Before the last fascicle of the OED was published, the scientific principles of its compilers had already started to influence other dictionaries. The editor of Chambers’s Twentieth Century Dictionary of the English Language acknowledged the indebtedness of his etymologies to ‘the magistral New English Dictionary of Dr James A. H. Murray and Mr Henry Bradley, so far as completed’ (Reference DavidsonDavidson 1901: iii). Cassell’s New English Dictionary forcefully concurred, ‘In these days it would be an impossibility to put forth an English Dictionary that had any pretence to philological accuracy but for the labours of […] Sir James Murray with his co-workers’ (Reference BakerBaker 1919: xi). In The Universal Dictionary of the English Language, Reference WyldHenry Cecil Wyld (1932: v) hailed the OED as a ‘wonder of the age’. Though none of these British lexicographers claimed to have derived their entries from quotational evidence in the same manner as the OED, models of descriptivist, citation-based lexicography were quickly adopted by several American dictionaries, all of them also published in Britain.

When the first part of The Century Dictionary appeared, its editor Reference WhitneyWilliam Dwight Whitney (1889 I: v) declared that the ‘first duty of a comprehensive dictionary is collection, not selection’ – echoing the sentiments of the London Philological Society’s Proposal thirty years before. Yet selectivity would unavoidably slip into The Century Dictionary as it did into the OED. It is present in Whitney’s own boast that The Century Dictionary’s definitions were informed by ‘a special collection of quotations selected […] from English books of all kinds and of all periods of the language, which is probably much larger than any that has hitherto been made for the use of an English dictionary’ – except ‘that accumulated for the Philological Society of London’ (xi).Footnote ¹ A similar commitment to empiricism was expressed (again with a touch of equivocality) in Funk & Wagnalls’s Standard Dictionary of the English Language, which agreed that ‘A dictionary, says Archbishop Trench, should be an inventory of the language’, the main function of which was ‘to record usage; not, except in a limited degree, to seek to create it’ (Reference FunkFunk 1893 I: vii). Thus, the usages recorded in the Standard Dictionary had been drawn by nearly 500 readers from ‘the works of several thousand representative authors, of the various ages of English literature from Chaucer’s time to the present’ (vii). As the editor of Webster’s New International Dictionary would later observe, the OED had ‘caused the public to expect larger things of other dictionaries’ (Reference HarrisHarris 1909: vii).

Webster’s New International was clear about its own ‘inestimable debt’ to the OED, and reiterated the descriptivist credo that ‘it is the function of a dictionary to state the meanings in which words are in fact used, not to give expression to the editors’ opinions as to what their meanings should be’ (Reference HarrisHarris 1909: vii). Though the New International’s first edition did not explain the basis of its entries, its second was founded on a collection of over one million quotations drawn from thousands of sources (Reference NeilsonNeilson 1934: vii). The collection had grown to six million by the time the third edition was published, and its editor-in-chief, Philip Babcock Gove (Reference Gove1961: 6a), was even firmer in his rejection of ‘prescriptive and canonical definitions […] unless confirmed by independent investigation of usage borne out by genuine citations’.Footnote ² Merriam-Webster had at last surpassed the quotation files of the OED, which had been based on ‘some five millions of excerpts from English literature of every period amassed by an army of voluntary readers and the editorial staff’, as stated in the preface to its 1933 reissue (Reference 289Craigie, Onions, Murray, Bradley, Craigie and OnionsCraigie and Onions 1933a: v). Still, history has borne out Craigie and Onions’s boast about the OED’s empirical ideals: ‘It is a fact everywhere recognized that the consistent pursuit of this method has worked a revolution in the art of lexicography’ (v). Descriptivist principles have dominated the profession ever since.

The rise of highly collaborative, quotation-based dictionaries has coincided with a decline in the personal prominence of lexicographers. A century ago, George Philip Reference KrappKrapp (1925: 375) witnessed ‘the disappearance of the individual in the making of modern dictionaries, and the emergence of what may be called the syndicate or composite dictionary’. Whereas ‘older dictionaries depended for their value on a name’ such as Johnson, the modern dictionary is ‘the work of numerous scholars, specialists and compilers whose names are altogether unknown to the persons who use the dictionaries’ (375). While this shift means that the names of a few editors are less likely to overshadow the labour of assistants and volunteers, it has had a depersonalizing effect. Raymond Reference WilliamsWilliams (1983: 18) observed that there is an ‘air of massive impersonality’ about the OED, albeit it is ‘not so impersonal, so purely scholarly, or so free of active social and political values as might be supposed’. The same is true of other syndicate dictionaries: their definitions are effectively anonymous, no longer visible as the work of individuals whose particular biases will inflect their treatment of the language whatever the evidence at their disposal.

In this concluding chapter, I want to move beyond the temporal scope of the rest of the book to look at how descriptivist ideology has developed since the early twentieth century, initially for print dictionaries and more recently online. I will also consider some of the competing models of dictionary-making that have emerged with the rise of virtual crowdsourced dictionaries as alternatives to professional reference works. In all cases, my aim is to make clear the representational limits that remain inherent in even the most modern of projects – limits that must be borne in mind by any researcher who relies on dictionaries as a window onto the linguistic present or past.

Common Usage and Minority Speakers

In On Some Deficiencies in Our English Dictionaries, Richard Chenevix Reference TrenchTrench (1857: 6) famously described the ideal dictionary as ‘an historical monument, the history of a nation contemplated from one point of view’. But whose point of view is to be memorialized at the crossroads of language and nationhood? Trench had already answered this question in one of his earlier books, English, Past and Present. There, he wrote in praise of ‘the true instinct of the national mind’, which had across the history of English ‘rejected and disallowed’ from the language ‘words [that] were not idiomatic, or were not intelligible, or were not needed, or looked ill, or sounded ill’ (Reference TrenchTrench 1855: 53–54). This singular perspective of the public genius had set ‘the stamp of popular allowance’ on what was and was not English (53). Thus, Reference TrenchTrench (1857: 7–8) went on to write in Deficiencies, it is not up to lexicographers themselves to apply ‘the stamp of non-allowance’ when determining the content of a dictionary; rather, it is ‘for those who use a language to sift the bran from the flour, to reject that and retain this’. Lexicographers need not be selective, for the nation has selected for them.

It goes without saying that Trench’s view of a harmonious and homogeneous speech community was wilfully naive. Real communities are marked by heteroglossia, the cacophonous ‘multiplicity of social voices’ – a concept that Mikhail Reference BakhtinBakhtin (1981: 263) was incidentally developing in the same years the OED’s first Supplement was being compiled. OED1’s quotation banks achieved an unprecedented degree of multivocality, and yet the views expressed on sexuality in its selected quotations and associated definitions differed little from the lexicography of prior centuries: same-sex acts, actors, and desires were immoral, unnatural, criminal, exotic, pathological, and unspeakable.

The inclusion of the socially and linguistically marginalized remains a problem for present-day dictionaries – although, since OED1, the authority of descriptivist lexicography has been enhanced by the promises held out by electronic data processing. In 1969, the editor of the American Heritage Dictionary announced that it had been compiled with the aid of a machine-readable corpus of texts amounting to about a million words, in addition to traditional paper citation files (Reference MorrisMorris 1969: 201). Two decades later, the British Collins COBUILD English Language Dictionary signalled a turn from corpus-assisted to corpus-based lexicography by making an electronic database of twenty million words its empirical foundation (Reference SinclairSinclair 1987: xv). Other dictionaries soon followed suit. The digital revolution came too late for the second Supplement to the OED, published in four volumes from 1972 to 1986, but by the time the final volume was released, the ‘computerization of the OED’ was already underway (Reference Burchfield and BurchfieldBurchfield 1986: xii). The product of this computerization was OED2 (1989), an amalgam of the first edition and its two Supplements which added about five thousand new words and senses of its own. For the last of these, the editorial team supplemented the OED’s traditional quotation slips with ‘large computer databases containing research abstracts, newspaper and periodical texts, and legal reports’, which were ‘invaluable in […] providing examples to complete the lexicographical record’ (Reference Simpson and WeinerSimpson and Weiner 1989 I: xxii).

The ‘corpus revolution’ has allowed lexicographers to analyse the semantic and grammatical contexts in which words are used across a range of text types on a hitherto inconceivable scale. As the decades have passed, their size has increased exponentially. Merriam-Webster now combines a physical citation file of 15.7 million quotations with an electronic database of ‘more than 70 million words drawn from a great variety of sources’ (‘Help’ n.d.). Collins’s current dictionaries are based on the Bank of English – a 650-million-word subset of the 4.5-billion-word Collins Corpus – which is built from ‘a carefully chosen selection of sources, to give a balanced and accurate reflection of English as it is used today’ (‘The History of COBUILD’ n.d.). At almost 2.5 billion words, the Oxford English Corpus used by lexicographers at Oxford University Press ‘represents all types of English’ from ‘all parts of the world’ (‘The Oxford English Corpus’ n.d.).Footnote ³

Paradoxically, with the rise of corpus-based lexicography, the dictionary’s construction of authority appears to have become circular. The public defers to lexicographers to learn how language is (or should be) used, and lexicographers learn how language is used by deferring to the public. Yet what makes the linguistic knowledge of the lexicographer more definitive than that of the layperson is the volume of information at the lexicographer’s disposal (Reference CameronCameron 1995: 48). An individual language user might be unaware of the consensus about a particular word’s meaning, spelling, or pronunciation, and lexicographers, with the help of corpus evidence, provide access to that consensus. Ostensibly, this has allowed dictionaries to shrug off the last vestiges of partiality and become truly objective reflections of English as it is used by the majority of speakers. The authority of dictionaries has, in a way, shifted from ipse dixits to argumenta ad populum: gestures to the prestige of canonical authors in earlier centuries have been replaced with appeals to language users collectively.

Yet, as the American Heritage editor Reference GrahamAlma Graham (1974) long ago observed, corpora are not neutral. Any claims about their representativeness are unsettled by Stuart Reference Hall, Gurevitch, Bennett, Curran and WoollacottHall’s (1982: 64) argument that representation is not a passive duplication of reality but ‘implies the active work of selecting and presenting, of structuring and shaping: not merely the transmitting of an already existing meaning, but the more active labour of making things mean’. As such, attempting to represent the common usage of a language is a performative act. Consensus is hardly ever unanimous, and a corpus will inevitably amplify certain perspectives over others – depending on what texts are available, which of these are earmarked for inclusion, and how the data are searched and interpreted by the lexicographer.

What is more, when a dictionary is intended to document a ‘majority view’ of language, the position of minority groups is necessarily precarious. In an eloquent meditation on the social power of pejoratives, the lexicographer Kory Reference StamperStamper (2017: 151–52) asks, ‘when the force of a profane word or slur is felt and perceived differently person to person, how can lexicographers possibly concisely communicate […] its full range of use?’ We could further ask: if a word used to refer to a minority group is considered a slur by many of the group’s members but not by the majority of speakers in a speech community (however that community is defined), should lexicographers side with popular usage, or should they respect the views of some minority members and label the word ‘offensive’ or ‘derogatory’? How do dictionaries adjudicate among diverging popular, radical, and reactionary uses of words whose meaning is subject to intense political debate, such as marriage, family, sex, gender, woman, and man? Do lexicographers have a duty to record the identity terms of marginalized groups if those terms and identities are ignored or rejected by mainstream society? How large or loud does a minority have to be before the language used by and about it warrants a place in a general dictionary?

In a paper delivered at New York’s International Conference on Lexicography in English, June 1972, the sociolinguist David W. Maurer made this prediction:

Maurer did not identify the groups responsible for this social ferment, but his listeners must have drawn up their own lists: hippies, environmentalists, Black Power activists, second-wave feminists, gay liberationists. Incidentally, the conference was held at the Hilton Hotel – two and a half miles from the Stonewall Inn, and two and a half weeks before campaigners would take part in New York’s third annual gay pride march.

For queer scholars, 1969 is less memorable as the year that heralded the corpus revolution than as the year that witnessed the Stonewall riots. In the decades since the riots gave strength and visibility to the modern LGBTQ+ rights movement in North America, Britain, and beyond, activists have pushed back against heterosexism in all areas of society, including language. As Bruce Reference RodgersRodgers (1972: 11) noted in the introduction to his gay slang dictionary, The Queens’ Vernacular, words are a means of both imposing and resisting power: ‘Slang is social protest [… it] is also the expression of the underdog – it is always aimed against the establishment.’ The establishment’s reaction was ambivalent. Robert W. Burchfield, editor of the OED’s second Supplement, wrote a review of The Queen’s Vernacular for The New York Times in which he praised Rodgers for braving the ‘unattractive aspects of modern life’ to collect the ‘sad language of sexuality’ (Reference Burchfield1972b: BR22). And yet, shortly before travelling to the 1972 lexicographers’ conference in New York, Burchfield sent a letter to Oxford University Press in which he expressed the wish that the attendees might devise ‘a policy of concerted action against the various pressure groups that plague the dictionary houses’.Footnote ⁴ Compiling alternative lexicons was one thing; criticizing mainstream dictionaries was another. The greatest thorns in Burchfield’s side were public reactions to the handling of racial, religious, and national terms (see Reference BurchfieldBurchfield 1989: 109–15), but other lexicographers began to reconsider their treatment of same-sex vocabulary as well.

In the 1970s, the makers of the American Heritage Dictionary asked their usage panel of 150 language pundits to decide whether using gay to mean ‘homosexual’ was ‘appropriate’ in formal contexts. Authors Annie Dillard and Sheridan Baker were in favour; journalist Russell Baker and science fiction writer Isaac Asimov were opposed (Reference ShenkerShenker 1979: 149). In the eighties, dictionaries’ labelling of homophobic slurs as ‘derogatory’ or ‘offensive’ was patchy; by the end of the nineties, it was fairly consistent (Reference McCluskeyMcCluskey 1989: 112–13; Reference NorriNorri 2000: 103). Implementing changes like these has become easier as dictionaries have gone digital: updates that would once have taken years to appear in a new print edition can now be published online in an instant. Over the last two decades, in response to changing civil laws and social attitudes, online dictionaries have removed gender restrictions from definitions of marriage (Reference StamperStamper 2017: 230–54) and revised entries for same-sex lexis to avoid negative evaluations such as ‘unnatural’ (Reference BrewerBrewer 2013: 112–14). But in other, less overt ways, dictionaries still echo the discursive asymmetries of the past.

Early English lexicographers’ disregard of Latin loanwords for sexuality between women finds a contemporary echo in certain French and German dictionaries’ failure to admit queer and coming out as common loanwords from English (Reference Nossem, Baer and KaindlNossem 2018: 182–83). Current definitions of sex and sexual intercourse can elide same-sex acts as effectively as did eighteenth-century definitions of copulation (Reference Pakuła and PakułaPakuła 2021a; Reference TurtonTurton 2021b). Although the application of ‘offensive’ labels to slurs may appear to defer to the feelings of the people they target, these labels tend to centre the degrading usage of mainstream speakers rather than the words’ ironic reuse or affirmative reclamation by marginalized speakers (Reference RussellRussell 2021: 237). In general dictionaries, the often pathologizing word homosexual is still presented as the default term for same-sex attraction in English, while identity labels championed by in-groups, such as lesbian and gay, are positioned as socially marked (Reference TurtonTurton 2020). Even as the formats of dictionaries have changed, old prejudices continue to circulate in them. Descriptivist principles do not eliminate this social and linguistic inequality: they simply conceal it.

At the same time, the Web has enabled methods of dictionary-making that depart from the ‘majority rule’ approach of professional lexicography. Crowdsourced dictionaries, to which anyone with an internet connection can contribute, avoid some of the problems of representing the diversity of English by letting diverse speakers represent themselves. A lexicon with an abundance of virtual space and volunteers will almost certainly contain material that is absent from dictionaries maintained by smaller teams of workers. At present, OED Online includes entries for around 500,000 words and phrases (‘About the OED’ 2023). The English version of Wiktionary (2002–), one of the largest user-generated dictionaries, has over one million entries (‘Statistics’ 2023). Nonetheless, open collaboration comes with its own hazards. Amateurs have neither the resources and training of professionals nor the accountability of employees, and the crowdsourcing of dictionaries has not transformed them into egalitarian utopias. There remain representational biases among volunteers and, by extension, in the content they create: see, for instance, the gender gap on Wikipedia, Wiktionary’s encyclopaedic counterpart (Reference Hargittai and ShawHargittai and Shaw 2015; Reference Wagner, Graells-Garrido, Garcia and MenczerWagner et al. 2016). Yet diversity on its own is not a perfect solution. The more contributors a project has, the greater number of viewpoints it will need to contend with. One need only look at Wiktionary’s editorial discussion page for the headword marriage – where semantic arguments have ranged over issues of gender, religion, miscegenation laws, polygamy, and consensual extramarital relationships – to see that disagreement is an inherent part of language use.

Some online lexicons have avoided presenting the illusion of consensus by abandoning the conventional dictionary structure of one entry per headword, instead allowing multiple users to post different entries for the same word. One such dictionary is Queer Undefined, founded by Sara Goldstein-Weiss in 2017, whose ‘About’ page explains that it is meant to be an antidote to mainstream dictionaries that ‘give an impersonal dictionary definition’ of ‘LGBTQ+ terms’. Instead, Queer Undefined ‘gather[s] informal definitions from LGBTQ+ people – to give multiple personal perspectives’. This is borne out by the dictionary’s page for bisexual, which currently lists fifty-nine entries. A note at the top of the page says, ‘Disagree with these definitions? Add your own.’ However, even this conscious heteroglossia offers no escape from questions of linguistic authority and regulation. Rather than a corpus or a literary canon, the legitimacy of entries on Queer Undefined comes to rest on what Sally Reference McConnell-GinetMcConnell-Ginet (2020: 240) calls ‘first-person semantic authority’: it is speakers’ identification with a particular term that validates their definition of it. The site’s submission form not only invites users to ‘Define [a] term and/or say what it means to you’, but asks ‘What has this term meant in your experience? How have you embodied this term?’ and ‘How do you identify?’ Once an entry has been submitted, it is reviewed by the site administrator. If it is found not to adhere to the dictionary’s guidelines and ‘generally disrespects the LGBTQ+ community’ or ‘dismisses/erases a particular LGBTQ+ identity’, it is rejected. In order to uphold civil discourse, language is still subject to gatekeeping.

Of the crowdsourced projects that have dispensed with the need for civility, none is more notorious than Urban Dictionary (1999–). Chiefly intended to record slang and other informal usage, the website was ‘set up in opposition to established authority and is, by its nature, irreverent’ (Reference Peckham, Coleman and ColemanPeckham and Coleman 2014: 186). As on Queer Undefined, Urban Dictionary locates its authority in its contributors, ‘the speakers of the informal language it seeks to document’ (186). Unlike on Queer Undefined, editorial policing on Urban Dictionary is ostensibly decentralized. While the actual workings of the website are opaque, on the surface, just as anyone who signs in can submit an entry, so anyone can act as an editor by viewing new submissions and voting on whether they should be published. Although Urban Dictionary claims not to tolerate definitions that ‘harass, discriminate, and/or directly incite violence against others’ (Reference Guidelines‘Content Guidelines’ n.d.), the site is riddled with abusive content often directed at women and minorities (Reference TurtonTurton 2021a). Rather than alleviating linguistic and social ostracism, crowdsourced lexicography may exacerbate it.

For researchers of present-day English, while these user-made dictionaries are not scholarly reference works, they can be rich primary sources when handled with care. Urban Dictionary’s unfiltered spontaneity offers ‘raw materials that are invaluable to any serious slang lexicographer’, Julie Reference ColemanColeman (2012: 291) observes. Meanwhile, Queer Undefined could supply scholars of language, gender, and sexuality with exemplary data on how the same identity labels may be differently understood by the varied people who use them. For the researcher of English before the twenty-first century, there can be no real equivalent to these extremely multivocal, grass-roots lexicons. The closest it is possible to come is in the curated quotation banks of contemporary dictionaries with a historical scope, such as the OED, the Middle English Dictionary (1952–2001), and the ongoing Dictionary of Old English (1986–); corpora connected to the latter two projects have been available online since 1997 (DOE) and 1998 (MED). But because of professional dictionaries’ advertising of themselves as descriptive works, researchers have not always approached them with as much caution as they might show to other texts.

I will give two examples relating to the OED. In the 1970s, the dictionary’s second Supplement recorded that the earliest known use of homosexual in English was in 1892, when Richard Krafft-Ebing’s sexological treatise Psychopathia Sexualis was translated out of German. Shortly afterwards, John Reference BoswellBoswell (1980: 42) pointed out that the OED’s evidence could be antedated: homosexual had appeared in the 1891 edition of A Problem in Modern Ethics, John Addington Symonds’s plea for the social tolerance of same-sex desire. Nevertheless, a decade later, David Reference HalperinHalperin (1990: 15) cited the OED to argue that before the 1892 translation of Krafft-Ebing ‘there was no homosexuality’.

The OED’s second Supplement also registered, at last, the same-sex sense of lesbian; it traced the adjective to 1890 and the noun to 1925. Emma Reference DonoghueDonoghue (1993b), conscious that ‘Entire theories rest on the OED dates’ (270), showed that both the adjective and noun could be pushed back to 1732, when they were used in The Toast, a poetic diatribe against the Irish aristocrat Lady Frances Brudenell (3). Yet twenty years later, Laura Reference DoanDoan (2013: 138) continued to rely on the OED’s dates as proof that lesbianism had not reified into a stable sexual identity in England before the 1920s.

Halperin and Doan cite Boswell and Donoghue among their references, and so it is an open question whether the former overlooked the latter’s antedatings or simply decided that the OED was the more authoritative source. Either way, were Halperin or Doan to consult the dictionary today, they would need to modify their arguments. Since 2018, the ongoing revision of the OED for its third, online edition has seen extensive updates to the dictionary’s coverage of gender and sexuality (Reference DentDent 2018). Following Boswell and Donoghue, OED3 now gives 1891 and 1732 as the earliest known dates of use for homosexual and lesbian.

Though all four of the above authors have written influentially on the history of sexuality, they have come to it from different academic backgrounds. Halperin is a classicist and queer theorist; Boswell was a medieval historian. Doan specializes in modernist literature, Donoghue in literature of the eighteenth and nineteenth centuries. It is to the great and lasting credit of the OED that it is one of the few reference works that English-speaking researchers across specialisms and disciplines hold in common. Any criticism of the dictionary must be tempered by the acknowledgement that ‘Few inquiries into particular words end with the great Dictionary’s account, but even fewer could start with any confidence if it were not there’ (Reference WilliamsWilliams 1983: 18). However, it is precisely because of the OED’s status as the foremost authority on the history of English that the gaps in its record must not be taken for granted.

Representing Past Sexualities in the Present

Nicholas Reference Lo VecchioLo Vecchio (2021) has shown that many of the updated entries for sexual terms in OED3 could yet be supported by more robust historical evidence. Here, I am especially interested in the dictionary’s citation of texts written by same-sex-attracted people. In light of the 1933 Supplement’s preference for pathological depictions of homosexuality over liberatory ones (see Chapter 5), it is notable that the ongoing revision of the dictionary has introduced a number of quotations from Sapphic and Uranian authors who were active in the late nineteenth and early twentieth centuries. Collecting data of this kind has been made easier in the last few decades by the reissuing of previously banned books and the publication of private pamphlets, letters, and diaries.Footnote ⁵ For instance, at present, OED Online cites five quotations from Symonds’s Problem in Modern Ethics, nine from Radclyffe Hall’s Well of Loneliness, twenty-five from Oscar Wilde’s letters, and sixty-nine from Virginia Woolf’s diaries. Nonetheless, these authors have not always been read as carefully as they could have been.

Though OED3 has shifted its earliest attestation of homosexual (adj.) from Krafft-Ebing’s Psychopathia Sexualis to Symonds’s Modern Ethics, the dictionary still cites Psychopathia Sexualis for the first known use of homosexuality. Here again it could be antedated by Symonds. The translator’s preface to Psychopathia Sexualis (Reference Krafft-Ebing1892b: ix) is dated November 1892; Symonds had used homosexuality in a letter to Havelock Ellis in October of that year. Unaccountably, the OED’s second Supplement had cited Symonds’s letter (erroneously placed after the quotation from Krafft-Ebing), but the citation has been taken out of OED3. From the second Supplement onwards, the OED has also included an entry for Symonds’s coinage of Uranian to describe same-sex desire. In OED3 – which defines the adjective as ‘Homosexual; pederastic; spec. (of poetry) expressing an admiration for male youth’ – the earliest citation is taken from the 1883 printing of A Problem in Greek Ethics, though Symonds had used the word in letters as far back as 1866. OED3 defines the noun Uranian as ‘A homosexual man; a pederast’, and attests it first from The Intermediate Sex (Reference Carpenter1908) by Edward Carpenter, rather than from the second edition of Love’s Coming-of-age (Reference Carpenter1906). Carpenter’s and other writers’ use of Uranian to include women-desiring women, discussed in Chapter 5, is elided in the dictionary. Finally, OED3’s earliest attestation of intersex in the erotic sense remains Xavier Mayne’s Intersexes (Reference Maynec1910), despite his use of it in Imre: A Memorandum (Reference Mayne1906). These antedatings are given in Table 6.1.

Amendments such as these, however slight, are significant not only for the historical basis of the dictionary but for the bearing they have on the array of ideas embedded in it. As Jack Reference HalberstamHalberstam (2011: 19) notes, ‘The history of alternative political formations is important because it contests social relations as given […] while not necessarily successful in the sense of becoming dominant, [they] do offer models of contestation, rupture, and discontinuity for the political present.’ Antedatings modify our understanding of how and when words for same-sex desire began circulating in Victorian and Edwardian England, and how institutional discourses about that desire were accepted, adapted, and challenged.

Interdatings can also be valuable for introducing alternative perspectives to culturally dominant points of view. Charlotte Reference Brewer, Tieken-Boon van Ostade and van der WurffBrewer (2009a, Reference Brewer2009b, Reference Brewer2011, Reference Brewer2015) has shown that throughout its history, the OED has consistently overrepresented the writing of male literary authors at the expense of other texts. The reading programme for OED3 has attempted to remedy this by deliberately ‘cover[ing] women’s writing and non-literary texts which have been published in recent times, such as wills, probate inventories, account books, diaries, and letters’ (‘Reading Programme’ n.d.), but quotations from female authors still make up a small fraction of the whole. This imbalance is particularly striking in OED3’s treatment of terms for sexuality between women. Although the early twentieth century witnessed the reclamation of lesbian and Sapphic by certain women-attracted women – as well as masculine-identified figures assigned female at birth – in Britain and North America, this act of linguistic self-making is hardly visible in the dictionary’s quotation banks for the terms.

Some wider historical context is offered by Table 6.2, which tallies all the author attributions for quotations under lesbian, Sapphic, and related headwords in OED3, sorted by gender. ‘Attribution’ in this case means the name (or names, for works by more than one author) appearing at the beginning of each quotation in small capitals. Multiple attributions to the same author are counted separately. In instances where a quotation has been taken from a text for which OED3 gives no author (which is the case for most newspapers, academic journals, and other periodicals), the attribution is counted as anonymous. In all, sixty-four attributions were found for sixty-two quotations.Footnote ⁷

Just 15.6 per cent of the attributions are to names gendered female. If the anonymous citations are discounted and only the thirty-seven named attributions are considered, female names still make up 27.0 per cent against male names’ 73.0 per cent. While the conventional gendering of a name does not always reveal the gender of the person referred to, it is fair to say that in most of these quotations women are being spoken about rather than speaking for themselves – whatever their sexuality.

If we consider the period from 1900 to 1933 (the year the first Supplement was published), we find that of the ten quotations in this subset, only one is attributed to a woman. A 1923 diary entry by Virginia Woolf is cited under Sapphist: ‘She is a pronounced Sapphist, & may..have an eye on me.’ OED3 has no citations at all for lesbianism or Sapphic (adj.) in this period, which happens to fall during a larger gap in the quotational record for each of these words: there are no quotations for lesbianism after 1897 and before 1971, and none for Sapphic (adj.) after 1895 and before 1950. These gaps could be filled, and the records of most of the other entries expanded, by a greater attention to the writings of Woolf and her contemporaries.

The published works of Katherine Mansfield and Mary MacLane considered in Chapter 5 might be supplemented with the now-public letters of several figures in or around the Bloomsbury Group. Reference Woolf and RosenbaumWoolf (1995: 54) recollected that sex became a topic of open conversation for Bloomsbury around 1909: before then, ‘We had known everything but we had never talked. Now we talked of nothing else.’ In the following years, talk of sex between women turns up in the correspondence of Woolf, her sister Vanessa Bell, Woolf’s sometime lover Vita Sackville-West, and the artist Dora Carrington. Beyond Bloomsbury, but existing in an attenuated network with it, were the gender-nonconforming American poet H. D. (Hilda Doolittle), who was acquainted with Sackville-West, and the boyish English novelist Bryher, who was intimate with H. D. Poetry by H. D. was published in the American magazine The Little Review, founded by Margaret Anderson and co-edited by her lover Jane Heap. One of Anderson’s less well-heeled contacts was Eve Adams (born Chawa Zloczewer), a Polish Jewish immigrant who travelled around the United States selling anarchist literature (Reference KatzKatz 2021: 34). While no edition of her letters exists, Adams privately printed a collection of short stories titled Lesbian Love in 1925; there is a single surviving copy whose whereabouts are known, and it was not reprinted until 2021. Uses of lesbian, Sapphic, and related words by these writers are contrasted with the current OED3 quotations in Tables 6.3 and 6.4.Footnote ⁸

Private writing allowed for greater freedom in using explicit terms for same-sex desire – although the never quite banished need for secrecy, or simply brevity, could still lead to words being coded or clipped. Sapphic might be reduced to ‘S’ (Reference Carrington and ChisholmCarrington 2017: 296) and Lesbian(s) to ‘L.’ or ‘Llllllllllls’ (H. Reference Bryher and FriedmanD. and Bryher 2002: 50, 325). However, unabridged forms are also well attested. The only words of which I have found no examples are the noun Sapphic (‘A lesbian’), which OED3 has not traced further back than 1985, and the adverb Sapphistically.

I call the quotations in the tables ‘alternative’ to highlight the divergent visions of lesbianism and Sapphism they represent beside the texts in OED3. However, I am not arguing that OED3’s existing quotations and the prejudices they embody – Aldous Huxley’s contempt for ‘English sodomites and middle-aged Lesbians’, Ladislaus Slominski’s assertion that ‘Lesbian practices’ are prompted by women’s fear of venereal disease, The Lancet’s claim that ‘Sapphism’ itself is a disease – should all be struck from the record. The preceding chapters have made clear that attitudes like these are important parts of the discursive histories of same-sex lexis. In any case, negative bias is also palpable in some of the alternative quotations, such as Bryher’s use of lesbianism while wryly passing on an acquaintance’s psychoanalytic concern for Bryher’s lack of children, or Carrington’s mockery of starchy morals and collars in an anecdote about another woman’s Sapphic affairs. Yet these quotations also capture usage that is playful (H. D.’s compliment to Bryher on looking ‘most dashing […] and Lesbic in that hat and that Cat’), amused (Sackville-West’s remark that the poet Else Lasker-Schüler nursed a ‘Sapphic passion for the Lady of Shalott’), earnest (Adams’s description of the masculine clothing that ‘means so much in the life of a Lesbian’), and delighted (Heap’s reaction to her ‘delicious’ experience in ‘the Lesbians home’, Natalie Clifford Barney’s house in Paris). These alternative quotations are clearly as ideological as those in the OED; their ideologies just run in different directions. But we ought to ask why certain directions – certain semantic possibilities – have fewer signposts in the dictionary than others.

The OED is a history of language in use, and as Sara Reference AhmedAhmed (2019: 20) reminds us, ‘A history of use is also a history of that which is not deemed useful enough to be preserved or retained.’ The question of who ‘counts’ in the dictionary, which authors are legitimated as authorities, is not trivial. At the start of this book, I argued that OED3’s treatment of the sexual sense of queer leaves unclear that the word seems to have started out not as a slur but as an in-group term in the early twentieth century. It is equally inaccurate to suggest, as the current evidence in OED3 does, that lesbian and Sapphic were not used as self-descriptors in the same period, or that the clinical and insulting connotations often assigned to them by men were never overridden. Of course, the alternative roll-call of authors in Tables 6.3 and 6.4 is biased as well. It has been determined by factors such as the uneven levels of wealth and education among English speakers in the early twentieth century, the means at individual writers’ disposal to establish lasting literary reputations for themselves, the judgements of later executors and scholars about whose papers are worth archiving and publishing, and my selective reading of those publications. That Virginia Woolf alone accounts for six of the twenty-two alternative quotations (27.3 per cent) is due to the comparatively high number of lesbian and Sapphic uses preserved across six volumes of her edited letters. With the exception of the working-class Jewish immigrant Eve Adams – who was deported from America for writing Lesbian Love, arrested in Nazi-occupied France, and killed in the Auschwitz concentration camp in Poland – the authors cited above reflect a tendency among scholars of Sapphic modernism to focus on the lives of ‘privileged white middle- or upper-middle-class lesbians precisely because they are the best documented’ (Reference Doan, Garrity, Doan and GarrityDoan and Garrity 2006: 8). The same could be said of the male Victorian and Edwardian authors I have quoted.

Aside from the matter of who had access to words like Sapphist and Uranian, and whose usage of them has been archived, there is the question of who had the most to gain from using these words in the first place. Kadji Reference AminAmin (2023: 95) argues that the scientific project of sexual taxonomy (and thus the reclamation of taxonomy as a tool for self-making) was ‘a white distinction – a testament to the individuality, complexity, and value of white bodies and psyches’; conversely, ‘Indigenous and racialized peoples were not considered deviant as sexological individuals […] but rather en masse, as constitutively and polymorphously perverse populations’. Beyond the white queer canon above stand figures like the writer Angelina Weld Grimké (1880–1958), mentioned in Chapter 5, whose love poems and letters to women have since her death been published only in part. Maata Mahupuku (1890–1952), a member of a prominent Māori family in New Zealand and the sometime lover of Katherine Mansfield, kept a journal in 1907 which she later gave to Mansfield, but it has not been made public (Reference AngusAngus 1996). Other self-accounts by black and brown women who desired women are preserved in retrospective documents. In interviews given towards the end of her life, the activist Mabel Hampton (1902–1989) remembered that she and other masculine lesbians in Harlem in the twenties and thirties had called themselves by names far removed from any institutional taxonomy: ‘Bulldykers and ladylovers, stud and butch, and the other ladies [i.e. their partners], “This is my friend, my wife”’ (Reference NestleNestle 1993: 934). Historical texts evidencing in-group terms like these may yet come to light through ongoing archival projects. (Hampton herself was a long-time contributor to New York’s Lesbian Herstory Archives.)Footnote ⁹ But there are limits to what can be recovered. José Esteban Reference MuñozMuñoz (1996: 9) observes that Western archives of the multiply minoritized – such as queer people of colour – are often ‘anecdotal and ephemeral’ rather than ‘traditionally evidentiary’ in the ways required by ‘official histories’. If we attempt to write a history of language anyway, then we must accept that while we can never offer a perfect reflection of the past, we can at least be transparent about the partiality of our sources and methods.

The Future of Queer Linguistic History

If even modern lexicography is not a colourless description of reality but a performative reconstruction of it, then it follows that, as Eve Kosofsky Reference SedgwickSedgwick (1990: 155) once argued in a different context, ‘the most productive questions we can ask’ about the definitions of words are ‘not “What is the true meaning, the accurate assignment of these labels?” but, rather, “What are the relations instituted by the giving of these labels?”’ To put it another way, what does the definition of a word reveal about the relations between a dictionary, its maker, its audience, and their society? Following Wenge Reference ChenChen (2019), this book has stressed the importance of reading a dictionary not as a text alone but as a discursive practice moulded by its production, transmission, and reception, and as a social practice taking place in a particular culture at a particular time. The outlines of this milieu may be more difficult to perceive in a dictionary that occupies the same temporal and cultural space as the researcher, for what is familiar is apt to pass unnoticed. Here, the value of historical research lies in its capacity to make the here and now seem strange again. As Reference MuñozMuñoz (2009: 30) points out, ‘a turn to the past’ can be a means of ‘critiquing the present’ and imagining a better future. We should not only recognize where our current prejudices come from, but recover the ways in which those prejudices have already been challenged by people who came before us.

The preceding chapters have shown that the sexual norms encoded over the history of English lexicography have been shaped not only by the words that are being (or not being) defined – buggerer, sodomite, pederast, tribade, confricatrix, invert, homosexual, lesbian, Sapphist, or Uranian – but by a host of paratextual, intertextual, and extratextual factors. These include the sources to which the lexicographer has access, the genre, size, and price of the resultant dictionary, and the gender, education, and profession of the target usership. They include, too, the readers who actually used the dictionary, whether they accepted, resisted, or rewrote what they found in it. All of these factors interact with wider ideologies about sexuality and language. Some ideologies involve weighty ontological or epistemological questions about the natures of these two social behaviours, while others centre on smaller, more pragmatic matters that exist on a spectrum between the erotic and the linguistic, shading from one into the other. They encompass what sexuality is and is not; which expressions of sexuality are permissible in general society, and which are appropriate subjects for the law, the church, or science; what qualifies a text as obscene; what a dictionary ought and ought not to be (polite or forthright, selective or exhaustive, intuitive or empirical, prescriptivist or descriptivist); which words do or do not belong to the language; what language is and is not; and which ways of thinking about these things should be acknowledged to be knowledge at all. The answers to these questions vary across texts and over centuries, but they continue to inform and be informed by dictionaries today. Whatever their ideological scaffolding, all dictionary-makers and users are part of the construction of social knowledge. In this respect, searching for a sharp epistemological rupture between the subjective lexicography of the past and the objective lexicography of the present is futile.

Indeed, the trouble with staring too deeply into any historical rift is that it may come to seem untraversable to the researcher as well. This is as true of the study of sexuality as it is of lexicography. Writing this book has required a kind of disciplinary itinerancy: I have moved between the camps of dictionary scholarship, queer theory, literary criticism, and linguistics, borrowing tools and concepts from each. But the field my training has led me to feel most at home in – queer linguistics – is also the one that has so far had the least to do with historical research. Although the inaugural anthology of queer linguistics, Anna Livia and Kira Hall’s Queerly Phrased (Reference Livia and Hall1997), contained several essays devoted to the history of sexuality, in the quarter-century since then queer linguists’ interest in the past has flagged. At present, our research agenda risks being swept along by a broader interdisciplinary current in queer studies that Valerie Reference TraubTraub (2016: 269) argues has favoured ‘the cultural production of the last century and the current moment: these temporal frames become the occasion for theory, while everything prior to the twentieth century is positioned as simply history’ – and thus not worth theorizing about. In fact, the sights of queer linguistics have been trained more often on the current moment than the last century.

But change is on the horizon. William Reference LeapLeap (2020) has called for a return to ‘queer historical linguistics’ in his own work on the decades immediately prior to the 1969 Stonewall riots (see also Reference MotschenbacherMotschenbacher 2020). Other researchers, among them Nicholas Reference Lo VecchioLo Vecchio (2020), Silvia Reference RomanoRomano (2022), and David Reference Peterson, Leap and PetersonPeterson (in press), are now pushing the temporal frame back before 1900 as well. This book is a contribution to that effort. While queer linguistics currently finds much of its purpose in unravelling the discursive fabric of heteronormativity, the regulation of desire clearly predates the fashioning of heterosexuality in the late nineteenth century. Given that any articulation of identity depends on the reworking of earlier norms – the elaboration of a pattern that refers back but never precisely repeats – then to understand better the sexual discourses of the present, we must trace them into the past.

Lexicography offers a way backward. If one of our reasons for doing the history of sexuality is to learn about figures who somehow seem to prefigure our own erotic desires, then doing queer linguistic history might lead us into the workrooms of writers who shared our fascination with language – even those with whom we appear to have little else in common. If we are really lucky, reading dictionaries will guide us to the sides of other, earlier users who scanned the same pages and found a term for themselves, or for something almost like themselves, or an ambiguity, gap, or silence into which they could speak their own meaning. Looking in the dictionary teaches us that the history of words before queer is far queerer than we might think.