)closing the bisexual harassment exemption-bisexualinvisibility ...... 442 3. (Coherently)closing the bisexual harassment exemption-bisexualvisibility (again) ...... 444 C. Recognizingand Closingthe Horseplay Exemption ...... 446 1. Understandinghorseplay- thehomosocial and thehomoerotic ...... 448 2. Recognizingthe horseplay exemption- bisexualinvisibility ...... 450 3. Closingthe horseplay exemption- bisexualvisibility ...... 451 D. SexualHarassment at a Crossroads ...... 454 1. Thepost-Oncale status quo ...... 454 2. Bisexualityas goad ...... 457 E. OtherApplications ...... 458 CONCLUSION...... 460

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 356 STANFORDLA WREVIEW [Vol. 52:353

INTRODUCTION

Teachinga seminaron SexualOrientation and the Law, I facedan old inconsistencyso frontally that it became difficult toavoid giving it sustained attention.I began the course in what appears to be a commonway,l by pos- ingbasic questionsabout sexual orientation. I askedwhy contemporary Americansociety2 organizes people according to theirsexualities;3 why we do so onthe basis of sexual orientation inparticular;4 and why, when classi- fyingby sexual orientation, we insiston doingso withthe binary system of heterosexualand homosexual.5 In discussingthe last question, I adduced the

1. A leadingcasebook on sexualorientation and the law, for example, begins with such an in- troductoryunit. See WILLIAMB. RUBENSTEIN,CASES ANDMATERIALS ON SEXUALORIENTATION ANDTHE LAW 1-40 (2d ed. 1997) (presentingbackground materials discussing sexual classifications). 2. Thesetemporal and geographical restrictions apply to thisentire analysis. These restric- tionspartially recognize that the concept of orientation is culturally specific. See note77 infraand accompanyingtext (describing Latin bisexuality). That recognition, however, will unfortunately be incomplete,insofar as itignores variations within modem American society based on, for example, culture,race, and class. See, e.g.,Will Roscoe, How to Becomea Beardache:Toward a Unified Analysisof GenderDiversity, in THIRD SEX, THiRD GENDER:BEYOND SEXUALDIMORPHISM IN CULTUREAND HISTORY 329, 330-49(Gilbert Herdt ed., 1994) (describingthe Native American berdache,an identitybased on genderatypicality that cuts across the orientation categories consid- eredhere, as an identitythat is moresocially salient in manyNative American societies than those orientationcategories). 3. Resistanceto theclassification ofpersons according to theirsexualities can be seenin that usageof the word "queer" which refers to individuals who fall outside of the realm of the "normal," eitherbecause of theirsexuality or forsome other reason. See MichaelWarner, Introduction to FEAROF A QUEERPLANET: QUEER POLITICSAND SOCIALTHEORY vii, xxvi-xxviii(Michael War- nered., 1993). Thisusage's "aggressive impulse of generalization," id. at xxvi,resists the reifica- tionof sexualityas an axis of demarcationby recastingthe conflict between sexual deviance and sexualnormalcy as onebetween social deviance and social normalcy. 4. Resistanceto theclassification of personsaccording to theirsexual orientations can be foundin Eve Sedgwick'sprovocative list of alternativeclassifications. See EVE KoSOFSKY SEDGWICK,EPISTEMOLOGY OF THECLOSET 25-26(1990). Sedgwicknotes that even if we setout to distinguishbetween people based on theirsexualities, many other axes besidessex of object choiceare available. See id. at 25. Forexample, "[s]exuality makes up a largeshare of theself- perceivedidentity of somepeople, a smallshare of others"';"[s]ome people spend a lot of time thinkingabout sex, others little"; and "[s]ome people like to have a lotof sex, others little or none." Id. Indeed,Sedgwick contends that the rise of sex ofobject choice as thecritical axis of definition was a contingentand puzzling historical development out of an era in whichsuch other axes had analogousdiacritical force. See id.at 8-9. 5. Resistanceto theclassification of personsaccording to a binarysystem of sexualorienta- tioncan be foundin Alfred Kinsey's classic statement: Males do notrepresent two discrete populations, heterosexual and homosexual.The worldis not to be dividedinto sheep and goats. Not all thingsare blacknor all thingswhite. It is a fundamentalof taxonomythat nature rarely deals withdiscrete categories. Only thehuman mindinvents categories and triesto forcefacts into separated pigeon-holes. The livingworld is a continuumin each and everyone of itsaspects. The soonerwe learnthis concerning hu- mansexual behavior the sooner we shallreach a soundunderstanding ofthe realities of sex. ALFRED C. KiNSEY, WARDELLB. POMEROY& CLYDE E. MARTIN,SEXUAL BEHAVIOR IN THE HUMANMALE 639 (1948) [hereinafterKINSEY ET AL.,MALE].

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 357

view-powerfulin modem American culture from at least the publication of theKinsey studies6 onward-that sexual orientation arrays itself along a continuumfrom exclusive heterosexuality to exclusive homosexuality.7 I notedthat this view encouraged us tothink of the straight/gay binary as de- finingthe ends of a continuumthat could be stretched,accordion-like, toac- commodateever finer gradations of cross-sexand same-sexdesire. This meantrecognizing a group-oftencalled bisexuals-on the intermediate stretchof thecontinuum, as well as thepossibility of a group-sometimes calledasexuals-not represented on thecontinuum at all.8 Indeed,I argued

6. ALFRED C. KINSEY, WARDELLB. POMEROY,CLYDE E. MARTIN& PAUL GEBHARD, SEXUALBEHAVIOR IN THEHuMAN FEMALE (1953) [hereinafterKINSEY ET AL., FEMALE];KINSEY ET AL.,MALE, supra note 5. 7. See Carol Queen, Sexual Diversityand BisexualIdentity, in BISEXUAL POLITICS: THEORIES,QUERIES, & VISIONS 151, 152 (Naomi Tuckered., 1995) (notingthe importanceof the Kinseycontinuum). 8. It is withsome regretthat I have decidednot to attempta systematicdiscussion of asexuals in thisarticle, especially since asexuals are, if anything,more likelythan bisexuals to be erased in sexualitydiscourse. To concede thatthere are two formsof desire-cross-sex and same-sex desire-is to recognizethe analyticpossibility of at least fourkinds of persons. These include: (1) those who harborcross-sex but not same-sexdesire; (2) those who harborsame-sex but not cross- sex desire; (3) those who harborboth formsof desire; and (4) those who harborneither form of desire. Yet even thosewho acknowledgethat orientation arrays itself on a continuumspanning the firstthree categories often ignore the factthat the continuumfails to representthe fourth.This is somewhatsurprising, as thenumber of individualsin thiscategory is not insignificant.See Naomi Mezey,Dismantling the Wall:Bisexuality and thePossibilities of SexualIdentity Classification Based on Acts, 10 BERKELEYWOMEN'S L.J. 98, 106-07 (1995) (citingthe Kinsey reportand noting thatthe number of asexuals is "notnegligible"). Kinsey,for example, found that asexuals included 14 to 19 percentof unmarriedwomen between the ages of twentyand thirty-five.See KINSEY ET AL., FEMALE,supra note6, at 499. My regretis made keen by theconvergences between bisexual and asexual erasure,most nota- bly the refusalby bothself-identified straights and self-identifiedgays to acknowledgeeither category. Thus asexuals, like bisexuals,are proneto being accused of duplicityor false consciousness, or, more specifically,of being closeted gays. See JEFFREYS. NEVID, Lois FICHNER-RATHUS& SPENCERA. RATHUS,HUMAN SEXUALITY IN A WORLD OF DIVERSITY 302 (1995) (describing asexuals as a subset of homosexuals). See generallyALAN P. BELL & MARTINS. WEINBERG, HOMOSEXUALITIES:A STUDY OF DIVERSITYAMONG MEN ANDWOMEN (1978) (discussingasex- uals throughoutas subsetsof male and femalehomosexuals). The decision to defera discussionof asexuals foranother day, however,is supportedby the undertheorizeddivergences between bisexuality and asexuality,which suggestthat the two topics deserve separateanalysis. While both doubled and absentdesire appear to threatenstraights and gays, theydo so in quite differentways. To takeone crudecut at thatdifference, consider the dis- parate ways in which the time-honoredconflation of sexualityand sin ramifiesacross bisexuality and asexuality. If thisconflation leads some to view bisexuals as particularlyculpable because of their"promiscuous" desire forboth sexes, see notes 353-393 infraand accompanyingtext, it leads some of the same people to view asexuals as particularlypure. See Lucinda J. Peach, From Spiri- tual Descriptionsto Legal Prescriptions:Religious Imagery of Womanas "Fetal Container"in the Law, 10 J.L. & RELIGION 73, 76 (1994) (notingthat the two templatesof thefeminine in the Chris- tiantradition are Eve and Mary,and thatMary, in contrastto Eve, is portrayedas pure,celibate, and asexual). But while such purityis oftenascribed to celibacy,see, e.g., GabrielleBrown, The Celi- bate LifeIs Fulfilling,in HUMAN SEXUALITY 393, 395 (Bruno Leone et al. eds., 1985) (notingthat celibacy "can be thoughtof as the desire forsomething more eternal,more permanent"),it is not obvious whetherthat ascription applies equallyto thesubset of celibateswho are asexual. Celibacy

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 358 STANFORDLAWREVIEW [Vol. 52:353 thatsexual orientation classifications that only used the two "monosexual" terms9"heterosexual" and "homosexual" were unstable and naive. As soonas theintroductory unit was over,however, the inconsistency occurred.I foundmyself and the class falling back into the very "unstable" usagesI had workedhard to retire-specificallythe usages of thewords "heterosexual"and "homosexual"as mutuallyexclusive, cumulatively ex- haustiveterms.10 While we sometimesrallied by usingthe word "queer" insteadof "gay,""1or by adding the rider "or bisexual" to "gay,"12 these ef- fortswere token and fitful. In theface of legal discussionsl3 and academic commentary14thatwere relentless in reifying the straight/gay binary, it was difficultto holdthe bisexual steadily visible, even as a spectralpossibility. And whilethis failure to resistwhat I had criticizedas a distortionwas strikingin a classthat sought to treat the issue of sexual orientation with sophistication,it was simultaneouslyall too recognizable as an inconsistency thatriddles more quotidian discourse. Many who would not deny that bi- sexualsexist when the subject of bisexuality arises can nonetheless revert to thestraight/gay dichotomy when the topic shifts.15 I myself can speakat lengthabout bisexuals at onemoment and then, in the next, field a question

may be pure because it constitutesa conquestof the baser desires of the body, see id.; if so, the celibate asexual's claim to purityis attenuatedbecause his licentiousdesire is not overcome,but ratherabsent. And even if describedas pure,the absence of desiremay be viewed as a disquieting purity,insofar as our hedonicpleasure in othersis viewedby some as a generative,fecundating, and humanizingforce even (or perhapsespecially) when sublimated. See Leon R. Kass, The Wisdomof Repugnance:Why We Should Ban theCloning of Humans, 32 VAL.U. L. REv.679, 691-92 (1998) ("Sexual desire ... is thussublimated into erotic longing for wholeness ...."). Thus,while bisexualityand asexualitymay in some senses be viewed as simpleopposites (oversexed v. undersexed), theyshare negative connotations. But theseconnotations, in tum,are differentlynegative. 9. Technically,"monosexuality" denotes the state of possessing one of the two traditional "sexes" (male or female),as opposed to havingone of thetwo traditional"orientations" (heterosexual or homosexual). See 9 OXFORDENGLISH DICTIONARY 1029, (2d ed. 1989) [hereinafterOED]. In thisarticle, however, I followthe rising practice of using the termin the lattersense. See, e.g., RUTH COLKER, HYBRID: BISEXUALS,MULTIRACIALS, AND OTHER MISFITS UNDER AMERICAN LAW 16 (1996) [hereinafterCOLKER, HYBRID]. 10. See Mezey, supra note 8, at 98 (describingthe categoriesas exclusive and comprehen- sive). 11. See MARJORIEGARBER, VICE VERSA:BISEXUALITY AND THE EROTICISMOF EVERYDAY LIFE 62-66 (1995) (discussing"queer" as a means of includingbisexuals withina coalitionmove- mentof sexual minorities). 12. See Liz A. Highleyman,Identity and Ideas: Strategiesfor Bisexuals,in BISEXUAL POLITICS,supra note 7, at 73, 83-86 (describingthe "lesbigay" strategy of namingbisexuals alongside lesbiansand gay men). 13. See, e.g.,notes 48-52 infraand accompanyingtext; notes 503-534 infraand accompanying text. 14. See,e.g., notes 60-62 infra and accompanying text. 15. See ChristopherJames, Denying Complexity: The Dismissal and Appropriation ofBisexu- alityin Queer,Lesbian, and Gay Theory,in QUEER STUDIES: A LESBIAN,GAY, BISEXUAL,AND TRANSGENDERANTHOLOGY 217, 226 (BrettBeemyn & MickeyEliason eds., 1996).

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 359 suchas "IsX straightor gay?" without instinctively feeling as ifan important possibility-thebisexual possibility-has been elided. Whatis happeninghere? Why is bisexualityso invisible?If we inter- pretthat invisibility as the product of erasure, why does that erasure occur? Whyis bisexualitynow becomingsufficiently visible that commentators havebegun to theorizeits invisibility as the result of erasure? How might contemporarysexual orientation politics and law lookdifferent ifthis trend towardvisibility continues? This article will occupy itself with these questions. Framingthe questions in this way itself raises definitional issues. One of thedifficult things about writing about sexuality is thatthere are too many distortionsto correct at once. Someof these distortions must be accepted, althoughnever without the fear that they will infect the result. I addresstwo suchdistortions atthe outset. Thefirst is a qualificationofthe definition of"bisexual." In askingwhy bisexualsare invisible and/or erased, I assumethat there is a categoryof individualswho can usefully be denominatedas bisexuals. I provisionallyde- finethat category below.16 I emphasize,however, that I do notpresent the categoryof bisexuals (or heterosexuals, homosexuals, orasexuals) as a natu- ralkind awaiting detection. While individuals harboring both cross-sex and same-sexdesire clearly exist, it doesnot necessarily follow that such indi- vidualsshould be classifiedtogether as a group.Strictly speaking, then, the realquestion is notwhy bisexuals are erased: one cannoterase something thatdoes not have material existence outside of one's writing it, and which onehas never written. Rather the question is whywe havedivided the world oforientation into categories that tend to suppressthe existence of bisexual desire. Evenbearing this in mind, however, I find it difficult todiscuss that sup- pressionwithout isolating the various constituencies that have interests in effectuatingorresisting it. Whilesocially constructed, the constituencies of heterosexual,homosexual, bisexual, and asexual have political and material consequence.Underscoring their contingency takes nothing away from their socialimportance orheuristic utility. I thereforeself-consciously retain the terms. Thesecond issue is thedefinition of"sex." To possess"bi"-sexual de- sireimplies the existence of two sexes-male and female. While bisexuality is sometimesseen as erodingthe salience of the sex binary,17 itthus nomi- nallyreifies the premise that there are only two sexes. That premise has been contestedon the grounds that a significantportion of the population is inter- sexedat birth, with genitalia that do notconform to thetwo conventional sex

16. See textaccompanying notes 71-109 infra. 17. See textaccompanying notes 314-352 infra.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 360 STANFORDLAWREVIEW [Vol. 52:353

categories.18Yet a manwho is attractedto womenand intersexed individuals,but not to men,is notconsidered a bisexual, because the intersexed do notcount as a sex forthese purposes.'9 Bisexuality also suggeststhat these twosexes are defined biologically rather than culturally.20 This is theclassic divisionmade between anatomical sex andsocial gender.21 Thus a lesbian whois attractedtoboth "butch" and "femme" women will fail the classical definitionof a bisexual,because even though she is attractedto twoindi- vidualswho might be saidto havedifferent "genders," both of them are of thesame "sex." I willagain leave both these premises-that there are two sexes (male and female)and that anatomical sex can be coherentlydistinguished from socialgender-unchallenged forthe purposes of this analysis. I wishto rec- ognizethese premises, however, as deeplycontestable ones, and to gesture brieflytoward the costs of leaving them unassailed. First, the intersexed oc- cupya placebetween the two conventionally ordained sexes (male and female)that the bisexual occupies between the two conventionally ordained orientations.Making bisexuality visible on thegrounds that intermediate categoriesdeserve social attention while letting intersexuality remain invisi- blethus creates an ironicasymmetry. More importantly, theelision of inter- sexualityrepresents a missed opportunity tointerrogate orientation. For just as bisexualitychallenges our conceptions of sex,22so toomight intersexual- itybe usedto challenge our conceptions oforientation. To see this,think of theman who is attractedmore to intersexed individuals than to either men or to women.23What is thisman's orientation? What if he is attractedonly to theintersexed? Is thebisexual, who is sometimesunderstood to be "pan- sexual,"also understood tobe attractedtothe intersexed? Morefamiliarly, the distinction between sex and gender has been under increasingfire in contemporaryqueer theory, such that its use requiresde-

18. See SUZANNE J.KESSLER, LESSONS FROM THE INTERSEXED 135 n.4 (1998) (notingesti- matethat the "frequency ofintersexuality may be as highas 2 percentof live births"). 19. See id. at 5 (arguingthat the popular fascination with intersexed individuals may stem in partfrom the fact that one's heterosexual or homosexualstatus is notcalled into question by attrac- tionto an intersexedperson). 20. See, e.g., MARTINS. WEINBERG,COLIN J. WILLIAMS& DOUGLAS W. PRYOR, DUAL ATTRACTION: UNDERSTANDING BIXSEXUALITY 4 (1994) (definingbisexuality according to sex ratherthan gender). 21. See Biddy Martin,Sexualities Without Genders and OtherQueer Utopias,24.2-3 DIACRITICS104, 104 (1994). 22. See textaccompanying notes 314-352 infra. 23. Thereis a notableamount of pornography that depicts intersexed individuals, leading one tobelieve that some consumers may be particularlyattracted to theintersexed. See KESSLER,supra note18, at 160n.82 ("The attraction to intersexedbodies is farmore common than what would be revealedin countingpartners of the intersexed, as is evidentin theabundant pornography available featuringactors with vaginas and standard-sized penises ....").

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 361

fense.24The defense is purelypragmatic-bisexuals are simply not currently understoodas people who are attracted toboth genders, as opposedto sexes. Whilenot attempted here, however, that extension should be theorized.This is becauseit is clearlytrue that people have gender orientations as well as sexualorientations-one can be attractednot to all men,but only to butch men,or not to all women,but only to femmewomen. Why are certain people attractedtoboth kinds of gender performance within one sex, while oth- erslimit themselves to peoplewho perform their gender only in one way? Theanswer is doubtlesslinked to theways in whichthey perceive and ap- preciate"biological" sex. Indeed,it is thestrength ofthis link that has led theoristslike Judith Butler to posit that there is no pre-discursivebiological substrateof sex that can be distinguishedfrom gender.25 Be thatas itmay, thisarticle leaves the work of theorizing the gender orientatation-bi- oroth- erwise-foranother day. Thatsaid, I cannow assail the question of whybisexuals are invisible and/orerased in contemporaryAmerican culture. In PartI, I demonstrate thatbisexuality is invisible relative to homosexuality and that this invisibility is betterexplained by bisexual erasure than by bisexual nonexistence. After definingbisexual invisibility, I show that bisexuals (under any plausible definitionof bisexuality)are much less sociallyand politically visible than homosexuals.I acknowledge,however, that this discrepancy in visibility doesnot necessarily betoken bisexual erasure, as it couldsimply mean that thereare fewer bisexuals than homosexuals in thepopulation. Because de- terminingthe merit of this position requires a more precise definition of"bisexuality,"I generate and defend a provisionaldefinition of bisexualityas theability to feelmore than incidental sexual desire for both sexes. Using thisdefinition, I look at whatthe major sexuality studies say about the inci- denceof bisexuality and homosexuality in the population. Two thingsare surprisingabout such an investigation.First, to myknowledge, no onehas previouslymade such a systematiccomparison. Second, when such an in- vestigationis actuallymade, it revealsthat each ofthe major sexuality studies demonstratesthat the number of bisexuals is greaterthan or comparable tothe number of homosexuals. This suggests that bisexual invisibility is not a reflectionof the fact that there are fewer bisexuals than there are homosexualsin the population, but is rathera product of social erasure. Havingdemonstrated erasure in PartI, I seekto explainit in PartII. I suggestthat erasure occurs because the two dominantsexual orientation groups-self-identifiedstraights and self-identifiedgays-have shared in-

24. See, e.g., JUDITHBUTLER, GENDER TROUBLE: FEMINISMAND THE SUBVERSIONOF IDENTITY7 (1990) (describingthe claim, central to herwork, that "this construct called 'sex' is as culturallyconstructed as gender;indeed, perhaps it was alwaysalready gender, with the conse- quencethat the distinction between sex and gender turns out to be no distinctionatall"). 25. See id.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 362 STANFORDLAWREVIEW [Vol. 52:353 vestmentsin thaterasure. It is as if thesetwo groups, despite their other virulentdisagreements, have agreed that bisexuals will be madeinvisible. I callthis the epistemic contract of bisexual erasure. To supportthe existence of sucha contract,I adduce evidence that self-identified straights and self- identifiedgays both deploy the same three strategies of bisexualerasure: classerasure, individual erasure, and delegitimation. In PartIII, I describethe investments that both self-identified straights andself-identified gays have in bisexual erasure. These are: (1) an interest instabilizing sexual orientation; (2) aninterest inretaining sex as a dominant metricof differentiation; and(3) an interestin defendingnorms of monogamy.I disaggregateeach interest into its three components: (1) thecompo- nentshared by bothstraights and gays;(2) thecomponent held only by straights;and (3) thecomponent held only by gays. Thefirst investment monosexuals have in bisexual erasure is an interest in stabilizingsexual orientation. The component of that interest shared by bothstraights and gays is an interestin knowingone's place in the social order:both straights and gays value this knowledge because it relieves them of theanxiety of identity interrogation. Straights have a morespecific interest inensuring the stability of heterosexuality because that identity is privileged. Less intuitively,gays also have a specificinterest in guardingthe stability of homosexuality,insofar as theyview that stability as thepredicate for the "immutabilitydefense" or foreffective political mobilization. Bisexuality threatensall ofthese interests because it precludesboth straights and gays from"proving" that they are either straight or gay. Thisis becausestraights (forexample) can only prove that they are straight by adducing evidence of cross-sexdesire. (They cannot adduce evidence of the absence of same-sex desire,as itis impossibleto prove a negative.)But this means that straights cannever definitively prove that they are straight ina worldin which bisexuals exist,as theindividual who adducescross-sex desire could be either straightor bisexual, and there is no definitiveway to arbitratebetween those twopossibilities. Bisexuality is thus threatening to all monosexualsbecause itmakes it impossible to prove a monosexualidentity. The secondinterest monosexuals have in bisexualerasure is an interest in retainingthe importanceof sex as a distinguishingtrait in society. Straightsand gays have a sharedinvestment inthis because to be straightor to be gayis to discriminateerotically on thebasis of sex. Straightshave a specificinterest in preservingthe importance of sexbecause sex norms are currentlyread through a heterosexual matrix: to be a manor a womanin contemporaryAmerican society is in partdefined by one's sexualattractive- nessto the opposite sex. Gaysalso havea particularinterest in sexdistinc- tions,as homosexualityis often viewed as a wayto engagein completesex separatism-thatis, as a meansof creating single-sex communities that are bondedtogether erotically as well as sociallyand politically. Bisexuality

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 363 endangersall ofthese interests because it posits a worldin whichsex need not(or should not) matter as muchas monosexualswant it to matter. Indeed, bisexualsand asexuals are the only sexual orientation groups that have at leastthe capacity not to discriminate onthe basis of sex in any aspect of their lives. Thefinal interest that monosexuals have in bisexual erasure is an interest in defendingnorms of monogamy. Both straights and gays share this interest,as thedominant ethic of contemporary American society favors dyadic relationships.Straights may have a particularinterest in thisinsofar as the formof nonmonogamy associated with bisexuals has been connected to HIV infection,with bisexual "promiscuity" acting as a bridge(phantasmatically if notactually) between the "infected" gay population and the"uninfected" straightpopulation. Gays may have a particularinterest in monogamy insofaras theyseek to assimilateinto "mainstream" society. Bisexuality threatensall ofthese interests because bisexuals are often perceived to be "intrinsically"nonmonogamous. Thus,along at leastthree different axes, both gays and straightshave distinctbut overlapping interests that are threatened by theconcept of bisexuality.It is thusunsurprising that both of these sexual orientation groups colludein bisexual erasure. In PartIV, I examinehow self-identified bisexuals both capitulate to and resistthe epistemic contract. I first show how the contract retards the degree to whichindividuals can articulatea bisexual identity. I thenturn to the waysin whichindividuals articulating such an identityhave begun to dis- solvethe contract. I hypothesizethat this visibility will only increase over time,and I endorsethis trend. In PartV, I lookat how bisexual invisibility and visibility affect practical outcomesin the legal realm. Taking the case of same-sex sexual harassment, I notethat bisexuals have remained largely invisible in thatjurisprudence, althoughthey are more visible in this doctrinal area than in others. I thenask how thatjurisprudence might be transformedif bisexuals were rendered morevisible.

I. THE ERASUREOF BISEXUALS Thatbisexuals are being erased will appear self-evident to many. Yet numerousconversations I have had on the subject of bisexuality have called intoquestion the nature of the proposition (e.g., under what definition ofbi- sexualityis erasureoccurring?), ifnot its veracity (e.g., under a particular definition,are bisexuals really being erased?). Answering these questions is useful,not only to assuage the doubts of skeptics, but also to ensure that non- skepticsare talking about the same thing when they agree that bisexuals are beingerased. I thereforetreat bisexual erasure as a controversialproposition

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 364 STANFORDLAWREVIEW [Vol. 52:353

thatmust be provenrather than assumed. That proposition consists of two claims:first, that bisexuals are invisible; second, that this invisibility stems notfrom nonexistence, but rather from erasure.

A. BisexualInvisibility To showthat bisexuals are invisible, I must specify and defend a definitionof bisexual invisibility (as opposedto a definitionofbisexuality),26 and thenproduce evidence that bisexuals are invisible under this definition. De- finingbisexual invisibility is somewhat difficult, as bisexuality exists at the intersectionof many different kinds of socialinvisibility. Even when nar- rowlydefined, however, bisexual invisibility is easy to prove.

1. Bisexualinvisibility defined. Thereare at least three causes of bisexual invisibility, and it will be usefulto speakof each as creatinga different kind of invisibility.27 The three invisibilitiescan be seen as nestedwithin each other: the firstaffects straights,gays and bisexuals; the second affects only gays and bisexuals; and thethird affects only bisexuals. In thisarticle, I focus on the last kind of invisibility. Thefirst kind of invisibility arises from a generalurge to keepall sex- ualitiesinvisible, which leads to theinvisibility not only of bisexuals,but also of homosexuals,and evenof heterosexuals.As David Halperinhas noted,sexuality discourse has always been haunted by "the ancient and per- sistentspecter of sexual despecification."28 Despite recent efforts to demys-

26. Itmay seem strange that I define"bisexual invisibility" before defining "bisexuality." See notes71-109 infra and accompanying text. The order arises from the fact that I believethat bisexuals areinvisible under any definition of bisexuality,such that I neednot specify the definition of bisexualityindemonstrating bisexual invisibility. 27. In contrastto the way I haveused the words in previous work, "invisibility" and "visibility"in thisarticle refer to socialrather than corporeal (in)visibility. See KenjiYoshino, Assimila- tionistBias in Equal Protection:The VisibilityPresumption and theCase of "Don'tAsk, Don't Tell,"108 YALE L.J.485, 497-98 & n.44(1998) (using the word "invisibility" throughout torefer to "corporealinvisibility"). Corporeal visibility refers to "the perceptibility oftraits such as skincolor thatmanifest themselves on thephysical body in a relativelypermanent and recognizable way." Id. at497. Socialvisibility, incontrast, "designates the perceptibility ofnonphysical traits." Id. Social visibilityincludes "'declarative visibility,' which arises when a corporeallyinvisible trait is made visiblethrough speech," id. at497 n.44;"[p]olitical visibility, [which] 'arises when a personclaims groupmembership as a centraland constitutivefeature of heridentity,"' id. (quotingKathryn Abrams,The Supreme Court, Visibility, and the "Politics of Presence," 50 VAND.L. REv.411, 414 (1997)); and"programmatic visibility," which "'arises from group members' efforts to connect their group-basedidentities with a particularpolitical interest or program,"' id. (quotingAbrams, supra, at 414). Forthe most part, the term "visibility" in thisarticle will refer to thesubset of social visibilitythat is politicalvisibility. 28. DAVID M. HALPERIN, SAINT FOUCAULT: TOWARDS A GAY HAGIOGRAPHY 65 (1995).

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 365

tifysex, "the erotic still arouses acute moral anxiety and confusion."29 Thus itis stilltrue that, as RichardPosner has noted, "[a]nyone in our society who wantsto write about sex ... hadbetter explain what the source of his interest in thesubject is."30 In thelaw, squeamishness about sexuality causes even heterosexualsexual subjects to be shroudedin euphemism.For instance, a numberof state statutes criminalizing sodomy denote heterosexual as well as homosexualsodomy no morespecifically than as a "crimeagainst nature."31 As onelegal encyclopedia puts it, the terms of these statutes "reflect the leg- islators'reluctance to setout in detail the elements of sodomy because of its loathsomenature."32 That these chronically vague provisions33 have been upheldagainst void-for-vagueness challenges34 suggests that the judiciary sharesthe legislativesqueamishness about sexuality. Judicial reticence aboutnaming deviant heterosexual sexual practices can also be seenin the contextsof obscenity regulation35 and sexual harassment.36

29. JEFFREYWEEKS, SEXUALITY1 (1986); see also JEFFREYWEEKS, SEXUALITYAND ITS DISCONTENTS:MEANINGS, MYTHS & MODERN SEXUALITIES44 (1985) ("Sexuality is a fertile sourceof moral panic, arousing intimate questions about personal identity, and touching on crucial socialboundaries.... This is whatmakes sex a particularsite of ethical and political concern-and offear and loathing."). 30. RICHARDA. POSNER,SEX ANDREASON 1 (1992). 31. See, e.g.,ARIz. REV.STAT. ? 13-1411(1999) ("A personwho knowingly and without forcecommits the infamous crime against nature with an adultis guiltyof a class 3 misde- meanor.");IDAHO CODE ? 18-6605(1999) ("Everyperson who is guiltyof theinfamous crime againstnature, committed with mankind or withany animal, is punishableby imprisonment in the stateprison not less thanfive years."); MASS. GEN. LAWS ANN. ch. 272, ? 34 (West1992) ("Who- evercommits the abominable and detestable crime against nature, either with mankind or witha beast,shall be punishedby imprisonmentin the state prison for not more than twenty years."); MICH.COMP. LAWS ANN. ? 750.158(West 1999) ("Any person who shall commit the abominable and detestablecrime against nature either with mankind or withany animal shall be guiltyof a felony...."); MISS.CODE ANN. ? 97-29-59(1998) ("Everyperson who shall be convictedof the detestableand abominable crime against nature committed with mankind or witha beast,shall be punishedby imprisonmentin the penitentiary for a termof not more than ten years."); N.C. GEN. STAT.? 14-177(1999) ("If anyperson shall commit the crime against nature, with mankind or beast,he shallbe punishedas a Class I felon.");OKLA. STAT. ANN. tit. 21, ? 886 (West 1999) ("Anyperson who is guiltyof the detestable and abominable crime against nature, committed with mankindor with a beast,shall be guiltyof a felony.").Unlike other "crime against nature" statutes, noneof the above statutes goes on todefine what constitutes a "crime against nature." 32. 70A AM.JUR. 2D Sodomy? 3 (1987)(authored by Lonnie E. Griffith,Jr.). 33. For a discussionof thevagueness of thesestatutes, see RUTHANNROBSON, LESBIAN (OUT)LAw:SURVIVAL UNDER THE RULE OF LAW 47-57 (1992). 34. See, e.g.,Rose v. Locke,423 U.S. 48 (1975) (percuriam) (upholding Tennessee "crime againstnature" statute against vagueness challenge); Wainwright v. Stone, 414 U.S. 21 (1973) (per curiam)(same for Florida statute); Hogan v. State,441 P.2d 620 (Nev. 1968) (same forNevada statute);Wamer v. State,489 P.2d526 (Okla.Crim. App. 1971) (same for Oklahoma statute). 35. See A BookNamed "John Cleland's Memoirs of a Womanof Pleasure" v. Massachusetts, 383 U.S. 413,441 (1966) (Clark,J., dissenting) (noting that "quotations from typical episodes [of an allegedlyobscene book] would so debaseour Reports that I willnot follow that course"). 36. See Oncalev. SundownerOffshore Servs., Inc., 523 U.S. 75, 77 (1998) (omittinggraphic factsin sexual harassment case inthe "interest of both brevity and dignity").

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 366 STANFORDLAWREVIEW [Vol. 52:353

Thesecond kind of invisibility relates specifically to same-sexdesire. It can be seenin thetreatment of same-sexdesire as unspeakable.As Eve Sedgwicknotes, the nonmedical terms for same-sex desire in theChristian traditionincluded "'that sin which should be neithernamed nor committed,' the'detestable and abominablesin, amongst Christians not to be named,' ...'things fearfulto name,' . . . [and] 'the love thatdare not speak its name.... "'37 In thelaw, the code of silence about same-sex desire can be seenin so-called"no promo homo" statutes,38 which prohibit public educa- tionlikely to promote homosexuality.39 It can also be seenin the military's current"don't ask, don't tell" policy, in whicharticulations ofsame-sex de- sireare chilled.40 Unlike the first kind of invisibility,this invisibility does notpertain to cross-sexdesire. Thus, heterosexuality may be promotedun- derthe "no promo homo" statutes,41 and articulated under "don't ask, don't tell."42In bothcases, however, bisexuals are made as invisibleas homosexuals.This is because"unspeakable" same-sex desire is alsoa component ofbisexuality. Despite their name, the "no promo homo" statutes also pro- hibitthe promotion ofbisexuality, either by explicitly encompassing bisexu-

37. SEDGWICK,supra note 4, at202-03 (footnotes omitted). 38. See, e.g.,Nan D. Hunter,Identity, Speech, and Equality,79 VA. L. REv. 1695,1702-06 (1993) (describingthe genesis of these statutes). 39. See, e.g.,ARIZ. REV. STAT. ? 15-716(c)(1998) ("No districtshall include in itscourse of studyinstruction which: (1) Promotesa homosexuallife-style. (2) Portrayshomosexuality as a positivealternative life-style. (3) Suggeststhat some methods of sex aresafe methods of homosexual sex."); CONN.GEN. STATE. ? 46a-81r(1997) ("Nothingin [theenumerated sections] shall be deemedor construed... to authorizethe promotion of homosexualityor bisexualityin education institutions...."); LA. REv. STAT.ANN. ?17-281(A)(3) (West 1999) ("No sex educationcourse offeredin thepublic schools of the state shall utilize any sexually explicit materials depicting male or femalehomosexual activity."); MiNN. STAT. ? 363.021(1999) ("Nothingin thischapter shall be construedto ... authorizeor permitthe promotion of homosexualityor bisexualityin education institutionsor require the teaching in educationinstitutions of homosexuality or bisexualityas an acceptablelifestyle . . . ); TEX.HEALTH & SAFETYCODE ANN. ? 85.007(b)(West 1992) ("The materialsin theeducation programs intended for persons younger than 18 yearsof age must... statethat homosexual conduct is notan acceptablelifestyle . . ."). 40. See 10 U.S.C. ? 654(b)(2)(1998) (notingthat, in theabsence of furtherfindings, a mem- berstating "that he or she is a homosexualor bisexual"will be separatedfrom the armed forces). BecauseI alluderepeatedly to thepolicy in thisarticle, I briefly describe it here. The policy imme- diatelypredating "don't ask, don't tell" contained a categoricalexclusion of homosexuals,noting that"[h]omosexuality [was] incompatible with military service." 32 C.F.R. pt.41 app. A (1998) ("EnlistedAdministrative Separations: Standards and Procedures").After winning his firstpresi- dentialelection, then President-elect Bill Clintonmaintained that he wouldlift the ban on gaysin themilitary and permit gays to serve openly. See JANETE. HALLEY,DON'T: A READ-ER'SGUIDE TO THE MILITARY'SANTI-GAY POLICY 20 (1999). Thisengendered massive resistance on thepart ofthe military establishment. See id. at 21. Thecurrent policy, colloquially known as "don'task, don'ttell," was framedas a compromise.See WilliamN. Eskridge,Jr. & PhilipP. Frickey,The SupremeCourt, 1993 Term-Foreword:Law as Equilibrium,108 HARV.L. REv. 27, 92 (1994). Underthis policy, gays can servein themilitary, but they may not serve openly without risking separation.See 10U.S.C. ? 654(b)(2)(1998). 41. See note39 supra. 42. See note40 supra.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 367 alitywithin their interdictions orby prohibiting homosexual acts.43 And the "don'task, don't tell" policy explicitly muzzles bisexuals as wellas homosexuals-the"statements" portion of the statute asserts that the presumption ofhomosexual conduct is triggeredif"the member has stated that he orshe is a homosexualor bisexual."44 The finalkind of invisibilitypertains only to bisexuality.Bisexual in- visibilitymanifests itself in the studied omission of bisexuality indiscussions ofsexual orientation. One example among many45 is an anthologyof essays on sexualorientation entitled Homosexuality/Heterosexuality.46 Thiselision carriesover into the law, where discussions of sexual orientation almost in- variablyprivilege the straight/gay binary. One example among many47 is the recentSupreme Court opinion in Romer v. Evans.48In thatcase, the Court foundthat Amendment 2 of Colorado's state constitution violated the federal EqualProtection Clause. Amendment2 stated that there would be "No Pro- tectedStatus Based on Homosexual, Lesbian or Bisexual Orientation."49 The Amendmentthus took the class of bisexuals seriously, probably because the municipalordinances itoverrode explicitly protected bisexuals.50 The Court, however,subsumed bisexuals into the homosexual category, noting that it wouldrefer to the "named class" protected by the ordinances "as homosexual personsor gays and lesbians."51 Thus, the only references tobisexuals in the opinionoccur in thequoted language of the ordinances and Amendment 2; whenthe Court speaks for itself, it speakssolely about homosexual persons.52 BecauseI wishto consider how bisexuals qua bisexuals are being erased, I focuson the third kind of invisibility. Thus, bisexual invisibility is defined forthe purposes of this article as thatsocial invisibility that affects only bisexuals.In orderto determine how significant that invisibility is,I mustiso- lateits effects from the effects of theother kinds of invisibility.I can ac- complishthis isolation by comparing the relative visibility of bisexuality and

43. See note39 supra. 44. 10U.S.C. ? 654(b)(2)(1998) (emphasis added). 45. Forother examples, see notes54-70 infra and accompanying text. 46. HOMOSEXUALITY/HETEROSEXUALITY: CONCEPTS OF SEXUAL ORIENTATION (David P. McWhirter,Stephanie A. Sanders& JuneMachover Reinisch eds., 1990). 47. Forother examples, see textaccompanying notes 503-534 infra. 48. 517 U.S. 620 (1996). 49. Id. at624 (quotingCOLO. CONST.,art. II ? 306 (1992)). 50. See id. (citingDEN. REV. MUN. CODE, art.IV, ?? 28-91to 28-116(1991); ASPENMUN. CODE ? 13-98(1977); BOULDER REV. CODE ?? 12-1-1to 12-1-11(1987)). 51. Id. 52. Thisdoes not indicate that the Court was intendingto excludebisexuals from its protec- tions,but rather that the Court permitted the homosexual category to absorbbisexuals. Yet this absorption,ofcourse, is itselfan exampleof bisexual invisibility.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 368 STANFORDLA WREVIEW [Vol. 52:353

homosexuality.53To demonstrate that bisexuality is invisible under my definition,I must demonstrate that bisexuality is less socially visible than homosexuality.

2. Evidenceof bisexual invisibility.

Sucha demonstrationis easily made. On-linesearches for the words "homosexuality"and "bisexuality" inmainstream newspapers, newsmagazi- nes,and academic abstracts reveal a strikingdiscrepancy inthe incidence of thetwo terms. In theperiod from January 1, 1990to November30, 1999, theLos Angeles Times had 2790 documents mentioning "homosexuality" and 121documents mentioning "bisexuality";54 USA Todayhad 1768 documents mentioning"homosexuality" and twenty-ninedocuments mentioning "bisexuality";55and The WallStreet Journal had 396 documentsmentioning "homosexuality"and ninedocuments mentioning "bisexuality."56 In the sametime period, Time magazine had 240 documentsmentioning "homosexuality"and fifteen documents mentioning "bisexuality";57 US. Newsand WorldReport had 120 documentsmentioning "homosexuality" and three documentsmentioning "bisexuality";58 and The New Republichad 144 documentsmentioning "homosexuality" and threedocuments mentioning "bisexuality."59While I expectedmuch less of a discrepancyin moving frompopular to academicsources, this proved not to be thecase. In the sametime period, the Social Sciences Abstract Database on Wilson Web had 1122documents mentioning "homosexuality" and eighty-seven documents mentioning"bisexuality";60 theGeneral Sciences Abstracts had 221 documentsmentioning "homosexuality" and six documentsmentioning "bisexuality";;61the Humanities Abstracts had 962 documentsmentioning "homosexuality"and twenty-six documents mentioning "bisexuality."62 Thediscrepancy between the relative visibility of homosexuality and bisexualitycan be describedsociologically as well as statistically.Robin Ochs

53. A comparisonof bisexuality and heterosexuality would not isolate this kind of invisibility becausebisexuality could be moreinvisible than heterosexuality due to thesecond kind of invisibility. 54. Searchof WESTLAW, ALLNEWS Library, LAT File(Mar. 22, 1999). 55. Searchof WESTLAW, ALLNEWS Library, USATD File(Mar. 22, 1999). 56. Searchof WESTLAW, ALLNEWS Library, WSJ File (Mar. 22, 1999). 57. Searchof WESTLAW, ALLNEWS Library, TIMEMAG File (Mar. 22, 1999). 58. Searchof WESTLAW, ALLNEWS Library, USNWR File (Mar. 22, 1999). 59. Searchof WESTLAW, ALLNEWS Library, NEWREPUB File (Mar. 22, 1999). 60. Searchof GeneralScience Abstracts Database, Wilson Web (Mar. 18, 1999) . 61. Searchof GeneralScience Abstracts Database, Wilson Web (Mar. 18, 1999) . 62. Search of HumanitiesAbstracts Database, Wilson Web (Mar. 18, 1999) .

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 369

hasargued that bisexuals are invisible not only relative to straights,but also relativeto gays.63She points out that while we sometimessuspend the gen- eralpresumption that all individualsare straight,the presumption that re- placesit is thatall individualswithin that context are gay.64 Thus, "[i]n most families,for example, members are presumed to be heterosexual;conversely, ata women'sbar all thewomen present are presumed lesbians."65 There are fewcontexts, however, in whichan individualis presumedbisexual. In a similarvein, Marjorie Garber has observed that bisexuals have few recognizable symbolsof their identity66-the pink and blue "biangle" is one of the fewsymbols specifically denoting bisexuality67 and is muchless culturally visiblethan the gay symbols68 ofthe pink triangle69 orthe rainbow.70 Thus, evenwhen the heterosexual presumption that all individualsare straightis suspended,itis replacedby the monosexual presumption that all individuals arestraight orgay. Itis understandablethat the heterosexual presumption has received more theoreticalattention than the monosexual presumption. The erosionof the presumptionthat all individualsare heterosexual was obviously necessary to identifythe fallback presumption that all individualsare heterosexual orho- mosexual.The evidence above demonstrates, however, that the monosexual presumptionis powerful, and urgently requires such attention. I willtherefore take it to be relativelyuncontroversial that bisexuals, howsoeverdefined, are less socially visible than homosexuals in contempo- raryAmerican culture. What remains to be exploredis thecause of this invisibility.

63. See Robin Ochs, Biphobia:It Goes More than Two Ways,in BISExuALITY: THE PSYCHOLOGYAND POLITICSOF AN INVISIBLEMiNoRITY 217, 225 (Beth A. Firesteined., 1996) [hereinafterINVISIBLE MINORITY]. 64. See id. 65. Id. 66. See GARBER,supra note 11, at 25. 67. See id. (notingthat "lavender and blue 'biangles' seem to go unrecognized"). 68. It is truethat these "gay" symbolshave been appropriatedfor multiple other purposes. See id. ("Pink triangles. . . have been takenover as 'the provinceof everyliberal-leaning Clinton supporter."'). But the inauguralgay meaningsof thesesymbols still remain intact and widely understood. 69. See, e.g., Gays Revise Nazi Symbol,S.F. CHRON.,June 29, 1992, at D5 (notinghigh visibilityof pink trianglesduring Gay Pride Month); Deb Price, ConcentrationCamps' Pink Triangle Recalled at Holocaust Museum,STAR TRIB., May 5, 1993, at 3E (notingthat "[t]he pink triangleis sportedaround the world . . . as a signof gay pride"). 70. See, e.g., ThomasJ. Brady,How Some Gays Choose to Show the Colors,PHILA. INQUIRER,Nov. 19, 1997, at DI (notingrising popularity of rainbowas symbolof homosexuality); Deb Price,Rainbow Flag is Symbolof a UnitedGay People, STAR TRIB., Apr. 19, 1995, at 4E (de- scribingutility of rainbowsymbol in recognizinghomosexuals).

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 370 STANFORDLAWREVIEW [Vol. 52:353

B. BisexualErasure

In discussingthis project with peers, I was struckby how many of them agreedthat bisexuals were invisible relative to homosexuals without believ- ingthat bisexuals were being erased. In theirview, the scant visibility of bisexualsrelative to homosexualsdid not signify their "erasure," but rather accuratelyreflected their smaller relative numbers. Call thisthe "relative nonexistencethesis." The merit of this position depends on thenumbers of bisexualsand homosexuals inthe population. Those numbers in turn depend on thedefinitions of "bisexual"and "homosexual." I thus provide and de- fenda provisionaldefinition of"bisexuality" (and, by doing so, provide and defenddefinitions of"homosexuality" and "heterosexuality"). I then apply thisdefinition to a numberof major sexuality studies to compare the relative incidencesof bisexuality and homosexuality.

1. Bisexualitydefined.

I discussthe definition ofbisexuality with some care because the validity ofmy thesis depends up onthis definition. Whether bisexual invisibility de- notesbisexual erasure turns on howexpansively or narrowlybisexuality is defined.To adoptwhatever definition ofbisexuality undergirds the conven- tionalwisdom that "' [e]veryoneis bisexual,'`71for example, is to demonstratebisexual erasure at the moment of definition. Similarly, toadopt what- everdefinition of bisexualityundergirds the countervailing conventional wisdomthat "' [t]hereis no suchthing as bisexuality,"'72is simultaneously to demonstratebisexual nonerasure. Neitherdefinition is particularly satisfactory. The universalizing definitionof bisexualitymaximizes the number of personserased, but only by droppingthe threshold ofbisexual ascription so lowas arguablyto empty the termof diacriticalforce.73 The eradicatingdefinition of bisexualitymini-

71. GARBER, supranote 11, at 16 (quotingconventional wisdom); see also PAULA C. RUST, BISEXUALITY AND THE CHALLENGE TO LESBIAN POLITICS: SEX, LOYALTY, AND REVOLUTION 1 (1995) [hereinafterRUST, CHALLENGE] (quotingview that "[e]veryone is inherentlybisexual"). 72. GARBER, supra note 11, at 16 (quotingconventional wisdom); see also RUST, CHALLENGE,supra note 71, at 1 (quotingview that bisexuality "does not exist"). 73. Thiscan be seenin theintuitive distance between the statement "Everyone is bisexual" andthe statement "I am bisexual"-thatis, in theability of individualsto say theformer without concedingthe latter. At the level of formal logic, this makes no sense. If I ampart of "everyone," and"everyone is bisexual,"then I, too,must be bisexual.At thelevel of praxis, however, it is in- telligiblebecause the word "bisexual" in thefirst sentence is seento havea sufficientlydilute sig- nificance(precisely because it is spreadso thinlyacross the mass of humanity) as tohave no consequence. In thesecond sentence, the word "bisexual" is seento havea sufficientlypotent signifi- cance(precisely because it refers to some, but not others, in thehuman population) as tohave serious ramifications.Put in Sedgwickianterms, this is thedifference between a universalizingdis- courseof bisexuality (in whichbisexuality is a latenttrait in all ofus) anda minoritizingdiscourse of bisexuality(in whichbisexuality is a traitthat some, but not others, possess). See SEDGWICK,

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 371 mizesthe number of persons erased but only by raising the threshold of defi- nitionto dizzyingheights.74 Eliminating these two endpoints, however, still leavesa vastspectrum ofchoice. In stakingout a definitionalposition, two issues must be defensiblyre- solved. Thefirst is theaxis or axesalong which sexual orientation will be defined.The second is howexpansively or narrowly to definea particular sexualorientation along the chosen axis or axes. Thereare three axes along which sexual orientation is conventionally defined:desire, conduct, and self-identification.75Definitions can relyon a singleaxis, or on a combinationofthe axes.76 The number of bisexuals will varydramatically according to the permutation used. Consider,for example, thistaxonomy of bisexualities constructed bya studyusing a pureconduct- baseddefinition: [1] "DefenseBisexuality" (defending against homosexuality in societies where itis stigmatized),[2] "LatinBisexuality" (the insertive role in certain"Mediter- raneancultures" is notregarded as homosexual,so thatmen who participate in same-sexencounters may consider themselves nonetheless heterosexual), [3] "RitualBisexuality" (as withthe Sambiaof Papua-NewGuinea, in which youngermales fellate older men in orderto ingesttheir "masculinizing" semen, a practicethat is partof a riteof initiation, may continue for years, and is appar- entlyreplaced by exclusiveheterosexuality after marriage), [4] "MarriedBi- sexuality,"[5] "SecondaryHomosexuality" (more frequently called "situational bisexuality"-sexwith same-sex partners in prisonsor othersingle-sex institutions,in publicparks or toilets, or formoney), [6] "Equal Interestin Male and FemalePartners" (so-called true bisexuality), [7] "ExperimentalBisexuality," and[8] "TechnicalBisexuality" (with partners who may be dressedas members ofthe other sex, or have had someform of genderreassignment: transsexuals ormembers of a "thirdsex" in somecultures).77

supranote 4, at 1 (usingthese terms in discussingthe definitions ofhomosexuality and heterosexuality). 74. Indeed,there is a way in whichthe twopositions converge in theirextremity. Cf GARBER, supranote 1 1, at 105(noting that bisexual invisibility can be "producedas a startlingby- productof omnipresence"). For if "everyone is bisexual"then the definition of bisexualitymust be so anemicas to leaveopen the possibility that "no one is bisexual"in a less anemicsense. Con- versely,if "no oneis bisexual"then the definition ofbisexuality must be so robustas toleave open thepossibility that "everyone is bisexual"in someless robustsense. See Michaeldu Plessis,Bla- tantlyBisexual; or, UnthinkingQueer Theory, in REPRESENTING BISEXUALITIES 19,30 (DonaldE. Hall & MariaPramaggiore eds., 1996) (noting that "[i]f everyone is bisexual,'bisexuality' can no longerbe a specificor pertinent feature"). 75. See EDWARD 0. LAUMANN, JoHN H. GAGNON, ROBERT T. MICHAEL & STUART MICHAELS, THE SOCIAL ORGANIZATION OF SEXUALITY: SEXUAL PRACTICES IN THE UNITED STATES 290 (1994) [hereinafterLAUMANN ET AL.] (notingthat "homosexuality" can referto "same-genderbehavior, desire, self-definition, or identification or some combination of theseele- ments"). 76. See id. 77. GARBER, supranote 11, at 30 (numberingadded).

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 372 STANFORDLAWREVIEW [Vol.52:353

This arrayappears so vertiginouslydiverse that it mightfairly be asked whetherthese individuals should all be characterizedbythe same term.78 In actuality,however, this array gives an oversimplifiedpicture of bisexuality, insofaras itonly uses one axis of definition. Forif a differentaxis were used, many of the individuals who fall into manyof the above categories would not be characterizedas bisexuals, and viceversa. Thus, if we reliedon desirealone, no "defensebisexuals" would be consideredbisexual, given that they only profess cross-sex desire to mitigatethe stigma of possessing same-sex desire.79 Some "ritual bisexuals" and "situationalbisexuals" might also fall out of the desire-based category. This is becausethese individuals could be engagedin same-sexconduct for rea- sonsother than sexual desire (such as initiationinto adult status in the case of theritual bisexual8O orfinancial remuneration inthe case ofthe situationally bisexualsex worker8l). At the same time, however, we mightexpect many individualswho would not be characterizedas bisexuals under a conduct- baseddefinition to be includedunder a desire-baseddefinition. This would encompassall individualswho nursed unacted same-sex desires. Similarly,ifwe definedbisexuality according to self-identification alone, "defensebisexuality" would probably be theonly category that was not seri- ouslydiminished. "Latin bisexuals" engage in sexual conduct with both men andwomen but self-identify as heterosexual rather than as bisexual.82And a significantportion of themen who engage in extramaritalsame-sex conduct,83or who viewtheir same-sex conduct as "experimental"84or "situational,"85are likely to self-identify as straight. Becausethe populations captured by the different axes are so different, thechoice of axis (or combination ofaxes) is crucial.The choice cannot be madeacontextually, butwill depend on the purpose for which the definition is used. Forexample, the study that generated the taxonomy described above useda conduct-baseddefinition because its purpose was to examinethe role

78. Cf id. at 31 (speculatingabout theusefulness of usingthe single term "bisexuality" to de- scribea similarlydiverse array of activitiesand fantasies). 79. See RonaldC. Fox,Bisexuality in Perspective:A Reviewof Theoryand Research,in INVISIBLEMINoRITY, supra note63, at 3, 22-23. 80. See id. at 23. 81. See FRITZKLEIN, THE BISEXUALOPTION 19 (1978) ("Male prostituteswho allow themselves to be fellatedfor money often do not considerthis a homosexualact, or considerthemselves bisexual."). 82. See Fox, supra note79, at 23. 83. See JohnJ. Brownfain, A Study of the Married Bisexual Male: Paradoxand Resolution, in BISEXUALITIES:THEORY AND RESEARCH173, 173 (Fritz Klein & TimothyJ. Wolf eds., 1985) (notingevidence that"individuals who choose to live withina conventionalmarriage while simul- taneouslyhaving significantlove-sex relationswith members of theirown sex .... are not rare, onlyrarely identified"). 84. See KLEIN,supra note81, at 18. 85. Seeid.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 373 of bisexualityin the transmissionof the HumanImmunodeficiency Virus ("HIV").86 This definitionmakes obvious sense because HIV is spread throughsexual conductrather than through desire or self-identification.87 However,neither this nor any other axis of definitionwill be validacross all contexts. So whichaxis or combinationof axes is bestsuited for our purposes? I believe thatthe answeris a pure desire-baseddefinition. This may seem self-servingas the desire-baseddefinition is likelyto yieldmore bisexuals thana conduct-basedor self-identification-baseddefinition,88 thereby making theclaim of erasure easier to sustain. But the purelydesire-based definition of bisexualitycan be defended againstthis criticism. One can framethat defense by (1) askingwho is bisexual undera desire-baseddefinition who wouldnot be bisexualunder a conduct-basedor self-identification-baseddefinition, and then (2) asking whetherit is fairto call thoseindividuals bisexual for the purposes of deter- miningerasure. The analogyto homosexualityis useful. For whenwe ask thosequestions about homosexuality, the answers are thatthe desire-based definitionincludes those who have unactedsame-sex desires and thatit is clearlydefensible to call thoseindividuals homosexual. I assumethat these individualsdo notact outtheir desires through conduct or self-identification because of the stigmaof homosexuality,and I assumethat the inabilityto convertdesire to conductor self-identificationis consequential to thatindividual(at leastinsofar as thedesire is morethan incidental).89 I wouldcall theindividual who possessesunexpressed same-sex desire a closetedhomo- sexual ratherthan a heterosexual.Indeed, I would view the exclusionof such individualsfrom the definitionof homosexualsfor the purposesof demonstratingerasure to be ironic,as it wouldpermit erasure to controlthat definition,to leterasure erase itself. And I wouldnot expect much resistance to theseanalytic moves. I believethat these moves will be muchmore strongly resisted when ap- plied to bisexuality,but thatthey are nonethelessvalid. The desire-based definitionwill includeindividuals who desireboth sexes butwho do notex- pressthat dual desireeither through conduct or throughspeech. The resistance ariseswhen this suppression is characterizedas consequential.It is harderto arguethat bisexuals are givingup an importantaspect of their

86. See GARBER,supra note 11, at 30. 87. See JosephP. Stokes,Kittiwut Taywaditep, Peter Vanable & DanielJ. McKirnan, Bisex- ual Men,Sexual Behavior, and HIV/AIDS, in INVISIBLE MINORITY, supra note 63, at 149,150-51. 88. See, e.g.,COLKER, HYBRID, supra note 9, at 15 (notingthat "[o]nly about 1 percentof the adultpopulation identifies as bisexual,"but that "nearly 4 percentacknowledge that they are attractedto people of both sexes") (citing Trip Gabriel, A NewGeneration Seems Ready to GiveBi- sexualitya Place inthe Spectrum, N.Y. TIMES,June 12, 1995, at CIO). 89. I discussthe restriction that the desire be morethan incidental below. See notes99-108 infraand accompanying text.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 374 STANFORDLA WRE VIEW [Vol. 52:353

identitysince they are not being denied all satisfactoryinterpersonal sexual expression(as a homosexualin thesame position arguably is) butmerely sexualexpression with one sex.90In fact,the stereotype of bisexualsas "greedy"or "promiscuous"implies that bisexuals are asking for more than theirdue, such that they should be forcedto giveup sexualexpression with one of thetwo sexes.91Yet forthe purposes of erasure,the inquiry is whetherthey are forgoingsomething rather than whether they should be for- goingit. Andthe answer to that positive, rather than normative, question is clearlyaffirmative. To theextent that bisexuals are not permitted toexpress theirdual desires, they might fairly characterize themselves as harmed.To excludethe individual who has unexpressed dual desires again permits era- sureto erase itself. For that reason, I believe that a purelydesire-based definitionof bisexuality is the most appropriate. Evenafter desire is justifiedas thepertinent axis, we mustdetermine howexpansively or narrowlyto definebisexuals based on desire.At one extreme,there is Freud'sview that bisexual desire is a universalcondi- tion92-perhapsthe most prestigious version of the"everyone is bisexual" position.At the other extreme is theview that only "true" or sex-blindbi- sexuals-thosewho have equal desire for men and women93-are bisexual, whichasymptotically approaches the view that "no one is bisexual"depend- ingon how seriously one takes the requirement ofequality.94 Isolating desire as theaxis of definition thus still leaves the full spectrum of bisexual definition-frombisexual universality tobisexual nonexistence-available.

90. See ArthurA. Murphy& JohnP. Ellington,Homosexuality and theLaw: Toleranceand ContainmentII, 97 DICK. L. REv.693, 709-10 (1993) (arguing that existing sodomy laws should be modifiedto exonerate "true homosexuals" but not bisexuals, on theground that bisexuals can more easilymodify their behavior). 91. See notes353-357 infra and accompanying text. 92. See, e.g.,Sigmund Freud, Analysis Terminable and Interminable,in 23 THE STANDARD EDITION OF THE COMPLETEPSYCHOLOGICAL WORKS OF SIGMUNDFREUD 211, 243-44(James Stracheytrans., Hogarth Press 1973) (1937) [hereinafterWORKS OF SIGMUNDFREUD] ("We have cometo learn,however, that every human being is bisexualin thiscase andthat his libidois dis- tributedeither in a manifestor latent fashion, over objects of both sexes."); Sigmund Freud, Civili- zationand Its Discontents, in 21 WORKSOF SIGMUNDFREUD, supra, at 57, 105(1930) ("Manis an animalorganism ... withan unmistakablybisexual disposition."); Sigmund Freud, The Psycho- genesisof a Case ofHomosexuality ina Woman,in 18 WORKSOF SIGMUNDFREUD, supra at 145, 158 (1920) ("In all of us, throughoutlife, the libido normally oscillates between male and female objects...."); id.at 157(noting "the universal bisexuality of human beings"). 93. Whatit meansto have"equal" desire for men and women is of coursea deeplyfraught question.Perhaps this is whytrue bisexuality is framedas "sex-blind"bisexuality-implying that, ifbisexuals do noteven "see" sex,they cannot disburse their desire unequally on thebasis of sex. But,as discussedbelow, the rhetorical formulation of "sex-blindness" is somewhat misleading, as bisexualscan in facttell the difference between men and women. And if the bisexual can differen- tiate,it seemsunlikely that this differentiation willnot affect her desire asymmetrically, given the powerof sex as a diacriticaltrait in our society. See notes314-318 infra and accompanying text. 94. See RUST, CHALLENGE,supra note 71, at 48-49(describing this definition of bisexuality andnoting its restrictiveness).

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 375

HereI choosea narrowview of bisexuality. I do nottake this position becauseI rejectthe view that bisexuality is universal-I remain agnostic as to thispoint. Rather, I adopt this position because assuming universal bi- sexualitywould make my argument too easy (and therefore too easy to dis- miss). I placetwo restrictions onmy desire-based definition ofbisexuality. The firstrestriction pertains to thenature of desire.Desire is a classicfloating signifier,that is, a termwithout a clear referent. Especially in postmodem times,its ambiguity has beenresolved in favorof an expansivedefinition. Thus,under a Foucaultianreading, sexual desire should not be distinguished fromother pleasurable uses of the body.95 And under a Sedgwickianread- ing,desire can incorporate itsputative opposite, as inthe theory that the ho- mophobiaof "homosexualpanic" is in actualityfueled by an occluded, guilty,and surprised desire.96 But if the desire-based definition ofbisexual- ityencompasses the nonsexual pleasure (and sexual displeasure) we feelin thecompany of other men and women, we areall indeedbisexuals. To avoidsuch a dilutionof the bisexual definition, I restrict the definitionof desire to itsmore everyday meaning of "sexual appetite or lust."97 I thustake the categories of the "homosocial" and the "homosexual" (or, alternatively,the "hostile" and the "desirous") to be analyticallydistinct.98 Thesecond restriction pertains to the frequency with which the desire is experienced.To be classifiedas a bisexualunder my definition, an individual's sexualdesire for each sex must be morethan incidental. To understand thisrestriction, itis helpfulto examinethe analogous restrictions commonly placedon homosexuality.If Kinsey is tobe believed,in themiddle of this century,a single same-sex sexual experience was sufficient tobrand an individualas a homosexual.99That cultural consensus, if it everexisted, has

95. See 1 MICHEL FOUCAULT,THE HISTORYOF SEXUALITY:AN INTRODUCTION157-59 (RobertHurley trans., Vintage Books 1990) (criticizing the privileging of sex overother economies of"bodies and pleasure"). 96. See SEDGWICK,supra note 4, at 20. Sedgwickcontends that "[j]udicially, a 'homosexual panic'defense for a person(typically a man) accused of antigay violence implies that his responsi- bilityfor the crime was diminishedby a pathologicalpsychological condition, perhaps brought on byan unwantedsexual advance from the man whom he thenattacked." Id. at 19. Sedgwickgoes on to observethat "the reason why this defense borrows the name of the(formerly rather obscure and little-diagnosed)psychiatric classification 'homosexual panic' is thatit refersto thesupposed uncertaintyabout his own sexual identity of the perpetrator ...." Id. at20. Thus,the word "homosexual"in "homosexualpanic" is aptlyambiguous, representing both the same-sex desire in the victimand the same-sex desire in theperpetrator. The ambiguity represents the transitivity (imag- inedor real) of same-sex desire, which could be describedas thereal cause of the panic. 97. 4 OED, supra note 9, at 522. 98. Thatthese distinctions are madein manyquarters, as in thecase of sexualharassment, willbe demonstratedbelow. See textaccompanying notes 430-534 infra. 99. In 1948,Kinsey stated: Everywhereinour society there is a tendencytoconsider an individual'homosexual' ifhe is knownto have had a singleexperience with another individual ofhis own sex. Underthe law

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 376 STANFORDLAWREVIEW [Vol. 52:353

shiftedin certainfora to onethat holds that adventitious homosexual desire, evenif manifested in explicit same-sex conduct, is notsufficient to strip an individualof heterosexual status. This latter view is perhapsbest seen in the legalsystem-probably the preeminent cultural site in whichpublic institu- tionaldeterminations of sexual orientation are currentlymade. Congres- sionalsupport for this view can be seenin the"don't ask, don't tell" statute.'00The statute generally makes homosexual conduct grounds for exclu- sion fromthe military.101 Yet it also containsan exemptionthrough which an individualwho has engaged in same-sex sexual conduct can remain in the militaryifhe candemonstrate that such conduct is unlikelyto recur.102To theextent that we viewCongress as attemptingto weed out homosexuals, thisso-called "queen-for-a-day" exemptionl03 may be seenas a self-imposed checkon labelingindividuals as gaybased on isolatedsame-sex sexual conduct.The judiciary also uses its interpretive faculty to restrict the definition of"homosexual" in thisway. Judicial"queen-for-a-day" exemptions can be seen in thecontexts of thecivil service,104adoption,105 and marriage.106

an individualmay receive the same penalty for a singlehomosexual experience that he would fora continuousrecord of experiences. In penal and mental institutions a male is likelyto be rated'homosexual' ifhe is discoveredtohave had a singlecontact with another male. In soci- etyat large, a malewho has worked out a highlysuccessful marital adjustment is likely to be rated'homosexual' ifthe community leams about a singlecontact that he has had with another male. KiNSEYET AL.,MALE, supra note 5, at647, 650. 100. 10 U.S.C. ? 654 (1998). 101. The statuteasserts that "[a] memberof thearmed forces shall be separatedfrom the armedforces" if it is found"[t]hat the member has engagedin, attempted to engagein, or solicited anotherto engage in a homosexualact or acts ..." 10 U.S.C. ? 654(b)(1)(1998). 102. Evenif a findingis madethat a servicememberhas engagedin homosexualconduct, the membercan evadeseparation if he or she can show,inter alia, that"such conduct is a departure fromthe member's usual and customary behavior," 10 U.S.C. ? 654(b)(1)(A)(1998); "suchcon- duct,under all thecircumstances, is unlikely to recur,"id. ? 654(b)(1)(B);and "themember does nothave a propensityorintent to engage in homosexual acts," id. ? 654(b)(1)(E). 103. See HALLEY,supra note 40, at 39-48. 104. In Dew v. Halaby,317 F.2d582, 583 (D.C. Cir. 1963),for example, the court accepted thata veteranwas nothomosexual even though he hadhad four homosexual encounters as an ado- lescent.The court quoted the psychiatric evaluation of the appellant William Dew: "I decidedlydo notbelieve [Dew] to have a homosexualpersonality disorder. I believethat theseveral incidents which he gavea historyof having engaged in while age 18 anda fresh- manin collegewere isolated incidents primarily the result of his curiosity. I do notfeel that isolatedepisodes of homosexual behavior can be consideredabnormal or perverted, but rather a processof normal sexual investigation and curiosity. There is considerableevidence in the psychiatricliterature that this allegation is true." Id. at583 n.3(quoting testimony ofpsychiatrist). 105. See In re Opinionof theJustices, 530 A.2d21, 24 (N.H. 1987)(excluding from definitionof homosexualin a proposedadoption statute those persons who, for example, had had one homosexualexperience during adolescence, but who thenengaged in exclusivelyheterosexual behavior). 106. See Freitagv. Freitag,242 N.Y.S.2d643, 644 (N.Y. Sup.Ct. 1963)(deciding that man's confessionof homosexual experiences prior to marriage did not mean he had"a truecase ofhomosexuality").

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 377

Theseexemptions can be readas precludingaberrant acts of same-sex desire fromtainting an individual with a homosexualidentity. Sincebisexuality is also a stigmatizedidentity, we wouldexpect that isolatedhomoerotic conduct would not earn an individuala bisexual ascription. The evidencealready adduced supports this, as it suggeststhat indi- vidualsmaintain heterosexual (rather than bisexual or homosexual) identities despiteisolated homosexual conduct.107 My definitionof bisexualitythus also excludesindividuals who only have incidental or occasionalsame-sex desire.108In theinterest of symmetry,my definition further excludes individualswho only have incidental or occasional cross-sex desire. This elimi- natesfrom the bisexual category those who, for example, experimented with homosexualactivity in their young adulthood but then engaged in completely heterosexualactivity for the rest of their adult lives (or vice versa). To summarize,my analysis adopts a restricteddesire-based definition of bisexuality.I first chose and defended desire as theaxis of definition. I then placedtwo restrictions onthe desire-based definition: (1) I defineddesire as sexualappetite; and (2) I requiredmore than incidental desire for both sexes beforeclassifying an individualas bisexual.The definition is not presented as perfect,even for these purposes, but simply as workable.109Having definedbisexuality inthis way, I nowexamine what the major sexuality studies say aboutthe relative incidence of bisexualityand homosexualityin the population.

2. Thestudies.

Howdo themajor sexuality studies describe the relative numbers of bisexualsand homosexuals in thepopulation? I expected to findmany secon- darysources that set forth the answer to this question. To mysurprise, this provednot to be thecase.110 This might be explainedby a certainreluctance to use theexisting studies for any purpose, given that they suffer from a numberof problems. These include problems with methodology in particular studies,problems with obtaining truthful accounts in any study,and problemswith calibrating orientation even assumingtruthful accounts.

107. See notes100-106 supra and accompanying text. 108. I addressthe question of whatconstitutes "incidental or occasionaldesire" below. See note146 infra and accompanying text. 109. The definition,for example, makes no attemptto calibrate"the erotic appeal of trans- gression,"GARBER, supra note 11, at 29; theintensity of desire,see generallyMichael D. Storms, Theorieso/Sexual Orientation,38 J. PERSONALITY & SOC. PSYCHOL. 783 (1980); or the ways in whichdesire is connectedto other aspects of an individual'sself-conception (such as herpolitics or self-identification),see SEDGWICK, supra note4, at25-26. 110. Thesource that came closest to a systematiccomparison of the relative incidence of bisexualityand homosexuality was Fox,supra note 79, at 16-17. Eventhis source, however, left out manyof the major studies.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 378 STANFORDLAWREVIEW [Vol. 52:353

Theseare seriousissues that anyone deploying the studies should address, and I do so below.",1 The beliefthat these problems are keepingpeople from utilizing the studies,however, is probablyincorrect. The famousten percent figure for homosexuals,for example, arises from the Kinsey studies.112 This suggests thatit is nota genericuse of the studies, but rather specific uses of them, that arebeing suppressed. The starkdiscrepancy between the popularization of theKinsey findings on homosexualityand the lack of popularization ofthe Kinseyfindings on bisexualityl13 suggests that bisexual invisibility provides an alternativeexplanation for the absence of such comparisons. I therefore thoughtit usefulto go throughthe major sexuality studies and to makea systematiccomparison. In evaluatingthe "studies that purport to tellabout sex in America,"114 RobertT. Michaeland his colleagues discussed seven major studies115: the Kinseystudy (which is actuallycomprised of separate studies for men16 and womenl17),the Playboy study,118 the Redbook study,l19 the Hite study,120the Mastersand Johnson study,121 the Janus and Janus study,122 and theLaumann study.123They also attestedto thereliability of twounnamed studies from Franceand England, which are identifiable based on theirother work as the

11 1. See notes 177-186 infraand accompanyingtext. 112. See,e.g., BRUCE BAWER, A PLACEAT THE TABLE: THE GAY INDIVIDUALIN AMERICAN SOCIETY 82 (1993) ("From the appearanceof the Kinsey Reportsin 1948 and 1953 untilvery recently,it was a truismthat about 10 percentof Americansare homosexual."); JenniferGerarda Brown,Competitive Federalism and the Legislative Incentives to Recognize Same-Sex Marriage, 68 S. CAL. L. REV. 745,776 (1995) ("Since 1948, when AlfredC. Kinsey and his associates released their path-breaking study, Sexual Behavior in the Human Male, the oft-citedstatistic is thatten percentof thepopulation is gay."). 113. See MARTIN S. WEINBERG,COLIN J. WILLIAMS & DOUGLAS W. PRYOR, DUAL ATTRACTION:UNDERSTANDING BISEXUALITY 4 (1994) (notingthat Kinsey's findingsabout bisexuality"have been generallyignored"). 114. ROBERT T. MICHAEL, JoHN H. GAGNON, EDWARD 0. LAUMANN & GINA KOLATTA, SEX IN AMERICA: A DEFINITIVE SURVEY 15 (1994). It bears note that this is a companion volume to the Laumann study, see LAUMANN ET AL., supra note 75, written by the same researchers who conducted that study. 115. See MICHAEL ET AL., supra note 1 14, at 15-25. 1 16. K[NSEY ET AL., MALE, supra note 5. 1 17. KINSEY ET AL., FEMALE, supra note 6. 118. The Michael study does not provide a citation for the Playboy study. As there are a number of Playboystudies to which it could refer,I do not attempt a citation here. 119. CAROL TAVRIS & SUSAN SADD, THE REDBOOK REPORT ON FEMALE SEXUALITY (1975). 120. SHERE HITE, THE HITE REPORT: A NATIONWIDE STUDY ON FEMALE SEXUALITY (1976). 121. WILLIAM H. MASTERS & VIRGINIA E. JOHNSON, HOMOSEXUALITY IN PERSPECTIVE (1979). 122. SAMUEL S. JANUS & CYNTHIA L. JANUS, THE JANUS REPORT ON SEXUAL BEHAVIOR (1993). 123. LAUMANN ET AL., supranote 75.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 379

Spirastudy124 and the Wellings study.l25 Other sources corroborate that this listis a fairlycomprehensive rendering ofthe major sexuality studies.126 I narrowedthis list of nine studies in twoways. First,I eliminatedthe studies-theRedbook study,127 theHite study,128 and the Spira study129-that didnot include data sufficient to permit a comparisonof therelative incidencesof bisexuals and homosexuals inthe population. I then eliminated the Playboystudy because I thoughta magazine study was likely to be lessaccu- rateand less credible than the other studies.130 This left five studies that I felt usefullycompared the incidences of bisexuality and homosexuality. Thefive studies defined sexual orientation differently, each using some pernutationof thethree axes of desire,conduct, and self-identification.131 Wherepossible, I adjusted the definition used by the study to conform to my restricteddesire-based definition ofbisexuality. As thestudies used different definitions,and as evenstudies using the same definition used different scales132or measures,133 it is notsurprising that the studies differ widely in

124. See ALFREDSPIRA, NATHALIE BAJOS & THE ACSF GROUP,SEXUAL BEHAVIOR AND AIDS (1994). 125. KAYE WELLINGS, JULIA FIELD, ANNE JOHNSON & JANE WADSWORTH, SEXUAL BE- HAVIOR iN BRITAIN: THE NATIONAL SURVEY OF SEXUAL ATTITUDES AND LIFESTYLES (1994). 126. See, e.g.,Tamar Lewin, Monogamy More Prevalent than Believed, Survey Finds, L.A. DAILYNEWS, Oct. 7, 1994,at NI (notingpopularity of Hite,Redbook, and Playboy studies); Eliza- bethValk Long, To OurReaders, TIME, Oct. 17, 1994,at 4, 4 (indicatingthat the Hite, Laumann, andMasters and Johnson reports were featured as coverstories in Timemagazine). 127. TheRedbook survey asked a sampleof womenwhether they had had a "sexualexperi- encewith another woman" since the age of 18,without asking about whether the women who had hadsuch experiences had also hadsexual experiences with men. TAVRIS& SADD,supra note 119, at 163. 128. The HiteReport asked a sampleof womenthe question "Do youprefer sex withmen, women,yourself, or notat all?"HITE, supra note 120, at 261. Thereport noted: "A hundredand forty-fourwomen in thisstudy (8 percent)said they preferred sex withwomen. Another seventy- threeidentified themselves as 'bisexual,'and eighty-fourmore women had had experienceswith bothmen and women but did not answer as topreference (another 9 percent)."Id. Thisstatement couldbe readto meanthat 8% of womencould be characterizedas lesbian,while 9% couldbe characterizedas bisexual. The question, however, is infelicitouslyworded for my purposes, given thatpreference for women as sex partnersmight not be an exclusiveor evena strongpreference. Similarproblems arise when one triesto characterizethe remaining 9% as bisexual,given that it probablyincludes individuals with only incidental experiences with one sex (whoare not consideredbisexual under my definition). Thus, while the Hite Report is widelyknown, I do notrely on it here. 129. TheSpira study does not clearly distinguish between homosexuals and bisexuals, making a comparisonof the incidence of thesetwo groups impossible. See generallySPIRA ET AL., supra note124. 130. Had I notalready eliminated it, the Redbook study would also havebeen excluded on thisbasis. 131. See notes136-176 infra and accompanying text. 132. CompareKiNSEY ET AL., MALE,supra note 5, at 639-41(arraying individuals according to thecriteria of desire or behavior on a seven-pointscale), with WELLrNGS ET AL.,supra note 125, at 179(arraying individuals according to the criteria of desire or behavior on a five-pointscale).

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 380 STANFORDLA WREVIEW [Vol. 52:353

manyof their results. For example, according to these studies, bisexuals are estimatedto compriseanywhere from 0.2 percent134to fifteen percent135 of thetotal population. What is surprising,however, is thateach study found thatthe incidence of bisexuality was greaterthan or comparableto theinci- denceof homosexuality.

a. Kinsey(1948 & 1953).

Beginat thebeginning,136 with Alfred Kinsey's foundational studies of sexualbehavior in the human male'37 and female.138 In both studies, Kinsey definedsexual orientation according to a combinationofdesire ("psychosex- ualresponse"'39) and conduct ("overt sexual experience"I40), without reliance on self-identification.141Based on thisdefinition, he createdthe famous "Kinseyscale," which limned a sexualorientation continuum spanning from 0 to6.142 Thenumbers denoted the following orientations: 0-No homosexualcontacts involving psychic or physicalresponse; exclusive heterosexualcontacts 1-Incidentalhomosexual contacts; frequent heterosexual contacts 2-More thanincidental homosexual contacts; but more frequent heterosexual contacts 3-Equal homosexualand heterosexual contacts 4-More thanincidental heterosexual contacts; but more frequent homosexual contacts 5-Incidentalheterosexual contacts; frequent homosexual contacts 6-No heterosexualcontacts; exclusive homosexual contacts143

133. Comparethe definition of a Kinsey 1 in KINSEYET AL., MALE,supra note 5, at 639-41, and KiNSEYET AL., FEMALE,supra note 6, at 471,with the definition of a Kinsey 1 in MASTERS& JOHNSON,supra note 121, at 14-15. For moreon thiscomparison, see note 154 infra. 134. WELLINGS ET AL., supra note 125,at 183. 135. See KiNSEY ET AL., MALE, supra note5, at 650-51. 136. This analysismarches through the studieschronologically in orderto providea sense of the timelineon whichthe studies were conducted. It is probablyworth pointing out, however, that the firsttwo studies,which were conductedbefore 1980, are fundamentallydifferent from the last three,which were all conductedin the 1990s. For example,it is somewhatsurprising that the Kin- sey studiesremain the most widely cited of thestudies, despite the factthat they were conductedin an era wheresexual mores were dramaticallydifferent. See Mezey, supra note 8, at 104 (noting thatKinsey's workis stillthe most "widely cited research on sexualityin theUnited States"). 137. KINSEYET AL.,MALE, supra note 5. 138. KINSEY ET AL., FEMALE, supra note6. 139. KiNSEY ET AL., FEMALE, supra note 6, at 471; KINSEY ET AL., MALE, supra note 5, at 647. 140. KiNSEY ET AL., FEMALE, supra note 6, at 471; KINSEY ET AL., MALE, supra note 5, at 647. 141. See KiNSEY ET AL., FEMALE, supra note6, at 471-72;KiNSEY ET AL., MALE, supra note 5, at639-41. 142. See KiNSEY ET AL., FEMALE, supra note6, at 471-72;KiNSEY ET AL., MALE, supra note 5, at639-41. 143. See KiNSEY ET AL., FEMALE, supra note6, at 471-72;KINSEY ET AL., MALE, supra note 5, at639-41.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 381

Thisscale has provedimmensely influential, not only in subsequentstud- ies,144but also in more popular literature.145 WhileKinsey resisted breaking this continuum into the three categories ofheterosexual, bisexual, and homosexual,146 it will be usefulfor me to do so. I definedbisexuality earlier as theability to experience desire in a more thanincidental way for both men and women. Because Kinsey himself cre- atedcategories of incidentalsame-sex desire or behavior(Kinsey 5s) and incidentalcross-sex desire or behavior (Kinsey Is), it is easyto restricthis scaleto fit my definition. I therefore categorize Kinsey Os and Is as "heterosexual,"Kinsey 2s, 3s, and 4s as "bisexual,"and Kinsey 5s and6s as "homosexual."However, because Kinsey did not disaggregate desire and conduct, a completeadaptation of his standard to mineis impossible.To theextent thathis definitioncaptured individuals who had sexualconduct with both sexeswithout desire for both sexes, it will be overinclusive. Underthis definition, what is theratio of bisexuals to homosexuals in the Kinseystudies? Kinsey found that "25 percent of the male population ha[d] morethan incidental homosexual experience or reactions (i.e., rate[d]from 2-6) forat least three years between the ages of 16 and55,"147 while"10 per centof the males [were] more or less exclusively homosexual (i.e., rate[d]5 or 6) forat leastthree years between the ages of 16 and55."148 Underhis timeand age restrictions,Kinsey thus found that fifteen percent of the male populationfell within the bisexual range (Kinsey 2s, 3s, and4s) whileonly tenpercent of the male population fell within the homosexual range (Kinsey 5s and6s). Forwomen, an averageof ten percent of unmarried females and 2.5 percentof married females rated from 2 to 6 "in eachof theyears be- tweentwenty and thirty-five years of age."149 In contrast,only an averageof fourpercent of unmarried females and less than one percent of marriedfe- malesrated 5 or6 "ineach of the years between twenty and thirty-five years

144. See, e.g.,MASTERS & JOHNSON,supra note 121, at 8 (notingits deployment of theKin- seyscale); WELLINGS ET AL., supranote 125, at 179 (notingits deployment of a modifiedKinsey scale). 145. See, e.g., VeronicaGroocock, Alfred C. Kinsey:A Public/PrivateLife, NEW STATES- MAN,Nov. 21, 1997,at 49 (reviewingbook about Kinsey and his reports); William A. Henry,Born Gay?,TIME, July 26, 1993,at 36, 37 (notingKinsey's estimate of the incidence of homosexuality in thepopulation); John Leland, Bisexuality, NEWSWEEK, July 17, 1995,at 44, 47 (discussingKin- sey's findingson bisexuality);Jonathan Weiner, The GayScience, THE NEW REPUBLIC,Jan. 2, 1995,at 35, 36 (discussingmethodology and results of Kinsey report). 146. See KiNSEY ET AL., MALE,supra note 5, at 639 (notingthat sexual orientation is better describedas a continuum).Indeed, Kinsey cautioned that even his seven-point scale was a distor- tionof the continuum of sexual orientation. Id. at 647 ("Finally,it should be emphasizedagain that thereality is a continuum,with individuals in thepopulation occupying not only the seven catego- &es whichare recognized here, but every gradation between each of the categories, as well."). 147. Id.at650. 148. Id. at651. 149. KINSEYET AL., FEMALE,supra note 6, at473.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 382 STANFORDLAWREVIEW [Vol.52:353

ofage."150 Under these time and age restrictions, Kinsey thus found that six percentof unmarried women and more than 1.5 percent of married women fellinto the bisexual range, while only four percent of unmarried women and lessthan one percent of married women fell into the homosexual range. In theKinsey studies, we see a remarkableconsistency ofresult, in whichthe percentageofbisexuals (Kinsey 2s, 3s, and 4s) is 1.5times the percentage of homosexuals(Kinsey 5s and6s) forboth men and women.

b. Mastersand Johnson (1979).

TheMasters and Johnson study also deployedthe Kinsey scale.151 The study,however, involved only "homosexual" subjects-it admitted into its datapool only individuals whom the researchers deemed to be KinseyIs to Kinsey6s.'52 The limitation was almostcertainly too strictly enforced-for example,"[t]he Kinsey 1 representativeswere accepted into the program onlyif they could document that they were currently living in a homosexual relationshipofat leastthree months' duration."153 This implies that the re- searcherswere deploying the Kinsey scale in a differentway than Kinsey himselfdid, for it is extremelydoubtful that the Kinsey researchers would haveclassified as a Kinsey1 anyindividual who had beeninvolved in a same-sexrelationship ofthat duration.154 The Masters and Johnson classifi- cationwould thus underreport thenumber of individuals who would be bisexualunder my definition. Despite this, however, the Masters and Johnson studyalso found that the incidence of bisexuality was greater than or comparableto the incidence of homosexuality. For men, the percentage of bisexuals (Kinsey2s, 3s, and 4s) inthe sample was 46.8 percent, while the percentage ofhomosexuals (Kinsey 5s and6s) was48.9 percent.155 For women, the percentageof bisexualswas 59.8 percent,while the percentage of homo- sexualswas 36.6 percent.156Thus, while the ratio of bisexual men to gay menwas lower than that in the Kinsey study, the study still found that bisex-

150. Id. at 473-74. 151. MASTERS& JOHNSON,supra note 121, at 8. 152. See id. at 14. 153. Id. at 14-15. 154. See KiNSEY ET AL., FEMALE,supra note6, at 471 ("Individualsare rated as I's iftheir psychosexualresponses and/or overt experience are directed almost entirely toward individuals of theopposite sex, although they incidentally make psychosexual responses to theirown sex, and/or haveincidental sexual contacts with individuals of their own sex.... Suchpersons make few if any deliberateattempts to renewtheir homosexual contacts.") (emphases added); KrNSEY ET AL., MALE supranote 5, at 639 ("Individualsare rated 1's if theyhave only incidental homosexual contacts whichhave involved physical or psychic response, or incidentalpsychic response without physical contact.").The individuals described as KinseyIs byMasters and Johnson seem highly unlikely to meetthese criteria. 155. See MASTERS& JOHNSON,supra note 121, at 29. 156. See id at32.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 383 ual menexisted in comparable numbers to gay men. And the ratio of bisex- ualwomen to lesbians was greater than that found in the Kinsey study.

c. Janusand Janus (1993). TheJanus study relied on self-identificationalone, asking subjects to la- belthemselves as heterosexual, homosexual, orbisexual.157 It is thusimpos- sibleto ascertain how subjects in the Janus pool would be definedaccording to myrestricted desire-based definition. But it nonethelessbears note that thestudy reported a roughly one-to-one ratio of bisexuals to homosexuals for bothmen and women.158Among male respondents,five percent self- identifiedas bisexualwhile four percent self-identified as homosexual.159 Amongfemale respondents, three percent self-identified as bisexual, while twopercent self-identified as homosexual.160 The report also suggestedthat theremay have been some under-reporting ofbisexuality: "In our interviews, we found... respondentswho identifiedthemselves as heterosexualand re- portedhaving homosexual relations [and] respondents who identified themselvesas homosexualsand reported that they have heterosexual relations as well."161

d. Wellings(1994). In a studyof sexual behavior in Britain,162researchers used a five-point scale(adapted from the seven-point Kinsey scale)163 to ascertainsexual at- tractionand sexual experience among men and women. Attraction was not furtherdefined,164 butsexual experience was defined as "anykind of contact withanother person that [the subject] felt was sexual,"including "kissing or touching,or intercourseor anyother form of sex."165 Unlikethe Kinsey study,the Wellings study disaggregated attraction and experience, yielding twosets of data.166 The five-point scale, to which I haveassigned numbers, wasas follows:

157. See JANUS& JANUS,supra note 122, at 70. 158. See id. Whilethe Janus report obtained data on same-sexexperiences, it did not report similardata on bisexual experiences, making a comparisonof the two on thisaxis impossible. 159. Seeid. 160. Seeid. 161. Id. Theseindividuals could, of course,be Kinseyls and Kinsey5s respectively,in whichcase itwould not be underreportingofbisexuality under my definition. 162. WELLINGSET AL.,supra note 125. 163. Seeid. at 179. 164. See id. at 181. 165. Id. 166. See id.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 384 STANFORDLAWREVIEW [Vol. 52:353

Wellings0-only heterosexualattraction/experience Wellings1-mostly heterosexual attraction/experience Wellings2-both heterosexualand homosexual attraction/experience Wellings3-mostly homosexual attraction/experience Wellings4-only homosexualattraction/experiencel67 The Wellingsscale is botheasier and harder to adaptto mydefinition of sexual orientationthan the Kinseyscale. It is easier in thatthe Wellings scale disaggregatesdesire and conduct,thereby permitting the isolationof desire. It is harderin thatthe Wellingsscale does not clearlydistinguish betweenincidental desire (or conduct)and nonincidental desire (or conduct). To be consistentwith the Kinsey-based bisexual category above, the bisexual categoryshould include (1) all Wellings2s; and (2) some butnot all Wel- lingsIs; and (3) somebut not all Wellings3s. I thereforeindicate the numberof "bisexuals"as a rangeextending from a narrowunderinclusive definition(ust Wellings2s) to a broadoverinclusive definition (Wellings Is, 2s, and 3s). By the same logic,I indicatethe numberof "homosexuals"as a rangeextending from a narrowdefinition (ust Wellings4s) to a broaddefinition(Wellings 3s and4s). Insofaras attractionwas concerned,for men, 0.5 percent(narrow definition)to 5.0 percent(broad definition) were attracted to bothmen and women, while0.5 percent(narrow definition) to 1.0 percent(broad definition) were attracted(almost) exclusively to men.168For women,0.2 percent(narrow definition)to 4.2 percent(broad definition) were attracted to bothmen and women,while 0.3 percent(narrow definition) to 0.5 percent(broad definition)were attracted (almost) exclusively to women.169The Wellingsstudy thusalso foundbisexuals to existin numbersgreater than or comparableto thenumber of homosexuals.170

167. See id. at 183. 168. See id. 169. See id. 170. The ratioof bisexualsto homosexualsaccording to thebehavior-based definition was similarto thatgenerated under the desire-based definition. In itsbehavior inquiry, the Wellings studyasked two different sets of questions. The first set was similar to the attraction-based question describedabove. See id. at 182; note165 supraand accompanyingtext. It determinedthat for men,0.3% (narrowdefinition) to4.8% (broaddefinition) had sexual experience with both men and women,while 0.4% (narrowdefinition) to 1.0% (broaddefinition) had sexual experience (almost) exclusivelywith men. See WELLINGSET AL., supranote 125, at 183. Forwomen, 0.1% (narrow definition)to2.5% (broaddefinition) had sexual experience with both men and women, while 0. 1% (narrowdefinition) to 0.3% (broaddefinition) had sexualexperience (almost) exclusively with women.See id. Thestudy asked a separateset of questions pertaining tothe exclusiveness of same-sex experience. Menand women were asked about the sex of theirsexual partners over different time periods-(1) ever;(2) inthe last five years; (3) inthe last two years; and (4) in thelast year. Id. at209. Formen, the percentage of thosewith both male and female partners as opposedto thepercentage of thosewith exclusively male partners for these time periods was as follows:(1) 3.4% v. 0.3% ever;(2) 0.8% v. 0.6% inthe last five years; (3) 0.5% v. 0.6% in thelast two years; and (4) 0.4% v.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 385

e. Laumann(1994).

The Laumannstudy categorized sexual orientation separately according to all three of the common indicia-desire, conduct, and self- identification.171Subjects were asked to completethe phrase "In generalare you sexually attractedto . . ." withone of fiveresponses: "(1) only men; (2) mostlymen; (3) bothmen and women;(4) mostlywomen; and (5) only women."172The scale forattraction is thusidentical to the five-pointWel- lingsscale; I thereforefinesse the difficulties of mappingthis onto my defi- nitionin thesame way as above.173For men, 0.6 percent(narrow definition) to 3.9 percent(broad definition)were attractedto bothmen and women, while 2.4 percent(narrow definition) to 3.1 percent(broad definition)re- portedattraction to menonly.174 For women,0.8 percent(narrow definition) to 4.1 percent(broad definition) reported attraction to bothmen and women, while 0.3 percent(narrow definition) to 0.9 percent(broad definition)re- portedattraction to womenonly.175 Taking the narrow and broad definitions as settingthe endpoints of a range,the percentages of "bisexuals"are again greaterthan or comparableto thoseof "homosexuals."176

0.7% in thelast year. See id. Forwomen, the percentage of those with both female and male part- nersas opposedto thepercentage of those with exclusively female partners for these time periods was as follows-(1) 1.7%v. 0.1% ever;(2) 0.5% v. 0.2% in thelast five years; (3) 0.2% v. 0.2% in thelast two years; and (4) 0.2% v. 0.2% in thelast year. See id. Underboth sets of questions, the percentageof individuals having sex withboth men and women was greaterthan or comparable to thepercentage of individuals having sex only with their own sex. See id. 171. See LAUMANNET AL.,supra note75, at 31 1. 172. Id. at658. 173. See note168 supra and accompanying text. 174. See LAUMANNET AL.,supra note 75, at 31 1. Thisstatistic provides the most support for thenonexistence thesis of anyin thestudies surveyed. Under the narrow definition there are four timesmore homosexuals than there are bisexuals. If one takesthe midpoint of each range,however,the ratio of bisexuals to homosexuals is closeto parity (2.25: 2.75). 175. See id. 176. In studyingconduct, the Laumann study again asked a set of questionssimilar to that askedin the Wellings study. Subjects were asked about the sex oftheir sexual partners over differ- enttime periods-(1) sincepuberty, (2) sinceage 18,(3) in thepast five years, and (4) in thelast year.Id. Formen, the percentages of those with both male and female partners as opposedto the percentageof those with exclusively male partners for these time periods were as follows:(1) 5.8% v 0.6% sincepuberty; (2) 4.0% v. 0.9% sinceage 18; (3) 2.1% v. 2.0% in thepast five years; and (4) 0.7% v. 2.0% in thepast year. Id. Forwomen, the statistics were as follows:(1) 3.3% v. 0.2% sincepuberty; (2) 3.7% v. 0.4% sinceage 18; (3) 1.4%v. 0.8% inthe past five years; and (4) 0.3% v. 1.0%in the past year. Id. Whenasked to self-identify,2.0% of menself-identified as "homosexual" and 0.8% of men self-identifiedas "bisexual." Id. 0.9% ofwomen self-identified as lesbian, while 0.5% of women identifiedas bisexual.Id.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 386 STANFORDLAWREVIEW [Vol.52:353

f. Critiquesof the studies.

Thefive studies considered above all cameto roughly the same conclu- sionabout the relative incidence of bisexuality to homosexuality-namely, thatthe incidence of bisexuality was greaterthan or comparable to theinci- denceof homosexuality. While three problems with these studies merit discussion,none ultimately damages the validity of this conclusion for my purposes. The firstproblem concerns the flawedmethodology of someof the studies.The primary methodological flaw concerned sampling practices. As theLaumann researchers noted, the Kinsey studies, the Masters and Johnson study,and the Janus and Janus study all failedto engage in probability sampling.177Of thestudies canvassed above, only the Laumann study and the Wellingsstudy used such sampling.178 Another flaw related to the time variable. Again,all studiesbut the Laumann and Wellings studies failed to ask questionsthat adequately narrowed the timeframe inwhich individuals expe- rienceddesire. This led these studies to someversion of the "once a bisexual,always a bisexual"position, in which a womanwho had more than inci- dentaldesire for a womanin her adulthood but who had subsequently expe- riencedonly desire for men would still be consideredbisexual. The Laumannstudy and theWellings study, however, did notsuffer fromeither of theseflaws. Bothstudies engaged in probabilitysampling, andboth took the time variable into account.179 And the findings of these twostudies about the relative numbers of bisexuals and homosexuals were comparableto the findings of the other studies. Thus, even if we acceptthat theKinsey, Masters and Johnson, and Janus and Janus studies were illegitimate,this only means that there are fewer studies to supportthe same conclusion. Thesecond problem is oneendemic to all sexualitystudies, which is the problemof having to rely on subjectiveaccounts to determinesexual orientation.Because desire and conductwere not calibrated directly by these studies,researchers were forced to rely on individual accounts of both.180 In thissense, distinguishing desire and conducton theone handfrom self- identificationon the other is slightlymisleading, since all ofthe data relied on someform of self-identification. Thefact that all ofthe information col-

177. See LAUMANNET AL., supra note75, at 44, 46 (criticizingMasters and Johnson study andJanus study); MICHAEL ET AL.,supra note 1 14, at 17 (criticizingKinsey studies). 178. See LAUMANNET AL.,supra note75, at 46. 179. See notes 170, 176 supra. 180. Whileattempts have been made to ascertainarousal (which is takenas a proxyfor desire)directly, see ALANHYDE, BODIES OF LAW 173-74(1997), the studies canvassed here made no suchattempts.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 387

lectedwas basedon subjectiveaccounts raises the concern that individuals gaveuntruthful responses. Formy purposes, the most damaging way in whichsubjects could have dissembledwould have been for homosexual subjects to have said they were bisexual. If thesehomosexual subjects had engagedin such"defense bisexuality,"there would have been underreportingof homosexuality and overreportingofbisexuality in thesestudies.181 The confidentialityassured in all ofthe studies,l82 however, diminishes the motive to dissemblein this way. Moreover,bisexuals also haveincentives to characterizethemselves as homosexuall83or heterosexual,184so we could also expectsome overreport- ingof homosexuality and underreporting ofbisexuality. The thirdand mostgeneral concern is thatno studywill everbe able to tellus who is straight,who is gay,and who is bisexual. Thereare weak and strongversions of thisconcern. The weak versionis thateven if all indi- vidualstold the truth as theyknew it, some may not have known the truth.185 The studieswould then be flawedbecause they relied on somesubjects with incompleteknowledge. This will,of course,always be a problemwith survey data. Moreover,assuming that there is a truthto be known,I thinkit is reasonableto assumethat the studies would give a closerapproximation of theactual number of bisexualsthan more quotidian forms of information. Thisis becausethe studies provided a relatively self-reflective andconfiden- tialforum in which individuals could articulate a stigmatized desire. But this doesnot mitigate the strong version of the concern, which is thatexperiences ofsexual orientation are so differentas not to be commensurable.Even as- sumingthis, however, the studies might then be readnot to ascertainthe ac- tualnumber of bisexuals, but rather to underscore a disjunction between aca- demicand populardiscourse. In otherwords, we mightask whyacademic knowledgeabout the number of homosexualstrickled so quicklyinto main- streamculture,186 while academic knowledge (produced in thesame studies)

181. See notes77-79 supra and accompanying text. 182. See JANUS& JANUS,supra note 122,at 5 (discussingconfidential nature of study); KINSEY ET AL., FEMALE,supra note 6, at59 (same);KINSEY ET AL., MALE,supra note 5, at 44-45 (same);LAUMANN ET AL.,supra note 75, at 71-73(same); MASTERS & JOHNSON,supra note 121, at49 (same);WELLINGS ET AL.,supra note 125, at 15,20 (same). 183. Forexample, bisexuals have an incentiveto characterizethemselves as gayin orderto interactsocially with other sexual orientation groups. See RUST,CHALLENGE, supra note 71, at 95- 98 (notingthat some lesbians refuse to interactwith or dateopenly bisexual women). They may also self-identifyas gay simplybecause this is themost readily available template for same-sex desire.See textaccompanying notes 397-398, infra. 184. Bisexualshave an incentiveto characterizethemselves as straightto partakeof heterosexualprivilege. See Brownfain,supra note 83, at 173 (describingmarried bisexuals who pass as straight). 185. See, e.g., SEDGWICK,supra note 4, at 26 (notingthat Western romance tradition, in- cludingpsychoanalysis, is predicated on theassumption that our sexual desires are not transparent toourselves). 186. See note112 supra and accompanying text.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 388 STANFORDLAWREVIEW [Vol. 52:353

aboutthe number of bisexuals did not. This in itself is suggestiveof the erasureof the conceptual category of bisexuality, even if that category has no stablereferent.

C. BisexualErasure as a Cause ofBisexual Invisibility In thisPart, I madetwo claims. My first claim was that bisexuals are invisible.In makingthat claim, I specifiedthat while bisexuals are affected by at leastthree different kinds of invisibility, only one-the invisibility ofbi- sexualsqua bisexuals-isthe subject of this analysis. I maintainedthat the bestway to isolatethe extent of this kind of invisibility was to comparethe relativevisibilities of homosexuality and bisexuality. I thenadduced both statisticaland sociological evidence to demonstratethat bisexual invisibility is farmore pervasive than its homosexual counterpart. My secondclaim was thatthis invisibility betokens bisexual erasure ratherthan bisexual nonexistence. This required me to be moreprecise about mydefinition of bisexuality.I therefore adopted and defended a restricted desire-baseddefinition. Applying this definition (where possible) to five majorsexuality studies, I found that the studies all demonstratedthat the in- cidenceof bisexuals is greaterthan or comparable to the incidence of homosexuals. Puttingthese two claims together, I conclude that bisexuals are invisible in modemAmerican society and thatthis invisibility arises from erasure ratherthan from nonexistence. I now turn to the reasons for this erasure.

II. THE EPISTEMICCONTRACT OF BISEXUALERASURE Thereare at leastthree different explanations for why bisexuals are beingerased in popular culture. While I describeall three,I focus on only one ofthem-the political explanation. This political explanation posits that bisexualsare being erased because the two most powerful sexual orientation constituencies-self-identifiedstraights and self-identified gays-have mu- tualinvestments inthe erasure of bisexuals.187 I call this the epistemic con- tractof bisexual erasure. I thensupport the existence of such mutual inter-

187. Thishypothesis comporis with Naomi Mezey's insightful account of bisexuality, which contendsthat "those who walk on eitherside of theboundary between heterosexuality and homo- sexualitytirelessly repair and define"that boundary to denybisexuals any onticground. See Mezey,supra note 8, at 101. My analysis,however, extends beyond Mezey's in positingthat the investmentsthat straights and gays have in bisexualerasure subsume more than the desire to pre- servethe "ethics" of heterosexuality and homosexuality. Compare id. at 112-20(explaining bisexual erasureby noting heterosexual and homosexual ethics), with text accompanying notes 258-394, infra(explaining bisexual erasure by notinginvestments in stabilizing orientation, preserving the importanceof sexas a distinguishingtrait, and maintaining norms of monogamy).

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 389

estsby demonstrating how both straights and gays use the same strategies to erasebisexuals.

A. DifferentExplanations Thereare three explanations for why bisexuality is invisiblerelative to homosexualityand heterosexuality. These explanations-which I call ontic, cognitive,and political-are general explanations for why any intermediate categorythat straddles a binaryopposition might be renderedinvisible. Whileall ofthese explanations are deeply intertwined, I will argue that the politicalexplanation is the most powerful. Theontic explanation is that we see thingsas binariesbecause they are binaries.188In describing how his philosophical predecessors adopted oppo- sitesas principles,Aristotle noted that "'they give no reasonfor doing so, but are,as it were,compelled by thetruth itself."'189 A varietyof modemthink- ers have also maintainedthat binary classifications reflect "the twoness of reality,"190or real oppositionsin theworld.191 This is notnecessarily to say thatall aspectsof realityare composedof opposedprinciples, merely that some-such as night/day,male/female, and life/death-seeminglyare.192 Underthis theory, the binary opposition between gay and straightis such a "real"binary, and bisexuality is erasedbecause it does notexist. The ontic theorymay be assailed on empiricaland nonempirical grounds.The previous Part contested this theory by noting that studies have shownthat the category of bisexuality isjust as "real"as thecategory of homosexuality.193One could also makea nonempiricalcase forskepticism of thestraight/gay binary based on itssuspicious convenience. The straight/gay binaryis oftenseen as a simpleback formationfrom the male/female binary, insofaras thesexual orientation binary categorizes individuals according to thesex oftheir object choice.194 But there is no reasonto believethat be- causethere are only two sexes (assuming for the sake of argument that this is

188. See RODNEY NEEDHAM, COUNTERPOINTS 35 (1987) (notingposition that oppositions are "an unevadableresponse of consciousnessto its environment"). 189. G.E.R. LLOYD, POLARITYAND ANALOGY 15 & n.2 (1966) (quotingAristotle's Physics). 190. DavidMaybury-Lewis, The Questfor Harmony,in THE ATTRACTION OF OPPOSITES 1, 12 (David Maybury-Lewis& Uri Almagor eds., 1989) (quoting C.R. HALLPIKE, THE FOUNDA- TIONSOF PRIMITIVETHOUGHT 234 (1979)). 191. See Maybury-Lewis,supra note 190, at 12 (citingHALLPIKE, supra note 190); ROBERT HERTZ, The Pre-Eminenceof the RightHand: A Studyin Religious Polarity,in DEATH AND THE RIGHTHAND (1960); LLOYD, supra note189. 192. Maybury-Lewis,supra note 190, at 12. 193. See textaccompanying notes 110-186supra. 194. As I demonstratebelow, themale/female binary may also (counterintuitively)be seen as a back formationfrom the straight/gay binary. See notes339-341 infraand accompanyingtext.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 390 STANFORDLA WRE VIEW [Vol.52:353

true)l95individuals will themselves break into two discrete categories based on desirefor one sex or the other. If anything,itis counterintuitivethattwo setsof desired objects would lead to two corresponding sets of desiring subjects(each of which only desired one set of objects).196 Coffee and tea are twodifferent objects, but we do notassume that all individualsbreak down intomutually exclusive sets of coffee drinkers and tea drinkers. The cognitiveexplanation is thatthere is a humantendency to under- standthings in termsof binaries.197 In otherwords, even if the underlying realityis nota setof binaries, our limited cognitive capacity as humanbeings leadsus to apprehendcomplex phenomena in termsof "[t]heextreme sim- plicityof ... binarystructure," which "link[s] the fewest terms capable of sustaininga relation."198 This theory, which is a hallmarkof structuralan- thropology,199posits that binary thinking is "a spontaneousand necessary act ofthe mind, such that it is to be foundin everytype of culture."200 Under thistheory, bisexuality disappears as a categorybecause of a humanblind- nessto all intermediatecategories. Thecognitive theory can also be contestedon empiricaland nonempirical grounds.The empirical critique notes that our capture by binary episte- mologyis byno meanscomplete. The intensityof binarythinking varies acrossand within cultures,201 thereby demonstrating thatit is notan inescap- ablefact of human cognition. The existence of alternative cognitive models also substantiatesthis point-think of the infinite subdivision model of the continuum,202orthe triadic model of dialectical thinking ("thesis-antithesis- synthesis").203This critique is supplementedby a nonempiricalone. Bi- sexuality,after all, is not inimicalto all binaries-considerthe bisex- uallmonosexualbinary. Thus, even if we believethat the human mind has a tendencyto binarize,that does not explain why the straight/gay binary pre- vailedover the bisexual/monosexual binary. The answer to thatquestion, I contend,is political.

195. As notedabove, there is a growingliterature challenging the binary nature of sex. See note18 supra and accompanying text. 196. See note8 supra(discussing how two forms of desirelead analyticallyto at leastfour categoriesof persons). 197. Thisview is perhapsmost prominently associated with Claude Levi-Strauss. See generally CLAUDE LEVI-STRAUSS,THE SAVAGEMIND 217-44 (George Weidenfeld& Nicolson Ltd. trans.,1966) (observing that binary thinking manifests itself across a diversearray of societies). 198. NEEDHAM,supra note 188, at xi. 199. See UriAlmagor, Dual OrganizationReconsidered, inTHE ATTRACTION OF OPPOSITES, supranote 190, at 19,19. 200. NEEDHAM,supra note 188, at 1. 201. See Maybury-Lewis,supra note 190, at 12. 202. See NEEDHAM,supra note 188, at 230-31. 203. See MichaelH. Hoffheimer,Hegel's First Philosophy of Law, 62 TENN.L. REv. 823, 839 (1995).

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 391

Thepolitical explanation is that some intermediate categories are harder to see becausethey are caught in the middle of a politicalstruggle. Bisexu- alityis invisiblenot because we areinnately blind to intermediate categories, butbecause agonistic politics have bifurcated the continuum we wouldoth- erwisesee.204 Under this explanation, the clash between opposed forces has compelleda choosingof sides and a denialof intermediate ground. In the remainderof this article, I concern myself solely with this political explanation. In consideringhow political struggle has erasedbisexuals, I mustfirst identifythe relevant groups in thatstruggle. I believethat the relevant groupsare still heterosexuals, homosexuals, and bisexuals. But here a crucial shiftin definitionoccurs. When these groups define themselves politically,they define themselves according to self-identificationrather than throughconduct or desire.In speakingof the straight, gay, or bisexualin- vestmentsin bisexualerasure, I thus mean the investments ofseif-identified straights,self-identified gays, and self-identified bisexuals. My hypothesisis thatbisexuals remain invisible because both self- identifiedstraights and self-identified gays have overlapping political interestsin bisexual erasure. It is as ifself-identified straights and self-identified gayshave concluded that whatever their other disagreements, they will agree thatbisexuals do notexist. Put another way, the sexual orientation continuumthat runs from straight through bisexual to gayis a "loopified"one, in whichstraights and gays are actually closer to eachother on thisissue than eithergroup is to bisexuals.205Because of this, self-identified straights and self-identifiedgays enter into what I willcall an epistemiccontract of bisexualerasure.

B. TheEpistemic Contract Defined

As I defineit, an epistemic contract is a contractin thesense that a social contractis a contract.In otherwords, it is nota consciousarrangement be-

204. See generallyGARBER, supra note 11, at 80 (arguingthat political struggle is definedby oppositepoles which suppress the middle ground); Mezey, supra note 8, at 112-20(noting the strict politicalenforcement ofrigid categories of sexual identity). 205. Cf AkhilReed Amar& Vik Amar,President Quayle?, 78 VA. L. REv. 913, 934-35 (1992) (notingincentives for political parties at oppositeends of a politicalcontinuum to unite againstthe middle of thecontinuum). The term"loopification" appears to be DuncanKennedy's coinage. See Duncan Kennedy,The Stages of theDecline of thePublic/Private Distinction, 130 U. PA.L. REv. 1349,1354-57 (1982) (describingtheory of "loopification" inwhich the ends of a con- tinuumare closer to eachother than either is to themiddle). Other scholars have also notedthis politicalphenomenon. See, e.g. HerbertMcClosky & Dennis Chong,Similarities and Differences BetweenLeft-Wing and Right-WingRadicals, 15 BRIT. J.POL. Sci. 329, 343 (1985) (notingthat the "programmaticdifferences between the radical left and the radical right frequently act to obscure thecharacteristics that are shared by the two camps" including "intolerance of ambiguity").

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 392 STANFORDLAWREVIEW [Vol. 52:353

tweenindividuals, but rather a socialnorm that arises unconsciously.206 It mightbe fairlyasked why the epistemic contract (or, for that matter, the so- cialcontract) is expressed as a contractas opposedto a norm.My answer is thatI believethe contractual language better captures the fact that the ar- rangementarises between distinct groups who enter it becauseof overlap- pingbut not congruent interests. Norms appear to arisefrom society as a whole,while "social contracts" seem to ariseout of the constituencies into whichsociety is fractured."Don't ask, don't tell," for example, is a con- tractualformulation of the norm of gaysilence that lays bare the interests (andimplicitly the constituencies) involved: straights will not inquire if gays willnot flaunt.207 The choice,however, is ultimatelysemantic-one could easilyrephrase the substance of what follows in the language of norms. Theepistemic contract is epistemicinsofar as itrelates to thenature of knowledge.It is a socialarrangement about what can be acknowledgedor known.This arrangement arises between groups that have distinct but over- lappinginterests inthe promulgation orrepression of certain kinds of knowledge. In thisarticle, I focus on thelatter-that is, on an epistemiccontract thatrelates to what cannot be known. The epistemiccontract is a politicalmodel that could be profitablyap- pliedbeyond the sexual orientation context to otherintermediate identity categories.Multiracials, for example, can be readilyanalogized to bisexuals in manyregards.208 Multiracials, like bisexuals, are alternativelydefined accordingto universalizing definitions which state that all (oralmost all) individualsare multiracial,209 or according to minoritizingdefinitions which statethat a smallproportion ofthe population is multiracial.210 Multiracials, likebisexuals, are usually defined in thelatter sense, and accordingto an analogouslyvaried set of terms.211 As inthe bisexual context, the tendency

206. See, e.g., JEAN-JACQUESROUSSEAU, THE SOCIAL CONTRACT AND DISCOURSES 14 (G.D.H. Cole trans.,1950) ("The clauses of thiscontract ... have perhapsnever been formallyset forth,[but] they are everywherethe same and everywheretacitly admitted and recognized... 207. See Yoshino,supra note27, at 556. 208. RuthColker's book Hybridusefully analogizes the problematicbipolar model of orien- tationwith those that obtain in thecontexts of race,gender, and disability.See generallyCOLKER, HYBRID,supra note9. 209. See JohnA. Powell,The Colorblind Multiracial Dilemma: Racial CategoriesReconsid- ered, 31 U.S.F. L. REv. 789, 797-98 (1997) (notingthat if multiracialmeans thatan individualhas ancestryof differentraces, then "virtually everyone is multiracial"). 210. See,e.g., Luther Wright, Jr., Who's Black, Who's White, and WhoCares: Reconceptual- izingthe United States's Definition of Race and Racial Classifications, 48 VAND. L. REv.513, 557- 58 (1995) (notingthat biracial people, definedas people withparents of differentrecognized races, compriseroughly three percent of birthsin theUnited States). 211. Such shiftingdefinitions can be seen in theUnited States Census's categorizationof individuals withboth whiteand black ancestry. From 1790, when the Census began to distinguish (based on third-partyreports) between whites and blacks, to 1850, no intermediatecategory was recognized.See ChristineB. Hickman,The Devil and the One Drop Rule: Racial Categories,Afri- can Americans,and the US. Census,95 MICH. L. REv. 1161, 1182 (1997). This is similarto mod-

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 393

to adoptnarrow binarizing definitions ofrace can be tracedto anxietiesfelt byboth the dominant and the subordinate groups.212 Recently, self-identified multiracials,like self-identified bisexuals, have organized to resistsuch bi- narizingdefinitions.2l3 And theseself-identified multiracials, like self- identifiedbisexuals, face the dilemma of whether simply to add the interme- diategroup as a thirdcategory or to deployit to more deeply interrogate the principleof racial categorization.214 Atthe same time, however, the political model adduced here clearly does notdescribe all intermediatecategories. The contemporary American middle class,for example, is bothan intermediatecategory (as itsname would suggest)and much more visible than any other class.215 This shouldinspire cautionin applyingthe model too quickly even to groupsthat seem superfi- ciallyto be greatlyilluminated by it.216 I thereforeconstrain myself in this analysisto the context of sexual orientation. Theepistemic contract of bisexual erasure is a particularlypowerful one. Thiscan be seenby contrastingit to theepistemic contract of homosexual erasureexemplified by the military's current "don't ask, don't tell" policy.217 "Don'task, don't tell" satisfies both criteria of an epistemiccontract. It is a

els of orientationthat fail to recognizethe bisexual possibility. See notes 226-232, 238-242 infra and accompanyingtext. From 1850 to 1910, the Census recognizedan intermediatecategory of "mulattoes." See JOELWILLIAMSON, NEW PEOPLE: MISCEGENATIONAND MULATrOESIN THE UNITEDSTATES 24 (1995); Hickman,supra, at 1185. Indeed,in the 1890 census,enumerators were asked to categorizethe intermediatespan betweenblack and white accordingto a set of named fractions(octoroon, quadroon, mulatto). See WILLIAMSON,supra, at 112; Hickman,supra, at 1186. This partitionedcontinuum is reminiscentof the Kinsey scale. See note 143 supra and accompanyingtext. Afterthe census of 1920, however,the Census Bureau moved away fromthe named fractionsrule to the "one drop rule." See WILLIAMSON,supra, at 114; Hickman,supra, at 1187. This rule of hypodescentcould be analogized to the "one act rule" in the orientationcontext. See note99 supra and accompanyingtext. 212. See F. JAMESDAVIS, WHO IS BLACK? ONE NATION'SDEFINITION 137 (1991) (noting thatboth blacks and whitessupport a binarizing"one droprule"). 213. See Hickman,supra note211, at 1164-65. 214. See id. at 1165 ("One wingof thisnew multiracialmovement argues that a new 'multiracial box' should be made available forthe growingnumber of childrenof interracialmarriages. Anotherwing of thismovement, in books and law reviewarticles, suggests that the additionof this categoryshould be a partof a wholesaleredefinition of theracial identitiesof mostAmericans."). 215. This discrepancyin visibilitybetween the middle class and otherintermediate categories such as bisexuals and multi-racialsmay be due to thedifferent incentives felt by theprivileged class in each context. The "upperclass" may feelpressure to erase itselfin ways thatheterosexuals and whites may not. See PETERW. COOKSON,JR. & CAROLrNEHODGES PERSELL,PREPARING FOR POWER:AMERICA'S ELITE BOARDrNGSCHOOLS 28 (1985) ("Part of the socializationfor power is learninghow to conceal wealth,or at least minimizeits importanceby neveropenly referring to it ....'). I am indebtedto Bruce Ackermanfor raising the discrepancy,and to Zachary Potterfor suggestingthis explanation for it. 216. Multiracialsand bisexuals, for example, can be easily distinguishedon a numberof grounds,such as thefact that multi-racial identity is oftenthought to be geneticand heritable,while bisexual identityis oftenthought to be behavioraland notinherited. 217. 10 U.S.C. ? 654 (1998).

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 394 STANFORDLA WRE VIEW [Vol. 52:353

contractinsofar as it articulatesthe needs of twodistinct constituencies- crudely,self-identified straights and self-identified gays. As conventionally told,each side gets something out of the policy2l8-self-identified straights securea militaryin whichhomosexuality is repressed ("don't tell"); self- identifiedgays garner a militaryin whichgays can ostensibly serve without beingsubjected to witchhuntsor harassment ("don't ask").219 And "don't ask,don't tell" is epistemicinsofar as whatis reallyat issue in the contract is a jointrepression of knowledge. Self-identified straights and self-identified gayshave been led throughtheir distinct but overlapping interests to agree thatindividuals in themilitary do nothave to ownwhat they know-that homosexualsexist in their ranks. As I haveargued elsewhere, the existence of the "don't ask, don't tell" policysignifies both continuity and discontinuity with prior social norms. 220 It signifiescontinuity in that the policy did not spring full-fledged from the headof Congress, but rather grew out of the underlying culture.221 In other words,"don't ask, don't tell" arises out of a pervasivestrategy of silencethat haslong affected gays in public culture-a strategy that permits gays to live as equalcitizens if andonly if they agree to "pass"as straight.222At the sametime, however, "don't ask, don't tell" also signifies a breakdown of that underlyingculture, for if thatculture were still extant, there would be no needto codify it.223 In contrastto "don'task, don't tell," bisexual erasure as suchneeds no explicitlegal enforcement. This is becausebisexual erasure is stillsuffi- cientlyrooted in socialculture and thus appears natural, just as homosexual erasureappeared natural when "don't ask, don't tell" was so rooted.What explainsthe discrepancy? Theanswer is thatstraights and gays are the most powerful constituen- ciesin both of these contracts, and that gays feel differently about homosexual erasurethan they do aboutbisexual erasure. The epistemic contract of "don'task, don't tell" is less a case of realinterest convergence between straightsand gays than it is a caseof compromise. The policy splits the dif-

218. See, e.g., Eskridge& Frickey,supra note40, at 92 (describing"don't ask, don't tell" as a compromise);Diane H. Mazur,The Unknown Soldier: A Critiqueof "Gaysin theMilitary" Schol- arship and Litigation,29 U.C. DAVIS L. REV. 223, 227-28 (1996) (same); Francisco Valdes, Queers,Sissies, Dykes, and Tomboys:Deconstructing the Conflation of "Sex," "Gender,"and "SexualOrientation " in Euro-American Law andSociety, 83 CAL. L. REV.1, 370 (1995) (same). 219. There is evidence thatthe witchhuntshave persistedunder the "don't ask, don't tell" policy. See C. DIXON OSBURN& MICHELE M. BENECKE,SERVICEMEMBERS LEGAL DEFENSE NETWORK,CONDUCT UNBECOMING CONTINUES: THE FIRST YEAR UNDER "DON'T ASK, DON'T TELL, DON'T PURSUE" 11 (1995) (notingdocumentation of 15 actual or attemptedwitchhunts under "don't ask, don't tell"). 220. See Yoshino,supra note27, at 556-57. 221. See id.at556. 222. See id. 223. See id. at 556-57.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 395

ferencebetween the self-identified straight interest in theremoval of gays fromthe military and the self-identified gayinterest in the ability of gays to serveopenly (although it stronglyfavors the former).224 This is whythis particularepistemic contract is so unstable-eachside (and particularly the gayside) wants more than the status quo provides.225In contrast,the epistemiccontract of bisexual erasure is, at least insofar as gaysand straights are concerned,a case ofreal interest convergence. Both straights and gays, for differentreasons, want bisexuals to be invisible. Because thesetwo groups-whichagain are the most powerful sexual orientation constituencies-wantthis invisibility, this epistemic contract is much more stable. In- deed,it is so stableas notto be perceivedas a contractatall.

C. Strategiesof Erasure

Theexistence of a bilateralcontract of bisexual erasure is supportedby thefact that both straights and gays engage in the same strategies ofbisexual erasure.There are three such strategies: (1) classerasure, (2) individualerasure,and (3) delegitimation.These strategies imply the acceptance of suc- cessivelystronger claims about bisexual existence-the first strategy does notrecognize the category of "bisexuality";the second acknowledges the categorybut excludes an individualfrom it; and the third accepts individual bisexualityas a stableidentity but stigmatizes it.

1. Straightdeployments of the strategies.

Class erasureoccurs when straights deny the existence of the entire bi- sexualcategory. This can occur explicitly or implicitly. A fewdecades ago, explicitdenial appears to havebeen fairly common even among academic theorists,some of whom believed that all self-identifiedbisexuals were actu- allyhomosexuals in denial.226While the claim that "there is no suchthing as bisexuality"remains part of the common wisdom today,227 perhaps the more prominentform of denial is nowthe implicit one. Suchimplicit denial often occursthrough the use of the straight/gay binary as a completemeans of de- scribingall individuals.228As discussed above,229 this was thestrategy used

224. See COLKER,HYBRID, supra note 9, at64. 225. Gaydissatisfaction with the status quo canbe seenin thelegal challenges to thepolicy. See,e.g., Able v. UnitedStates, 155 F.3d 628, 631 (2d Cir.1998) (upholding "don't ask, don't tell" againstconstitutional challenge); Philips v. Perry,106 F.3d 1420, 1424 (9thCir. 1997) (same); Richenbergv. Perry,97 F.3d 256,260 (8thCir. 1996)(same); Thomasson v. Perry,80 F.3d 915, 927 (4thCir. 1996) (same). 226. See Fox,supra note 79, at 21. Onetheorist, for example, maintained that bisexuality was "a statethat has no existencebeyond the word itself," that is, "an out-and-outfraud." EDMUND BERGLER,HOMOSEXUALITY: DISEASE OR WAY OF LIFE 80-89 (1956). 227. GARBER,supra note 11, at 16. 228. See notes45-46 supra and accompanying text.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 396 STANFORDLAWREVIEW [Vol. 52:353 bythe Romer Court. A moresubtle strategy of implicit class erasure is the descriptionofbisexuality as "bisexualchic" in the mainstream (and thus pre- sumptivelystraight) press.230 As thephrase suggests, bisexuality is here madevisible as a phase,fashion, or fad-its appearance is inscribedwith its imminentdisappearance.231 Indeed, impending evanescence may be thepre- conditionfor visibility.232 Individualerasure recognizes that bisexuals exist as a class,but contests thata particularindividual is bisexual.Such challenges by straightsoften arisewhen a self-identifiedbisexual is accusedof being a duplicitousor nas- centhomosexual.233 In either case, the bisexual self-ascription is seen as a "phase"from which an individualwill ultimately emerge.234 Such erasure replicatesthe "phase" paradigm of bisexual chic at thelevel of the individual-ontogenyhere recapitulates phylogeny. Self-described bisexuality is thusseen not as a stableindividual identity but a placefrom which a stable monosexualidentity is acknowledged orchosen. Finally,delegitimation occurs when straights acknowledge the existence ofindividual bisexuals but attach a stigmato bisexuality. Itmay seem odd to characterizedelegitimation as a strategyofbisexual invisibility, given that it makesbisexuality more visible. But because it chillsthe expression of bisexualityby portraying itonly in a negativeway, I includeit here. Common straightstereotypes ofbisexuals portray them as promiscuous,as duplicitous, as closeted,and especially as bridgesfor HIV infectionfrom the "high risk" gaypopulation to the "low risk" straight population.235

229. See notes47-52 supra and accompanying text. 230. See,e.g., Bisexual Chic: Anyone Goes, NEWSWEEK, May 27, 1974,at 90,90 (notingthat bisexuality"was probablyinevitable," since "his-and-her clothes, hair styles and role assignments blurredthe line between the sexes until they overlapped, the only thing left to swap was sexitself'); TheNew Bisexuals, TIME, May 13, 1974,at 79 (describingfashionability ofbisexuality); see also GARBER,supra note 1 1, at 18-24(describing bisexual chic). 231. See GARBER,supra note 11, at 20 (askingwhether bisexual chic turns bisexuality into "a fashion-likeplatform shoes, bell-bottomed trousers, or double-breastedsuits-that appears and thendisappears, goes underground, only to be 'revived'with a difference?"). 232. An analogymay be madeto Leo Bersani'shypothesis that AIDS has madegays more visiblein partbecause it promises their ultimate disappearance-gay visibility is mademost palat- able at themoment its transience is underscored.LEO BERSANI, HOMOS 21 (1995). Whileobvi- ouslyless sinister,the concept of bisexualchic may operate in thesame way, reassuring straight culturethat it is safeto look at bisexuals by characterizing them as a passingfancy. 233. See GARBER, supranote 1 1, at 145. 234. Theinstability of bisexuals,which is framedhere in temporalterms as a "phase"argument,can also be framedin spatialterms, as in descriptionsof bisexualsas being"on thefence." See, e.g., AmberAult, Hegemonic Discourse in an OppositionalCommunity: Lesbian Feminist StigmatizationofBisexual Women, in QUEER STUDIES, supra note 15, at 204, 206. 235. See Ochs,supra note 63, at 227.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 397

2. Gaydeployments of the strategies.

Thatgays and lesbians would collude with straights in theerasure of bi- sexualsis superficiallysurprising.236 As sexual minorities who have them- selvesbeen rendered invisible, one wouldexpect that gays and lesbians wouldbe loathto assumethe role of oppressor relative to othersexual minorities.237Yet gaysdeploy all ofthe strategies of bisexual erasure used by straights.This supports the thesis that bisexual invisibility must be under- stoodnot as theunilateral erasure of bisexuals by straights, but the bilateral erasureof bisexuals by both straights and gays. Self-identifiedgays deploy all threestrategies ofbisexual erasure-class erasure,individual erasure, and delegitimation. Speaking of class erasure, ChristopherJames laments the phenomenon inwhich a queer,gay, or lesbian theorist"excludes bisexuality as a relevantcategory of sexual identity," and then"claims behaviorally bisexual people or texts with bisexual characters or contentas 'queer,''gay,' or 'lesbian."'238 James goes on to give examples of scholarlytexts that elide the bisexual category and unself-consciously cate- gorizeindividuals whom he believeswould be moreappropriately deemed bisexualas queer,gay, or lesbian.239This practice of elision,of course,is notlimited to theacademy. As MarjorieGarber notes, popular compendi- umsof gay "lists," such as The Gay Book ofDays240 and The Gay/Lesbian Almanac,241purport to listonly gays (as theirtitles would suggest), but in- cludemany bisexuals in their pages.242 Evenwhen gays formally recognize bisexuals as a category,they can peremptorilyevacuate individuals from that category.243 A classic example

236. See WEINBERG ET AL.,supra note 113, at 8 (notingwith surprise that bisexuals reported homosexualsto be just as negativeas heterosexualsabout bisexuality). 237. See id. at 117 (notingthat bisexuals, when confrontedwith hostility from gays, may be dismayedbecause theyview themselvesvictimized by thesame kindof prejudiceas gays). 238. James,supra note 15, at 228. To his credit,James acknowledges that some usages of the words "queer," "gay," and "lesbian" mightbe thoughtto encompassbisexuality. Id. at 228-29. (I thinkthis is mosttrue for the term "queer"-bisexuals, forexample, would seem to fitcomfortably withinthe usage of the word "queer" to denote all those who fall outside of the regimesof the "nornal," see supra note 3.) But Jamespoints out thatbecause thesewords signifyexclusive ho- mosexualityfor many others,theorists who use the termsexpansively should declare and justify such usage. James,supra note 15, at 229. 239. See id. at 229 (criticizingDAVID BERGMAN,GAIETY TRANSFIGURED:GAY SELF- REPRESENTATIONIN AMERICANLITERATURE (1991) for appropriatingHerman Melville, Henry David Thoreau,and HenryJames as gay men); James,supra note 15, at 229-30 (criticizingKARLA JAY& JOANNEGLASKOW, LESBIAN TEXTS AND CONTEXTS:RADICAL REVISIONS (1990) forap- propriatingEmily Dickinson, Virginia Woolf, H.D., and Djuna Barnesas lesbians). 240. MARTiN GREIF, THE GAY BOOK OF DAYS: AN EVOCATIVELY ILLUSTRATIVE WHO'S WHO OF WHO IS, WAS, MAY HAVE BEEN, PROBABLYWAS, AND ALMOSTCERTAINLY SEEMS TO HAVEBEEN GAY DuRINGTHE PAST 5,000 YEARS (1982). 241. JONATHANNED KATZ, THE GAY/LESBIAN ALMANAC: A NEW DOCUMENTARY (1983). 242. See GARBER, supra note 11, at 49-50. 243. See Ault,supra note234, at 208.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 398 STANFORDLAWREVIEW [Vol.52:353 of suchindividual erasure can be foundin thebeginning of lesbian theorist TerryCastle's The Apparitional Lesbian.244 Castle there invokes the figure of GretaGarbo, whom she viewsas exemplifyingthe book's thesis-that "[w]henit comesto lesbians. . . manypeople have troubleseeing what's in firontof them."245This social myopia seems to extendto bisexuals as well, becauseCastle characterizes Garbo as a "lesbianactress"246 even after ob- servingthat Garbo "occasionally had affairs with men as wellas women."247 Castle'ssubsequent concession that Garbo could be characterizedas a bisexual exoneratesher from the charge of class erasure.248 But her summary re- jectionof thatcharacterization implicates her in individualerasure-Castle believesit "moremeaningful to referto [Garbo]as a lesbian"249because "whileGarbo sometimes makes love to men, she would rather make love to women."250But how do we knowthat Garbo prefers women to men? And, evenassuming that we knowthat Garbo sexually preferred women to men, whydoes this make her a lesbian? Likestraights, gays can often engage in thiskind of erasure by charac- terizingindividuals who self-describe as bisexual as goingthrough a "phase" thatwill end in monosexuality.251Thebelief that bisexuals are protohomo- sexualsis a particularlyprevalent one among gays.252 The greater force with whichthis belief is expressedin thegay community may be partiallyex- plainedby experiential skepticism. Like straights, gays have observed "bisexuals"subsequently come out as gay.253But presumablyunlike self- identifiedstraights, some self-identified gays have gone through this phase themselves.254This experience may lead them to be suspiciousof those who claimbisexuality as a stableidentity. The claim that bisexuals are protoho- mosexualsmay also be moreintensely held in thegay community, perhaps

244. TERRY CASTLE, THE APPARITIONAL LESBIAN: FEMALE HOMOSEXUALITY AND MOD- ERNCULTURE (1993). 245. Id. at 2. 246. Id. 247. Id. at 15. 248. See id. 249. See id. 250. Id. 251. See WEINBERGET AL., supra note 113, at 117. This characterizationis particularly ironicgiven that gays themselvesare oftentold by straightsthat their homosexuality is just a phase on theirway to heterosexuality.See Teemu Ruskola,Minor Disregard: The Legal Constructionof theFantasy that Gay and Lesbian Youth Do NotExist, 8 YALE J.L.& FEMINISM269, 280 (1996). 252. See, e.g.,Paula C. Rust,Sexual Identity and BisexualIdentities: The Struggle for Self- Descriptionin a ChangingSexual Landscape, in QUEER STUDIES, supra note 15, at 64, 65. 253. See GARBER,supra note 11, at 145 (describinghow Boy George, David Geffen,and EltonJohn all self-identifiedas "bisexual"before "coming out" as gay). 254. See, e.g., RUST, CHALLENGE,supra note 71, at 44 ("Bisexual identityis oftenconsid- ered a steppingstone on theway to comingout as a lesbian. Slightlymore than half of [thelesbians in the study] called themselvesbisexual before they came out as lesbians; in hindsight,these womenmight well see theirearlier bisexual identityas a transitionalstage.").

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 399

becausegays have moreto lose thanstraights when bisexuals abandon them.255Politicized homosexuals may be particularlyprone to holdingthis view.256 Finally,gays delegitimate bisexuals in two ways. Again, the more obvious wayis simpledenigration-"the lesbian and gaycommunity abounds withnegative images of bisexuals as fence-sitters,traitors, cop-outs, closet cases,people whose primary goal in life is toretain 'heterosexual privilege,' [or] power-hungryseducers who use and discardtheir same-sex lovers , ,"257 Less intuitively,gays can also delegitimatebisexuals by imperson- atingthem. Some might contend that such "defense bisexuality" increases, ratherthan decreases, bisexual visibility, as gayswho say they are bisexual swellthe ranks of the bisexual category. But the transience and the quality ofthe visibility must also be considered-suchgays may later reveal them- selvesto be gay,thereby detracting from the credibility ofthose who experiencetheir bisexuality as a stableidentity.

D. TheEpistemic Contract as a Cause ofBisexual Erasure Whilethe erasure of thebisexual category can be explainedthrough a varietyof hypotheses, the political hypothesis is the most plausible. I there- forefocus on thatexplanation, maintaining that self-identified straights and self-identifiedgays have sharedpolitical interests that lead theminto an epistemiccontract of bisexual erasure. The existenceof sucha contractis demonstratedbythe fact that both straights and gays use the same strategies to erasebisexuals. This suggests that the erasure is beingdriven not just by straightinterests, but also by gay interests. I now explore the nature of those interests.

III. MONOSEXUALINVESTMENTS IN THE EPISTEMICCONTRACT In thisPart, I hypothesizethat bisexuals are erased because their visibil- ityis threateningtothree interests that both straights and gays possess: (1) an interestin thestability of sexual orientation categories; (2) an interestin theprimacy of sex as a diacriticalcharacteristic; and (3) an interestin the preservationofmonogamy. While these interests overlap in some ways, they arealso distinctin others.I thusdisaggregate each interest into three por- tions:(1) theportion of the interest that is sharedby both straights and gays;

255. See Mezey, supra note 8, at 118 ("As an oppressedsubculture, homosexual communities rely on an ethic of homosexual identityto maintainstrength and coherenceagainst the constant onslaughtof virulentattacks ...."). 256. See WEINBERGET AL.,supra note 113,at 117. 257. Lisa Orlando,Loving WhomWe Choose, in BI ANY OTHERNAME: BISEXUALPEOPLE SPEAKOuT 223, 224 (Loraine Hutchins& Lani Kaahumanueds., 1991).

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 400 STANFORDLA WRE VIEW [Vol. 52:353

(2) theportion of the interest that is distinctivelystraight; and (3) theportion ofthe interest that is distinctivelygay. It willbe usefulto entertwo qualifications before embarking on this analysis.First, I am mindful that separating the parties and attributing view- pointsto themin theway that I haverisks essentializing the opinions that membersof each category have about the others. After some thought, I have overriddenthis weighty objection. I do so becauseI believethat certain opinionsare morelikely to be heldby one group than another, although it willnot be heldby all membersor onlymembers of thatgroup.258 Thus, whenI speakof the "investment ofself-identified gays in bisexual erasure," I do notmean that all gaysor only gays have this investment. Rather, I indicatethe investment inbisexual erasure that is morelikely to be heldby self- identifiedgays than by any other group. Second,I wish to stressthe importanceof distinguishingbetween stereotypesofbisexuality and the realities underlying those stereotypes. The followinganalysis will sometimes identify a viewpoint as a stereotypebut nonethelessgive it weight in the analysis. That weight, however, arises only fromthe factthat the investmentsof self-identifiedstraights and self- identifiedgays will often be drivenby perceptions rather than by realities. Indeed,I willargue that the stereotypes that bisexuals are (1) politicallyun- reliable,(2) completelysex-blind, or (3) intrinsicallypromiscuous, are ex- tremelypowerful in shaping straight and gay attitudes about bisexuals. Such argumentation,ofcourse, should not be takenas accessionto the validity of thosestereotypes.

A. StabilizationofSexual Orientation Bisexualitydestabilizes sexual orientation bymaking it logically impossibleto provethat one has a monosexualidentity. Both straights and gays have sharedinvestments in stabilizing their identities, as membersof all groupsare likely to draw some comfort from rigid social orderings. Straights andgay, however, also have distinctive investments instabilizing orientation categories.For straights, itis an investmentinthe retention of heterosexual privilege;for gays, it is an investmentin the retention of theimmutability defenseand one in the ability to form an effective political movement.

1. Sharedinvestment. Bisexualitycalls into question the sexual orientation ofthe self. To see this,contrast the ease ofproving one is straightor gayin a worldin which

258. Compare,e.g., COLKER, HYBRID, supra note 9, at 21-24(noting that many gays have negativeopinions about bisexuality), with, e.g., id. at 24 (notingthat some gays were more tolerant ofColker's fluid sexuality than she herself was).

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 401 bisexualsare not acknowledged to existwith the difficulty of provingthe samething in a worldin which bisexuals are recognized. In a worldthat de- niesbisexual existence, cross-sex desire and same-sexdesire are mutually exclusive.259This means that the presence of cross-sex desire ipso facto ne- gatesthe presence of same-sex desire, and vice versa. Thus, demonstrating cross-sexdesire is sufficienttoprove that one is heterosexualand not homosexual. Thatwe livein thisworld-and under this analytic regime-is demon- stratedby howsexual orientation is proved here. As I willshow in some detail,alleged harassers in thesexual harassment context have adduced ex- actlysuch claims of cross-sex desire as dispositiveof their heterosexuality.260 Similarly,legal institutions assigning individuals a homosexual identity have presumedthat any individual who has manifested same-sex desire (and not deployedthe queen-for-a-day exception) is homosexual.261 In a worldthat recognizes bisexuals, a would-be heterosexual must show (1) thathe is notgay and (2) thathe is notbisexual. (While I takehetero- sexualityas myexample, it should be clearthat this analysis is equallyappli- cableto homosexuality.)The firsthe can do by showingcross-sex desire. Thesecond he canonly do byproving the absence of same-sex desire. But thisis impossibleto do,as itis impossibleto prove a negative.Thus, after thebisexual possibility is acknowledged, itis notonly harder, but logically impossible,to proveone's heterosexuality. Evidence of cross-sexdesire is nonresponsivetothe charge that one harbors same-sex desire, for no amount of lovingattention that a mangives to hiswife can prove that he doesnot lustin his heart for men. In orderto safeguarda regime in which"straightness"(or "gayness")can be proved,all monosexualsmust repress bisexual existence. Bothstraights and gayshave a sharedinvestment in stabilizingtheir identities,as membersof all groupstake comfort in knowingtheir place in thesocial order. This interest may seem less intuitive for gays, who occupy a disfavoredplace in that order. But as MaryMcIntosh has noted, rigid cate- gorizationmay be comfortingeven to those stigmatized within it, because "it appearsto forecloseon thepossibility of driftingback into normality and

259. Thisanalysis assumes, without endorsing, the erasure of asexuality as an analyticpossibility.See note8 supra. 260. See notes520-534 infra and accompanying text. See, e.g., N.H. REV.STAT. ANN. ? 170- B:2 (1992) (defininganyone who engages in same-sexoral or anal sex as a homosexual);Wood- wardv. UnitedStates, 871 F.2d 1068,1069 n.l (Fed. Cir. 1989) (quotingSecretary of theNavy Instruction1900.9A) (describing Navy's pre-1993 policy defining anyone with same-sex desire as a homosexual). 261. See COLKER,HYBRID, supra note 9, at 45-56(discussing the legal definition of homosexualityin adoptionand military contexts).

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 402 STANFORDLA WRE VIEW [Vol. 52:353

thusremoves the element of anxious choice."262 Robin Ochs provides anec- dotalsupport for this hypothesis: WhenI haveasked gay men to explaintheir fears about bisexuality and bisexual people,one theme has repeatedlyarisen. As one gayman put it, "Coming outas gaywas thehardest and most painful thing I haveever done in mylife. Now I'm finallyat a placewhere I havea solididentity, a community, a place to call home. Bisexualsmake me uncomfortablebecause their existence raises forme thepossibility that I mightbe bisexualmyself. And coming to terms withmy identitywas so hardfor me the firsttime around, I cringeat the thoughtof havingto go throughsuch a long,hard, painful process a second time."263 Thusgays, along with straights, share an interestin identitystabilization, sincesuch stabilization roots them in a communityand relieves them of the anxiouswork of identity interrogation.

2. Straightinvestment.

As membersof the privileged orientation class, straights have a distinc- tiveinterest in maintainingtheir identity. That identity must be assuredat boththe individual and at the collective levels. Atthe individual level, bisexuality threatens heterosexuals inthe manner describedabove, by making it impossible for them definitively toprove their heterosexuality.The investment in stabilizing their orientations at the indi- viduallevel leads self-identified straights to takedifferent stances toward homosexualsand bisexuals. Ironically,the investmentin heterosexual privilegecan lead straights toincrease the visibility ofhomosexuals. Prior to theintroduction ofthe concept of homosexuality, no such incentive existed, as theconcept of heterosexuality was so universalas tobe itselfinvisible.264 Sincethe introduction ofthe concept of homosexuality,265 however, heterosexuality'sstability has been predicated on oppositionto homosexuality,as

262. Mary McIntosh,The Homosexual Role, in FORMS OF DESIRE: SEXUAL ORIENTATION AND THE SOCIAL CONSTRUCTIONIST CONTROVERSY 25, 28 (Edward Steined., 1990). 263. Ochs, supra note 63, at 232. 264. See JOHN D'EMILIO, SEXUAL POLITICS, SEXUAL COMMUNITIES: THE MAKING OF A HOMOSEXUALMINORITY IN THE UNITED STATES 1940-70,at 10 (1983) (notingthat in colonial America,where "the existenceof lesbians and gay men was inconceivable,"the term "'[h]eterosexuality'remained undefined, since it was literallythe only way of life");GARBER, supra note11, at 40 (assertingthat "[b]efore people began to speakof 'homosexuals'as a kindof person,a socialspecies, there was no needfor a termlike 'heterosexual"'). The factthat the term heterosexualityis a conceptualback formation from the term homosexuality is reflected linguisti- callyin thefact that the word "heterosexual" post-dated the word "homosexual." See GARBER, supranote 1 1, at 40; DAVID M. HALPERIN,ONE HUNDREDYEARS OF HOMOSEXUALITY155 n.1 (1989). 265. See, e.g., 1 FOUCAULT, supra note 95, at 42-43 (datingcreation of conceptof the homo- sexualto the late nineteenth century).

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 403

"thedenotation of anyterm is alwaysdependent on what is exteriorto it."266 Thatopposition in turn is onlypossible if homosexuals are visible enough to be distinguishedaway. Thus,even attempts to exorcisehomosexuality be- yondthe pale of discoursedo so onlyby namingit.267 It is in thissense that D.A. Millerclaims that heterosexuals "unabashedly need" homosexuals.268 Whilestraights need gays, they do notneed bisexuals. To thecontrary, thevery system in whichstraights need homosexuals is undergirdedbybi- sexualnonexistence. This is becausestraights need gays to exorcise the possibilityof same-sex desire from themselves. But this exorcism can only oc- curin a worldwhere bisexuals do not exist. Two linkedobjections might be raised. The firstis thatmost hetero- sexualsactually do notspend much time worrying about whether they are straight.Like whiteness in therace context,269 heterosexuality in the orientationcontext is sufficientlyprivileged as to achievea kindof transparency.270It mightthus be contendedthat straights cannot be repressingbi- sexualityout of anyneed to provetheir straightness since they do notfeel thisneed. This contention,however, has grownharder to make,as the greatervisibility of sexualminorities (indicatively gays) has somewhatde- naturalizedheterosexuality. More to thepoint, the fact that straights can leavetheir sexual orientations undertheorized should itself be seenas a pre- ciousentitlement.271 Even straights who do notfeel as ifthey must prove theirsexual orientation may have an interestin nothaving to thinkabout theirsexual orientation. And bisexuality, more than homosexuality, requires themto do preciselythis. A relatedobjection is that,as a practicalmatter, we assessorientation by gestaltintuitions rather than by analytic proof. A womanmight not be able to "prove"that her husband is straight,but she may believe she can com- fortablyinfer this from any number of small gestures he makestowards her andother women and men. And if no analyticproof is required,itarguably doesnot matter that no suchproof exists. But this is debatable-beingtold thatthere is no logicalway to proveone is straightcould plausibly affect

266. Diana Fuss, Inside/Out,in INSIDE/OUT:LESBIAN THEORIES, GAY THEORIES1, 1 (Diana Fussed., 1991). 267. "Don'task, don't tell" is a goodinstance of this dynamic as itmakes homosexuality un- speakableby naming it for the first time in the United States Code. See 10 U.S.C. ? 654 (1994). 268. D.A. Miller,Anal Rape, in INSIDE/OUT:LESBIAN THEORIES, GAY THEORIES,supra note 266,at 135. 269. See BarbaraJ. Flagg, "Was Blind, But Now I See": WhiteRace Consciousnessand the RequirementofDiscriminatory Intent, 91 MICH.L. REv. 953,957 (1993). As Flaggnotes: "The moststriking characteristic ofwhites' consciousness of whiteness is thatmost of the time we don't have any. I call thisthe transparency phenomenon: the tendency of whitesnot to thinkabout whiteness,or about norms, behaviors, experiences, orperspectives that are white-specific." Id. 270. See,e.g., JANIS BOHAN, PSYCHOLOGY AND SEXUAL ORIENTATION 31-34 (1996). 271. See EVE KOSOFSKYSEDGWICK, Privilege of Unknowing:Diderot's The Nun, in TENDENCIES23, 23 (1993)(discussing this "privilege of unknowing").

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 404 STANFORDLA WREVIEW [Vol. 52:353 evenone's informal orientation ascriptions. Indeed, the popular press, when linkingbisexuality with AIDS, oftensought to affectorientation ascriptions in exactlythis way, suggesting that women could not know for certain that theirhusbands were straight.272 Moreover, something closer to therigorous formof proof may be requiredin specialized contexts. As I willshow,273 this includesthe legal realm, a contextin whichorientation determinations have someof their most serious consequences. Bisexualitychallenges the integrity ofheterosexuality atthe collective as well as at theindividual level. Individualstraights wish to stabilizethe straight/gaybinary because they wish to locatethemselves as belongingto theheterosexual group. What membership inthat group means, however, is determinedat the collective rather than at theindividual level. As Naomi Mezeyhas argued, straights collectively attempt to preserve a "heterosexual ethic,"that is, an ethicthat heterosexuality hasthe monopoly on sexualvir- tue.274In orderfor the heterosexual ethic to present itself as congruentwith virtue,however, the homosexual ethic must be presentedas congruentwith vice.275The survival of the heterosexual ethic is thusdependent on a binary worldview of "right and wrong, of health and sickness, of heterosexual and homosexual."276 Mezeydescribes the martial rhetoric of homophobia as arisingout of this dependence.277She then astutely notes that the real enemy to the heterosexual ethicis notthe named enemy of homosexuality, butthe unnamed enemy ofbisexuality.278 By deconstructingthestraight/gay binary, bisexuality revealsthat it cannot be isomorphicwith the virtue/vice binary. It thusthreat- ensthe heterosexual ethic at thecollective level as wellas threateninghet- erosexualidentity atthe individual level.

3. Gay investment. Gays,as well as straights,have distinctiveinvestments in stabilizing theirsexual orientation. Again, their interest in suchstabilization is less intuitive.I earliernoted that gays, like straights, have a genericinterest in simply"knowing their place" in the social order, even if that place is a stig- matizedone. Butgays also have interests in identitystabilization that arise

272. See, e.g.,Richard A. Knox,Bisexuals Put Womenat Risk StudiesSay, BOSTON GLOBE, June20, 1990,at 57 (describingdifficulty of knowingwhether a "straight" man is actuallybisex- ual); JonNordheimer, AIDS SpecterforWomen: The Bisexual Man, N.Y. TIMES,Apr. 3, 1987,at 1 (similar). 273. See notes503-534 infra and accompanying text. 274. Mezey,supra note 8, at 112-21. 275. See id.at 116. 276. Id. atllS. 277. See id. 278. See id.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 405

notin spiteof, but because of, their stigmatized status. Insofar as thesein- vestmentsare relatedto stigma,they will be heldby gaysbut not by straights.These interests may also be brokeninto individual an'd collective interests. Theindividual gay interest in identitystabilization arises out of a desire to retainthe immutability defense. Gays often defend their homosexuality by characterizingit as an immutabletrait.279 This defense can sometimes mitigatethe stigma associated with homosexuality, both in theeyes of oth- ers280and in theeyes of gaysthemselves.281 Immutability has exonerative forcebecause of the widely held belief that it is abhorrentto penalize indi- vidualsfor matters beyond their control.282 That belief has found widespread expressioninAmerican anti-discrimination jurisprudence.283 Bisexualsthreaten the immutability defense. The reasoning here is not immediatelyobvious. Formally, bisexuality and immutability arenot neces- sarilyinconsistent. There could, after all, be fourimmutable categories- immutableheterosexuality, immutable homosexuality, immutable bisexuality,and immutable asexuality.284 Indeed, there might be an infinitenumber ofcategories on (and off) the sex-of-object choice spectrum, all ofwhich are immutable. Butas a practicalmatter, the introduction of bisexuals complicates the immutabilitydefense through a two-step process. First, once the bisexual possibilityis introduced, itbecomes impossible for the self-identified gayto provehe is gay,in precisely the same way it becomes impossible for the self- identifiedstraight to provehe is straight.285Even assumingdiscrete immuta- blecategories, itwill be unclearwhether the individual manifesting same-sex desirewho says he is immutablygay is immutablygay, or whetherhe is in truthimmutably bisexual. Second, to the extent that an individualis immu-

279. See JanetE. Halley,Sexual Orientation and the Politics of Biology: A Critiqueof the Ar- gumentfromImmutability, 46 STAN. L. REv.503, 507 (1994) [hereinafterHalley, Biology]. 280. See id. at 567 (notingpotency of immutabilityargument to persuadenon-gays of the humanityof gays);see also id. at 518 n.62(citing studies showing that people who think homosexualityis immutableare more likely to disapprove of discnrmination against homosexuals). 281. See id. at 535 (describingresearcher's elation at discoveringan allegedbiological basis forhomosexuality, because the researcher had "'always felt that [he] was born gay"'). 282. See,e.g., ESSAYS IN MORALRESPONSIBILITY (John M. Fischered., 1986) (collecting es- saysdescribing relationship between moral responsibility and free will). 283. See,e.g., Frontiero v. Richardson,411 U.S. 677,686 (1973) (pluralityopinion) (quoting Weberv. AetnaCas. & Sur.Co., 406 U.S. 164,175 (1972)) (notingthat legally burdening immutable characteristicsviolates "the basic concept of oursystem that legal burdens should bear some relationshipto individualresponsibility"). The SupremeCourt has calledthis reliance on immuta- bilityinto question in subsequentdictum. See Cityof Cleburnev. CleburneLiving Ctr., 473 U.S. 432,442-43 n.10 (1985). 284. Cf WEINBERGET AL.,supra note 113, at 4 (asking"[i]f homosexuality and heterosexu- alityare genetically predetermined, does that mean we needto look for a bisexualgene as well?"). 285. See notes259-273 supra and accompanying text.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 406 STANFORDLA WRE VIEW [Vol. 52:353 tablybisexual, his immutabilitywill generally not work to exoneratehim. This is becauseimmutability offers absolution by implyinga lack of choice.286But even an immutablebisexual is perceivedto have a choice-he can chooseto fitinto the heterosexual matrix by selectinga partnerof the oppositesex.287 The gayinvestment in bisexual erasure may thus arise in partbecause it problematizes the immutability defense. Suchan investmentcould be challengedon normativegrounds, as im- mutabilityhas the vices of its virtues. The virtue of the immutability defense is thatit forecloses the question of validity.288 In so doing,however, itcon- stantlydefers the normative debate that could establish the identity's validity.289Indeed, avid use of the immutability argument may be readas an im- plicitconcession of theidentity's invalidity-for if an identitywere truly perceivedto be valid,many would not ask ifit was immutable.290Just as immutabilitymoots the question of validity,so too does validitymoot the questionof immutability. As thegay rights movement progresses, some theorists are privileging thevalidity argument over the immutability argument, contending that being gayis validregardless of whether itis chosen.291This may in part be fueled

286. See Halley,Biology, supra note 279, at 518-19(noting that immutability operates as an exonerationstrategy because it eliminates choice). 287. See id. at 528 (notingthat the immutability theory "does not explain why bisexuals-by hypothesiscapable of satisfactorysexual encounters with members of the so-called'opposite' sex-shouldnot be encouragedor forcedto do so"). To be sure,one couldsay thathomosexuals also havethis choice. Buta socialdistinction appears to be drawnbetween the bisexual and the homosexualbased on theperceived sacrifice each must make to chooseto live as a heterosexual. The immutablebisexual, unlike the immutable homosexual, is notgiving up all sexualpleasure in conformingtoheterosexual norms of conduct, but merely giving up sexualpleasure with one of two sexes. See notes90-91 supra and accompanying text. A similarkind of logic can be seenin judicial treatments ofbilingual individuals. In Garciav. Gloor,618 F.2d264 (5thCir. 1980), for example, the Fifth Circuit considered a Title VII employ- mentdiscrimination claim brought by a bilingualMexican-American employee to challengehis formeremployer's prohibition on speakingSpanish on thejob. Thecourt noted that "[t]o a person who speaksonly one tongue... languagemight well be an immutablecharacteristic like skin color." Id. at 270. It thenwent on to note,however, that "the language a personwho is multi- lingualelects to speak at a particulartime is bydefinition a matter of choice." Id. (emphasisadded). Thus,the court reasoned that while people who do notpossess a favoredtrait may be protectedon groundsof immutability, people who possess both a favoredand a disfavoredtrait will never be so protected,because they may always choose to manifest only the favored trait. 288. See VERAWHISMAN, QUEER BY CHOICE 11-13 (1996). 289. See id. at 6 ("Ifwe argueagainst only the 'possible to change'assertion, we leaveun- challengedthe more insidious assumption that it is desirableor necessary to do so. Andto theex- tentthat homosexuality is acceptable only if it is notchosen it remainsstigmatized, illegitimate, deviant."). 290. Cf.BERSANI, supra note 232, at 57 ("[T]hevery question of 'how we got thatway' wouldin manyquarters not be askedif it were not assumed that we endedup thewrong way, the purposebehind the question has generallybeen to learnhow we mightbest go backand right the wrong."). 291. See WHISMAN,supra note 288, at 30-32.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 407 bythe fact that the immutability defense is notavailable for gays who do not experiencetheir orientation as immutable.292 (As conventionallytold, lesbi- ansare more likely than gay men to fallinto this group because of lesbianism'snexus to feminism.)293To the extent that the immutability strategy it threatensbecomes less important, bisexual visibility will become more palat- able. Indeed,those wishing to hasten the move from immutability arguments to validityarguments might embrace bisexual visibility. This is becausebi- sexuals,in contrastto homosexuals or heterosexuals, will always be seenas havinga choice.294As such,they literally embody the fact that individuals mightchoose (and choose to acknowledge) same-sex sexuality. As a positivematter, however, the day when gays (or at leastgay men) renouncethe immutabilitydefense appears distant. Even commentary deeplycritical of theuse of immutabilityin the courts recognizes that im- mutabilityremains a potent argument inthe political forum.295 The immutabilityargument is often the only effective strategy for gays "seeking to per- suadetheir parents, coworkers, and neighbors that they can love someone of thesame sex andremain fully human."296 Continued gay reliance on the immutabilityargument may lead to continuedgay reliance on bisexualerasure. Theother gay interest in stabilizinggay identity is a collectiveinterest in effectivepolitical mobilization. Even (or perhaps especially) gays who believe thatall individualsare somewhatbisexual may feel that bisexuals shouldnonetheless ally themselves with gays in orderto combathomopho- bia. Thisis in partbecause bisexuals are seen as flightrisks-individuals whocould at anytime abandon the gay community to lead straightlives.297 Andeven if they do notactually leave, bisexuals may be seenas less com- mittedto fighting heterosexual privilege because of their ability to partake of

292. See Halley,Biology, supra note 279, at 520. 293. See RUST,CHALLENGE, supra note 71, at 163(noting that some feminists see a connec- tionbetween lesbianism and feminism,relying on theassumption that "lesbianism is a possible choicefor all women"). 294. Whetherthis is actuallythe case is opento question. In theWeinberg study, a groupof bisexualswere asked: "Is itpossible that someday you could behave either exclusively homosexual or exclusivelyheterosexual?" WEINBERG ET AL., supranote 113, at 33. Approximatelyeighteen percentof bisexuals said "no." Id. at app.A at 317 tbl.3.2. Thisdoes not necessarily, but might, reflectthe fact that some bisexuals do notexperience themselves as beingable to restrict themselves to one sex,such that they do notexperience themselves as havinga "choice"between sexes. An- otherway of putting this is thatsome bisexuals may actually fit the stereotypical view of bisexuals whichholds that bisexuals feel intrinsically incomplete in sexualrelations with only one sex. 295. See Halley,Biology, supra note 279, at 567-68(criticizing reliance on immutabilityin legalcontext, but not in extra-legal political context). 296. Id. at567. 297. See Ochs,supra note 63, at 228-29(noting that "[m]any lesbians and gaymen believe thatbisexuals have less commitmentto 'the community,'and thatwhatever a lesbian or gay man mighthave to offer to their [sic] bisexual partner will not be enoughto outweigh the external bene- fitsoffered to those who are in heterosexual relationships").

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 408 STANFORDLA WRE VIEW [Vol. 52:353

it. Moreover,even if bisexuals appear to be deeplycommitted to fighting heterosexism,they can be seenas inherentlytraitorous, insofar as solidarity requiresnot just supportinga group but also fighting(as opposedto con- sortingwith) the group's enemies.298 Thisinterest in bisexualerasure is thusan interestin ensuringthat the lineof battle is clearlydrawn. The conflictbetween sexual minorities and thosewho oppress them is framedas isomorphicwith the conflict between gays and straights,thus makingthe battleline the virgulein the "straight/gay"binary.299 Precisely because gays view themselves as a belea- gueredminority, their tolerance for those who might sap theirpolitical soli- darityby blurring that battle line is limited.300 Thisinterest may be morepowerful for lesbians than for gay men. This mayseem counterintuitive, as I earlier implied that because lesbians may haveless invested in the immutability defense than gay men, bisexuals might notbe as threateningtothem along the axis of immutability.301 Butit may be exactlybecause lesbians are more likely to view their orientation as fluid that bisexualsmay be moredeeply threatening tolesbians on the axis of political mobilization.If biologyis believednot to providea substratefor lesbian identity,then the fear that any given lesbian might become bisexual and therebyabandon the lesbian community looms larger.302 Likethe investment inimmutability, thisinvestment inpolitical stability can be challenged.Destabilization actually has fourpro-gay deployments, whichI speculativelyorder here from least to most radical. First, the bisexual possibilitysuggests that the straight category is notas monolithicor se- cureas it mayoriginally seem. To theextent that bisexuals are waverers, gaysshould court, rather than suppress, bisexuals who are living as straights. Thisis especiallytrue if we acceptthe results of the sexuality studies consid- eredabove,303 which indicate that the addition of bisexualscould signifi- cantlyswell the ranks of homosexuals. Second, gays could deploy bisexuality'stendency to destabilizeheterosexuality to enlist straights in thefight

298. See id. at 230 (describinglesbian hostility to bisexuals,and noting that part of the objec- tionwas thatsome lesbians "'believed that only lesbians had an antipatriarchalsexuality, which meantthat only lesbians were working against sexism. . . [and]conflated sexual practice and political actionand believed that what one did in bed, and with whom, had direct consequences for sup- portingor dismantling a patriarchal power structure"' (citation omitted)). 299. See Mezey,supra note 8, at 115(noting that "in a discourseof militant hetero- and ho- mosexualitythe divisions are clear; one knows who the enemy is andhence one knows oneself"). 300. See id. at 118 (notingthat "[a]s an oppressedsubculture, homosexual communities rely on an ethicof homosexualidentity to maintainstrength and coherenceagainst the constant onslaughtof virulent [homophobic] attacks"). 301. See note293 supraand accompanying text. 302. Anotherway in whichlesbians may be morethreatened by bisexuals on theaxis ofpo- liticalmobilization concerns the interest lesbians have in combattingpatriarchy. I considerthis interestbelow. See notes345-352 infra and accompanying text. 303. See notes110-176 supra and accompanying text.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 409 againstthe regulation of homoerotic conduct. Straights may be lesslikely to burdensuch minorities ifthey are worried about drifting into the stigmatized categorythemselves.304 Third, bisexuals could be seenas performinga me- diatingor conciliatoryoffice. If bisexuals are being erased in orderto pre- servethe binary logic of agonisticconflict, it seemsintuitive that bisexual visibilitymight challenge the paradigm of conflict itself.305 Finally, bisexuals couldbe deployednot just as a thirdparty alongside gays and straights, butas a categorythat permits gays to challengeany sexual orientation categorization.Something akin to this insight is neatlycaptured in thebisexual retortto being called a fencesitter-namely,"your fence is sittingon me."306 In thisformulation, it is notfencesitting, butthe fence, that is theproblem; andit is notbisexuality, but the line establishing binary categorization, that needsto be erased. The lastpossibility merits further discussion. A utopianversion of the argumentwould use bisexuality as a meansof retiring all sexualorientation classifications.307Theargument that the prevailing orientation classifications shouldbe retiredshould be a familiarone to gays,as it is an argumentthey espousedeven before the patterned visibility of bisexuality.308 Atthe begin- ningof the 1970s, gay liberationists "prophesied the disappearance of both 'thehomosexual' and 'the heterosexual' through the abolition of constraining categories."309 One theoryof whythis movement to retirecategories ulti- matelydid not reach fruition is that it was difficult to create a politicsout of it.310 Activistswere faced with the perennial problem-how does one resist a categorizationwithout organizing around the oppressed category?3 1 And

304. Cf JOHNHART ELY, DEMOCRACYAND DISTRUST 160 (1980) (describingage as a cate- goryon whichdiscrimination is less likelyto occurbecause individuals understand that they will "drift"into the stigmatized category of old age). 305. See Highleyman,supra note 12, at 90 (notingthat "[m]any bisexuals object in principle tothe gay and lesbian movement's 'us' versus'them' paradigm"). 306. RebeccaKaplan, Your Fence Is Sittingon Me: TheHazards of BinaryThinking, in BISEXUALPOLITICS, supra note 7, at267, 267. 307. See Mezey,supra note 8, at99. 308. See AmandaUdis-Kessler, Identity/Politics: A History of theBise-xual Movement, in BISEXUALPOLITICS, supra note 7, at 17, 19 (notingthat "[t]he word 'bisexual' existed before Stonewall,and there were people who lived as bisexuals,even famous ones at times"but that "the focuson bisexualityas a coreaspect of one's identitydoes not seem to havearisen in anypatterned wayuntil after Stonewall"). 309. StevenEpstein, Gay Politics, Ethnic Identity: The Limits of Social Constructionism,in FORMS OF DESIRE: SEXUAL ORIENTATIONAND THE SOCIAL CONSTRUCTIONISTCONTROVERSY 239,252 (EdwardStein ed., 1990). Accordingto Epstein,this post-Stonewall activism was a de- parturefrom homosexual politics in thepreceding decades, which stressed the goal of integration withoutchallenging the straight/gay dichotomy. See id. 310. See id. at253-54. 311. See id. at 254 ("Thisis a familiardilemma, and one thatis byno meanspeculiar to the gaymovement: How do youprotest a socially imposed categorization, except by organizing around thecategory? Just as blackscannot fight the arbitrariness ofracial classification without organizing

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 410 STANFORDLAWREVIEW [Vol. 52:353 howcan one organize around the category without reifying the very catego- rizationone seeks to retire? It is worthrevisiting these questions, as bisexualityprovides an ingen- iousanswer to them.This is becausebisexuality may be "notjust another sexualorientation but [also] a sexualitythat undoes sexual orientation as a category."312One paradox is answeredby another-the general paradox that one can onlychallenge a classificationthrough classes that reify it is answeredby the specific paradox that bisexuality is a classthat challenges the veryclassification that creates it. Preciselybecause bisexuality no sooner assertsitself as a categorythan it questions its own boundaries, bisexuality is arguablythe post-Stonewall activist's dream.313 Again,however, these normative objections may not sway many gays whohold this particular investment in bisexual erasure. Gays face signifi- cantpolitical oppression under the prevailing straight/gay paradigm. Given this,many gays feel that the energies of the movement should be devotedto arguingwithin, rather than against, that paradigm. This is particularlytrue sincegays have used that paradigm to createan effectivepolitics of opposition.

B. BisexualityDestabilizes the Primacy of Sex

Thesecond investment straights and gays have in bisexual erasure is that bisexualsare seento destabilizethe primacy of sex as a diacriticalaxis. Straightsand gays have a sharedinvestment in the primacy of sex because theirorientation identities rely on it. Straightshave a peculiarinvestment in theprimacy of sex becausesex identitiesare currentlydetermined by a straightmatrix. And gays have a peculiarinvestment because homosexuality is sometimesdeployed as a meansof sex separatism, which both reflects and reinforcesthe primacy of sex.

as blacks,so gayscould not advocate the overthrow of thesexual order without making their gay- nessthe very basis of their claims."). 312. GARBERsupra note 1 1, at 65; see also Mezey,supra note 8, at 99 ("Bisexualityas a critiqueof thehetero/homo paradigm actually facilitates [a] moreradical analysis, which concludes, ironically,that bisexuality works no betterthan the other two categories in accuratelydescribing concretesexual behavior, and that a newconceptualization of sexual identities, such as one based on acts,is needed."). 313. Oris it? Onecould contend that the conundrum is not so easilysolved, for bisexuality's instability,which prevents it fromreifying the categorization, also arguablymakes it a politically ineffectivecategory. Under this view, bisexuality offers the worst rather than the best of both worlds,being too inchoateto be an effectiveclass andtoo rigid truly to contestthe reification of orientationcategories.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 411

1. Sharedinvestment.

Bisexualitydisrupts the primacy of sex as a diacriticaltrait. One disqui- etingaspect of the bisexual is thatshe refuses to require that an objectof de- sirebelong to a particularsex. In thestereotypical view, this is becausethe bisexualis sex-blind,314refusing to distinguishat all betweenmen and womenin her erotic life.315 Under this characterization, theworld the bisexual sees is likeFoucault's "world in whichgrins h[a]ng about without the cat."316 In thisworld, sexual pleasure becomes a (literally)floating signifier, a sexlesssmile left by a fadingbody. Thestereotypical view appears to misdescribe the majority of bisexuals. Somebisexuals claim to be sex-blindin the sense that they fall in love with a personrather than with a sex.317But most appear to desiremen as menand todesire women as women.Thus in the Weinberg study, over four-fifths of thebisexuals interviewed distinguished between being sexual with a manand beingsexual with a woman,citing behavioral, bodily, and emotional differences.318That finding comports with common sense-sex distinctions are so pervasivein contemporaryAmerican society that it is hardto believethat bisexualscould evade such distinctions.

314. By "sex-blind,"I mean that a persondoes not take sex intoaccount in choosingerotic partners.Following Neil Gotanda'sanalysis of colorblindness,I distinguish between literal and figurativeforms of sex-blindness.See Neil Gotaida,A Critiqueof "Our ConstitutionIs Color- Blind,"44 STAN.L. REv. 1, 18-19(1991). The literalform posits that the person does noteven registerother people's sexes. Literalsex-blindness is like medicalcolor-blindness-just as the personwho has red-greencolorblindness cannot tell the difference between the two colors, so too does theperson who has literalsex-blindness find it impossibleto tellthe difference between the sexes. Cf id. The figurativeform, in contrast,notes that the person registers other people's sexes, butattaches no consequenceto thatperception. Figurative sex-blindness is likejuridical colorblindness.Just as thosewho espouse the colorblindness ideal say that they recognize differences in colorsbut attach no salienceto those differences, so too do thoseespousing figurative sex-blindness saythat they recognize differences between sexes but attach no salienceto thosedifferences. In my reviewof theliterature on bisexuality,I have not come across any instances of individualswho claim thatthey are literallysex-blind. By "sex-blindness"I therefore mean figurativesex- blindness. 315. Thisis theview that underlies the fears surrounding the "bisexual harassment exemption"discussed below. See notes464-474 infra and accompanying text. 316. Michel Foucault, Introduction to HERCULINE BARBIN, BEING THE RECENTLY DISCOVEREDMEMOIRS OF A NINETEENTH-CENTURYFRENCH HERMAPHRODITE vii, xiii (Michel Foucaulted., Richard McDougall trans., 1980). Foucaultdescribes this world as theworld seen by thehermaphrodite, butit seems more aptly the (stereotypical) view of the bisexual's world. 317. See COLKER,HYBRID, supra note 9, at 27 (describingbisexuals who state "'that when theyfell in loveit was witha personrather than a gender"')(quoting BETTY FAIRCHILD& NANCY HAYWARD,NoW THAT You KNOW: WHAT EVERY PARENTSHOULD KNow ABOUT HOMO- SEXUALITY75 (1989)); PaulaC. Rust,Who Are We and WhereDo WeGo from Here? Conceptual- izingBisexuality, in CLOSERTO HOME: BISEXUALITY& FEMINISM281, 298 (ElizabethReba Weise ed., 1992) (describingattitudes of lesbian women towardbisexuality ranging from "bisexuals are indiscriminate"to "bisexuals fallin love witha person,not a gender"). 318. See WEINBERGET AL.,supra note 113, at 50-53.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 412 STANFORDLAWREVIEW [Vol. 52:353

Despitethe fact that "sex-blind" bisexuals may comprise only a small minorityofall bisexuals,bisexuality still presents a deep threat to sex norms. Evenif bisexualsare not completely sex-blind, they still do notpresump- tivelyeliminate one sex from their fields of erotic possibility. For all bisexuals,whether sex-blind or not, sex matters less.319 Whyis a challengeto the priority of sexso threateningtomonosexuals? Thereappear to be twoanswers-challenging theprimacy of sex destabilizes monosexualidentity and underscores a tension between public and private attitudestoward sex.

a. Destabilization.

Challengingthe primacy of sex destabilizesnot only monosexual identity,but human identity. Judith Butler posits that in contemporary American culture,we arenot viewed as humanuntil we have a sex,that "the moment in whichan infantbecomes humanized is whenthe question, 'is it a boyor girl?'is answered."320Thus, "[t]hose bodily figures who do notfit into ei- thergender fall outside the human, indeed, constitute the domain of the de- humanizedand the abject against which the human itself is constituted."321 SuzanneKessler's study of theintersexed corroborates Butler's claim.322 Kesslerdescribes how the medical profession preserves the existing binary at all coststhrough surgical "reconstruction" shortly after the birth of the inter- sexual323-literallydoctoring the data to fitthe binary sex hypothesis. Even in situationswhere the intersexed infant is healthy,324thesurgery is castas a "neonatalpsychosexual medical emergency,"325 supporting the claim that the infant'sstatus as a humanbeing is endangeredby the mere fact of intersexu- alityitself. In sucha culture,not to possess a sexis notto be human.

319. MarorieGarber recounts the story of a woman,"Elizabeth," who fell in love with someone on theinternet who was purportedto be a man. See GARBER,supra note 1 1, at 33-34.When theyfinally met and the "man" revealed herself to be a woman,Elizabeth overcame her initial re- sistanceover the course of the hour's conversation and then made love to her. See id. at34. Garber notesthat Elizabeth does notself-identity as a bisexual,id., and she certainlywould not seem to qualifyas a "sex-blind"bisexual. Yet in herrefusal to permit sex of object choice to be "theover- ridingconsideration," id., she perhaps represents the anxiety of the bisexual possibility. 320. BUTLER,supra note 24, at 111. Linguisticsupport for this claim can be foundin the fact thatthe infant who is nota "he" or a "she"must be an "it,"a pronounreserved for nonhuman subjects. 321. Id. 322. See generallyKESSLER, supra note 18. 323. See id.at 12-32. 324. Kesslerdistinguishes between three kinds of genitalsurgery-(1) that which is lifesav- ing;(2) thatwhich improves the quality of life;and (3) thatwhich is aesthetic.See id. at 34. By "healthy"I mean a childwho does not require the first two kinds of surgery. 325. Id.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 413

Butif this is true,does bisexuality, which stereotypically does not pass desirethrough the lens of sex, threaten human identity? Much more would needto be saidbefore answering in theaffirmative. Even without that discussion,however, it can be notedthat bisexuality, like intersexuality, suggeststhat the question "Is it a boyor a girl?"is thewrong question to be asking.And if this is thequestion that determines our humanity, it should comeas no surprisethat the capacity of bisexuals to undermine the sex cate- goryis deeplythreatening toindividuals of all categories.326 Butwhile the bisexual's ability to underminesex categoriesis generi- callytroubling for all individualsin contemporaryAmerican society, it is arguablydistinctively troubling for monosexuals. This is becausemonosex- uals,unlike bisexuals, define their orientation identities according to that distinction-thatis,by their attraction tomen but not women, or vice versa. Anindividual's orientation can be affectednot only by challenging the fixity ofhis orientation, but also by challenging the fixity of the sex ofhis object choice. Withouta clearand privilegeddistinction between "man" and "woman,"there is no clearand privileged distinction between "straight" and "gay."In theprevious analysis, we sawthat bisexuality directly challenged sexualorientation categories by destabilizing them.327 Here we see thatbi- sexualityindirectly challenges sexual orientation bydestabilizing sex.

b. Thetension between public and private treatments of sex. Thereis a secondreason why monosexuals might be peculiarlydisturbed bythe ability of bisexuals to callthe primacy of sex into question. This re- latesto a tensionin how sex is treatedin the erotic and nonerotic realms. It is notquite true, as I saidbefore, that it is alwaysthreatening to suggest that the"Is ita boyor a girl?"question is thewrong question. In thenonerotic realm,this question is widelymarked as immoral.Thus, American antidis- criminationlaw prohibits the state and many employers from attaching consequenceto the answer to this question.328 To someextent these norms carry overinto the nondesirous private realm-those who discriminate onthe basis of sex in theirprivate affiliations are increasingly subjected to moral,if not legal,censure.329 But when we enterthe erotic realm, the norms shift dra-

326. Thatbisexuals may also be threatenedby bisexuality'scapacity to undermine sex catego- riesmay be seenin theinsistence of many bisexuals, as notedabove, that they are attracted to men as menand women as women.See note318 supra and accompanying text. 327. See notes259-313 supra and accompanying text. 328. See generallyCASES AND MATERIALSON SEX-BASED DISCRIMINATION(Herma Hill Kay& MarthaS. Westeds., 4th ed. 1996). 329. See, e.g.,Harvard Woman Sues Male 'Flies,' S.F. CHRON.,Mar. 3, 1988,at B6 (de- scribinglegal challenge to all-malestudent clubs); Andrea Estes & ErinN. O'Leary,Judge Orders Clubto Put Womenon Par withMen, BOSTON HERALD, Dec. 2, 1998,at 4 (describingcourt order mandatingsex integrationof tournamentsat Massachusetts golf club); PatriciaWen, Single-Sex

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 414 STANFORDLAWREVIEW [Vol. 52:353 maticallyin theother direction, such that sex distinctionsare not only permitted,but expected. Monosexuals routinely discriminate onthe basis of sex in choosingtheir erotic partners; indeed, that practice constitutes them as monosexuals. Thereis thusa fundamentaltension in the treatment ofsex in the nonde- sirousand desirous realms. If I discriminatebetween male and female em- ployeesin theworkplace, I can expect in manyquarters to be criticizedfor politicalinsensitivity. But if I discriminatebetween male and female part- nersin my erotic attachments, I am much less likely to encountersuch criticism. Indeed,to the extent that I discriminatein favor of a same-sexerotic partner,itis mycritic who can expect (in somequarters) to be criticizedfor politicalinsensitivity. Thistension requires justification, not only because of itsfacial inconsistency,but also becauseof its deep consequentiality. Specifically, accep- tanceof discrimination inthe desirous private realm may render it impossible to fullyreject discrimination inthe public realm. In thelaw, this problem is perhapsbest seen in thejudicial refusal to mandatethat legal distinctions betweenthe sexes be completelyabolished. Thus,while the Supreme Court'sequal protection jurisprudence subjects race-based classifications to strictscrutiny,330 it subjects sex-based classifications only to intermediate scrutiny.331Even as theactual nature of intermediatescrutiny has tacked closerto strictscrutiny,332 theSupreme Court has retained the rhetorical distinctionbetween race-based and sex-based classifications.333 We have thus notbeen able to push jurisprudential norms against sex discrimination as far as we havepushed norms against race discrimination. Whilecommentators have justifiedthis distinction in a numberof ways,334the justification most pertinent to thisanalysis relates to desire. GeorgeRutherglen explains the lower tier of scrutiny associated with sex by notingthat sex-based "classifications arevalued for their own sake within an

HealthClubs Get Protection, BOSTON GLOBE, Feb. 7, 1998,at Bi (notingNOW's oppositionto billlegalizing single-sex health facilities). 330. See, e.g.,Bush v. Vera,517 U.S. 952,952 (1996); AdarandConstructors, Inc. v. Pena, 515 U.S. 200,201 (1995). 331. See, e.g.,United States v. Virginia,518 U.S. 515,515 (1996); J.E.B.v. Alabamaex rel. T.B.,511 U.S. 127,127 (1994). 332. See Cass R. Sunstein,The Supreme Court, 1995 Term-Foreword:Leaving Things Un- decided,110 HARV. L. REv.4, 75 (1996) (notingthat the Court in Virginia"did not merely restate theintermediate scrutiny test but pressed it closer to strict scrutiny"). 333. Cf Virginia,518 U.S. 515. 334. Forexample, they have noted that women, unlike blacks, are not a minority,see JOHN HARTELY, DEMOCRACYAND DISTRUST164 (1980); thatwomen, unlike blacks, have historically beenintegrated with the dominant group, see id.; and thatwomen, unlike blacks, exhibit "real" physicaldifferences that distinguish them from the group to whichthey have been historically sub- ordinated,see SylviaA. Law,Rethinking Sex and theConstitution, 132 U. PA. L. REv. 955, 1007 (1984).

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 415 importantarea of human life, namely sexual conduct, whereas racial classifi- cationsare neither so widelynor so openlyvalued in everydaylife."335 Sex is differentfrom race insofar as wenot only permit, but expect, individuals to havedesires vectored toward only one sex. The extentto whichthe monosexual presumption in the desirous realm blocksan advanceto a sex-blindregime in the nondesirous realm should not be underestimated.Our erotic relationships, after all, are oftenviewed as simultaneouslyconstituting and reflecting our most important emotional attachments.If we routinelystructure these crucial relationships by discrimi- natingon the basis of sex, it should come as no surprisethat we (andour ju- dicialinstitutions) experience a failure of nerve when asked to categorically abolishall sex-baseddistinctions. Monosexualscould defend the distinction between public sex-blindness andprivate sex-consciousness ina varietyof ways.336 I do notseek to evalu- atethese rationales here, as mypoint is simplythat bisexuals may occasion anxietybecause they hold out the conceptual possibility that these rationales arenot dispositive. Bisexuality implies that sex need not be as importantin ourdesirous lives as we havemade it.337 Bisexuals and asexuals are thus the onlyindividuals who at leasthave the capacity not to discriminateon the basisof sex in any aspect of their lives.338 As such,they have the potential to evadethe public/private tension with regard to sex in whichmonosexuals findthemselves.

2. Straightinvestment. Straightshave a distinctiveinvestment inbisexual erasure relating to the primacyof sex. Thisis becausesex is currentlyunderstood through a heterosexualmatrix; that is, straightshave a monopolyon sexnorms. And bisexuality,unlike homosexuality, hasthe potential to disrupt that monopoly.

335. GeorgeRutherglen, Sexual Equality in Fringe-BenefitPlans, 65 VA. L. REv. 199,209 (1979). 336. First,they could point to the fact that they experience their orientations tobe immutable, suchthat the desire is beyondtheir control. Second, they could rely on autonomyarguments relat- ingto theirright to discriminatein an immenselyintimate and consequential sphere of theirlife. Third,they could contend that their private norms of sex-discriminationrely on "realdifferences" betweenmen and women, which are recognized as permissibleeven by public anti-discrimination norms.Fourth, they could contend that differentiation between the sexes fordesirous purposes shouldnot be considered"discrimination" because it does not have subordinating effects. There are doubtlessmany other possible rationales. 337. See COLKER,HYBRID supranote 9, at 30 (notingthat "embracing the category of bi- sexualitywould help society recognize that one can findan organizingprinciple other than biological sexto define sexual attractiveness"). 338. I thankIan Ayresfor this point.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 416 STANFORDLA WRE VIEW [Vol. 52:353

To see this,begin with the increasingly accepted view that we all performour sex to somedegree.339 Sex, no lessthan any other performance, is alwaysshaped by its audience. When we performour sexes, for whom are we performing?The answer is at leastpartially that we areperforming for thosewho might potentially desire us, and that the performance is an attempt to convertthat potentiality into an actuality.The performanceof sex will thusalways be affectedby prevailing codes of desire. When (as now)heterosexualityis the prevailing code of desire,women will be encouragedto performtheir sex in a waythat is attractivetomen (and vice versa, although thesymmetry is not complete). To be a "woman"is tobe attractivetomen, tobe a "man"is tobe attractivetowomen. 340 Itis thereforean errorto accede to the conventional wisdom that sex is a stable,prediscursive substrate from which heterosexuality is a simpleback formation.341While heterosexuality is a back formation from sex, it is also paradoxicallytrue that sex is a backformation from heterosexuality. Current normsof sex andcurrent norms of heterosexuality arethus implicated in a feedbackloop in which each shores up the other. Homosexualitydoes not present much of a challengeto prevailingsex normsin thisregard. Even if men begin to lookwith desire at othermen, thismay not significantly change sex performance, as that desirous look is readas thefamiliar look of theheterosexual woman now displacedin a man'sbody. This is themodem vestige of the older trope of the homosexual as "invert"-a"woman trapped inside a man'sbody" or viceversa.342 The logicof the invert is themeans through which homosexuality is read back intoa straightparadigm.343

339. See, e.g., BUTLER,supra note24, at 24-25. While Butlerrefers to genderrather than sex in thisquotation, she elsewherehypothesizes that there is no real distinctionbetween gender and sex. See id. at 7 (describingthe distinction as "no distinctionat all"). 340. See DUNCANKENNEDY, SEXY DRESSING,ETC. 161 (1993) ("So long as I am a straight man, a partof my being is hostageto women: I wantthem to exist as women,not men,as bearers of thepossibility of myown sexual excitement."). 341. See note 194 supra and accompanyingtext. 342. See TimEdwards, Beyond Sex and Gender:Masculinity, Homosexuality and Social The- ory,in MEN,MASCULINITIES & SOCIALTHEORY 110, 112 (JeffHearn & David Morganeds., 1990) (notingcharacterization of male homosexualsas thosewho sufferfrom having a "femininesoul in a male body"). 343. As Bersaninotes, sometimes the infliction of the male gaze on a male performeris read back intothe heterosexualmatrix by invertingthe sex of the performerrather than that of the ob- server: The New YorkTimes reported on April3, 1993,that a radarinstructor who chose not to fly withan openlygay sailor,Keith Meinhold, feared that Meinhold's "presence in the cockpit woulddistract him from his responsibilities." The instructor"compared his 'shock' at learning therewas a gay sailorin his midstto a womandiscovering 'a man in theladies' restroom."' Note thecurious scatological transsexualism in ourradar instructor's (let us hopemomentary) identificationof his cockpitwith a ladies' restroom.In thisstrange scenario, the potential gay attackerbecomes the male intruderon femaleprivacy, and the "original"straight man is

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 417

Thebisexual, on theother hand, deeply destabilizes sex norms because herdesirous look cannotbe as easilycollapsed into a straightparadigm. Howdoes a manmake a bisexualdesire him? Not by adopting aspects that havetraditionally been described as "masculine"as thebisexual is also po- tentiallyattracted to the"feminine." The logicof inversionis insufficient hereto protecttraditional sex normsfrom bisexuality because bisexuality doesnot simply invert cross-sex desire, but rather supplements itwith same- sexdesire. Bisexuality, then, contests heterosexuality's current monopoly on sexperformance ina waythat homosexuality cannot.

3. Gay investment.

Gays,as wellas straights,have investments inpreserving the primacy of sex as a diacriticaltrait. Unlike straights, gays can deploy their homosexu- alityto engagein a sex separatismthat is notonly social, but also erotic. Bisexualsthreaten this sex separatismby creating a bridgeto theopposite sex. All gayscan engage in a moretotal form of sex separatism than is possible fornoncelibate heterosexuals. Gays have at leastthe capacity to livea fullerotic (as wellas a socialand a political)life in a communityfrom which theother sex is completelyexcluded. Thus, at least along the axis of separatism,lesbians and gay men are just as, ifnot more, likely than straights to privilegethe importance ofsex as a diacriticaltrait.344 Thebest example of gay sex separatism is when women's political solidarity(feminism) is hypothesizedto be connectedto theirsame-sex erotic attachment(lesbianism). As articulatedin theold slogan"Feminism is the theory,lesbianism is thepractice,"345 theview is thatthe authenticity of a

metamorphosed,through another man's imagined sexual attention, into the offended, harassed, or evenviolated woman. BERSANI,supra note 232, at 16-17(citations omitted). 344. Thisinsight may be somewhatcounterintuitive, given that homosexuality is often seen as underminingthe importance of sex. It is of coursetrue that homosexuality is often seen as sub- vertingsex expectations,given that many stereotypically deem gay men to be less masculineand lesbiansto be less femininethan their straight counterparts. It is also truethat homosexuality can be seenas diminishingcertain forms of hierarchy between the sexes to the extent, for example, that lesbianismcan be seenas empoweringfeminism. But diminishing sex-based hierarchy and dimin- ishingsex-based consciousness are twodifferent things, and I wouldcontend that homosexuality cando theformer not in spiteof, but because of, the fact that it does not do thelatter. 345. The slogan"Feminism is thetheory, lesbianism is thepractice," is usuallyattributed to Ti-GraceAtkinson. See, e.g., NancyChater & LilithFinkler, "Traversing Wide Territories": A Journeyfrom Lesbianism to Bisexuality,in PLURAL DESIRES: WRITINGBISEXUAL WOMEN'S REALITIES14, 15 (Leela Acharya,Nancy Chater, Dionne Falconer, Sharon Lewis, Leanna McLen- nan& SusanNosov eds., 1995) (attribution ofquotation to Atkinson by Finkler). As NancyChater notes,however, the original quote was actually"Feminism is a theory,lesbianism is a practice."Id. at 36 n.1; see also GARBER,supra note 1 1, at 44 (quotingAtkinson correctly as saying"Feminism is a theory,lesbianism is a practice").As Chaternotes, the alteration in theslogan is "subtlebut

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 418 STANFORDLAWRE VIEW [Vol.52:353 politicsis determinedbyhow it plays out in erotic practice. While this is an extremeposition, one way of reading it is tonote that, as AdrienneRich has famouslyargued, this distinction between political and eroticsame-sex at- tachmentsmay itself falsely binarize a continuumof practicesin which womenidentify with other women.346 Likelesbians, gay men can also createcommunities that are totally sex segregated.Indeed, given the dominance of men in almostevery sphere of publiclife (perhaps most pertinently employment) itis likelyto be easierfor a gayman than for a lesbianto live a lifethat entirely excludes those of the othersex. Butwhile gay men have more opportunity toengage in sexsepa- ratism,they may have less motive to do so. Forif "Masculinism is the theory,gay maleness is thepractice," does not trip off the tongue, it must be in partbecause men as a groupare not disempowered.347 Lesbians may draw synergisticconnections between their disempowered status as womenand theirdisempowered status as homosexuals.In contrast,the gay man's privi- legedstatus as a mancan be experiencedas conflictingwith his stigmatized statusas a homosexual.Patriarchy makes the social meaning of lesbian separatismdifferent from the social meaning of gay male separatism. Thisdifference inturn may make bisexuals more threatening to lesbians thanthey are to gaymen. If homosexualitypermits a dreamof a social worldwithout the other sex, bisexuality disrupts that dream by constantly holdingout the possibility of cross-sex attachment. But the awakening may be ruderfor lesbians than for gay men, precisely because cross-sex attach- mentmay signify capture by patriarchy. Oneway of understanding thisis toconsider so-called lesbian pornography,that is, the "[p]ortrayals ofallegedly lesbian 'scenes"' that "are a staple of heterosexualpomography."348 It might be askedwhy the straight male consumersof this pornography find the scene so titillating,as it could actu- allysignify an eroticworld in whichthere is no needfor men.349 Many of thepossible answers cluster around the power of the male gaze to override crucial,"Chater, supra at 36 n.1, arguablyshading an ostensiblyneutral distinction between feminismand lesbianism into a hierarchicalformulation favoring the latter. 346. See AdrienneRich, Compulsory Heterosexuality and LesbianExistence, in POWERSOF DESIRE: THE POLITICSOF SEXUALITY177-205 (AnnSnitow, Christine Stansell & SharonThomp- soneds., 1983). Thisbreakdown of the distinction between the political and the personal may par- tiallyexplain why lesbians are less likelythan gay men to deployimmutability as an etiological explanationfor their homosexuality. Whether itis inactuality chosen or not, lesbians may feel as if theirsame-sex attachments are moreconstructed than essential, given that such attachments are politicalas wellas erotic. 347. See SEDGWICK,supra note 4, at36 (describingthis position). 348. JohnStoltenberg, Pornography, Homophobia and Male Supremacy,in PORNOGRAPHY: WOMEN,VIOLENCE, & CIVIL LIBERTIES145, 155 (CatherineItzin ed., 1992). 349. See Rich,supra note 346, at 187 (describingthe male fearthat "women could be indif- ferentto themaltogether, that men could be allowedsexual and emotional-therefore economic- accessto women only on women'sterms").

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 419 anyintimation that the women are unavailable to straightmen.350 If, for example,the male spectator believes that the women are both available to him, theirnumerosity enlivens rather than excludes: He sees two odalisques ratherthan one couple. The sameinability to imaginea womanwho does notdesire men may lead to thebelief that the absence of men in thesexual scenesignifies that he himselfis thatman, that the invisibility ofmen in the mise en scene is theinvisibility of subjectposition. Yet again,the spectator maybelieve that the scene depicts women out of controlof their sexuality, suchthat a man'sentry would lead them to directgreater attention toward him.Undergirding all of the above interpretations is the assumption that the malegaze is so powerfulthat there is no femalebody that is notultimately arrayedfor male delectation. And this in turn means that there is no woman whodoes not care about the male gaze; there is no womanwho does not desiremen. One mighttherefore think that the male gaze readsall womeninto straightwomen. But the denomination ofthe scene as "lesbian"pornography,as well as thescene's surface depiction of a woman'sdesire for a woman,would seem to undercut this. I thinkwe mustentertain the possibilitythat part of theappeal of these scenes lies in themale gaze's abilityto overridethe desire that women have for each other. In otherwords, even whenwomen are acknowledged to desire each other, that desire is nonetheless onlythere to be trumpedby the desire that men have for them. In the contextof "lesbian" pornography, themale gaze may be an opticwhere sex- ismand heterosexism are powerfully fused-men sexually conquer women at thesame time as heterosexualityconquers homosexuality. Under this reading,the charge of the scene derives from the man's triumph over both feminismand lesbianism,sexual conquest is politicalconquest. For that chargeto exist, then, the women cannot be straight,as the charge depends on an overriddensame-sex desire. Butthis means that the "lesbians" in the pornography are beingread not as straightwomen, but as "bisexuals."351Straights in thiscontext may thus be usingprecisely the set of analyticmoves described earlier, in whichbi- sexualitydestroys the ability of monosexuals to provethat they are monosexual.352No matterhow ardently the women express their desire for each other,they are unable to exorcisethe possibility raised by the male gaze, in whichthat desire not only coexists with, but is supersededby, a desirefor

350. See ANDREADwoRKJN, PORNOGRAPHY: MEN POSSESSINGWOMEN 47 (1989) (describing"lesbian" pornography as "the ultimate tribute to male power: the male is notin theroom, yetthe women are there for his pleasure"). 351. Thismight be seenas an anomalousway in whichstraight males make bisexuals more, ratherthan less, visible. Whatis curious,however, is thatthey do notmake bisexuals visible as such-recallthat the pornography is still viewed as "lesbian"pornography, rather than as "bisexual"pornography. 352. See notes259-263 supra and accompanying text.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 420 STANFORDLA WREVIEW [Vol. 52:353 men. Bisexualityin thisformulation may threaten lesbians as thetrope throughwhich women's attachments to women are imaginatively destroyed andfalsified. Indeed, if I amright, the destruction and falsification are what inpart stimulate male desire. The analysismay be generalizedbeyond pornography-bisexuals may be threateningtolesbians as a culturaltrope for the phantasmatic subversion oftheir same-sex desire by straight men. Patriarchy, here especially the male gaze,may thus make bisexuals more threatening along this axis to lesbians thanto gay men.

C. BisexualityDestabilizes Norms of Monogamy Evenin its nomenclature, bisexuality is framed as excess. To be "bi"is to be doubled,353to be "mono"is to be one.354To defamiliarizethese formulations,we mightreturn to the Freudian conception of originary bisexuality,which frames bisexuality as an prelapsarianwholeness from which we fallaway.355 Rather than denominating the monosexual as wholeand the bisexualas surfeit,we mightdenominate the bisexual as wholeand the mon- osexualas fragment.Viewed in this light, the flaw might be seennot as the bisexual'sexcess, but rather the monosexual's lack.356 Thatwe do notview things in this light can be seennot only in the way we namebisexuals, but also in the way that we otherwiseportray them. Bi- sexualdesire is seennot as a completion,but rather as an excess. Thisis perhapsmost clearly seen in the stereotype ofbisexuals as promiscuous.357 It shouldtherefore not surprise that the final investment straights and gayshave in bisexualerasure stems from their perception that bisexuality threatensnonns of monogamy.The investmentin thatnonn shared by

353. See 2 OED, supranote 9, at 165. 354. See 9 id.at 1008-09. 355. See note92 supraand accompanying text. 356. A similarparadigm shift was madein TheSymposium with regard to sex. See PLATO, SYMPOSIUM41-45 (Benjamin Jowett trans., Hayden Pellicia rev., 1996). ThereAristophanes re- countsthe story of how each man and woman was originallyconjoined at theback to anotherman orwoman. As punishmentfor arrogance, each man and woman was cutaway from his or her other half.This is whymen and women now go aboutsearching for other men and women, in thehopes thatby findingtheir lost other halves, they can regaintheir prelapsarian wholeness. Again, this challengesour view thatindividual men and womenare wholeunits and thatpaired men and womenare doubled units. It insteadencourages us to viewthe couple as theunit and individual menand womenas incompletehalves of thatunit. Thatview, of course,has morecredence in contemporaryAmerican culture (think of descriptionsof couplesas "unions"or descriptionsof spousesas "betterhalves") than the view that bisexuals represent a desired wholeness. 357. See WEINBERGET AL., supranote 113, at 69. The self-conceptionof some bisexual groupsdemonstrates the strengths of thisstereotype. The Washington,D.C. Bi-Women's(and Men's)Network, for example, was formedfor "more traditionally oriented bis intomonogamous or duogamousrelationships only." See PoliticalActivism: A BriefHistory, in BI ANY OTHERNAME, supranote 257, at 359, 361.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 421 straightsand gaysis thesexual jealousy both groups experience in nonmonogamous(or potentiallynonmonogamous) relationships. The distinctivelystraight investment relates to theperception that bisexual nonmonog- amyleads to the transmission ofHIV fromthe gay community tothe straight community.The distinctivelygay investment relates to thedesire of some gaysto evade the portrayal ofgays as promiscuous.

1. Sharedinvestment. Forboth straights and gays, bisexuality can raise deep issues of sexual jealousythat may ultimately require its denigration ordisavowal. I takethe straightman who is romanticallyinvolved with a bisexualwoman as my paradigmexample. However, it should be clearthat the anxieties discussed arenot specific to his orientation orsex. Atthe most basic level, the straight man is threatenedbyhis female partner'sbisexuality because he must compete not only with other men, but also withother women. One might say he is madeto understand Woody Allen's famousquip that bisexuality doubles one's chancesof gettinga dateon a Saturdaynight358 from the other side. Thisfear, understood purely quantita- tively,seems illogical. As a preliminarymatter, it is basedon bad math- shiftingfrom being a "straight"woman to beinga "bisexual"woman does notnecessarily "double" your dating pool, as manystraight men and lesbians willnot date avowed bisexuals.359 More importantly, itseems unlikely that a significantpart of the jealousy engendered by bisexuality can be explained bythe anxiety of enteringa larger dating pool. Does it makea difference whetherone is competingwith five billion people rather than two-and-a-half billionpeople? Or even fifty rather than twenty-five people? Thesequestions suggest that the threat of bisexuality is qualitative rather thanquantitative. Itis notthat one's bisexual partner can leave you for twice as manypeople, but that she can leaveyou for a differentkind of person. Theanxiety is arousednot by rivals who might also offer what you possess, butby rivals who might offer what you do notpossess. Whilethis qualitative concern is moreintelligible, itis notbeyond criticism.This is becausethe qualitative concern is basedin part on the assump- tionthat if one desiresboth sexes, one mustconsummate that desire with bothsexes. As PaulaRust has noted, "the bisexual's ability to form relation- shipswith members of both sexes is interpretedas a needfor relationships

358. See JosephP. Kahn, TheNew Book on Bisexuality, BOSTON GLOBE, Sept. 6, 1995, at 75 ("To be bisexual,Woody Allen once said, immediatelydoubles yourchances fora date on Saturday night."). 359. See, e.g., RuST, CHALLENGE,supra note 71, at 101 (notingreluctance of lesbians to as- sociate withbisexuals).

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 422 STANFORDLAWREVIEW [Vol. 52:353 withmembers of both sexes."360 Rust then criticizes this logic, asking read- ersto "imagine concluding that a personwho finds both blue and brown eyes attractivewould require two lovers, one with each eye color, instead of con- cludingthat this person would be happywith either a blue-eyedor a brown- eyedlover."361 Evenif one retires the assumption that those who desire both sexes must consummatetheir desires for both sexes, the qualitative concern still sur- vives. Thisis becausethe straight man may be certainof his bisexual partner'sfidelity, but may still be concernedthat because of her stated bisexuality,he is notsatisfying all ofher desires. Her articulation ofher desire for womenis readas anarticulation ofhis erotic inadequacy. Buteven this refined defense seems inadequate. As one commentator has noted,it wouldbe incrediblyhubristic to believethat any individual couldever satisfy all ofanother individual's desires, whether erotic or otherwise.362In emphasizingthis limitation, bisexuality could thus sound a healthynote of realismabout the incompleteness of all humanrelation.363 Moreover,one could argue that to be foundinadequate because of one's sex wouldbe less threateningthan to be foundinadequate on some other ground,364as all ofus areinadequate in possessing only one sex.365 A finalresponse might be thatthere is a differencebetween inarticulate andarticulate bisexuality. To returnto the earlier example, one might see a differencebetween a womanwho mentions to her brown-eyed husband that she is also particularlyattracted to menwith blue eyes and a womanwho keepsreiterating this to her husband. The difference between saying that you areattracted toothers once and saying it repeatedly is the difference between realismand sadism; the first comports with norms of courteous monogamy whilethe second does not. Evenif the woman continues to be attractedto menwith blue eyes, she is notonly supposed to refrain from acting on that attraction,but also to refrain from repeatedly articulating it. Carrying oneself as a bisexualmust be readagainst this norm of courtesy. A manmight acceptthat it is irrationalto feelparticularly inadequate because his partner

360. Paula C. Rust,Monogamy and Polyamory:Relationship Issues for Bisexuals,in BISEXUALITY:THE PSYCHOLOGYAND POLITICSOF AN INVISIBLEMINORITY 127, 128 (BethA. Firesteined., 1996). 361. Id. One responseto thisis thatan individual'ssex is so muchmore important than an individual'seye color that the analogy is inapt.As discussedabove, however, bisexuals may not necessarilyhold the view that an individual'ssex is moreimportant than eye color. See notes314- 319 supraand accompanying text. 362. See GARBER,supra note 11, at 451-52 (quoting interview with Brian Ford). 363. Id. at 451. 364. See WEINBERGET AL.,supra note 113, at 108(noting that bisexuals who dated bisexuals weremore jealous of "outside"partners of theirown sex because"a personof thesame sex as themselvescould meet similar needs and thus replace them."); see also GARBER,supra note 11, at 449-50(discussing study). 365. I assumehere that intersexuals also possess one, rather than two, sexes.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 423 statedto him once that she was attracted towomen. He might,however, find itrational to feel inadequate ifhis partner kept reminding him of this. Therub here is thatthe man is interpretinghispartner's carrying herself as a bisexualto be sucha reminder.But this is notnecessarily the case, at leastin the sense that his partner might carry herself as a bisexualfor reasons otherthan to occasion jealousy. As ever,there are normative rebuttals to eachof thepotential reasons whybisexuals might occasion jealousy. As ever,these normative rebuttals do notmean that the jealousies will diminish. To theextent that they endure, theywill act as anincentive for bisexual erasure.

2. Straightinvestment.

Underthe monogamy heading, the distinctively straight investment inbi- sexualerasure relates to AIDS. Bisexualityhas been prominently portrayed as a bridgethat transmits HIV fromthe gay community tothe so-called gen- eralpopulation.366 "Inthe minds of many heterosexual Americans, bisexualityhas cometo be stronglyidentified with images of married,dishonest, closetedmen sneaking out on theirunsuspecting wives, contracting AIDS throughunsafe sex with other men, then infecting their innocent wives and children."367In thelate 1980sand early90s, the mainstream press ran a flurryof storiesconcerning the AIDS threatbisexuals posed to heterosexuals.368 I earliershowed that bisexuals were viewed as an avenuethrough which heterosexualitywas calledinto question.369 Bisexuals are thus perceived to be a bridgealong which two undesirable traits-same-sex desire and HIV- pass froma gayminority to a straightmajority. In thesexual orientation context,the majority is characterized as straight and HIV-negative, while the minorityis characterized as gay and HIV-positive. The bisexual challenges bothcharacterizations-the firstepistemologically, thesecond epidemiologi- cally.His mere existence calls the heterosexuality ofthe majority into question.And his sexual practices call its HIV-negative status into question. Alongsidethis convergence, however, is an importantdivergence. We saw thatcharacterizing thebisexual as a carrierof same-sexdesire plunged thebisexual into obscurity. In contrast,the characterization ofthe bisexual as a carrierof HIV loftedthe bisexual into temporary visibility. The discrep-

366. See Ochs, supra note63, at 227. 367. Id. 368. See, e.g.,David Gelman,A PerilousDouble Love Life: In theAIDS Era, BisexualsAre Becomingthe Ultimate Pariahs, NEWSWEEK, July 13, 1987, at 44; Knox, supranote 272, at 57; Katie Leishman,Heterosexuals and AIDS, ATLANTIC MONTHLY, Feb. 1987, at 39, 48; Nordheimer, supra note272, at 1. 369. See notes264-278 supra and accompanyingtext.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 424 STANFORDLA WRE VIEW [Vol. 52:353

ancycan be explainedby the perceived role of social visibility in perpetuat- ingeach. Homosexualityis generally regarded to be spreadthrough its visibility-themere utterance of the"contagious word" of same-sexdesire is thoughtto perpetuateit.370 HIV, on theother hand, is generallyseen as spreadthrough its invisibility-makingHIV more visible,figuratively througheducation,371 or literally through tattooing,372 is viewed to be pro- phylactic.Straights who wish to controlboth homosexuality and HIV are thusin a doublebind vis a vis bisexualvisibility. To theextent that they wishto safeguardtheir own heterosexuality, they will wish to erasebisexuals,as theexistence of the bisexual makes it impossible for them to conclu- sivelyprove their heterosexuality. To the extent that they wish to safeguard theirHIV-negative status, however, they will seek to keepthe bisexual visible.373HIV has forced straights toconfess the bisexual they would otherwise deny.374

370. See JUDITHBUTLER, EXCITABLE SPEECH: A POLITICSOF THE PERFORMATIVE103-26 (1997). 371. See CINDY PATTON,INVENTING AIDS 52 (1990) ("In privateconversations, in public healthcampaigns, and in thescientific and para-scientificliterature, 'knowledge' is promotedas the essentialingredient in theeffort to slow thetransmission of HIV. To the fearfulcitizen, knowledge means informationwhich proves that she/he will notdevelop AIDS."). 372. See LARRYGROSS, CONTESTEDCLOSETS 59 (1993) (notingWilliam F. Buckley,Jr.'s proposalthat men who testHIV-positive be tattooed). 373. This double bind cinchesparticularly tightly because one way in whichboth homosexualityand HIV mightproperly be describedas infectiousis thatone individualcan affectan entire community.Leo Bersani says thatthe numberof homosexualsin a populationmay be irrelevant insofaras homophobiais fueledby the "fearfulexcitement at the prospectof becomingwhat one alreadyis." BERSANI,supra note 232, at 28. "One gay man ... deep in a distantprovince of a nationotherwise totally successful in its genocidalcampaign against gays would be enoughto hold up, forhis appalled and expectantcountrymen, the mirror in whichthey could not escape recognizing theiralready recruitedselves." Id. (I am assuming that this analysis would apply just as stronglyto bisexuals). Similarly,it has been said that"'[i]t also takes only one bisexual to intro- duce theAIDS virus[sic] intothe heterosexual community."' Jan Zita Grover,AIDS. Keywords,in AIDS: CULTURALANALYSIS/CULTURAL ACTIVISM 17, 21 (Douglas Crimp ed., 1988) (quoting ART ULENE,SAFE SEX IN A DANGEROUSWORLD (1987)). Thus, straightsmust ensure that not a single bisexual entersthe culturalimagination while simultaneouslyensuring that every single bisexual remainsvisible. 374. No similardouble bind occurs relativeto homosexualvisibility, where a genericinvisi- bilityrule has worked. This may seem counterintuitive,insofar as thestraight goal would stillbe to cabin the two "diseases" of homosexualityand HIV, and insofaras the two "diseases" have opposed relationshipsto discourse,with homosexuality being spreadthrough discourse and HIV being spread by a lack of it. The significantdifference between the bisexual and the homosexual,however, is thathomosexuals are not perceivedas spreadingHIV into the straightcommunity. Thus one strategyto evade the double bind has been to spread HIV discourseonly insofaras it does not pertainto same-sex sexual conduct. See PATTON,supra note 371, at 55-56 (describingstate and federalfunding restrictions on AIDS educationthat forbid funding of projectsthat "promote homosexuality"). This cabins HIV among straightswhile decreasingthe spread of homosexuality,not only throughdampening discourse about homosexuality,but also by leaving homosexualsmore at riskfor HIV.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 425

Itmay seem ironic to list the visibility ofbisexuals as carriersof AIDS in themainstream press as a symptomofbisexual erasure. But that case can be defendedon two grounds. First, visibility about the bisexual AIDS threathas notbeen sustained, perhaps suggesting a straight investment in repressing straightvulnerability to AIDS.375 In otherwords, bisexuality may be re- pressednot in spiteof, but because of, the perception that it is an epidemiologicalbridge. Such denial about risk can be seenin manyother contexts, suchas theminoritizing discourse of AIDS as a diseasethat only affects certainsocially marginal groups.376 Second, to theextent that the bisexual hasbeen made visible as an AIDS threat,that visibility has been so narrow andso negativeas tobe entirelydelegitimating.377 Andsuch delegitimation,

375. An alternativeexplanation for why the role of bisexualsin thetransmission of HIV has notachieved sustained visibility is thatresearch has demonstrated that this threat is notsignificant. Thisexplanation seems weak, as thestudies on therole of bisexualityin cross-sexHIV transmis- sionhave been indeterminate. Some studiesminimize the importance of bisexuality'srole in suchtransmission. See, e.g., MariaL. Ekstrand,Thomas J. Coates, Joseph R. Guydish,Walter W. Hauck,Linda Collette & Ste- phenB. Hulley,Are Bisexually Identified Men in San Franciscoa CommonVector for Spreading HIV Infectionto Women?,84 AM. J. PUB. HEALTH915, 919 (1994) (concluding"that single, bisexuallyidentified men in San Franciscodo notappear to be a commonvector for spreading HIV diseaseto women");James G. Kahn,Jill Gurvey, Lance M. Pollack,Diane Binson& JosephA. Catania,How Many HIVInfections Cross the Bisexual Bridge? An Estimatefrom the United States, 11 AIDS 1031,1031 (1997) (concluding"that [cross-sex] transmission via bisexualityis a rela- tivelyminor component of the estimated 40,000 annual HIV infectionsin theUSA"). Otherstudies, however, contend that bisexuality's role in cross-sextransmission of HIV has beenunderestimated. See, e.g.,Theresa Diaz, SusanY. Chu,Margaret Frederick, Pat Hermann, AnnaLevy, Eve Mokotff,Bruce Whyte, Mary Herr, Patricia J. Ehecko,Cornelis A. Rietmeijer, FrankSorvillo & QaiserMukhtar, Sociodemographics and HIV RiskBehaviors of Bisexual Men withAIDS: Resultsfrom a MultistateInterview Project, 7 AIDS 1227,1231 (1993) (documenting "multiplerisk behaviors that may be morecommon among behaviorally bisexual men"); Thomas Lerner& MaryAnn Chiasson, Seroprevalence ofHuman Immunodeficiency Virus Type I andSex- ual Behaviorsin BisexualAfrican-American and HispanicMen Visitinga SexuallyTransmitted Disease Clinicin New YorkCity, 147 AM.J. EPIDEMIOLOGY 269, 271 (1998) (notingthat "[i]t is conceivablethat HIV transmissionfrom bisexual men to their female sexual partners plays a greater rolein heterosexualtransmission in African-American andHispanic communities than was previ- ouslyrecognized"). I makeno attemptto weighthe merits of these studies. I simplyregister the conflict between themas an indicationthat there has never been a scientificconsensus that bisexuality has a minimal rolein cross-sexHIV transmission.Indeed, studies reviewing the existing literature note that not enoughresearch has beendone to supportany such view. See, e.g., LyndaS. Doll & Carolyn Beeker,Male SexualBehavior and HIV Riskin the United States: Synthesis of Research with Impli- cationsfor Behavioral Interventions, 8 AIDS EDUC. & PREVENTION205, 219-220 (1996) (review ofexisting research which notes that the role of bisexuals in HIV transmissionmay be significantin certaincontexts, but that "[flurther epidemiologic and behavioral research is urgentlyneeded to morefully describes the[se] contexts"). 376. See PaulaA. Treichler,AIDS, Homophobia, and Biomedical Discourse: An Epidemic of Signification,inAIDS: CULTURALANALYSIS/CULTURAL ACTIVISM, supra note 373, at 31, 65-67. 377. Thisparticular genre of visibilityshould not be takenas solicitudefor bisexuals, but ratheras solicitudefor the way in whichbisexuals can infectthe straight population. In this,the bisexualoccupies a parallelposition to theprostitute. As Sedgwickhas recognized,"[i]t has been notable.. . thatmedia coverage of prostitutes with AIDS hasshown no interestin thewomen them-

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 426 STANFORDLAWRE VIEW [Vol. 52:353

as notedabove,378 contributes tobisexual erasure by chilling bisexuals from expressingtheir identities.

3. Gay investment. The distinctivelygay investment in bisexual erasure pertaining to mo- nogamyhas two elements. First, many gays and lesbians wish to retirethe stereotypeofgays as promiscuous.Second, some lesbians view bisexuals as carryingHIV intothe relatively unaffected lesbian community. Self-identifiedgays, of course, do notspeak univocally against promiscuity.Some in thegay community view monogamy as an oppressivecul- turalinstitution of the heterosexual establishment.379 Butover time, those seekingto disestablishthe stereotype (and anyunderlying reality beneath it)380have increased in number.381 One potential explanation for this shift is the AIDS epidemic,which has made manygays reconsidertheir non- monogamoussexual practices.382 Another is thatthe growing acceptability ofhomosexuality has made assimilation more of a possibility. selves, but only in theirpotential for infecting men." SEDGWICK,EPISTEMOLOGY OF THE CLOSET, supra note4, at 5 n.8. Similarly,media coverageof bisexuals withHIV has shownlittle interest in thebisexuals themselves, but only in theirpotential for infecting straights. See GARBER,supra note 11, at 93-96. 378. See note235 supra and accompanyingtext. 379. See,e.g., Pat Califia, UnMonogamy: Loving Tricks and TrickingLovers, in PUBLICSEX: THE CULTUREOF RADICALSEX 199, 199-204 (1994) (criticizingmonogamy as utopianand repres- sive); Douglas Crimp, How to Have Promiscuityin an Epidemic, in AIDS: CULTURAL ANALYSIS/CULTURALACTIVISM, supra note 373, at 237, 272 ("Gay male promiscuityshould be seen . . . as a positivemodel of how sexual pleasuresmight be pursuedby and grantedto everyone if those pleasures were not confinedwithin the narrowlimits of institutionalizedsexuality."); DOUGLASSADOWNICK, SEX BETWEENMEN: AN INTIMATEHISTORY OF THE SEX LIVES OF GAY MEN POSTWARTO PRESENT86 (1996) (notingarguments against monogamy made by gay activists in the 1960s). 380. This stereotypefinds more support in the gay male communitythan in the lesbian community.See Donna Binder,Monogamy: Is It For Us?, THE ADVOCATE,June 23, 1998, at 29, 30. 381. See, e.g., Bruce Bawer,Sex-Negative Me, in BEYONDQUEER: CHALLENGINGGAY LEFT ORTHODOXY171, 172 (Bruce Bawer ed., 1996) [hereinafterBEYOND QUEER] ("To suggestthat gays are more definedby theirlibidos [thanstraights] is to collaboratein the widespread,dehu- manizingview thatgay sex is invariablymechanical, impersonal, even bestial,while straightsex is an integralpart of the complex web of human feeling,connectedness, and commitmentbefore God."); MICHELANGELOSIGNORILE, LIFE OUTSIDE: THE SIGNORILEREPORT ON GAY MEN: SEX, DRUGS,MUSCLES, AND THE PASSAGESOF LIFE 208-65 (1997) (discussingthe prevalenceof mo- nogamousrelationships in thegay male community). 382. See, e.g., JohnW. Berresford,A Gay RightAgenda, in BEYOND QUEER,supra note 381, at 105, 108 (notingthat "[i]n a curiousway, AIDS itselfmay be helping[gay men] findsocial acceptance,"since it "has broughtto a screechinghalt" the gay male promiscuityof the 1970s and 1980s);Frederick R. Lynch,Nonghetto Gays: An Ethnography ofSuburban Homosexuals, in GAY CULTURE iN AMERICA:ESSAYS FROMTHE FIELD 165, 166 (GilbertHerdt ed., 1992) (notingthat "the adventof AIDS musthave had some degree of adverse effecton the lure of the 'swinging single' life-stylefor homosexuals and heterosexualsalike"). Others-most prominentlyrepresented in the Sex Panic movement-have arguedthat AIDS should not push gays towardmonogamy, but

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 427

Theclassic situs of the gay debate over promiscuity is the marriage context.Marriage is seenas deeplyconnected to monogamy, insofar as itis the state'ssanction of an ostensiblylife-long and erotically exclusive relation- shipbetween two individuals.383While this nexus remains imperfect in practice,the fiction of themarriage as bothreflecting and reinforcing mo- nogamouscommitment remains strong.384 Many gays have rejected mar- riagein thesame way that they have rejected monogamy, as exemplifying heterosexist(and sexist) norms.385 Here as elsewhere,however, there is a growingassimilationist faction in thegay community that seeks the entitle- mentof marriage. Members of this faction express disdain for promiscuity in generaland polygamy in particular(since the latter is oftenviewed as a potentialconsequence of sanctioning same-sex marriage).386 To theextent that bisexuals are viewed as intrinsicallypromiscuous, as- similationistgays will be loathto recognize them as politicalallies, as such an affiliationmight corroborate, rather than rebut, the stereotype of gaysas promiscuous.Bisexuals may thus pose thegay rights movement with a milderversion of the question posed by pedophiles387 orpolygamists.388 On the one hand,solidarity with other sexual minorities suggests that they shouldbe includedwithin the movement; on the other, realism suggests that thelegitimacy ofthe movement may be underminedbysuch affiliations. An additionalconcem voiced by lesbians against the ostensible promiscuityof bisexuals is thatbisexuals might carry HIV intothe lesbian community.Despite the fact that HIV is oftencharacterized as a "gay"disease, les- biansare less at risk for sexual transmission ofHIV thaneither gay men or

rathertoward the practice of safe sex. See, e.g., JosephSonnabend & RichardBerkowitz, Safer-Sex Panic, in SEX PANIC! 13, 14 (1997) (arguingthat "it is the call forsexual restraintand monogamy withits inherentdeemphasis of safersex thatnow constitutesthe greater hazard to thehealth of gay men"); JimEigo, TheMonogamy Code, in SEX PANIC!,supra, at 15, 16 (arguingagainst the "attack on safersex" and thepromotion of "global gay male monogamy"). 383. See WILLIAMN. ESKRIDGE,JR., THE CASE FORSAME-SEX MARRIAGE 1 (1996). 384. See id. atI. 385. See, e.g.,Paula Ettelbrick, Since When Is Marriagea Pathto Liberation?, in SAME-SEX MARRIAGE:PRO ANDCON: A READER(Andrew Sullivan ed., 1997). 386. See AndrewSullivan, Three'sa Crowd,in SAME-SEXMARRIAGE, supra note 385, at 278-79. 387. See, e.g., Duncan Osborne,The Troublewith NAMBLA, THE ADVOCATE,Dec. 14, 1993, at 40 (reportingthe InternationalLesbian and Gay Association's requestthat the NorthAmerican Man-Boy Love Associationresign from the organization); Joyce Price, Pedophiles ResistingExpul- sionfromGay Umbrella Organization, WASH. TIMES, Nov. 27,1993, at A4 (similar). 388. See, e.g.,David L. Chambers,Polygamy and Same-SexMarriage, 26 HOFSTRAL. REv. 53, 53 (1997) ("During the hearingsand debates thatled to the Defense of Marriage Act, many membersof Congress and many witnessesdrew comparisonsbetween polygamyand same-sex marriage.");Sullivan, supra note 386 (arguingthat polygamy and same-sex marriageare easily distinguishable).

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 428 STANFORDLA WRE VIEW [Vol. 52:353

straights.389It is thusunsurprising that some lesbians, like someheterosexu- als,390fear the potential of bisexualsto serveas an epidemiologicalbridge froma "high-risk"to a "low-risk"community. One bisexualcommentator notesthat "gay women are hesitant to becomeinvolved with bi-the-way-girls becauseWE SPREAD AIDS or will leave themfor a man,"391while another suggeststhat she is perceivedas an "AIDS-carryinghigh-risk parasite on the movement."392The propinquityof the chargeof HIV transmissionto the chargeor capitulationto patriarchy(framed either individually as the ulti- materejection of a womanfor a manor politicallyas theinability of thebi- sexualparasite to rejectheterosexual privilege) may be purelycoincidental. It may,however, suggest that just as the straightcommunity occasionally appearsto makeHIV standfor the "disease" of homosexuality,393so too does thelesbian community make HIV standfor the "diseases" of patriarchyand capitulationto heterosexualprivilege.

D. OverlappingMonosexual Investments as a Cause ofthe EpistemicContract

This Parthas suggestedthat straights and gays have enteredinto the epistemiccontract of bisexual erasure because of three different investments: (1) an investmentin stabilizingsexual orientation;(2) an investmentin retainingthe primacy of sex; and(3) an investmentin preservingnorms of monogamy. Straightsand gays have an investmentin stabilizingsexual orientation categories.The sharedaspect of thisinvestment is thesecurity that all indi- vidualsdraw from rigid social orderings.The distinctivelystraight invest- mentis the retentionof heterosexualprivilege. And the distinctivelygay investmentrelates to theperception that bisexuality endangers the immutabilitydefense and effectivepolitical mobilization. Bothstraights and gaysmay also wishto erasebisexuals because bisexualityhas disturbingconsequences for the current sex regime. All monosexuals are createdthrough a regimethat privileges sex, and theythus have an investmentin bisexual erasure that relates to their own constitution. Straightshave a specificinvestment in bisexualerasure because bisexuality disruptsthe power heterosexuality has to determinesex performance.And

389. See William N. Eskridge,Jr., Multivocal Prejudices and HomoEquity, 74 IND. L.J. 1085, 1118 (1998) (notingthat while gays are oftencharacterized as carryingHIV, "thedata makes it clear that... lesbiansseem to have thelowest rates of AIDS . .. of anyof thegroupings"). 390. See notes366-368 supra. 391. Karen and Miranda Augustine,What Some Call Community... OthersCall Clicks, in PLURALDESIRES, supra note 345, at 37, 37. 392. Michelle Spring-Moore,Queergirl, in PLURAL DESIRES, supra note 345, at 226, 243. 393. JUDITHBUTLER, BODIES THATMATTER: ON THE DISCURSIVELIMITS OF "SEX" 64 (1993) (describingthe conflation of homosexualityand AIDS as pathologicaldiseases).

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 429

gayshave a specificinvestment because bisexuality interferes with complete sex separatism. Finally,both straights and gaysare disquietedby bisexualsinsofar as bisexualsare thoughtto representnonmonogamy. Straights and gays have a sharedinvestment in decreasingsexual jealousy. Straightsare peculiarly threatenedinsofar as theybelieve that bisexual nonmonogamy bridges the gap betweenthe HIV-infectedgay populationand the uninfectedstraight population.And gays are distinctively threatened by bisexual nonmonogamy to theextent that they wish to retirethe stereotype of gays as nonmonogamous. These multipleand overlappinginvestments in bisexualerasure explain the longevityof the epistemiccontract. But theyare not insurmountable. The veryfact that bisexual erasure has beenrecognized indicates this, for the paradoxof statushierarchy is thatthe oppressed category must have some powerto be recognizedas such.394But thismeans that even as thereare investmentsin bisexualinvisibility, there are simultaneousand countervailing investmentsin bisexualvisibility. The groupwith the mostpowerful investment in bisexualvisibility is composedof self-identifiedbisexuals. In thenext Part, I look at howbisexu- als have respondedto theepistemic contract. That narrative is one of both capitulationand resistance,testifying to boththe strengthof the epistemic contractand the possibility of its gradual dissolution.

IV. SELF-IDENTIFIEDBISEXUALS AND THE EPISTEMICCONTRACT

The longevityof the epistemiccontract can be explainedin two ways. The firstis thatbisexuals are notharmed by thecontract, or, more strongly, thatbisexuals benefit from their own erasuresuch that the contract could be characterizedas trilateralrather than bilateral. The secondis thatdespite the factthat bisexuals are harmed by the contract, it is so powerfulthat bisexuals are preventedfrom fully conceiving of themselvesas bisexuals,or fromef- fectivelydissolving that contract even if theydo. The two explanationsare not mutuallyexclusive, insofar as bisexualsare not a monolithicgroup. Some bisexualsfit the stereotype of thosewho takeadvantage of heterosexual privilegewhile engaging in clandestinehomosexual activity.395 Others experiencebisexual erasure as a harmand militatefor bisexual visibility.396

394. Cf Martha Minow, The Supreme Court,1986 Term-Foreword: JusticeEngendered, 101 HARV. L. REv. 10, 68 (1987) ("Poweris at itspeak when it is leastvisible, when it shapes preferences,arranges agendas, and excludesserious challenges from discussion or evenimagina- tion."). 395. See notes83-85 supra and accompanying text. 396. See notes402429 infraand accompanying text.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 430 STANFORDLAWREVIEW [Vol. 52:353

Whichexplanation we viewas dominant,however, has significantconse- quencesfor how serious we viewthe problem of bisexual erasure to be. Whileit is difficultto arbitratebetween the two explanations, I will ar- guein this Part that we shouldseriously entertain the viability of the second one. I beginthis argument by observing that the epistemic contract prevents thearticulation ofbisexual identity at everyphase of bisexual development, impedingthe formation ofa bisexualsubject that could meaningfully consent to itsown erasure. I thennote that despite the demands of the contract, an increasinglyvisible bisexual movement has articulated its desire for visibility.

A. BisexualCapitulation to theEpistemic Contract Theepistemic contract obstructs every phase of bisexual development. I considerthree phases here: (1) pre-comingout as bisexual;(2) post-coming outto some others; and (3) post-comingout to "all" others. Theprocess of coming out as a bisexualmay be retardedby the fact that no robusttemplate of bisexual identity exists. Even in a timewhere there is an increasinglyavailable model of gay identity, there is no equivalentone for bisexualidentity. Many individuals who might otherwise identify as bisexual mayrefrain from doing so onlybecause they cannot imagine that identity.397Possible support for this hypothesis can be foundin a recentstudy showingthat bisexuals come out as bisexualrelatively later than homosexuals comeout as homosexual.398 Evenafter a bisexualcomes out to someothers, she encountersenor- mouspressure to present herself to the rest of the world as straight.A study ofthe relative degrees of "outness" in the workplace among (1) gays,(2) bisexuals,and (3) straightswith unorthodox sexual practices found that gays weremost likely to be out,followed by bisexuals, followed by heterosexuals withunorthodox sexual practices.399 It is surelyno accidentthat this exactly tracksthe ascending degree to which these identities can be hiddenwithin a heterosexualmatrix. Moreover, even after a bisexualcomes out to some others,she will be underpressure to be selectivelycloseted. This is perhaps mosttrue in thedating context, where bisexuals find that both straights and gayscan often be reluctanttodate bisexuals. Finally,even a bisexualwho is outto "all" otherswill have a difficult timecommunicating this. The monosexualpresumption means that if a

397. See GARBER, supra note 11, at 48-49 (givingaccounts of bisexuals who were introduced to thebisexual categoryrelatively late); WEINBERG ET AL., supra note 113, at 121 (notingthat bisexuals generally"come out" later thanhomosexuals "because thereis a less clear identitywith whichto 'come out"'). 398. See WEINBERGET AL., supranote 113, at 121. 399. See id. at 188-89.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 431

womanwalks down the street holding hands with a man,we are likelyto thinkof heras straight;if shedoes the same thing with a woman,we are likelyto think of her as lesbian.400This presumption leads us to codeindi- vidualsas monosexualunless they evidence concurrent bisexuality. This synchronicapproach to orientationassignation can obtaineven whena diachronicaccount of the individual's erotic life is available.Ruth Colkersupports this point by noting the reflexive use ofthe term "hasbian" to describewomen who shift from a sustainedsame-sex relationship to a sustainedcross-sex relationship.401 Such a shiftcould equally be described as sequentialbisexuality-the pattern of conduct is thusoverdetermined and requiresnarratological resolution. That this ambiguity is resolvedbefore beingrecognized suggests a synchronicapproach to orientationattribution: The"hasbian" classification reads the current conduct as the"true" one and thepast conduct as therepudiated one, rather than reading them both as an embraceof sequential bisexuality. Such a synchronicapproach will tend to eliminateall but concurrent bisexuals from the bisexual category.

B. BisexualResistance to theEpistemic Contract The argumentthat bisexuals capitulate to bisexual erasure because they are notharmed by it is also undercutby an increasinglyvisible bisexual movement.The movement's development can be crudelydivided into three phasesby decade. The bisexualmovement did notbegin in a patternedway untilafter Stonewall.402One could do worsethan date its inauguration tothe formation of theNational Bisexual Liberation (NBL) groupin New YorkCity in the early1970s.403 The NBL beganpublication of a magazine,Bisexual Expres- sion,in 1975,roughly coinciding with the first books and mainstream press about bisexuality.404The magazineprovided an antidoteto the de- legitimatingtreatments of "bisexual chic" in thepopular press (including Time405and Newsweek406) that suggested that bisexuality was theperquisite

400. See Ochs,supra note 63, at 225. 401. See RuthColker, A BisexualJurisprudence, 3 LAW & SEXUALITY127, 129-30(1993) [hereinafterColker, Bisexual Jurisprudence]. 402. San Francisco'sSexual Freedom League, founded in 1967,might be seenas a precursor to thebisexual movement in thatit encouraged experimentation with both sexes. It was not,however,explicitly a bisexual organization. 403. See StephenDonaldson, The Bisexual Movement's Beginnings in the 70s: A Personal Retrospective,inBISEXUAL POLITICS, supra note 7, at31, 42-43. 404. See Donaldson,supra note 403, at 43. 405. See TheNew Bisexuals, supra note 230, at 79. 406. See BisexualChic, supra note 230, at 90.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 432 STANFORDLAWRE VIEW [Vol.52:353

of androgynousrock stars such as David Bowie or EltonJohn.407 The NBL's meetingsconsisted of monthlysocial eventsand consciousness-raisingexer- cises.408In itssocial cast, the organization was similarto manythat followed in thatdecade 409 such as New York City's Bisexual Forum(founded in 1975),410 San Francisco'sBisexual Center(founded in 1976),411 and Chi- cago's Bi-Ways(founded in 1978).412Bisexual activists were also occasionally prominentin earlygay-identified endeavors. For example,Alan Rock- way,a "gay-identifiedbisexual," co-authored the well-known Dade County, Floridaordinance that became the first ordinance to protectgay rights,and whichspawned the national anti-gay "Save OurChildren" crusade.413 In the 1980s,the bisexual movement shifted toward claims forgreater recognitionas a politicalmovement,414 with activists attempting to distin- guishbisexuality from sexual swinging.BiPOL, thefirst bisexual political actiongroup, formed in San Franciscoin 1983.415And manydate the birth of thenational bisexual movement to thesecond March on Washingtonfor Lesbian and Gay Rightsin 1987,where a nationalbisexual network, later knownas BiNET USA was formed.416The shiftmight have been expected as a simplebyproduct of the increased maturity and self-consciousnessof the bisexualmovement, with increasing numbers of bisexualorganizations pro-

407. See JayP. Paul,San Francisco'sBisexual Center and theEmergence of a Bisexual Movement,in BISEXUALITIES: THE IDEOLOGY AND PRACTICE OF SEXUAL CONTACT WITH BOTH MEN AND WOMEN 130, 132 (ErwinJ. Haeberle& RolfGindorf eds., 1998);see also GARBER, supranote 11, at 18-19(describing Time and Newsweek articles); Donaldson, supra note 403, at 35 (same). It was atthis time that the term "AC/DC" became popularized to describe bisexual people. See GARBER, supranote 11, at 38. 408. See Donaldson,supra note 403, at 42-43. 409. See Udis-Kessler,supra note 308, at 22-23. 410. See ChuckMishaan, The Bisexual Scene in New YorkCity, 11 J.HOMOSEXUALITY 223, 224 (1985). Theorganization folded in 1983. See id. 411. See DanielleRaymond & Liz A. Highleyman,Brief Timeline of Bisexual Activism in the UnitedStates, in BISEXUAL POLITICS, supranote 7, at 333,334. TheBisexual Center of San Fran- ciscoheld its first press conference on June30, 1977. Speakersincluded Dr. BenjaminSpock, Dr. PhyllisLyon and Ruth Falk. Manyof thesespeakers challenged Anita Bryant's efforts to barho- mosexualsfrom employment as school teachers and discussed other civil rights issues. See Maggi Rubenstein& CynthiaAnn Slater,A Profileof the San FranciscoBisexuality Center, 11 J. HOMOSEXUALITY 227,228-30 (1985). 412. See GeorgeBarr, Chicago Bi- Ways: An Informal History, 11 J.HOMOSEXUALITY 231, 232-34(1985). In theearly 1980s, Action Bi-Women was formedin the Chicago area to develop a feminist-orientedbisexual group. See id.at 234. 413. See PoliticalActivism: A BriefHistory, in BI ANYOTHER NAME, supranote 257, at 359, 363. 414. See Udis-Kessler,supra note 308, at 26 (discussingincreasing political activity of bisexual movement). 415. See id.;Raymond & Highleyman,supra note 411, at 334. 416. See Paul,supra note 407, at 137;Political Activism, supra note 413, at 364; Raymond& Highleyman,supra note 411, at 335.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 433

liferatingacross the country.417But the shiftcan also be attributedto the AIDS crisis,which came intopublic visibility in theearly years of thedec- ade. In 1987,Time4l8 and Newsweek4l9 would again publish articles touching uponbisexuality, but this time not about "bisexual chic," but about bisexualityand AIDS.420 Bisexualityhad now acquireda seriousedge: Newsweek calledbisexuals the "ultimate pariahs" of the AIDS crisis.421 In the 1990s,bisexual politics has risento greaternational and interna- tionalprominence. The FirstNational Bisexual Conference convened in San Franciscoin 1990,with Bisexual Pride Day proclaimedby theSan Francisco Boardof Supervisors.422The followingyears saw theFirst International Bi- sexual Conferenceheld in Amsterdam,423the publication of two majorbi- sexual anthologies,424and the inaugurationof a nationalbisexual magazine.425By 1993, the March on Washingtonfor Lesbian and Gay Civil Rightshad become the March on Washingtonfor Lesbian, Gay and Bi Equal Rightsand Liberation.426Whether to nameor notto namebisexuals alongside gaysand lesbianshas becomean explicitissue of debate.427That debate has also trickledinto the law. As evidencedby Colorado'sAmendment 2428 and themilitary's "don't ask, don't tell" policy,429 bisexuals are increasingly nominallyvisible, even if politically repressed.

417. Theseincluded the Boston Bisexual Women's Network in 1983,the East Coast Bisexual Networkin 1985,the Seattle Women's Bisexual Network in 1985,and the Washington D.C. Bi- Women's(and Men's) Network in 1989. See PoliticalActivism, supra note 413, at 360-61. 418. See MarthaSmiglis, Scott Brown, Dave Morrow& LeslieWhitaker, The Big Chill:Fear ofAIDS, TIME, Feb. 16,1998, at 50, 52. 419. See Gelman,supra note 368. 420. See Paul,supra note 407, at 137-39(describing articles). 421. Gelman,supra note 368, at 44; see alsoGARBER, supra note 11, at 93 (describingarticle). 422. See Raymond& Highleyman,supra note 411, at 335-36. The conferencewas attended byover four hundred people from twenty-two states and five countries. See Paul,supra note 407, at 137. 423. See KLEIN,supra note 81, at 169. Thesecond and third international bisexual confer- encesoccurred in London and New York in the mid-1990s. See id. 424. See Bi ANY OTHER NAME, supra note 257; CLOSER TO HOME: BISEXUALITYAND FEMINISM(Elizabeth Rebea Weise ed., 1992). 425. TheBay Area Bisexual Network began publishing the magazine Anything That Moves in January,1991. See GARBER,supra note 11, at 54. 426. See Raymond& Highleyman,supra note 411, at 337. 427. See,e.g., GARBER, supra note I 1,at 80-81 (describing controversy over removal of word "bisexual"from Northampton, Massachusetts Lesbian and Gay PrideMarch in 1990); id. at 53 (describingcontroversy over inclusionof word"bisexual" in mastheadof San FranciscoBay Times,which then read, "The Gay/Lesbian/Bisexual Newspaper & Calendarof Eventsfor the Bay Area"). 428. See Romerv. Evans,517 U.S. 620, 624 (1996) (quotingColorado's Amendment 2, whichprohibits protection of bisexualstatus); notes 48-52 supra and accompanying text. 429. See 10 U.S.C. ? 654(b)(2) (1994) (makingunrebutted statements of bisexualself- identificationgrounds for separation from service); supra note 44 andaccompanying text.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 434 STANFORDLA WRE VIEW [Vol. 52:353

The bisexualmovement's aims are diverse, and both converge and di- vergewith those of the gay rights movement. Thus, bisexuals, like gays, can be seendeploying bisexuality both as a meansof retiring all sexualorienta- tioncategorizations and as a meansof contestingplacement within those categorizations.On theother hand, bisexuals also haveinterests in their visibilitythat are specificto them:that is, intereststhat combat biphobia ratherthan homophobia. These interests include the retirement ofstereotypi- cal viewsabout bisexuality, such as theperceptions that they are indecisive, duplicitous,and promiscuous.

C. TheDissolution of the Epistemic Contract

Therise of the bisexual movement is probably the main factor leading to thedissolution of the epistemic contract of bisexual erasure. But as I have shownin critiquingthe investments ofboth straights and gaysin bisexual erasure,monosexuals also have undertheorized investments inbisexual visibility.The gay investments inbisexual visibility are perhaps easier to see- bisexuality,for example, can operate as a usefulheuristic for evaluating the immutabilitydefense, as wellas a meansof erodingheterosexual privilege byproblematizing theconcept of heterosexuality itself. Straights, however, have investmentsin bisexual visibility as well,such as challengingsex- separatismin thegay community or defending conventional norms of monogamy. Giventhe rise of the bisexuality movement, itseems safe to saythat bi- sexualswill only grow more visible in the future. It is thereforetime to considerhow a worldin whichthe epistemic contract no longerobtained might look. In thenext Part, I considerhow a particularworld-the legal world- mightbe transformedbyenhanced bisexual visibility.

V. BISEXUALITYAND SEXUAL HARASSMENT LAW

Despitethe fact that it is one ofthe main public sites in whichformal determinationsofsexual orientation are made, the legal system is complicit incultural ignorance about sexuality.430 It should therefore not be surprising thatbisexuals are for the most part as invisiblein the law as theyare in our generalculture. Indeed, if academic commentary is any indication, bisexualityis farmore undertheorized inthe legal realm than it is inthe social sci- encerealm. In 1993,Ruth Colker began an essayentitled A BisexualJuris- prudencewith the statement: "A bisexualjurisprudence? Until I decidedto writethis essay, there was no suchthing as a bisexualjurisprudence."431 A

430. Cf POSNER,supra note 30, at 1 (notingthat "judges know next to nothingabout the subject[of human sexuality] beyond their own personal experience"). 431. Colker,Bisexual Jurisprudence, supra note 401, at 127.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 435 computersearch of law review articles that deal with the subject of bisexual- itybears out her claim: Only a half-dozenarticles take bisexuality inthe law as theirprimary subject.432 Howmight increased bisexual visibility affect the law? To beginthink- ingabout this question, I take the case of sexual harassment law. I do so for threereasons. First, bisexuals are more visible in thisarea of jurisprudence thanin others.433This visibility arises because liability under Title VII only liesif the sexual harassment occurs "because of ... sex." Underone inter- pretation,this doctrinal formulation permits bisexuals to evade liability when theysexually harass both men and women, because no victimcan claim that theharassment occurred "because of' thevictim's "sex." Bisexualsare thus notonly distinguished from heterosexuals and homosexuals, but are rhetori- callyprivileged above both. As we shallsee, this exemption has beenclosed as a practicalmatter. Butboth in recognizing and in closing the exemption, the courts have had to grapplewith the reality of bisexuality in a waythat they have not had to in otherareas. Thus,while the bisexual harassment exemption is unimportant as a practicalmatter, ithas made bisexuals uniquely visible in antidiscrimi- nationdoctrine. Second,bisexual visibility has the normative consequence of narrowing anotherpernicious exemption in the sexualharassment jurisprudence.434 Thisexemption, which I call the"horseplay exemption," arises when self- identifiedstraights defend against charges of same-sex sexual harassment by statingthat the alleged homosexual harassment was in factheterosexual horseplay.Because sexual harassment jurisprudence (at leastin thepast decade)has directedmuch more solicitude toward sexual harms than non- sexualharms, an allegedharasser who is deemedto have engaged in heterosexualhorseplay is muchmore likely to be exemptedfrom liability than one whois deemedto have engaged in homosexual harassment. This horseplay exemptionis pernicious because it makes liability turn on statusrather than on conduct-morespecifically, on thesexual orientation ofthe actors rather

432. I foundthree articles, two essays, and one student note that took the role of thebisexual in thelaw as theirmain subject. See RuthColker, Bi: Race,Sexual Orientation, Gender and Dis- ability,56 OHIOST. L.J. 1 (1995) (article);Ruth Colker, An EmbodiedBisexual Perspective, 7 YALE J.L.& HUMAN.163 (1995) (article);Mezey, supra note 8 (article);Colker, Bisexual Juris- prudence,supra note 401 (reviewessay); Douglas S. Miller,Rumpole and theEqual Opportunity Harasser(or JudgeBork's Revenge), 20 J.LEGAL PROF. 165 (1996) (essay);Robin Applebaum, The "UndifferentiatingLibido ": A Needfor Federal Legislation to Prohibit Sexual Harassment by a BisexualSexual Harasser, 14 HOFSTRALAB. L.J. 601 (1997) (note). As thislist shows, Ruth Colkerhas donethe most writing in thisarea; herarticle entitled Bi: Race, Sexual Orientation, Gender,and Disability has alsobeen expanded into a prescientbook. See COLKER,HYBRID, supra note9. 433. I omitfurther citation in this paragraph as wellas theones that follow as thepoints made hereare explored more deeply with citations below. See textaccompanying notes 464-502 infra. 434. See textaccompanying notes 503-534 infra.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 436 STANFORDLA WRE VIEW [Vol. 52:353

thanon thenature of the allegedly harassing acts. Andunlike the bisexual harassmentexemption, the horseplay exemption is wide open. Forreasons I willdescribe more fully, the horseplay exemption as itcur- rentlystands is predicatedon bisexualinvisibility. Bisexual visibility thus hasthe salutary consequence of significantly narrowing this exemption. And becausethe courts have already made bisexuality visible in the context of the bisexualharassment exemption, this argument has a certainelegant force. Whilethe courts have sought to keepthe bisexual harassment exemption closedand thehorseplay exemption open, this status quo is inconsistent. Narrowingthe bisexual harassment exemption through bisexual visibility logicallycompels courts to significantlynarrow the horseplay exemption. Callthis the conservative critique. Finally,closing the horseplay exemption has its own spillover effects on thedegree to whichthe sexual harassment jurisprudence will continueto fetishizedesire as determinativeofliability.435 The conservative critique is conservativein the sense that it accepts, rather than contests, the aforemen- tionedpremise that sexual harm is moreserious than nonsexual harm. As VickiSchultz has recently and persuasively argued, however, this premise is deeplyproblematic.436 Her argument is timely, as theSupreme Court's re- centdecision in Oncale v. SundownerOffshore Services437 suggests that sexual harassmentjurisprudence is at a cross-roads.It can eithercontinue to privilegesexual harm over nonsexual harm or breakwith that practice to treatsexual and nonsexual harm equally. Bisexualvisibility might act as a goadto forcecourts down the correct pathafter Oncale. Closingthe horseplay exemption leaves many putative straightsmore at riskof earningthe ascriptions of homoeroticismand harassment.Bisexual visibility might thus encourage straights, as well as gays, toresist a regimein which liability turns solely on desire. Call thisthe radical critique. Afterbriefly summarizing the development of sexualharassment jurisprudence,this Part considers each of these reasons for considering bisexual visibilityin sexualharassment law. Itthen briefly considers the implications bisexualvisibility might have outside the sexual harassment context.

A. A Summaryof the Development of the Sexual Harassment Jurisprudence TitleVII ofthe Civil Rights Act of 1964makes it "an unlawful employ- mentpractice for an employer ... todiscriminate against any individual with respectto his compensation, terms, conditions, orprivileges of employment,

435. See textaccompanying notes 535-554 infra. 436. See VickiSchultz, Reconceptualizing Sexual Harassment, 107 YALE L.J. 1683 (1998). 437. 523 U.S. 75 (1998).

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 437

becauseof such individual's ... sex."438This language has been interpreted to prohibitsexual harassment in theworkplace. As KatherineFranke has noted,however, courts have long been unclear about why sexual harassment is sex discriminationprohibited by TitleVII, and havegenerated at least threedifferent theories.439 The firstis thatsexual harassment is actionable insofaras itwould not have occurred but for the victim's sex; that is, insofar as itoccurred "because of . . . sex"under the language of the statute.440 The secondis thatit is actionablebecause it is sexualin nature.441The third is thatsexual harassment is actionable because it enactsone sex's subordina- tionof another.442 The firsttwo theories have been the most dominant. The strugglebe- tweenthem can be characterizedas passing through four conceptual phases. In thefirst phase, the "because of . . . sex"theory was the ascendant theory. Indeed,the desire-based theory was not even acknowledged-early claims of desire-basedharassment were rejected as notraising a cognizableclaim un- derTitle VII. Thus,in Corne v. Bausch& Lomb,Inc.,443 the judge dismissed allegationsof desire-based harassment by reasoning that no TitleVII claim could lie whenthe allegedharasser was merely"satisfying a personal urge,"444as opposedto acting pursuant to a companypolicy.445 And in Tom- kinsv. PublicService Electric and Gas Co.,446 thedistrict court refused to recognizedesire-based claims under Title VII becauseotherwise "[a]n invi- tationto dinner could become an invitationtoa federallawsuit."447 Charac- terizingthe sexualconduct as idiosyncratic,or as raisingslippery slope problems,the courts essentially sidestepped the question of whether this was discrimination"because of. . . sex."448

438. 42 U.S.C. ? 2000e-2(a)(1994). 439. See KatherineM. Franke,What's Wrong with Sexual Harassment?, 49 STAN.L. REv. 691,692-93 (1997). 440. See,e.g., Rabidue v. OsceolaRefining Co., 805 F.2d611, 620 (6thCir. 1986);Bundy v. Jackson,641 F.2d 934,942 n.7 (D.C. Cir. 1981);Barnes v. Costle,561 F.2d 983,989 n.49(D.C. Cir.1977); see also Franke,supra note 439, at 705-14. 441. See, e.g.,Yeary v. GoodwillIndus.-Knoxville, Inc., 107 F.3d 443, 445 (6thCir. 1997); Matternv. EastmanKodak Co., 104 F.3d 702,706 (5thCir. 1997);Fuller v. Cityof Oakland,47 F.3d 1522,1527 (9th Cir. 1995); see also Franke,supra note 439, at 714-25. 442. See Vandeventerv. WabashNat'l Corp.,887 F. Supp. 1178, 1181 (N.D. Ind. 1995); Goluszekv. Smith,697 F. Supp.1452, 1456 (N.D. Ill. 1988);see also Franke,supra note 439, at 725-29. 443. 390 F. Supp.161 (D. Ariz.1975), vacated, 562 F.2d55 (9thCir. 1977). 444. Id. at 163. 445. Forsimilar reasoning, see Millerv. Bankof America,418 F. Supp.233, 235-36(N.D. Cal. 1976)(finding that no TitleVII claimcould lie forthe isolated and unauthorized sex miscon- ductof one employee to another), rev 'd, 600 F.2d211 (9thCir. 1979). 446. 422 F. Supp.553 (D.N.J.1976), rev'd, 568 F.2d 1044(3d Cir.1977). 447. Id. at557. 448. See Franke,supra note439, at 700 (notingjudicial concern over the slippery slope problem).

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 438 STANFORDLA WRE VIEW [Vol.52:353

In thesecond phase, courts rejected this evasion as illogical,finding that desire-basedharassment typically occurred "because of ... sex." The im- petusfor this belief came from feminist activists such as KerriWeisel, who contendedthat desire-based harassment was sex-basedharassment because an individualusually directs desire only at one sex or the other. 449 Thefirst court to adopt this theory was the D.C. Circuit,in the 1977 case ofBarnes v. Costle.450The districtcourt in Barneshad grantedsummary judgmentto thedefendant because it foundthat Paulette Barnes "'was dis- criminatedagainst, not because she was a woman,but because she refused to engagein a sexualaffair with her supervisor.'`451 A panelof the D.C. Cir- cuit reversed,noting that desire-based claims and "because of ... sex" claimscould not be dissociatedin this way. It stated: We cannotaccept this analysis of the situation .... Butfor [Barnes's] woman- hood,from aught that appears, her participation in sexual activity would never havebeen solicited.... [Bames]became the target of her superior's sexual de- siresbecause she was a woman,and was askedto bow to his demandsas the pricefor holding her job. Thecircumstance imparting high visibility to the role of genderin theaffair is thatno maleemployee was susceptibleto suchan ap- proachby appellant'ssupervisor. Thus gender cannot be eliminatedfrom the formulationwhich appellant advocates, and that formulation advances a prima faciecase ofsex discrimination within the purview of Title VII.452 Othercircuits soon adoptedthis theory.453 And in 1986,in MeritorSavings Bank v. Vinson,454the Supreme Court noted that: "Without question, when a supervisorsexually harasses a subordinatebecause of thesubordinate's sex, thatsupervisor 'discriminate[s]' on thebasis of sex."455 In thethird phase, as Schultzhas shown,a chiasmusoccurred between thetwo theories.456 The desire-basedtheory, which had beenthe subordinate theory,became the dominant theory; the "because of ... sex" theory,which hadbeen the dominant theory, became the subordinate theory. Desire was nowoften framed as eithera preconditionor an importantcomponent of a successfulTitle VII sexualharassment claim. The ascendancy of the desire- based theorymanifested itself in a numberof ways,perhaps most promi-

449. See Schultz,supra note 436, at 1702 (citingKerri Weisel, Title VII: Legal Protection AgainstSexual Harassment, 53 WASH.L. REv. 123, 133-35 (1977)). 450. 561 F.2d 983 (D.C. Cir. 1977). 451. Id. at 986 (quoting Barnes v. Train, Civ. No. 1828-73, slip op. at 3 (D.D.C. Aug. 9, 1974)). 452. Barnes, 561 F.2d at 990. 453. See, e.g., Miller v. Bank of America,600 F.2d 211 (9th Cir. 1979); Tomkinsv. Public Serv. Elec. & Gas. Co., 568 F.2d 1044 (3d Cir. 1977). 454. 477 U.S. 57 (1986). 455. Id. at 64. 456. See Schultz,supra note 436, at 1690 ("Ironically,courts that once refusedto recognize thatsexual advances may occur because of sex now insiston such advances and fail to perceive manyother problems that confront women workers as sex-based.").

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 439

nentlyin thedenomination of the cause of actionas "sexualharassment" ratherthan as "sex-based"or "gender-based" harassment.457 It could also be seenin theEEOC guidelines,which advanced a desire-basedtheory of harassment.458That ascendancy could also be seenin thecase law. Some courtsheld desire to be a prerequisitefor a successfulTitle VII claim,459and evencourts that considered both sexual and nonsexual claims often ignored orsubordinated the latter.460 TheSupreme Court's recent opinion in Oncale can be seenas potentially usheringin a fourthphase. In Oncale,the Supreme Court had to ascertain whethersame-sex sexual harassment was actionableunder Title VII, andif so,under what restrictions. In so doing,it had to choose a theoryfrom a set thatincluded the desire-basedtheory461 and the "because of ... sex" theory.462 The SupremeCourt embraced the "because of ... sex"theory.463 This holdinghas obvious implications for all ofsexual harassment jurisprudence, insofaras it rejectedthe view that desire is a requirementfor a successful TitleVII suit. As I demonstratebelow, however, whether Oncale can truly be seenas breakingfrom the third phase remains to be seen.

B. BisexualVisibility-The Recognition and Closingof the Bisexual HarassmentExemption

Sexualharassment jurisprudence is distinctive in that it not only names bisexualitybut treats it differently(and rhetoricallymore favorably) than eitherheterosexuality orhomosexuality. It would be a mistake,however, to

457. See id. at 1692. 458. See id. at 1704& nn.95& 97 (notingthat that EEOC guidelinescould be andwere read toprivilege sexual harm) (citing EEOC Guidelineson DiscriminationBecause of Sex, 62 Fed.Reg. 63,622(1980) (codifiedat 29 C.F.R.? 1604.11(1997))). 459. See Schultz,supra note 436, at 1718& n.167(citing, inter alia, Yearyv. GoodwillIn- dus.-Knoxville,Inc., 107 F.3d 443,445 (6thCir. 1997);Mattern v. EastmanKodak Co., 104 F.3d 702,706 (5thCir. 1997); Fuller v. Cityof Oakland, 47 F.3d 1522,1527 (9th Cir. 1995)). 460. See Schultz,supra note 436, at 1710-29(citing, inter alia, Harrisv. ForkliftSys., Inc. 510 U.S. 17 (1993); Kingv. Boardof Regents, 898 F.2d 533 (7thCir. 1990);Scott v. Sears,Roe- buck& Co., 798 F.2d210 (7thCir. 1986); Reynolds v. AtlanticCity Convention Ctr., 53 FairEmpl. Prac.Cas. (BNA) 1852(D.N.J. 1990), affd, 925 F.2d 419 (3d Cir. 1991); Raleyv. Boardof St. Mary'sComm'rs, 752 F. Supp.1272, 1280-81 (D. Md. 1990);Graham v. AmericanAirlines, 731 F. Supp. 1494,1500-01 (N.D. Okla. 1989);Ross v. DoubleDiamond, Inc., 672 F. Supp.261 (N.D. Tex. 1987);Sapp v. Cityof Warner Robins, 655 F. Supp.1043, 1049-50 (M.D. Ga. 1987)). 461. CompareMcWilliams v. FairfaxCounty Bd. Of Supervisors,72 F.3d 1191,1195 (4th Cir.1996) (ruling that same-sex sexual harassment claim was not actionable where defendants were heterosexual),cert. denied, 519 U.S. 819 (1996),with Wrightson v. Pizza Hut,99 F.3d 138,143-44 (4thCir. 1996) (rulingthat same-sex sexual harassment claim was actionablewhere defendants werehomosexual). 462. See Quickv. DonaldsonCo., Inc., 90 F.3d 1372,1377 (8th Cir. 1996). 463. See Oncalev. SundownerOffshore Servs., Inc., 523 U.S. 75,78 (1998).

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 440 STANFORDLA WRE VIEW [Vol. 52:353

takethe court's recognition ofthe bisexual harassment exemption as an indi- cationof bisexual privilege. As a practicalmatter, the bisexual harassment exemptionhas been closed. Nonetheless, in order to closethe bisexual harassmentexemption in a coherentway, courts have been forced into theoriz- ingbisexuality with some sophistication. Thus, despite the fact that the exemptionis notjurisprudentially important, ithas compelled a significantad- missionabout bisexual existence.

1. Therecognition of the bisexual harassment exemption- bisexualvisibility.

Bisexualswere first made visible in the sexual harassment jurisprudence whenthe shift from the first phase to thesecond phase was beingcontem- plated.In Corne,464the court stated that it would be "ludicrous"to call desire-basedharassment sex discrimination, "because to do so wouldmean that ifthe conduct complained of [male sexual advances on females]was directed equallyto males there would be no basisfor suit."465 Thebisexual possibil- itywas thus adduced to problematize the extension of Title VII's coverageto desire-basedclaims. Whileit made the shift from the first phase to thesecond phase despite thisproblem, the D.C. Circuitin Barneswas notable to solveit. Barnes simplyburied the unresolved issue in a footnote,noting that "[i]n the case of thebisexual superior, the insistence upon sexual favors would not constitute genderdiscrimination because it would apply to male and female employees alike."466This left the Barnes theory vulnerable toattack, as couldbe seenin a subsequentcase in thecircuit adopting that theory. In dissentingfrom a denial of rehearingin Vinsonv. Taylor,467 then-judgesScalia, Bork,and Starrargued as follows: [T]hiscourt has twicestated that Title VII does notprohibit sexual harassment by a "bisexualsuperior [because] the insistence upon sexual favors would ... applyto maleand female employees alike." [citing Barnes v. Costle,561 F.2d 983,990 n.55(D.C. Cir.1977); Bundy v. Jackson,641 F.2d934, 942 n.7 (D.C. Cir. 1981)).] Thus,this court holds that only the differentiating libido runs afoulof TitleVII, andbisexual harassment, however blatant and however of- fensiveand disturbing,is legally permissible. Had Congressbeen aimingat sexualharassment, it seems unlikely that a womanwould be protectedfrom unwelcomeheterosexual or lesbianadvances but left unprotected when a bisexual attacks.That bizarre result suggests that Congress was not thinking of indi- vidualharassment at all butof discriminationin conditions of employmentbe- causeof gender. If it is properto classifyharassment as discriminationforTitle

464. 390 F. Supp. 161 (D. Ariz. 1975). 465. Id. at 163. 466. Barnes v. Costle,561 F.2d 983, 990 n.55 (D.C. Cir. 1977). 467. 760 F.2d 1330 (D.C. Cir. 1985) (dissentingfrom denial of rehearingen banc).

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 441

VII purposes,that decision at leastdemands adjustments in subsidiarydoc- trines.468 Thesejudges thus contended that Title VII couldnot have been intended to reachdesire-based discrimination because doing so wouldpermit bisexuals toevade liability. The embarrassmentthat bisexuality causes sexual harassment jurisprudenceis clear. An individualwho would be liablefor engaging in certain conductcan evade liability for that conduct by engaging in moreof the con- ductdirected at theopposite sex. I call thisthe "double for nothing" problem-by doublingthe proscribed conduct, the harasser lowers his liability to nothing.469This result is so counterintuitivethatcommentators who usually seemfar apart on the political spectrum-such as Robert Bork and Catharine MacKinnon-canagree that this result is anomalous.470 Theproblem of the bisexual harasser demonstrates themanner in which bisexualitydestabilizes norms that privilege sex-based distinctions. Title VII privilegessex as a diacriticalaxis-only discrimination that occurs on this axis(or on anotherenumerated axis) is actionable.As such,the statute can- notencompass bisexuals who are truly sex-blind within its prohibitions. I earliernoted that the sex-blind bisexual (or asexual) might be celebratedas theonly type of individualwho could credibly say that she never discrimi- natedon the basis of gender.471 But the sex-blind bisexual may express hos- tilityto othersindiscriminately as well. Andwhen she does so, ournorms againstsex discrimination will be insufficienttohold her liable. Thefact that bisexuals have garnered this "advantage" over both hetero- sexualsand homosexuals explains their relative visibility in thesexual harassmentjurisprudence. In areas of the law where differential treatment ofthe sexesis notrequired, bisexuals are neither privileged nor visible. A bisexual

468. Id. at 1333n.7. 469. It bearsnote that this "double for nothing" problem is notjust a problemin thedesire- basedharassment context, insofar as a harasserwho harassed both men and women in nonsexual wayswould also evadeliability under a strictapplication of the "because of .. . sex"test. Thebi- sexualharassment exemption is thus a subsetof an "equalopportunity" harassment exemption. The factthat "double for nothing" is notspecific to thedesire-based context should not be surprising, becausethe problem arises from the asymmetrical treatment ofsex requiredby the "because of ... sex"language in the statute, rather than anything specific to desire. Butthis means that the minatory statements made about "double for nothing" in Vinsonlack subtlety.These judges argued that Title VII couldnot have been intended to reach the desire-based contextbecause of "doublefor nothing." But the fact that the problem is notspecific to desire meansthat it cannot-withoutfurther argument be usedby courts to sustainthe proposition that TitleVII coversnondesire-based harms but not desire-based harms. 470. CompareVinson, 760 F.2d at 1333n.7 (notingthat a bisexualharassment exemption wouldentail a "bizarreresult"), with Brief amicus curiae of CatharineA. MacKinnon,at 22 n.6, Oncalev. SundownerOffshore Servs., Inc., 523 U.S. 75 (1998)(No. 96-568)(stating, in thecourse ofa descriptionof thebisexual harasser, that "equal discrimination is a clearoxymoron"). 471. See textaccompanying note 338 supra.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 442 STANFORDLAWREVIEW [Vol. 52:353 couldnot defend against a same-sexsodomy charge by sayingthat he had alsoengaged in cross-sexsodomy. This means that there is no incentivefor commentatorstomake bisexuals visible in thesodomy context-there is no privilegeagainst which critics must inveigh. Thevisibility of bisexuals in sexualharassment jurisprudence also suggeststhat bisexuals may in someinstances be morevisible in the law than in societyin general.Courts noted the bisexual possibility in thelate 1970s,472 wellin advanceof a nationalbisexual movement.473 This is somewhatsurprising,as legalknowledge of sexuality generally lags behind social knowl- edgeof sexuality.474 The exception to the general rule, however, may be explainedby theformalism of thelaw. The logic of the"because of . .. sex" formulationforced courts to hypotheticallyenvision the bisexual harasser longbefore he evermade an actualappearance. Logic thusopened up imaginativespaces of possibility before such spaces were politically occu- piedin any robust way.

2. (Incoherently)closing the bisexual harassment exemption- bisexualinvisibility.

One strategyfor dealing with the bisexual harassment exemption-class erasure-couldhave been predicted from the previous discussion of bisexuality.When Vinson finally made its way to theSupreme Court, the Court statedthat desire-based harassment was cognizable under Title VII.475 In so stating,the Courtimplicitly rejected the argument made by then-judges Bork,Scalia, and Starr in thatcase. Thebasis for that rejection, however, wasnot clear; the Court evaded giving one by simply ignoring bisexuals.476 Andwhen the Oncale Court reaffirmed that sexual harassment was cognizable insofaras it was discrimination"because of ... sex,"477it onceagain sidesteppedthe problem of thebisexual harasser by erasingbisexuality.478 The invisibilityofbisexuality in Oncalewas particularlyironic, given that JusticeScalia-who as a circuitjudge had insistedon thevisibility of bisexuality-wrotethe opinion for the Court.479

472. See, e.g., Barnes, 561 F.2d at 990 n.55. 473. See note 416 supra and accompanyingtext (dating inaugurationof national bisexual movementto late 1980s). 474. See notes431-432 supra and accompanyingtext. 475. MeritorSavings Bank v. Vinson,477 U.S. 57, 64 (1986). 476. See id. (makingno mentionof thebisexual harassment exemption). 477. Oncale v. SundownerOffshore Servs., Inc., 523 U.S. 75, 78-80 (1998). 478. See id. (makingno mentionof thebisexual harassment exemption). 479. Compare Vinson v. Taylor, 760 F.2d 1330, 1333 n.7 (D.C. Cir. 1985) (Bork, Scalia, StarrJJ., dissenting from denial of reh'g en banc) (raisingthe possibilityof the bisexual harasser), withOncale, 523 U.S. 75 (ignoringthe possibility of bisexualharasser).

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 443

As thecontrast between Justice Scalia's positions in Oncaleand in Vin- sonillustrates, elision is nota particularlypersuasive strategy in thisarea of jurisprudence,as courts that have already confessed the existence of bisexualityin onecase cannotlogically deny it in another.Perhaps in recognition of this,other circuit courts have engaged in milderforms of class erasure, engagingnot in elisionbut in explicit(and incomplete) denial of theexis- tenceof the class. Thus,in Henson v. Cityof Dundee,480 the Eleventh Circuit statedthat "[e]xcept in theexceedingly atypical case ofa bisexualsupervisor,it shouldbe clearthat sexual harassment is discriminationbased on sex."481That 1982 opinion's language about the "exceedingly atypical case of a bisexualsupervisor" has become a boilerplatephrase in sexualharass- mentcases.482 But neither the Henson court nor any other court employing itsphrase ever explains why bisexual supervision is so rare. In fairnessto thesecourts, actual defendants who deploy the bisexual harassmentdefense are rare. Such defendants appear more often in fiction483 or hypotheticals484thanactual litigation485; thedefense has beenraised in onlytwo cases and has been rejected in both.486 Defendants may understand thatthe "double for nothing" reasoning underlying the defense violates eq- uitableintuitions,487 and may fear that courts will privilegethat intuition abovethe exemption. Alternatively, bisexual harassers who self-identify as straightmay fear coming out as bisexualmore than they fear potential liability.Such harassers could rationally assume that the potential stigma of being

480. 682 F.2d897 (11thCir. 1982). 481. Id. at905n.11. 482. See,e.g., Doe v. Cityof Belleville, 119 F.3d 563, 576 (7thCir. 1997) (quoting Henson v. Cityof Dundee,682 F.2d 897,905 n.11 (1 IthCir. 1982));Fredette v. BVP ManagementAssocs., 112F.3d 1503,1506 n.5 (11th Cir. 1997) (same); Jones v. FlagshipInt'l, 793 F.2d714, 720 n.5 (5th Cir.1986) (same); McCoy v. JohnsonControls World Servs., Inc., 878 F. Supp.229, 232 (S.D. Ga. 1995) (same);Demele v. Belle of Orleans,L.L.C., 1997WL 411558at *7 n.7 (E.D. La. July21, 1997)(same). 483. See Miller,supra note 432, at 165. 484. See,e.g., Applebaum, supra note 432, at 601-02 & n.1. 485. See AmicusBrief of CatharineMacKinnon, supra note 470, at 21-22("The so-called bisexualharasser, eluding equality snares by indiscriminatelysexually harassing men and women alike,stalks the judicial imagination, cutting quite a figurein legalhypotheticals." (footnotes omitted)). 486. Thebisexual harasser defense has been raised, but not accepted, in two cases. See Raney v. Districtof Columbia, 892 F. Supp.283, 287-88 (D.D.C. 1995);Ryczek v. GuestServs., Inc., 877 F. Supp.754, 762 (D.D.C. 1995)(granting summary judgment to defendanton othergrounds). A similar"equal opportunity harasser" defense has beenraised to no avail in twoother cases. See Steinerv. ShowboatOperating Co., 25 F.3d 1459,1464 (9th Cir. 1994); Chiapuzio v. BLT Operat- ingCorp., 826 F. Supp.1334, 1336-38 (D. Wyo.1993). In othercases, defendants have failedto raisethe bisexual harasser defense where it might have been expected. See, e.g.,Griswold v. Fre- seniusUSA, Inc.,978 F. Supp.718, 723 (N.D. Ohio 1997)(involving married man who allegedly sexuallyharassed other men and failed to raise bisexual harasser exemption); Hopkins v. Baltimore Gas & Elec. Co., 871 F. Supp.822, 824 n.2,835 (D. Md. 1994)(same). 487. Cf.Saulpaugh v. MonroeCommunity Hosp., 4 F.3d 134,144 (2d Cir.1993) (noting, in ? 1983context, that it would take "audacity" for a defendanttoraise a bisexualharassment defense).

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 444 STANFORDLAWREVIEW [Vol. 52:353

labeledharassers was lessthan the definite stigma of self-identifyingas bisexuals.488 Thefortuity that not many bisexuals have pressed the defense, however, avoidsrather than answers the problem of the bisexual harassment exemption.It alsodoes not suffice in those situations inwhich the defense is actu- allyraised. Perhaps in response to both points, some courts have proffered a differentway of dealingwith the problem of thebisexual harassment exemption.This approach, which is ultimatelymore coherent, trades on bisexualvisibility, rather than invisibility.

3. (Coherently)closing the bisexual harassment exemption- bisexualvisibility (again). A morecoherent way of closing (or at leastsignificantly narrowing) the bisexualharassment exemption notes that while bisexuals may desire both menand women,they will not necessarily manifest that desire in a sex- neutralway. Thispremise significantly mitigates the "double for nothing" problem,for if bisexuals treat men and women in sex-differentiatedways, theywill still be liableunder a "becauseof. .. sex"test. The courts, albeit in a haphazardand undertheorized way, have hypothesized a number of ways in whichbisexual desire may have sex-differentiated manifestations. In so doing,they have significantly narrowed the bisexual harassment exemption by makingthe bisexual more, rather than less, visible. First,some courts have stated that a bisexualharasser will not evade li- abilityif he onlyharasses members of one sex. In otherwords, the harasser cannotescape liability simply by desiring both sexes, he mustalso express thatdesire in harassing ways toward both sexes. Thisrequirement ofactual harassmentofboth sexes distinguishes thebisexual harasser exemption from thebisexual harassment exemption. In Raneyv. Districtof Columbia,489the districtcourt noted that Barnes v. Costle490could be readas providingfor the formerexemption, while the subsequent case of Bundy v. Jackson49' could be readas providingfor the latter. The Raney court itself read Bundy as a clari- ficationof Barnes, and thus required actual harassment ofboth sexes.492 It notedthat while the defendant inthat case had raised the defense that he was bisexual,he was allegedto haveonly harassed men, and the court therefore rejectedhis defense.

488. This is particularlytrue given that,as Colker has pointedout, an employee who self- identifiesas bisexual can be terminatedon thatbasis withoutincurring Title VII liabilityfor the employer.See Colker,BisexualJurisprudence, supra note401, at 135-36. 489. 892 F. Supp. 283, 287-88 (D.D.C. 1995). 490. 561 F.2d 983, 990 n.55 (D.C. Cir. 1977). 491. 641 F.2d 934, 942 n.7 (D.C. Cir. 1981). 492. See Raney,892 F. Supp. at 288.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 445

Second,one courthas statedin dictathat a sequentialbisexual cannot deploythe exemption even if he harassesboth men and women. In Tietgen v.Brown 's WestminsterMotors, Inc. ,493 thecourt noted that the bisexual harassmentexemption "may not prohibit sexual harassment by a bisexualharasser"insofar as oneassumes that he or she "chooses whom to solicit for sex accordingto somecriteria other than gender."494 It went on, however, to contestthis assumption, noting that "it may be thata bisexualsolicits a per- sonfor sex based on theperson's gender, which gender the bisexual prefers at thatmoment."495 The court thus raised the possibility of, for example, an individualwho directs his sexual attention solely to womenfor a periodof time,then switches over to directinghis attention solely to men for a period oftime.496 Even if he harassedthe same numbers of men and women, one couldstill argue that at thetime he was harassingany individual man, no womanwas in danger,and vice versa. If one acceptedthis argument, his harassmentwould have occurred "because of . . . sex."497 Finally,courts have implied that even if the harasser had concurrent sexual desirefor both men and women and harassed them both, he couldstill be liableif he expressedthat desire in differentways. In Chiapuziov. BLT Op- eratingCorp.,498 a district court noted that the defendant's claim that its em- ployeeEddie Bell harassedboth men and women "raised the specter of the 'bisexualharasser."'499 The courtexorcised that possibility by contending thatalthough Bell harassedboth men and women, he harassedthem in differentways; specifically, by makingremarks about sex actshe wantedto

493. 921 F. Supp.1495 (E.D. Va. 1996). 494. Id. at 1501n.10. 495. Id. 496. The Weinbergstudy documented the existence of bisexualswho were attracted to one sexor the other depending on their"mood" at a giventime. See WEINBERGET AL.,supra note 1 13, at 55. One suchbisexual stated: "I don'tthink it has much to do withpitting a good-lookingman againsta good-lookingwoman. I thinkit has moreto do withmy own feelingsof whetherI'm attractedto menor women more at a particularpoint." Id. 497. As Garbernotes, the distinction between sequential and concurrentbisexuality is less thanclear: Cliniciansthese days tend to characterizebisexuality as either "sequential" or "concurrent," dependingupon whether the same-sex/opposite-sex relationships are goingon at thesame time.But although this will at firstseem useful in making gross distinctions, itis finallyless clearthan it appears. For one thing, what, precisely, is "the same time"? Alternate nights? Thesame night? The same bed? GARBER,supra note 1 1, at 147. Ifone collapses the distinction between sequential and concurrent bisexualityin thisway, one can completelyclose thebisexual harassment exemption, but only, perhaps,at the cost of absurdity. 498. 826 F. Supp.1334 (D. Wyo.1993). 499. Id. at 1336.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 446 STANFORDLAWREVIEW [Vol. 52:353 performwith women but not with men.500 Therefore, the court concluded, Bell'sharassment occurred "because of.. . sex."Sol Chiapuziois notreally an exampleof a courtfinding a defendant liable forconcurrently harassing both sexes in a desire-basedbut sex-differentiated way. Thisis becausethe court determined that while Bell's harassmentof womenwas sexual,his harassment ofmen was not. Thecourt thus deemed thatBell wasnot a bisexualharasser, but rather an "equalopportunity" harasser-thatis, someone who harassed both men and women, but who did not harassboth in sexualways. Indeed,the fact that Bell harassedone sex but notthe other in a sexualway provided the sex-based distinction needed for TitleVII liability. The logic of Chiapuzio,however, may be easilyextended to sex- differentiateddesire-based harassment ofboth sexes. To see this,consider thatthe court could have reasonably read Bell's commentstowards men as beingsexual in nature, and could still have found Bell liable.Bell's conduct is entirelyconsistent with what a manwho had desirefor both men and womenmight do in a societythat stigmatized same-sex but not cross-sex desire-thatis, expresshis desire for women openly, but find more closeted waysof expressing his desire for men. Even if the court had found this to be thecase, however, Bell wouldstill have been liable. If desire(even sex- neutraldesire) manifests itself in a systematicallysex-differentiated way,that is logicallysufficient tomeet the "because of . . . sex"requirement. Theseanalytic moves do notcompletely close the bisexual harassment exemption-theremight be a bisexualharasser who was genuinely"sex- blind"such that the sex ofthe victim had no salienceto him.502In sucha case,the bisexual harassment exemption would still obtain. Nonetheless, makingthe bisexual more visible has the effect of significantly narrowing the bisexualharassment exemption. By painting a subtler picture of the bisexual, we see thatnot all bisexualsare "sex-blind," and to theextent that they are not,they will be liableeven under a "butfor" test.

C. Recognizingand Closingthe Horseplay Exemption Bisexualvisibility does not have much normative bite in thecontext of thebisexual harassment exemption. The exemption has already been closed as a practicalmatter; bisexual visibility only provides a morecoherent justi- ficationfor that closure. In contrast,bisexual visibility has significantnor-

500. See id. at 1337-38. 501. See id. at 1338. 502. As notedabove, however, I am dubious about the robustness of the"sex-blind" bisexual category.See notes317-319 supra and accompanying text.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 447

mativeconsequences for another exemption in thesexual harassment jurisprudence:the horseplay exemption. The horseplayexemption occurs in the same-sexsexual harassment context.The exemption arises when harassers (almost invariably males) de- fendagainst accusations of homosexual harassment byrecasting the conduct as heterosexualhorseplay. Because the jurisprudence emphasizes sexual overnonsexual harm,S03 this re-characterization ofthe conduct, if successful, oftenabsolves the harassers of liability. Unlikethe bisexual harassment exemption, the horseplay exemption is wideopen as a matterof practice.504 This discrepancy can be explainedby therelationship that each exemption has to judicial intuitions about fairness. The bisexualharassment exemption was closedbecause the "double for nothing"problem it raised violated a judicial intuition of equity.The horseplayexemption, incontrast, is consistent with such a judicialintuition. Spe- cifically,the exemption reflects the intuition that the homosocial and the homoeroticare sufficiently intertwined that ambiguity should be resolvedin favorof theformer, lest heterosexuals be mistakenlyre-characterized as nonheterosexuals.SosThe status quo in whichthe bisexual harassment exemptionis closedwhile the horseplay exemption remains open thus is so- ciologicallyintelligible. Thatstatus quo, however, also defies logic. This is becausethe recogni- tionand closureof thebisexual harassment exemption relied on bisexual visibility.The horseplay exemption as itnow stands, however, relies on bisexualinvisibility. Closing the bisexual harassment exemption thus logically requiresat least significantly narrowing the horseplay exemption. In makingthis argument, I first examine the nature of the judicial intui- tionthat keeps the horseplay exemption open. I thennote the analytic moves thecourts make to protectthe horseplay exemption, pointing out that these movesrely on bisexualinvisibility. Finally, I considerhow the enhanced visibilityof bisexuality would foreclose these moves, requiring a significant narrowingofthe exemption.

503. Some mayquestion whether this is stillthe case afterOncale v. SundownerOffshore Servs.,Inc., 523 U.S. 75 (1998). I addressthis question below. See notes535-549 infra and accompanyingtext. 504. See, e.g., McWilliamsv. FairfaxCounty Bd. Of Supervisors,72 F.3d 1191 (4thCir. 1996); Martinv. NorfolkS. Ry. Co., 926 F. Supp. 1044 (N.D. Ala. 1996); Tietgenv. Brown's WestminsterMotors, Inc., 921 F. Supp.1495 (E.D. Va. 1996). 505. Thisintuition is probablythe impetus behind the various "queen-for-a-day" exemptions describedabove, in whichadventitious homosexual conduct is notseen as damagingan otherwise heterosexualstatus. See notes100-106 supra and accompanying text.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 448 STANFORDLA WRE VIEW [Vol.52:353

1. Understandinghorseplay-the homosocial and thehomoerotic.

Intimateactions among men are ambiguous-they can be eitherhomo- socialor homoerotic. This ambiguity is peculiar insofar as thetwo ways in whichit can be resolvedare diametrical opposites. Usually when an ambi- guitymust be resolved,the two ways of resolving itare quite similar to each other;indeed, it is theirsimilarity that leads to theconfusion. Intimate acts amongmen, however, are not ambiguous in thisway; the social meanings andconsequences of homosociality are squarely opposed to thoseof homoeroticism. Oneway of construing the opposition is to understand homosociality as a meansof makingmen, and homoeroticism as a meansof unmakingthem. Homosocialityis a meansof making men insofar as themen who can take (anddish out) hazing, razzing, or horseplayare constituted as "real"men, whilethose who cannot (or who choose to opt out) are constituted as "failed" men. Theserites are oftenexplicitly formulated as initiationswhere the manhoodof those hazed is testedand the manhood of those hazing is reaffirmed.506Even when they are not so formulated,homosocial behaviors of- tenperform that function. Over a courseof homosocial conduct-however inchoate-averdict is reachedas to whetherthe new worker is "oneof the guys"or not.507 Homoeroticism,in sharp contrast, is a meansof unmakingmen.508 As can be seenin theearliest formulations of homosexual men as "inverts"- thatis, as womentrapped inside men's bodies509-a homosexual male was a manwho would reveal himself to be a woman(and thus a failedman). That gaymen are viewed as failedmen can also be seenin the characterization of failedmen as gaymen.510 In case aftercase, the workplace outcast is cast

506. See HANK NUWER,THE DEADLY RITE OF HAZING204-06 (1990) (describinghighly sexualized hazingritual through which Coast Guardinitiates are inducted-or not-into thegroup). 507. See MichaelS. Kimmel,Masculinity as Homophobia:Fear, Shame,and Silencein the Constructionof GenderIdentity, in THEORIZINGMASCULINITIES 119, 128-29 (HarryBrod & Mi- chael Kaufmaneds., 1994) ("Othermen watch us, rankus, grantour acceptanceinto the realmof manhood.... Masculinityis a homosocialenactment. We testourselves, perforn heroic feats, take enormousrisks, all because we wantother men to grantus our manhood."). 508. This is, of course,a culturallycontingent claim. In othercultural contexts, homoeroticism could be seen as a means of initiatingyouths into manhood. See, e.g., EVA CANTARELLA, BISEXUALITYIN THE ANCIENTWORLD 6-8 (Cormac 6 Cuilleanaintrans., 1992) (describingsuch initiationsin ancientGreek society). 509. See Edwards,supra note342, at 112. 510. See Gary Kinsman,Men Loving Men: The Challenge of Gay Liberation,in BEYOND PATRIARCHY:ESSAYS BY MEN ON PLEASURE,POWER, AND CHANGE 103, 104 (Michael Kaufman ed., 1987) ("'Real' men are intrinsicallyheterosexual; gay men,therefore, are not real men."); Jo- seph H. Pleck,Men 's Power withWomen, Other Men, and Society:A Men 's MovementAnalysis, in THE AMERICANMAN 417, 424 (Elizabeth H. Pleck & JosephH. Pleck eds., 1980) ("Our society uses themale heterosexual-homosexualdichotomy as a centralsymbol for all the rankingsof masculinity,for the division on any groundsbetween men who are 'real men' and have power and

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 449

"4out"as a homosexual,regardless of whether there is evidencethat he mani- festedany same-sex desire.511 Thefact that intimate intra-male acts can be interpretedin diametrically opposedways makes that interpretive actextremely fraught.512 A misreading ofhomosociality as homoeroticism will have the devastating consequence of brandinga real man as a failedman. In resolvingthat ambiguity, the risk- aversedecisionmaker will err on theside of heterosexuality. This tendency couldbe defendedon the grounds that we believethe majority of the popula- tionto be straight.That defense, however, may be insufficient,in that the courtsare not dealing with the general population, but with the subset of that populationwhose same-sex conduct has given rise to a claimof harassment. Thissuggests that the premise here may be lessstatistical than equitable; courtsmay believe that where any ambiguity exists, it is betterto let the guiltyhomosexual go freethan to convictthe innocent heterosexual. That equitablepremise, however, will create problems of its own when the costs

maleswho are not. Anykind of powerlessnessor refusalto competebecomes imbued with the imageryof homosexuality."). 511. See,e.g., Johnson v. Hondo,Inc., 125 F.3d 408, 410 (7thCir. 1997) (worker's masculin- ityimpugned by allegations of homosexuality, despite lack of evidence of same-sex desire); Doe v. Cityof Belleville,119 F.3d 563, 566-67(7th Cir. 1997)(worker with earring called "queer" and "fag"despite lack of evidence of same-sex desire). 512. One mightask if it is just an unfortunatecoincidence for real menthat the modality throughwhich real men constitute themselves and other real men looks awfully like the modality throughwhich failed men constitute themselves and other failed men. I believethat it is no acci- dent. To see this,imagine a worldin whichsame-sex horseplay did not exist-where men consti- tutedthemselves and others as realmen by other existing methods-by work perfornance, say, or by salary.Real mennever touch or engagein sexualbanter. Thus, whenever a mantouches anotherman or attemptssexual banter with him, the import is unambiguous-thatman is gay. What do realmen lose when they enter this imaginary world? Atleast two things. First, they lose the ability to appropriate the signifier of the "private parts" to showhow close theyare to eachother. Even if we assumethe absence of desire,it stillmakes sensethat many homosocial hazing activities involve male genitalia. This is becausethe genitalia are beingdeployed here as signifiersof theprivate, rather than as signifiersof sex (althoughof coursetheir signification as the"private" stems in partfrom their use in sex). Whensomeone's genitaliaare used in a hazingritual, the message is thatthe hazed man's most private parts are availableto themen who haze him. Suchhazing can be interpretedas a fornof domination-of forcingthe hazed man to admit that he is powerlessto keep the hazers from violating any part of his body. Orit can be seenas a fornof community-of forcing the hazed man to admit that the group is so tightthat there is no realmof the "private" that he canwithhold from the group. In eithercase, thehazed man's constitution as a "real"man will depend on hisability to demonstratethat he can survivethis incident of powerlessness. Second,they lose theability to prove that they are so powerfulthat they can engagein taboo activitieswithout suffering the consequences of that taboo. Straightmen may be mostable to con- stitutethemselves as straightmen when they engage in homosocialactivity, because this permits themto demonstratethat even when they skate very close to theedge of homosexualascription, theynever fall over it. Anotherway of thinkingabout this is to considerhow the closeted homo- sexualmight be mostafraid of engagingin "parodic"demonstrations of homosexuality, lest that ostensibleparody be recognizedas revealinghis underlying homosexuality.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 450 STANFORDLAWREVIEW [Vol. 52:353 of lettinga bisexualor a homosexualgo freeas a heterosexualinclude the unredressedharms he has visited on others. This is theprecise issue raised in thesexual harassment context.

2. Recognizingthe horseplay exemption-bisexual invisibility. In thesame-sex sexual harassment context, courts are often confronted withmale-on-male intimate conduct and askedto determinewhether that conductis homosocialor homoerotic. The priority of desire-basedclaims in thesexual harassment jurisprudence means that the harasser will be much morelikely to be liableif the conduct is deemedsexual. Thus, the task of the courtsis doublyfraught; they are not only determining sexual orientation, butalso liability. To decidethat the harassser is heterosexualis often simul- taneouslyto decide that he is notliable; to decide that the harasser is homo- sexualis oftensimultaneously todecide that he is liable. The courts'reluctance to interpretambiguous conduct as homoerotic ratherthan as homosocialis oftencarried to extremes; so longas theharasser hasnot openly admitted his homosexuality orbisexuality, courts work furi- ouslyto assignhim a heterosexualidentity. For example, in McWilliams v. FairfaxCounty Board of Supervisors,513the courtnoted that the harassers engagedin the following conduct: On at leastthree occasions, coworkers tied McWilliams' hands together, blind- foldedhim, and forced him to his knees. On oneof these occasions, a coworker placedhis fingerin McWilliams'mouth to simulatean oralsexual act. During anotherof theseincidents, a coworker,Doug Witsman,and anotherplaced a broomstickto McWilliams'anus while a thirdexposed his genitalsto McWil- liams. On yetanother occasion, Witsman entered the bus on whichMcWil- liamswas workingand fondled him.514 This conductskates close to violatingprohibitions on sodomy(a quintessen- tially sexual category)in some jurisdictions.515Nonetheless, McWilliams's coworkerscharacterized it as "horseplay."516While theMcWilliams court didnot itself use thischaracterization, other courts have described the con- ductin thiscase as "razzingand hazing"517or "merelocker room antics, joking,or horseplay."518

513. 72 F.3d 1191 (4th Cir. 1996). 514. Id. at1193. 515. Texas's sodomystatute, for example, prohibits same-sex "deviate sexual intercourse," TEX.PENAL CODE ANN. ? 21.06(a) (West1994), and defines"deviate sexual intercourse" to include"the penetration of... theanus of another person with an object."TEX. PENAL CODE ANN. ? 21.01(1)(B)(West 1994). 516. McWilliamsv. FairfaxCounty Bd. ofSupervisors, 72 F.3d 1191,1194 (4th Cir. 1996). 517. Fredettev. BVP ManagementAssocs., 112 F.3d 1503,1507 (11th Cir. 1997). 518. Tietgenv. Brown'sWestminster Motors, Inc., 921 F. Supp.1495, 1501 (E.D. Va. 1996).

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 451

Courtsrationalize the resolution of ambiguity infavor of heterosexuality in twoways. First,courts reason from the premise that the harasser has manifestedcross-sex desire to theconclusion that he thereforeharbors no same-sexdesire. Second, they state that they will not impute sexual desire to a harasserin thesame-sex context unless the plaintiff has actuallyproven thatthe harasser is homosexual.519To say that the courts do notmention bisexualityin making either of these analytic moves does not seem to be say- ingmuch. The damage that bisexuality could do tothese moves is nonobvi- ous. Butwhen we examinethese moves more closely, we see thatboth of themcan be deeplyproblematized bybisexual visibility.

3. Closingthe horseplay exemption-bisexual visibility. Bisexualvisibility significantly narrows the horseplayexemption by challengingthe two analytic moves made to keep it open. Bisexual visibility disruptsthe first move-the assertion of heterosexual status as pre-emptive of homoeroticconduct-because it makesthis status impossible to prove conclusively.When the harasser is claimingheterosexual status, all he is reallyadducing is evidencethat he manifestscross-sex desire-that he hasa wife,that he is perceivedas straightby colleagues, etc. Whenthe plaintiff is claimingthat homoerotic conduct occurred, he is imputingsame-sex desire to theharasser. In findingthese allegations to be in tension,the courts are reallysaying that allegations of cross-sex desire and allegations of same-sex desireconflict. But as notedabove,520 the premise that bisexuality exists meansthat the harasser's allegations of cross-sexdesire and the plaintiff's allegationsof same-sex desire are not necessarily intension. If both the harassers allegationsof cross-sex desire and the plaintiff's allegations of same- sexdesire are plausible, the factfinder should not assume that either is lying or mistaken,but rather should entertain the possibility that they are both right,and that the harasser is thereforebisexual. This of coursemakes it morelikely that courts will deem the conduct to be sexual,which in turn makesit more likely that they will deem it to be a TitleVII violation. One exampleof thefirst analytic move can be foundin Johnsonv. Hondo,Inc.521 Johnson and Hicks worked in an all-maleenvironment forthe Coca-ColaCompany.522 Johnson's complaint alleged that Hicks subjected himto "homosexualadvances,"523 including statements such as "I'm going

519. A thirdstrategy that the courts use is simplyto denominatethe conduct in questionas horseplayby repressingthe homoerotic possibility. This strategy relies on fiatrather than argument-thedifficult interpretive problem of whetherthe defendant is engagedin homosocialor homosexualbehavior is wishedaway rather than confronted. I do notaddress it further here. 520. See notes259-313 supra and accompanying text. 521. 125F.3d 408 (7thCir. 1997). 522. See id.at 410. 523. Id. at413.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 452 STANFORDLAWREVIEW [Vol. 52:353

tomake you suck my dick," "come down to the carwash and suck my dick," "comeacross the street and suck my dick," and actions such as simulated masturbation.524 Thecourt found that there had been no harassmentbecause it found that Johnson'stestimony belied his belief that Hicks had made a "homosexual advance"on him.525 The court maintained: Forexample, Johnson testified that on oneoccasion, Hicks approached him and said: "I'm goingto getmy dicksucked," and then,apparently referring to Johnson'sgirlfriend, Hicks said "I thinkshe's probably watching TV now. I'll go byand have that bitch suck my dick." Hicksalso said,"that bitch ought to be gettingin theshower right now .... [T]hatredhead bitch got a niceass too. I oughtto go getmy dick sucked." It is extremelydifficult toreconcile remarks suchas thesewith Johnson's strained contention that Hicks was making"homosexualadvances" toward him.526 The courtthus used Hicks's expressionsof cross-sex desire as negatingthe possibilitythat he harboredsame-sex desire. This is logicalonly if the bi- sexualpossibility is elided. If the bisexual possibility is recognized, the allegationsof cross-sex desire are irrelevant. For while the allegations of cross- sex desirearguably negate the inference that Hicks is homosexual,they do notnegate the inference that he engagedin "homosexualadvances." And it is thelatter, rather than the former, which is relevantfor the purpose of de- terminingwhether Johnson has a desire-basedclaim. Notwithstandingthis problem,however, the logic of Johnson has been used in other cases.527

524. Id. at410 n.1. 525. Id. at413. 526. Id. 527. In Boltv. Norfolk S. Corp.,22 F. Supp.2d512 (E.D. Va. 1997),the court denied a same- sex sexualharassment claim because "the record provides no basis foran inferencethat Williams [thedefendant] was homosexualother than his harassingconduct." Id. at 518. Evidencedrawn fromthe record for the opposite conclusion included: (1) "Williamswas married during most of the timein question";(2) "Bolt [theplaintiffl saw Williamsout with a womansocially"; (3) "Bolt neversaw Williamson a datewith a man";(4) "Williamsdenie[d] that he is a homosexualand aver[red]that he has neverheld any sexual attraction for Bolt"; (5) "Boltsaid he did notknow if Williamswas homosexual,or how to tell if someone was homosexual,and admitted that he didnot knowwhether Williams subjectively intended to joke, engage in horseplayor was serious."Id. at 518-19. Thecourt then noted that "[t]he only evidence other than the harassing conduct itself is a swornaffidavit by a co-workerwho avers that he was toldby co-workers shortly after joining Nor- folkSouthern that Williams was homosexual,despite the fact he was married."Id. at 519. It then notedthat "[t]his statement, however, would be inadmissibleat trialto proveWilliams actually is homosexual."Id. Similarly,in Gibsonv. TanksInc., 930 F. Supp.1107 (M.D.N.C. 1996),the court rejected a same-sexsexual harassment claim in partdue to theabsence of evidencethat the defendant was homosexual,noting that "[i]n fact,[the plaintiffs] deposition indicates that [the defendant] was marriedduring the time the alleged sexual harassment occurred." Id. at 1109. Andin Eastonv. CrosslandMortgage Corp., 905 F. Supp.1368 (C.D. Cal. 1995),rev'd and remandedon othergrounds, 114 F.3d 979 (9thCir. 1997), the court found no sexualharassment of thefemale plaintiffs by theirfemale supervisors, noting the absence of "evidencethat the defen- dantsare lesbians," and pointing out that "[i]n fact, the record demonstrates just the opposite." Id.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 453

A secondstrategy the courts use to characterize the conduct as horseplay relieson thefact that the plaintiff has not proven that the harasser is homosexual. Thus,in McWilliamsv. FairfaxCounty Board ofSupervisors,528 the courtdismissed McWilliams's suit because "no claimis madethat any [of theharassers] was homosexual."529Other courts have approved its require- mentthat a plaintiffmust generally plead and prove the homosexuality ofthe harasserin order to prevail in a same-sexharassment suit.530 Evenassuming a straight/gay binary, making liability turn on one'ssexual orientationis a potentialnightmare, raising issues of privacy and propri- ety.531And once the bisexual possibility is introduced, the inquiry becomes evenmore complicated. In a simplestraight/gay regime, the plaintiffs attemptto prove same-sex desire would be blockedby credible evidence of the harasser'scross-sex desire. But again, once the bisexual possibility is introduced,such evidence of cross-sex desire represents almost no obstacleto the plaintifftrying to make out a claimof same-sex desire. The plaintiff seeking toprove that the harasser harbored same-sex desire would thus be entitledto continuedigging for that evidence no matterhow much evidence of cross- sexdesire the harasser adduced. To myknowledge, only one court has attached consequences to itsown acknowledgementofthe bisexual possibility. In thatcase, Griswold v. Fre- seniusUSA, Inc. ,532 thecourt observed that an allegedharasser who had ad- ducedevidence of cross-sexdesire could still have engaged in homoerotic conduct.533Based on thisobservation, the court permitted the plaintiff to continuesearching for evidence of the alleged harasser's bisexuality.534 Whenthe bisexual harassment exemption is readagainst the horseplay exemption,it becomes clear that we cannotclose the first without signifi- cantlynarrowing the second. Once courts admit in one context that bisexuals exist,they cannot then turn around in anothercontext and presumethat cross-sexdesire ipso facto negates the existence of same-sexdesire. If the at 1380. Examplesdrawn from the record included (1) "thekinds of decorationschosen for [one supervisor's]birthday party"; (2) one supervisor's"flirtatious questioning of theoutside services employee";(3) "numerousdiscussions [the supervisors] participated in concerningtheir own and otheremployees' heterosexual sexual experiences"; (4) theplaintiffs' admission that "they do not knowif [the supervisors] were even interested in havingsex withthe plaintiffs, with each other, or ifthey were lesbians"; and (5) thefact that both supervisors "were married during the relevant time period."Id. at 1374,1380. 528. 72 F.3d 1191(4th Cir. 1995). 529. Id. at 1195. 530. See, e.g.,Tietgen v. Brown'sWestminster Motors, Inc., 921 F. Supp.1495, 1502 (E.D. Va. 1996)(noting but waiving requirement). 531. See McWilliams,72 F.3d at 1198 (Michael,J., dissenting) (describing pursuit of harasser's"'true' sexual orientation" as "complicated, far-ranging and elusive"). 532. 978 F. Supp.718 (N.D. Ohio1997). 533. See id.at 727-28, 730. 534. See id. at730.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 454 STANFORDLA WRE VIEW [Vol. 52:353 horseplayexemption were to be madeconsistent with the portrait of the bi- sexualpainted in thecontext of thebisexual harassment exemption, many moreputative straights would be liable. Whetherconduct was nonsexual horseplayor sexual harassment would be an issueof fact. And in the resolu- tionof that issue of fact, alleged harassers would no longerbe permittedto categoricallynegate the possibility of homoerotic conduct simply by making plausibleprotestations ofcross-sex desire.

D. SexualHarassment at a Crossroads I havedemonstrated that closing the bisexual harassment exemption has implicationsfor the horseplay exemption. Closing the horseplay exemption inturn has implications for the rest of sexual harassment jurisprudence. The jurisprudenceis currently at a crossroadswhere courts must choose whether to continueprivileging desire or to breakfrom that tradition to embracea "becauseof . .. sex" standardthat stops privileging desire. Whilethe latter routeclearly makes better sense, it is uncleareven after Oncale that the courtswill take it. Becauseit problematizes the regulation of desire, bisexu- alitycould act as a goadthat prompts the jurisprudence to depart from its historicalfetishization ofdesire.

1. Thepost-Oncale status quo. In herarticle, Schultz trenchantly argues against the pre-Oncale status quo,in whichdesire-based claims were seen as morevalid than nondesire basedclaims.535 The cruxof herargument is that Title VII prohibitsdis- criminationonthe basis of sex; that sex discrimination takes many forms that havenothing to do withdesire; and therefore that interpretations ofTitle VII thatfetishize desire will be underinclusive.536When a malewelder sabotages a femalewelder's equipment because he believesthat women should not be welders,for example, it seems clear that sex discrimination is occurring even thoughsexual desire is notobviously at issue. Schultzproposes that sexual harassmentjurisprudence should return to the statutory language and protect againstany discrimination-sexual or otherwise-that occurs "because of . .. sex."537 Thisproposal has importantconsequences for same-sex sexual harassment. In thatcontext, a ruleprivileging desire has thepernicious conse- quenceof assigningdifferent sanctions for the same conduct based on the sexualorientation ofthe harasser. In consideringthe horseplay exemption, I demonstratedthat the same conduct could be characterizedeither as hetero-

535. See Schultz,supra note 436. 536. See id. at 1796-1805. 537. See id.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 455 sexualhorseplay or homosexualharassment, and that courts made that de- terminationbased on thesexual orientation of theharasser. We also saw that,under a desire-basedparadigm, an individualwas muchmore likely to be liablefor so-called homosexual harassment than for so-called heterosexual horseplay.But this effectively means that the sexual orientation ofthe partiescan be determinativeofliability; a straight person and a gayperson coulddo thesame thing and be punisheddifferently.538 Schutlz's proposal foreclosesthis inequitable result, encouraging courts to redirecttheir atten- tionto conduct rather than simply relying on desire-based status. As Schultz'sarticle was goingto press,539the Supreme Court handed downits decision in Oncale.540The OncaleCourt's holding is completely consistentwith Schultz's proposal. The Court not only stated that the plaintiffmust "always prove that the conductat issue ... constituted'dis- crimina[tion]... becauseof . .. sex,"'541but also clarifiedthat "harassing conductneed not be motivatedby sexualdesire to supportan inferenceof discriminationon the basis of sex."542 Thus the Court held that sexual har- assmentsubsumed all (sufficientlysevere and pervasive) discrimination that occurs"because of . .. sex," including,but not limited to, desire-basedharassment. It wouldbe naifve,however, to celebrate Oncale's holding too quickly as implementingSchultz's proposal. Prior courts have adopted rules analogous

538. It mightfairly be askedwhether the same actions undertaken by a nonheterosexualand a heterosexualare truly the "same" conduct. After all, desire could be characterizednot as a status, butas an intent.And there are multiple circumstances in which the same action is treatedas two differentforms of conductbased on intent.If twomen shoot their neighbors, but one does so in- tentionallyand the other does not, the first is deemedto havecommitted murder, the second manslaughter.JOSHUA DRESSLER, UNDERSTANDING CRIMINAL LAW ? 31.02 (1995) ("The common law definitionof 'murder'is 'thekilling of a humanbeing with malice aforethought.' Manslaugh- teris 'an unlawfulkilling of a humanbeing without malice aforethought."'). The same act- shootinga humanbeing-is deemedto be two legallydistinguishable conducts-murder and manslaughter-basedon somethingthat is entirelywithin the individual mind. Thisdistinction is arguably deployed in theCourt's hypothetical in Oncale, where the Court distinguishesbetween the coach smacking the buttocks of a male footballplayer on thefootball fieldand smacking the buttocks of a malesecretary back in theoffice. See Oncalev. Sundowner OffshoreServs., Inc., 523 U.S. 75, 81-82(1998). Thesmacks are the same act, but are treated as twolegally distinct forms of conduct. The first, in the Court's view, is clearlyhorseplay, the second is justas clearlyharassment. See id. at 81. The distinctionbetween them, the Court says, is context.See id. at 81-82. Butcontext here is shorthandfor desire-in the first instance the conduct is readas notdesirous, in the second it is readas theopposite. The flawin theanalogy is thatdesire should not be equatedwith bad intentfor at leasttwo reasons.First, desire may not always be withinindividual control. Second, desire can be seenas havingmany positive aspects, such that adding desire to an otherwiseacceptable act does notin itselfmake it worse. 539. See Schultz,supra note 436, at 1683n-t. 540. 523 U.S. 75 (1998). 541. Id. at81. 542. Id. at80.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 456 STANFORDLAWREVIEW [Vol. 52:353 to Oncale'sholding only to havethem ignored or underminedin practice. As Schultznotes,543 the D.C. Circuitheld in the1985 case ofMcKinney v. Dole544that harassment could be cognizableunder Title VII evenif it did not involve"sexual advancesor ... otherincidents with clearly sexual over- tones."545The McKinney court found that "any harassment orother unequal treatment... thatwould not occurbut forthe sex of the employeeor em- ployeesmay, if sufficiently patterned or pervasive,comprise an illegalcon- ditionof employment under Title VII."546 Seven other circuits adopted ana- loguesof this rule,547 which sounds remarkably like the rule adopted in On- cale. TheMcKinney rule, however, has hadlittle influence.548 Some subse- quentcourts have outright ignored it. Othercourts have undermined it by recognizingboth sexual harm and nonsexual harm, but by privilegingthe formerover the latter. In otherwords, while the rule articulated inMcKinney madeit clearthat both nonsexual and sexual harms were cognizable, it left openthe question of the relative emphasis courts should place on eachkind ofharm. Courts were thus not constrained by the McKinney rule from con- tinuingto privilege desire. Althoughthe authority ofthe Supreme Court makes it unlikelythat the Oncaleholding will be ignored,it couldbe underminedin exactly the way thatthe McKinney rule was undermined.Indeed, the norms articulated by theOncale Court seem almost to invitea continuedfetishization of desire. Thiscan be seenin itsdiscussions of the two requirements for a successful sexualharassment claim: the "because of . .. sex" requirementand the"se- verityand pervasiveness" requirement. In enunciatingits "because of . . . sex"requirement, theCourt began by statingthat "[c]ourts and juries have found the inference of discrimination easyto drawin most male-female sexual harassment situations, because the challengedconduct typically involves explicit or implicit proposals of sexual activity;it is reasonableto assumethose proposals would not have been madeto someone of the same sex."549 The Court here privileged the desire- basedroute to proving sex discrimination by:(1) characterizingdesire-based conductas the"typical" form of male-femaleharassment (without noting thatit may be typicalonly because of the historical fetishization ofsexuality bythe courts); and (2) notingthat the inference of discriminationis "easy to draw"in contexts where desire is present.

543. See Schultz,supra note 436, at 1733. 544. 765 F.2d 1129(D.C. Cir.1985). 545. Id. at1138. 546. Id. 547. See Schultz,supra note 436, at 1733n.250. 548. See id. at 1732-38. 549. Oncalev. SundownerOffshore Servs., Inc., 523 U.S. 75,80 (1998).

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 457

TheCourt also appeared to privilege desire-based claims in its discussion ofthe "severity and pervasiveness" requirement. It noted: "A professional footballplayer's working environment is not severely or pervasively abusive ... ifthe coach smackshim on thebuttocks as he headsonto the field-even if thesame behavior would reasonably be experiencedas abusiveby the coach'ssecretary (male or female) back at the office."550 But if the smack of a football'splayer's buttocks sounds different from the smack of a secre- tary'sbuttocks because the former smacks less of sexuality, this suggests that conductis morelikely to be deemedsevere and pervasive if it is desire- based. Thus,while the Supreme Court explicitly held that both desire-based harmsand nondesire-based harms should be cognizable,it arguablycontin- ued to privilegedesire-based harms over nondesire-based ones. Whatwill stoplower courts from doing the same?

2. Bisexualityas goad. Bisexualitymay be deployedto help the courts down the right path after Oncale. I earliernoted that the tendency of bisexualityto confusesexual orientationcategories had pro-gay deployments. One suchdeployment was thatbisexuality could threaten straights with falling into the stigmatized categoryof bisexuality.551 The theory was that the more likely straights were tobe threatenedwith homoerotic ascription, the less likely they would be to assignliability to that ascription. Oneway to see how this works is toask why the Court is so surethat the coachsmacking the football player on thebuttocks as he headsout onto the fieldis engagedin horseplay.The answer seems to be thatthis is football; thatsome homosocial acts are so transparentlyhomosocial that they cannot be readany other way. Butthis, as we havealready seen, is simplyuntrue- thehomosocial and the homosexual are inextricablyintertwined. Anyone whothinks that football is alwayshomosocial needs to review his Tennessee Williams.552The bisexual possibility powerfully exposes the illogic of this conclusorydefense. Whenthat defense is strippedaway, it becomesclear that the reason footballcannot be homoeroticis that some homosocial acts are so valuedthat theycannot afford to be taintedwith homosexuality.553 Thus, even if ambi-

550. Id. at 81. 551. See notes303-304 supra and accompanyingtext. 552. See generallyTENNESSEE WILLIAMS, CAT ON A HOT TIN ROOF (1955) (play turningon thehomoerotic undertones of "homosocial"interactions between football players). 553. Anotherexample of thisdynamic might be theNational Organizationof Women's hos- tilitytoward lesbians in the 1970s. See BARRY D. ADAM, THE RISE OF A GAY AND LESBIAN MOVEMENT97 (rev. ed. 1995) (notingthen-President Betty Friedan's descriptionof lesbians as a

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 458 STANFORDLA WRE VIEW [Vol. 52:353 guitybetween the homosocial and the homosexual is recognizedat one level, itmust be repressedon another. Butthis answer-that while the homosocial/homosexual binary is intrin- sicallyunstable, straights desperately try to preserveits stability-should serveas a cluefor those arguing against the regulation of desire. For if that binarycan be shownto be so unstablethat self-identified heterosexuals themselvesbecome threatened with potential homoerotic ascription, they too willhave an investmentinnot having liability turn on orientation.554 Thisis bisexuality'sfinal and most tentative critique of sexualharass- mentlaw. By destabilizingthe homosocial/homoerotic binary, bisexuality revealsthat sexuality is so fluidthat heterosexuality is impossible to prove. Historically,the reaction to thebisexual threat has beenrepression, but bisexualvisibility has becomesufficiently great that this repression has be- comemore difficult. Increased bisexual visibility makes the deployment ofa desire-basedparadigm more risky for straights, who will encounterchal- lengesnot only to theirconduct (horseplay) but also to theirstatus (heterosexual). To theextent that self-identified straights have investmentsin keepingthose statuses immured from scrutiny, a different form of repression is inorder. It is nownot the bisexual, but desire itself, that must be abjured. In sexualharassment jurisprudence, this means that the courts may wish to leavethe path they have been on, shifting from a desire-basedparadigm of sexualharassment toa broader"because of ... sex"paradigm.

E. OtherApplications

Althoughbisexuals are most visible in sexualharassment jurisprudence, bisexuality'spotential to effect changes in doctrine is notlimited to that area. I do no morethan touch on twoother areas in which it might make a difference-the"don't ask, don't tell" policy and same-sex marriage. The"don't ask, don't tell" policy raises issues analogous to thosein the sexualharassment context in that individual liability turns on sexualorientation.An individualdeemed homosexual will be evictedfrom the military if he engagesin homosexualconduct, such as same-sexsodomy.555 An individualdeemed heterosexual, however, can engage in the exact same conduct,

"lavendermenace" to thefeminist movement); ROSALIND ROSENBERG, DIVIDED LIVES: AMER- ICAN WOMENIN THE TWENTIETHCENTURY 208 (1992) (describingFriedan's view that"NOW couldnot survive ... iffeminism was identified with lesbianism"). 554. Whilethis strategy is admittedlyspeculative, it has parallelsin therace context in the ante-bellumera. As ArielaGross observes, abolitionists and fugitiveslaves writing in theNorth usedcases in whichindividuals who appeared "white" were enslaved to bringhome the horror of slaveryto white audiences. See ArielaGross, Litigating Whiteness, 108 YALE L.J.109, 127 (1998). Suchaccounts were meant to encourage "white readers to imagine themselves in theenslaved [person's]shoes." Id. 555. See 10U.S.C.A. ? 654(b)(1)(West 1994).

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 459 andnot be so evictedunder the "queen for a day"exemption.556 Here, the heterosexualascription performs an evenmore powerful role because it pro- tectsindividuals even from the consequences of their concededly homoerotic (ratherthan homosocial) acts. No one is arguingthat same-sex sodomy is simplyheterosexual horseplay, but sodomy will not have consequences for theindividual who can prove that he is "heterosexual."Again, sexual orien- tationstatus defines liability. Again, there is no reasonwhy bisexuality couldnot destabilize those statuses in sucha wayas to forcestraights to in- terrogatesuch a desire-basedparadigm. Atthis point, a bisexualactivist might comment on thefact that the two examplesI havethus far adduced-sexual harassment and "don't ask, don't tell"-linkincreased bisexual visibility to increasedbisexual liability. How shouldbisexuals feel about the fact that they are being rendered more visible simplyto be heldmore accountable? Is thisdeployment ofbisexual visibil- itytruly pro-bisexual? Or is itpro-gay and anti-bisexual? Insofar as these examplesare concerned, itwould certainly seem that bisexuals are being de- ployedeither to outself-identified straights as bisexuals or to threaten them withbeing outed. Of course, as canbe seenfrom the gay context, outing is notnecessarily against the interests ofthe group into which the individual is outed:even when costly to theindividuals involved, it has been viewed by manyas beneficialto thecommunity at large.557 It mightbe helpful,however,to thinkof a legalinstance in whichbisexual visibility might actually redoundto the benefit of the individual bisexual litigant. Marriagemay be suchan example.Both gays and bisexuals can argue thatthe cross-sex requirement ofmarriage violates sex discrimination norms becauseit permitsone sex butnot the other from marrying members of a givensex. Thebisexual, however, might express an additionalharm. This harmis thatthe state is contributingto sex-consciousness in society by dis- tinguishinginthis way between men and women. For the sex-blind bisexual, thisconsciousness is in itselfa harm,because it impedesher from seeing "through"sex to othertraits that she may find more important. Because the homosexual,just as muchas theheterosexual, does not purport to be blindto sexin this way, this is nota harmthat he will adduce.558 The claimthat the state harms the bisexual by forcingher to viewthe worldthrough the lens of sex is, ofcourse, a ratherutopian one. Thisis es-

556. See 10 U.S.C.A. ? 654(b)(1)(A)-(E) (West 1994). 557. See, e.g., WARRENJOHANSSON & WILLIAMA. PERCY, OUTING: SHATTERINGTHE CONSPIRACYOF SILENCE229 (1994) (notingthat while outing impinges on an individual'sprivacy, its neteffects can be beneficialfor the group). 558. Thereis anotherdistinction between sexual harassmentand "don't ask, don't tell" on the one hand and marriageon theother. While sexual harassmentand "don't ask, don't tell" primarily implicatebisexuality's influence on normsof sexual orientation,the marriagecontext primarily implicatesbisexuality's influence on normsof sex. This is because themarriage context is one that technicallydiscriminates on thebasis of sex ratherthan on sexual orientation.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions 460 STANFORDLA WRE VIEW [Vol.52:353

peciallytrue given my expressed doubt as to whetherany bisexual truly is sex-blind.559Yet a bisexualcould respond by noting that even if she is not completelysex-blind, she has a rightto aspire to lowering the salience of sex in herlife. She couldalso attemptan analogyto race,noting that the Su- premeCourt has ownedthe American judiciary's obligation not to increase race-consciousnessinsociety.560

CONCLUSION

"We havenot yet reached a point,"Garber notes, "though nothing is unimaginable,at which monosexuals, hetero- and homo-, band together to staveoff the advancement ofbisexuals."561 This article has arguedthat we have actuallynever left that point. It has maintainedthat bisexuality is erasedbecause self-identified straights and self-identified gays have entered intoan epistemiccontract of bisexual erasure. And it has suggested that the risingvisibility of bisexualitymight have transformative consequences for howwe thinkabout sexual orientation politics and law. The strengthof thishypothesis will depend in parton itsability to ex- plainthe erasure of bisexuality in dailylife. I thusend where I began,with theanecdote of my own erasure of bisexuals in my seminar on sexualorientation.Do theexplanations I provide intuitively explain why that erasure occurred?In onesense, the answer must be no. I do notthink that I erased bisexualsbecause of a fearof not being able to prove my orientation, orout of a fearof failingto retain sex as an importantdiacritical axis, or outof a fearof nonmonogamy. Rather, my erasure can be mosteasily explained by thefact that bisexuality is not part of our cultural or semantic stock: it is dif- ficultconstantly toread cases, articles, or popular accounts that erase bisexuals withoutfollowing suit. But this just displaces the question from the indi- viduallevel to the cultural level, for it is ultimatelynot that interesting simplyto point out that individuals erase bisexuals because they belong to cul- turesthat do. Andwhen the question arises of whatcultural investments

559. See notes317-318 supra and accompanying text. 560. See, e.g.,Adarand Constructors, Inc. v. Pena,515 U.S. 200, 228-29(1995) (approving prioropinion that criticized affirmative action program for delaying "'the time when race will becomea trulyirrelevant, or at leastinsignificant, factor"') (quoting Fullilove v. Klutznik,448 U.S. 448, 545 (1980) (Stevens,J., dissenting)); Shaw v. Reno,509 U.S. 630,657 (1993) (warningthat "[r]adicalgerrymandering, even for remedial purposes ... threatensto carryus furtherfrom the goalof a politicalsystem in which race no longermatters-a goal that the Fourteenth and Fifteenth Amendmentsembody, and to whichthe Nation continues to aspire");City of Richmondv. J.A. CrosonCo., 488 U.S. 469, 495 (1989) (criticizing"[t]he dissent's watered-down version of equal protectionreview" for ensuring "that race will always be relevantin Americanlife, and thatthe ultimategoal of eliminat[ing]entirely from governmental decisionmaking such irrelevant factor's as a humanbeing's race, will never be achieved"(citation omitted)). 561. GARBER,supra note 11, at 80.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions Jan.2000] BISEXUALERASURE 461 mightlead to theerasure of bisexuals, I believe the investments I have de- scribedgain in plausibility. Thedegree to which the hypothesis ofthis article is heldaccountable to livedexperience deserves more global comment. Colleagues have repeat- edlynoted this article's obsession with logic-its insistenceon analyzing bisexualityin strictlyCartesian terms. That observation has oftenbeen ac- companiedby the exhortation to give less to thereader's head and more to herexperience. The implication ofthe exhortation is that the logical account ofbisexuality leaves something important about bisexuality unconsidered. Thisis doubtlesstrue. The logical approach of the article may be readas compensationfor the often parlously imprecise terms in which debates about sexualityin general and bisexuality inparticular are conducted. Yet the fact thatit may also be readas overcompensationis important. Sexual identity hasalways struck me as a kindof illogic, given that sexuality is sucha pow- erfulsolvent of identity, a modality that expands the consciousness through shockand surprise. If this is right,then bisexuality may be thesexual identitythat best reflects the oxymoronic nature of all sexualidentity, insofar as bisexuality,too, is a contradiction,a class and its own dissolution. This may explainwhy explanations ofbisexuality that seek to tame bisexuality within thebounds of Cartesian reason will always feel anxiously incomplete. But thishas consequencesfor the law, whichis oftena projectthat privilegessuch reason. It maymean that if we are concernedabout the "logical"regulation of sexualityas failingto respectsexuality's fluid and narrativenature, we mightdo worsethan to beginby looking at thesexual identity-bisexuality-thatbestrepresents that nature. Properly harnessed, bisexuality'sdestabilizing force may be a powerfulmeans of contesting that regulation.

This content downloaded from 130.132.173.76 on Sun, 2 Jun 2013 16:33:04 PM All use subject to JSTOR Terms and Conditions