In 2013, in the days following the passage of the Marriage (Same Sex Couples) Act in England and Wales, a reporter for Gay Star News contacted the publishers of the Oxford English Dictionary. The reporter wanted to know whether the dictionary’s definition of marriage would be changed to show that the term was no longer legally restricted to the union of a woman and a man. The publishers, Oxford University Press, issued an affirmative – if rather careful – response: ‘We continually monitor the words in our dictionaries, paying particular [attention] to those words whose usage is shifting, so yes, this will happen with marriage’ (Reference MorganMorgan 2013). The Press’s statement was picked up by several international news outlets, and its measured tone was quickly lost amid a volley of dramatic headlines:

Although the news outlets’ stances towards the proposed change varied across the political spectrum, the above headlines were united in one respect: they did not specify whether the Oxford English Dictionary (or OED) was set to update its definition of marriage, or to alter the definition of marriage in the English language. None of the articles went on to make the second claim explicitly, but the fact that the headlines could be interpreted that way should not be dismissed out of hand as an accident. After all, the OED is often treated as if it were the regulatory authority of English. In the case of marriage, the close fit between the dictionary and the language could be difficult for some readers to prise apart, even when they encountered the story under a less ambiguous title. The Daily Mail’s headline – ‘Oxford English Dictionary will change the entry for “marriage”’ (Reference WebbWebb 2013) – was evidently about a dictionary entry rather than some abstract, universal meaning of marriage, yet it still attracted numerous online comments that charged the OED with taking part in an Orwellian conspiracy: its makers were condemned for colluding with the ‘thought police’, propagating ‘newspeak’, and attempting to prohibit dissenting views as ‘thoughtcrime’. These comments cast the dictionary as an instrument of extreme linguistic determinism. Altering a definition became a sinister attempt to alter the behaviour of English speakers at large, in the belief that dictating the ways in which speakers were ‘allowed’ to use the language would also restrict the political ideas they were capable of expressing in it.

This was not the first time the OED had been caught up in a debate about same-sex relations and the law. In 1942, Adelaide Mary, Lady John Kennedy, was brought before the King’s Bench Division of the High Court of Justice for calling Innes Margaret Kerr a lesbian. Lady Kennedy stood accused of defamation under the Slander of Women Act 1891, which made it an actionable offence to speak words in public that ‘impute unchastity or adultery to any woman or girl’. Kennedy’s barrister argued that the law could not have been intended to prohibit imputations of lesbianism, for

According to the defence, if the sexual sense of lesbian was not in the first edition of the OED in 1908, then it likely did not yet exist in English at all – in which case lesbian could not fall into the category of words that the Slander of Women Act was meant to control.Footnote ¹

Like certain Daily Mail readers’ dire predictions about marriage seventy years later, the argument of Lady Kennedy’s barrister conjures up a sympathetic link between language and the dictionary. However, the direction in which the link travels differs according to the viewpoint being defended. For the barrister, who believed that his client’s case was strengthened by the (lack of) evidence in the OED, language determined the dictionary: the OED was a transparent reflection of English as it had been used in the past. For the Daily Mail readers, who believed that the ‘traditional’ meaning of marriage would be undermined if its definition were changed in the OED, the dictionary threatened to determine language: the OED had the capacity to govern how speakers would use English in the future.

These polar images of the dictionary – impartial recorder or partisan regulator – have both featured prominently in the history of lexicography, as will be seen. Obviously, in the extreme forms of the polarity presented here, neither position bears up under scrutiny. No dictionary of a living language is ever complete or unbiased. The fact that the OED’s only sense of lesbian in 1908 was ‘Of or pertaining to the island of Lesbos’ did not mean that erotic uses of the word were non-existent at the time, and Lady Kennedy lost her case. Nor can any dictionary exert totalitarian control over a language. The OED did update its entry for marriage after 2013, so that sense 1a now begins, ‘The legally or formally recognized union of two people as partners in a personal relationship (historically and in some jurisdictions specifically a union between a man and a woman).’ But this has not prevented people from using the word to mean other things.

Rather, the truth lies somewhere between the extremes. A language and a dictionary are cultural objects; each is shaped by its culture, and each plays a role in shaping the culture in turn. This reciprocal relationship is not unique to dictionaries with longstanding prestige such as the OED. To some extent, all dictionaries are normative texts that help to authorize what is (or isn’t) legitimate language, what should (or shouldn’t) be expressible in that language, and who has (or lacks) the authority to determine where these boundaries lie. What is more, dictionaries don’t construct legitimate and illegitimate forms of language alone, but of other social categories and behaviours that language represents. Researchers working in fields including lexicology, discourse analysis, corpus linguistics, literary criticism, jurisprudence, and feminist, trans, and postcolonial studies have shown how dictionaries’ representations of race, ethnicity, gender, health, disability, politics, and religion have both responded to and reinforced cultural ideologies and prejudices over centuries of English lexicography.Footnote ²

The same is true of dictionaries’ representation of sexuality, which is the subject of this book. Among the questions it asks are: how did lexicographers writing English dictionaries between 1600 and 1930 make sense of human sexuality? How did they neglect or refuse to make sense of it? How did their attitudes towards it change over time, or vary across types of dictionary? And, as far as can be known, how did the dictionaries’ users respond? The reasons for selecting this 330-year timespan, and the range of dictionaries I have chosen to study, will be laid out in this Introduction. But first, I want to make clear how I am using the words sexuality and dictionary, which seem so dangerously matter-of-fact in this book’s subtitle, and what I mean by the queer of the main title. The latter partly signals the book’s theoretical orientation. Before the Word Was Queer takes its cue from the research tradition of queer linguistics, which critically examines the regulation of sexual bodies, behaviours, desires, and identities as they are realized through language.Footnote ³ However, while the main drive of queer linguistic scholarship has so far been directed at language and sexuality in the present day, my aim is to follow them back to a time before queer was widely understood to have anything to do with sex.

Looking for Queer

This book’s epigraph is taken from a review of the OED – specifically, the section of the dictionary dealing with words from Tombal to Trahysh – which appeared in The Scotsman in 1913. The reviewer made the ironic remark that ‘queer, outlandish words [were] quite common’ in the dictionary. When the OED’s section for Q had been published in 1902, the primary sense it had given to queer (a.¹) was: ‘Strange, odd, peculiar, eccentric, in appearance or character. Also, of questionable character, suspicious, dubious.’ Neither this nor the subsequent senses alluded to sexuality. Still, while queer may not have had the same resonance for the OED editors or the Scotsman reviewer that it has for many people today, it seems likely that the word had already been set on its modern semantic trajectory.

At present, in the OED’s third, online edition (OED3), the earliest written evidence of queer as an adjective meaning ‘Of a person: homosexual’ dates from an article published in the Los Angeles Times in 1914. The article reports on the discovery of a secret club ‘composed of the “queer” people’: ‘members sometimes spent hundreds of dollars on silk gowns, hosiery, etc. … At these “drags” the “queer” people have a good time.’

A similar quotation, also from 1914, appears in OED3’s entry for queer as a noun meaning ‘A homosexual’: ‘Fourteen young men were invited..with the premise that they would have the opportunity of meeting some of the prominent “queers”.’ This quotation is not from the LA Times: the dictionary cites it second-hand from a 1995 article in the Journal of the History of Sexuality by Sharon R. Ullmann. Though not mentioned in the dictionary, Ullman’s source was a private report by a correspondent for the Sacramento Bee, which like the LA Times was investigating an underground network of femme crossdressers and men-desiring men in Southern California (Reference UllmanUllman 1995: 593).

In November 1914, members of the network had been arrested by vice officers in Long Beach, and the statements they made to the police provided fodder for a cannonade of news stories which brought the in-group term queer into public use (cautiously flanked by quotation marks). Perhaps the most curious use of the word came from the Sacramento Bee, one of whose articles, ‘Miscarriage of Justice in the Long Beach Case’ (1914), featured the subheading ‘Look For “Queer”’. In context, this referred to a search for one particular queer who had evaded capture. But the phrase takes on new significance for the researcher on the hunt for words.

In fact, OED3 has looked for and found one earlier instance of queer as a noun for ‘homosexual’ – from a letter written in 1894 near Dumfries, Scotland, by John Sholto Douglas, the Marquess of Queensberry. The Marquess is now perhaps best known for lending his name to the Queensberry rules, a code of conduct for boxing. The quotation from his letter is certainly belligerent:

OED3 does not quote Queensberry’s letter directly but relies on a published transcription of it by Richard Reference EllmannEllmann (1987: 402). The letter was composed soon after the death of Queensberry’s eldest son, Francis Douglas, in what was reportedly a shooting accident, though Queensberry believed it was suicide. Rumours swirled that Francis had been having an affair with the Prime Minister, the Earl of Rosebery, and had taken his own life to escape a scandal. Queensberry accuses several people of leading his son to his death, including ‘Roseberry’, and the OED’s quotation of his letter has since been used by scholars as a jumping-off point in several linguistic histories of queer.Footnote ⁴ The jump is made all the more remarkable by the distance it clears: queer makes its written debut in the correspondence of a Scottish nobleman, disappears from the record for twenty years, and then re-emerges half a world away on the club scene of the American West Coast.

Queensberry’s manuscript letter is now held by the McFarlin Library at the University of Tulsa. My own reading of its queer passage, reproduced in Figure I.1, differs slightly from Ellmann’s transcription:

‘Queer’ is a modifier of ‘Liar’, not a plural noun. Perhaps the Marquess still intended the word to say something about Rosebery’s sexuality: later in the letter, he refers to the Earl as the ‘pimp liar Roseberry’ (Figure I.2), and it is tempting to see an erotic as well as syntactic symmetry between ‘Queer’ and ‘pimp’. But in the absence of other evidence, I think it is more plausible that Queensberry was using the adjective in a sense well-known to the Victorians and accusing Rosebery of being ‘of questionable character, suspicious, dubious’, to quote OED1’s first sense of queer (a.¹) again. The implication of untrustworthiness or pretence would place ‘Queer’ in good company with ‘Snob’ and ‘Liar’, as well as ‘Canting Christian hypocrite’.

Figure I.1 The Marquess of Queensberry’s 1894 letter (recto).

By permission of McFarlin Library, University of Tulsa.

Figure I.2 The Marquess of Queensberry’s 1894 letter (verso).

By permission of McFarlin Library, University of Tulsa.

The point of this digression is not just to call for a reappraisal of the earliest recorded use of the sexual sense of queer. It is a twofold caution. First, scholars can make misreadings which dictionaries reproduce, and vice versa. Second, when we set out looking for queers, we shouldn’t assume in advance that we will find the present readily legible in the past. It is a cliché that the past is another country with its own language and culture. The cultural and linguistic divide may not seem that great, but as any speaker of two related languages – or even two varieties of the same language – knows, it is important to be wary of false friends. Fag is slang for a cigarette in England and a slur for a gay man in America. What is queer in the twentieth or twenty-first century might not have been in the nineteenth.

This is not to claim that then is irrelevant to now. The queer literary critic Eve Kosofsky Reference SedgwickSedgwick (1990: 44) observed that part of the importance of studying earlier periods is that their otherness from our own time can ‘radically defamiliariz[e] and denaturaliz[e], not only the past and the distant, but the present’. Coming at the relation between the past and the present from a different angle and discipline, the sociolinguist Jan Blommaert stressed the need for analysts of contemporary discourse to take notice of the historical developments that gave rise to it. ‘Power and inequality have long histories of becoming,’ he wrote, and ‘so have the linguistic repertoires of people;’ as such, ‘part of the critical punch of what we do may ultimately lie in our capacity to show that what looks new is not new at all’ (Reference BlommaertBlommaert 2005: 37). These statements represent broadly divergent approaches to history, alterist versus continuist, but they can be compatible. Our differences from the past throw into relief how present beliefs about sexuality and language are neither eternal nor inevitable, while continuities between past and present enable us to trace the origins and development of still-living ideologies.

The dictionaries of bygone eras offer one rich trail of evidence for anyone studying the relationship between language and sexuality. The main value of this kind of research lies not in what dictionaries can tell us directly about how erotic practices and desires were enacted and felt in earlier times, but rather in what dictionaries reveal about past attitudes and anxieties around sexuality. This doesn’t mean that lexicography is utterly divorced from what people ‘really’ did. Dictionaries are written by people whose knowledge of the world is shaped by their own experiences. They are also instructive texts: one way or another, most of them seek to inform the understanding or behaviour of their users. Their guidance is more than just linguistic. Lexicons of medical and legal terms, for example, have long been intended to assist physicians and magistrates in the managing of public health and the policing of public behaviour. Beyond that, Linda C. Reference Mitchell and ConsidineMitchell (2010: 82) has argued that even dictionaries written for ordinary people in earlier centuries functioned as conduct books as well as language guides, imparting ‘prescribed cultural values for the rising middle classes’. In addressing the literate public, these lexicons enjoined readers to police themselves. This extended to their sexual comportment.

The history of sex in English lexicography is too vast to fit into one book, and this one comes with a number of restrictions. The first constraint is my use of dictionary. I follow Noel E. Reference Osselton and CowieOsselton’s (2009: 132) definition of the monolingual English dictionary as ‘a separate book solely for English words with English explanations’. This does offer considerable leeway between dictionary and encyclopaedia: do ‘English words’ include English (or anglicized) names? Do ‘explanations’ comprise technical, historical, and literary descriptions as well as semantic definitions? The distinction between lexical and encyclopaedic content is somewhat hazy in contemporary lexicography and was hazier still in earlier centuries, and I have not attempted to draw a clear line between them. I differentiate more sharply between dictionary and glossary, using the latter for a list of words and explanations that forms a supplementary part of a larger text. Discussion of glossaries in this book will be limited, but when I want to refer to glossaries and dictionaries interchangeably, I will rely on the term lexicons. These are my own uses; other scholars define these words differently.

The second constraint is my decision to focus on English dictionaries that were compiled and published in Britain. Lexicography and sexuality should be placed within their specific cultural contexts, and making this study diatopic as well as diachronic would have resulted in an unfeasibly long or unhelpfully generalist work. However, I will sometimes refer to bilingual, foreign-language, and American English dictionaries when their treatment of sexuality had a direct and significant influence on monolingual English lexicography in Britain.

The third constraint is what is included under sexuality. I use the word to encompass all erotic acts and desires, as well as any identities they constitute. Having said that, this book is mostly about acts, desires, and identities founded on relations between women and between men. When I discuss socially approved relations between a woman and a man, it is in the context of how same-sex relations were variably positioned as antithetical to, lesser than, or imitative of them. When I address other forms of proscribed sexuality (such as polygamy, adultery, and masturbation), it is to show how they were contrasted or conflated with same-sex intimacy. Likewise, when the book discusses the treatment of people marked by additional social differences in Britain – unfeminine women, unmasculine men, people whose gender performance evaded any easy binarization, people with ambiguously sexed or racially othered bodies, foreigners, heretics – it is in cases where these differences coincided with representations of sexuality between women or between men.

Although this book is foremost a study of language, I hope it will be of use to any scholars who have a stake in the history of sexuality, whether their disciplinary home is in linguistics, literature, anthropology, history, or elsewhere. As the literary critic Jeffrey Reference MastenMasten (2016: 16) points out, ‘There can be no nuanced cultural history of early modern sex and gender without spelling out its terms’, and the same could be said of all the periods covered here. Because I don’t want to take for granted the disciplinary knowledge that each reader brings to the book, this Introduction will give an overview of previous approaches to analysing the conjunction of lexicography and sexuality. I will then sketch the standard academic historiographies of Western sexuality and English dictionaries at large, together with some of the challenges that have been posed to them, and offer my own perspective on how to theorize the relationship between sex and lexis.

Researching Lexicography, Researching Sexuality

This study is not the first to venture into the lexicography of sex. In fact, my hope that the book might reach across disciplinary divides rests on the fact that a sizeable cross-disciplinary literature on this subject already exists. Almost a century ago, the linguists Alan S. C. Reference RossRoss (1934) and Allen Walker Reference ReadRead (1934) took an interest in English dictionaries’ inclusion (and exclusion) of sexual expletives. Since then, dictionary scholars and working lexicographers have considered the reticence or boldness with which dictionaries, old and modern, have treated not only profanity (Reference BurchfieldBurchfield 1972a) but contraception (Reference MugglestoneMugglestone 2007b), prostitution (Reference Stein, McConchie and TyrkköStein 2018), genitals (Reference Frank, Amodio and O’Brien O’KeeffeFrank 2003; Reference Mugglestone and GorjiMugglestone 2007a), same-sex acts (Reference ClaytonClayton 1991; Reference BrewerBrewer 2007, Reference Brewer2013), and even so prosaic a term as sexual intercourse (Reference LandauLandau 1974).

Feminist criticism of the sexism embedded in dictionaries has raised some of the same issues in a different light, pointing out how definitions of sexual verbs (Reference Strainchamps, Gornick and MoranStrainchamps 1971) or names for genitals (Reference WillinskyWillinsky 1987; Reference Braun and KitzingerBraun and Kitzinger 2001; Reference BrewerBrewer 2016) can play into or resist stereotypes of male agency and female passivity. Feminists have also called attention to dictionaries’ inadequate labelling of pejorative terms for lesbians and gay men (Reference RoseRose 1979; Reference Cannon and RobersonCannon and Roberson 1985), a matter that has since been taken up in studies of lexicographers’ handling of slurs more generally (Reference McCluskeyMcCluskey 1989; Reference NorriNorri 2000).

These academic critiques have found a practical counterpart in feminist dictionaries, which since the 1970s have offered a political antidote to the misogyny of standard dictionaries and the languages they represent (for overviews, see Reference PenelopePenelope 1990: 217–23; Reference CameronCameron 1992: 114–17; Reference RussellRussell 2012, Reference Russell2018: 183–210). Within this genre, gender interacts with sexuality most vividly in Monique Wittig and Sande Zeig’s Lesbian Peoples: Materials for a Dictionary, published in French in 1976 and in English in 1980. The dictionary imagines a utopian world in which all men have vanished and traditional nation states have been replaced by a global community of women-loving women, around whose culture the language has been reorganized. That Lesbian Peoples is in part a metatextual exploration of the possibilities and imperfections of lexicography itself is made clear by its entry for dictionary, which asserts that the book is a deliberately ‘lacunary’ representation of language and history, a selective ‘assemblage of words […] acting upon reality’ (Reference Wittig and ZeigWittig and Zeig 1980). By emphasizing the socially constructive nature of any dictionary, and the fragmented vision of reality it presents, the authors tacitly underscore the androcentrism and heterosexism of mainstream lexicography and historiography (see Reference WenzelWenzel 1981; Reference AndersonAnderson 1991, Reference Anderson1994).

A second alternative to the sexual conventions entrenched in standard dictionaries is offered by slang lexicography. Works on English ‘gay slang’ – or, in the early years, ‘homosexual argot’ – have appeared since the 1910s. At first these were brief lists of words that were claimed to be in use by same-sex subcultures in particular cities or countries, compiled mostly by psychologists and criminologists with varying levels of sympathy for their subjects. By the mid-twentieth century, however, the genre had given rise to longer glossaries and dictionaries avowedly by and for gay people – though most paid far more attention to men than women. Histories of the lexicography of gay slang (and of slang used about gay people by others) have been compiled by Gary Reference SimesSimes (2005), Reference ColemanJulie Coleman (2010, IV), and Jonathon Reference GreenGreen (2014, Reference Green2019). An earlier, more critical overview was given by Don Reference KulickKulick (2000), who took issue with how the genre had come to sustain the belief that there was such a thing as a singular gay slang associated with a uniform gay identity in Britain, in the United States, or internationally. The criticism is significant, for it came at a broader turn in the study of ‘gay and lesbian language’ towards analysis informed by theories of social constructionism, or what Reference KulickKulick (2000: 272) described as ‘shift[ing] the focus from gay and lesbian identities to the way in which language is employed to produce those subject positions’. Much of this shift occurred under the aegis of queer linguistics, which has a special interest in how language is used to normalize certain identities and behaviours at the expense of others.

A few queer linguists since Kulick have scrutinized how gay slang lexicons have promoted a view of same-sex identity in which whiteness and maleness are the unspoken rule (Reference McCormickMcCormick 2009; Reference LeapLeap 2020: 305–30). Others have highlighted how dictionaries produced by mainstream publishers have perpetuated heteronormativity, ignoring or erasing the perspectives of non-heterosexual speakers in their treatment of terms for sexual identities and practices (Reference GeorgeGeorge 1999; Reference Nossem, Baer and KaindlNossem 2018; Reference TurtonTurton 2020, Reference Turton2021b; Reference Pakuła and PakułaPakuła 2021a; Reference Lo VecchioLo Vecchio 2021). Still, queer linguistics’ engagement with dictionaries remains limited. So far it has concentrated on the twentieth and twenty-first centuries and overlooked the historical context from which contemporary norms of lexicography have arisen. Queer linguists have rightly encouraged dictionary-makers to approach their work from a diachronic angle in order to ‘challenge the naturalness of the [present] dominant discourse’ (Reference Nossem, Baer and KaindlNossem 2018: 184), and to ‘trac[e] the development of the struggle to create and control meaning’ (Reference GeorgeGeorge 1999: 55), but the same encouragement could be given to queer linguists themselves. To be sure, there are queer scholars in other fields, chiefly literary, whose interest in historical languages of sexuality has led them to read dictionaries published in the sixteenth and seventeenth centuries (Reference TraubTraub 2002, Reference Traub2016; Reference MastenMasten 2016; Reference GambleGamble 2023) or the eighteenth and nineteenth (Reference DonoghueDonoghue 1993b; Reference LanserLanser 2014). These studies are brilliant, incisive, and erudite, but the emphasis of their analyses understandably falls on more conventionally literary genres; in them, with few exceptions (Reference Bonnet, Livia and HallBonnet 1997; Reference BlankBlank 2011), the dictionary has tended to appear as a source of auxiliary evidence rather than as a text type of primary concern.

Queer linguists should not ignore early dictionaries as analytical objects in their own right, not least because we share with lexicography a disciplinary home. Linguistics already has a rich tradition of social research into language ideology and metadiscursive practice, which would lend itself well to studying the principles and methods of past lexicographers. Queer linguistics may have neglected this potential so far, but it has not been alone. Michael Reference Adams and AdamsAdams (2010: 20) has argued that socially minded linguists in general need to pay greater heed to the history of dictionaries and its relation to the history of language attitudes. The lack of communication runs both ways: scholars of lexicography have so far shown little interest in queer linguistics (Reference Pakuła and PakułaPakuła 2021b: 7). The upshot of this estrangement is that dictionary scholars and queer linguists have largely inspected the lexicography of sex at different focal lengths. Dictionary scholars tend towards microscopy, looking on individual lexicographers’ handling of sexuality as an idiosyncratic affair whose connections to the wider culture are left undertheorized. Conversely, in focussing on the bigger picture, queer linguists have sometimes lost sight of the finer details: a dictionary appears as though it were simply a machine-made product of its society’s prejudices, rather than a human artefact that has been shaped by the specific aims and beliefs of people working under particular constraints of time, space, and knowledge.

This book tries to find the middle ground between these points of view. In an earlier paper (Reference TurtonTurton 2020), I proposed a framework for analysing sexuality in dictionaries which built on the strengths of both lexicographic and queer criticism. I will elaborate the framework below, giving more attention to how the analyst might situate the local practices of dictionary-writers and users within more abstract social structures. However, as a precursor to that, I will briefly go over the standard histories of English lexicography and Western sexuality, with an eye especially to times when their courses appear to have run in parallel.

Doing the Histories of Sexuality and Dictionaries

The histories of Western sexuality and the English dictionary have each conventionally been divided into a ‘before’ period and an ‘after’ period, separated by a radical break in the late nineteenth century. In the case of sexuality, the break was the invention of homosexuality. The effect that the coining of this word had on the dominant cultural understanding of same-sex desire has been most famously articulated by Michel Foucault. Whereas previously same-sex intercourse, under the guise of sodomy, had been ‘a category of forbidden acts; their perpetrator […] nothing more than the juridical subject of them’, Reference FoucaultFoucault (1978: 43) claimed that with the arrival of the homosexual, the juridical subject was replaced with a more stable identity category – one which ‘became a personage, a past, a case history, and a childhood’. This conceptual shift was authorized by the emergence of a new psychological science, sexology. Foucault’s argument has been fundamental to constructionist accounts of Western history, which emphasize that contemporary discourses of sexuality are not transhistorical but timebound, and largely determined by ideologies emanating from powerful cultural institutions.

In early modern Britain, same-sex intercourse was a criminal and spiritual offence. While in Scotland it remained punishable at common law, in England the Buggery Act 1533 made ‘the detestable & abominable vice of buggery cōmitted with mankind or beast’ a felony by statute (Anno. XXV. Henrici VIII c1535: viii^v). Though it would prove a matter of debate whether ‘mankind’ was meant to include womankind, sex between women and between men alike was unambiguously condemned by religious doctrine. The King James Bible of 1611 rebuked the ‘vile affections’ of ‘women [who] did change the naturall vse into that which is against nature: And likewise also the men [who], leauing the naturall vse of the woman, burned in their lust one towards another’ (Romans 1: 26–27). Yet in the nineteenth century, the authority of the law and the church to dictate sexual morality was extended to the medical establishment, according to Jeffrey Reference WeeksWeeks (2003: 3): ‘doctors had adopted some of the attributes of a new priesthood, and many of its members seemed as certain of their views as the old’. The transformation of sexuality into an object of science is meant to have triggered a widespread ideological shift from what Reference SedgwickSedgwick (1990: 46–47) called a ‘universalizing’ model to a ‘minoritizing’ one. Whereas for centuries, same-sex activity had been understood as a crime or sin which any person might be liable to commit – just as anyone was capable of committing theft or arson – under the clinical gaze of sexology, same-sex desire came to be seen as an innate predisposition affecting a minority of people.

By contrast, the ‘before’ period of English dictionaries is represented less as a time of stasis than as one of gradual advancements – smaller paradigm shifts – before the great break. The period typically begins seven decades after the Buggery Act and seven years before the King James Bible, with the publication of A Table Alphabeticall by Robert Cawdrey in 1604. Cawdrey, a schoolmaster, was proceeded by a long line of bilingual (particularly Latin–) English dictionaries, but he inaugurated the tradition of the monolingual dictionary of ‘hard words’. A Table Alphabeticall was written, as its title page declares, to teach ‘the true vvriting, and vnderstanding of hard vsuall English wordes, borrowed from the Hebrew, Greeke, Latine, or French. &.’, and it was intended ‘for the benefit & helpe of Ladies, Gentlewomen, or any other vnskilfull persons’. Ostensibly, in purchasing the book, women of the upper ranks who were literate but had been denied the advantage of a classical education could still profit from a schoolmaster’s instruction. The extent to which this image reflected the genuine usership of early dictionaries, or the actual wants of highborn women, is questionable (Reference RussellRussell 2018: 32–37), but the target audience of later hard-word lexicographers soon grew to include other markets. Henry Cockeram’s English Dictionarie (1623: A2^r) was aimed at, ‘as well Ladies and Gentlewomen, young Schollers, Clarkes, Merchants, as also Strangers of any Nation’, while Thomas Reference BlountBlount (1656: A1^r) wrote the Glossographia simply for ‘all such as desire to understand what they read’.Footnote ⁶

At the same time, hard-word dictionaries began to embrace a wider range of vocabulary: obsolete words and terms of art (e.g. Reference BullokarBullokar’s 1616 English Expositor), names from classical mythology (Reference CockeramCockeram 1623), and scientific nomenclature (Reference BlountBlount 1656). But expansion would be tempered with discrimination. In The New World of English Words, Edward Reference PhillipsPhillips (1658: C2^r) informed his readers that, while his dictionary encompassed words lacking in ‘purity, or reputation’, for ‘the undistinguishing sort of Readers would take it very ill if they were not explained’, he would set an obelisk beside them to caution readers against their use. This practice, Reference Osselton and CowieOsselton (2009: 144) argues, ‘may be seen to mark the beginning of a clearly prescriptive tradition in the English monolingual dictionary’.

Alongside these works, there grew up a variety of specialist genres. Dictionaries of English, Latin, and French law terms (such as Rastell’s c1525 Exposiciones terminorum legum anglorum) and English, Latin, and Greek medical terms (Reference BoordeBoorde’s 1547 Breuiary of Helthe) predated the advent of hard-word lexicography, but they were now joined by dictionaries of theology (Reference WilsonWilson’s 1612 Christian Dictionarie) and etymology (Reference HogarthHogarth’s 1689 Gazophylacium Anglicanum). Other specialized genres which were aimed at a lay readership took the form of glossaries. Mary Evelyn’s ‘Fop-Dictionary’, appended to Mundus Muliebris: or, the Ladies Dressing-Room Unlock’d (Reference Evelyn1690), defines terms for cosmetics and fashion. Since the mid-sixteenth century, narrative exposés of England’s underworld had featured lists of secret words allegedly used by thieves and beggars. This criminal cant was first gathered into a standalone publication around 1698, when A New Dictionary of the Terms Ancient and Modern of the Canting Crew was put forth by ‘B. E. Gent.’.

While specialized lexicons proliferated towards the close of the seventeenth century, hard-word dictionaries began to expand beyond the ‘hard-word’ label. Reference K[ersey]A New English Dictionary (1702) by ‘J. K.’ (John Kersey) is often pointed to as the first general English dictionary – that is, the first to admit common as well as difficult words. Kersey was soon joined by Nathan Bailey, who gained enduring popularity in the eighteenth century with his octavo Reference BaileyUniversal Etymological English Dictionary (1721) and folio Reference BaileyDictionarium Britannicum (1730). Still, Bailey was eventually superseded by Samuel Johnson, whose intention to turn lexicography into a tool for standardizing English was heralded by A Plan of a Dictionary of the English Language in 1747. Johnson’s plan was, in some respects, a riposte to the two most celebrated dictionaries in Europe at the time – the Italian Vocabolario degli Accademici della Crusca and the French Dictionnaire de l’Académie françoise – both produced with the purpose of codifying a national standard language (see Reference ConsidineConsidine 2014). Likewise, Reference JohnsonJohnson (1747: 33) proposed to compile a work ‘by which the pronunciation of our language may be fixed, and its attainment facilitated; by which its purity may be preserved, its use ascertained, and its duration lengthened’. The dictionary would be systematic, educational, and edifying. Meanings of polysemous headwords would not be run together, as in earlier dictionaries, but disambiguated into discrete senses. Following the Vocabolario’s practice, Reference JohnsonJohnson (1747: 31) would affix ‘authorities’ to his definitions to illustrate headwords’ use in context; these would be quotations preferably taken from ‘writers of the first reputation’ and conveying ‘some elegance of language, or some precept of prudence, or piety’.

The realization of A Dictionary of the English Language eight years later departed from Johnson’s early ideals. As his preface admits, words had had to be ‘sought where they [were] used’, so that many of his quotations, rather than being elegant or pious, ‘serve[d] no other purpose, than that of proving the bare existence of words’ (Reference JohnsonJohnson 1755 I: B2^v). Johnson had also grown sceptical of the regulative potential of lexicography. As dictionary-makers could ‘produce no example of a nation that has preserved their words and phrases from mutability’, so they should not ‘imagine that [a] dictionary can embalm [a] language, and secure it from corruption and decay’ (C2^r). Instead, Johnson asserted that he did ‘not form, but register[ed] the language’; did ‘not teach men how they should think, but relate[d] how they have hitherto expressed their thoughts’ (C1^v). Despite these concessions, Johnson’s selected quotations were still largely confined to an elite male literary canon, while his definitions sometimes appeared to proscribe words by labelling them as ‘barbarous’, ‘low’, ‘bad’, or otherwise improper (cf. Reference Wild and AdamsWild 2010). Though Johnson did not succeed in standardizing the English language, his dictionary set a standard for English lexicography that lasted well into the nineteenth century. The first significant challenge to its authority came not from Britain but from the United States. Compiled by Reference WebsterNoah Webster, An American Dictionary of the English Language (1828) already declared its independence from Johnson and Britain in its title. As Johnson had sought to bring a degree of prestige to British lexicography that would rival that of the continent, so Webster wished to do the same for his own country. Ironically, when Webster’s dictionary was reissued in London in 1831–1832, with its title denationalized to A Dictionary of the English Language, the British print run of 3,000 copies exceeded the original American run of 2,500 (Reference MicklethwaitMicklethwait 2000: 272).

Despite its success on both sides of the Atlantic, Webster’s dictionary arrived at a time of growing debate in European and American scholarly circles over the nature of language and its study (Reference ReadRead 1966; Reference Adams, Ogilvie and SafranAdams 2020). Webster belonged to an older tradition which thought that the origins of words could be uncovered by a cursory acquaintance with other languages combined with reasoned introspection. In the early nineteenth century, this intuitive approach had gained fresh momentum in the wake of the philologist John Horne Tooke, who maintained – as the British lexicographer Charles Reference RichardsonRichardson (1836 I: 41) summarized – that ‘a word has one meaning, and one only; that from it all usages must spring and be derived; and that in the Etymology of each word must be found this single intrinsic meaning’. However, from the late eighteenth century onwards, a priori approaches to word history would be increasingly contested by the principles of a new strand of philology. Followers of the new philology maintained that the tracing of a language’s development should be empirically based on the study of its older forms (and their comparison to coeval forms of related languages) as preserved in early written sources. In Britain, many of these texts were at last becoming widely available through scholarly republications; yet it was in continental Europe that the new philology first gained general acceptance, and it was on the continent that it first informed lexicography (Reference MommaMomma 2013). In 1838, Jacob and Wilhelm Grimm began compiling their Deutsches Wörterbuch, a dictionary of High German that would chronicle the development of the language across centuries. Work commenced on historically informed dictionaries of French and Dutch in the next two decades (Reference Considine and DurkinConsidine 2016).

As Johnson had responded to the academy dictionaries of France and Italy a hundred years before, so a need was now felt for an English dictionary that would match the advances of German scholarship. What was wanted, as Richard Chenevix Reference TrenchTrench (1857) pronounced before the London Philological Society, was a ‘Lexicon totius Anglicitatis’: a complete record of the history of English (52). This lexicon would be free of the partialities and prejudices of earlier dictionaries, for the proper duty of a lexicographer to the language was to be ‘an historian of it, not a critic’ (4). Two years later, the Philological Society unveiled its plan to undertake such a project in its Proposal for the Publication of a New English Dictionary. It would produce ‘a new and more Scientific Dictionary than any at present existing’ (Philological Society 1859: 1), charting the development of English from the mid-thirteenth century to the present day with rigorous objectivity. ‘We entirely repudiate the theory,’ the Proposal asserted, ‘which converts the lexicographer into an arbiter of style, and leaves it to his discretion to accept or reject words according to his private notions of their comparative elegance or inelegance’ (2–3). Entries would be based on a groundwork of quotational evidence. Unlike Johnson, the compilers would not limit themselves to canonical literature but ‘admit as authorities all English books’, except treatises on ‘purely scientific subjects’ and post-Reformation works on ‘provincial dialects’ – regional English and technical jargon being beyond the scope of a dictionary of the ‘standard language’ (3).

The resulting New English Dictionary on Historical Principles – soon to become known as the Oxford English Dictionary – was published in instalments from 1884 to 1928. It was the work of countless hands: editors, sub-editors, assistants, volunteer readers, and specialist consultants. In the eyes of its chief editor, James A. H. Reference MurrayMurray (1900: 51), it signalled a major departure from scholarship of the past, ‘rendered possible’ by ‘the scientific and historical spirit of the nineteenth century’.Footnote ⁷ Lexicography would no longer be based on private introspection or received wisdom but on the facts of usage. After the dictionary’s completion, one of its surviving editors reaffirmed that they had ‘got beyond [the] stage’ of earlier lexicographers who thought it was ‘part of their duty to improve the English language’, but had instead made it their ‘business to […] record it as it was and as it is’ (Reference CraigieCraigie 1934: 26). This division between ‘improving’ and recording, like Trench’s distinction between critic and historian, encapsulates an opposition now generally known in linguistics as prescriptivism versus descriptivism. Though some early reviewers cavilled at what they saw as the OED’s lax attitude to linguistic correctness, the dictionary’s achievements were widely praised, and its scholarly ideals and execution have become a guiding light for subsequent lexicographers (see Reference Bailey and MugglestoneBailey 2000b). Indeed, Phil Reference BensonBenson (2001: 10) suggests that the break between prescriptivism and descriptivism marked by the OED – including the reconception of language as a system that may ‘be subjected to scientific inquiry’ – has been ‘crucial to the self-image of the modern dictionary’.

In this respect, the nineteenth-century emergence of scientific lexicography parallels the arrival of sexology. From the perspectives of their adherents, the theoretical turning point of each discipline – linguistic prescriptivism to descriptivism, sexual universalization to minoritization – might be understood as an epistemological rupture: a radical break separating the early, superstitious interpretation of a natural phenomenon from a later, scientific analysis (Reference BachelardBachelard 1968: 78). The two fields, sexological and lexicographical, would intersect in the publication of a single-volume Supplement to the OED in 1933, which added entries for homosexual and heterosexual, and which serves as the endpoint of this book.

However, as Reference SedgwickSedgwick (1990: 44) once cautioned, the ‘search for a Great Paradigm Shift’ – whether by field pioneers or their biographers – risks misrepresenting the complexities of both the past and the present. Her warning was particularly directed at the conceptual movement of the same-sex-desiring person from a ‘temporary aberration’ to a ‘species’, which she proposed was not as sharp as Reference FoucaultFoucault (1978: 43) implied. The universalizing view of the sodomite (and its religious and legal significance) did not simply cease to exist at the coining of homosexual; nor did the minoritizing view appear only at that moment. Rather, a plurality of universalizing and minoritizing models have co-existed and conflicted for centuries, though the names given to them, and their cultural salience relative to one another, have changed over time. Reference SedgwickSedgwick’s (1990: 47) insistence that scholars should not present history as a simple succession of one model ‘drop[ping] out of the frame of analysis’ to make way for the next has become an article of faith among historians of sexuality. Nonetheless, Valerie Reference TraubTraub (2016: 94–95) argues that subsequent research has still been too constrained by disciplinary periodization, leading to studies of sexuality in the Middle Ages, in the Renaissance, in the Victorian era, and so on, which emphasize the discreteness of these periods while overlooking their continuities, resemblances, and repetitions. In deciding on the timespan to be covered in this book, I have followed Reference TraubTraub’s (2016: 95) call to ‘speak across period divides’.

The history of dictionaries is no more impartial than that of sexuality. The conventional timeline I have recounted above – from Cawdrey to Bailey to Johnson to Webster to the OED – exemplifies what Reference Magnani and WattRoberta Magnani and Diane Watt (2018) call a ‘patrilinear’ view of textual history: one which presents a succession of famous male authors rivalling and then supplanting their forebears. Apart from side-lining other dictionary-writers, patrons, and users, including women (Reference RussellRussell 2018), the timeline reduces the rather erratic history of lexicography to a single, straightforward thrust. With the arrival of the OED, the prescriptivism of an earlier age drops out of the frame, ousted by the descriptivism of the modern profession. This narrative is still regularly encountered among dictionary-makers and critics. To take one recent example, The Cambridge Companion to English Dictionaries (2020) includes sections entitled ‘Eighteenth-century English Dictionaries: Prescriptivism and Completeness’ and ‘Nineteenth-century English Dictionaries: Descriptivism’. Still, this view of history has had to weather many disputes over how the goals of descriptivism may best be achieved, and to what extent they are achievable at all (see Reference BrewerBrewer 2010; Reference Mugglestone and DurkinMugglestone 2016; Reference Finegan and OgilvieFinegan 2020). That a truly neutral dictionary is impossible was asserted decades ago by the lexicographer Rosamund Reference MoonMoon (1989: 75): ‘Definitions are necessarily subjective, presenting socialized views of word-meaning, howsoever reinforced: objectivity […] is ultimately a myth.’ Reference BensonBenson (2001: 10) agrees that the ‘important distinction’ between prescriptivism and descriptivism ‘lies not so much in whether the modern dictionary prescribes or not as in the basis on which it claims authority’.

The untenability of the prescriptive–descriptive binary, not only in lexicography but in linguistics at large, has been argued at length by scholars of language ideology, among them Talbot J. Taylor and Deborah Cameron. These authors propose that descriptivism is as much a normative activity as prescriptivism; they simply differ in how overt they are about it. Reference Taylor, Joseph and TaylorTaylor (1990: 9) observes that because words depend for their meaning on how individual people use them in individual contexts, and because people do not use language uniformly, words cannot have objective, universal senses that are open to empirical verification. As such, definitions of words do not have a classical truth value – that is, they ‘are neither true nor false in the sense of corresponding or not to some state of affairs’ (24). Regardless of their writers’ intentions, dictionary entries are ‘not descriptions of facts, but rather citations of norms’ (24). A somewhat similar idea has since been put forward by the lexicographer Patrick Reference HanksHanks (2013: 73), who believes that word meanings are ‘best viewed as dynamic events that involve cooperation between speakers and hearers or writers and readers, rather than static, finitely bounded abstract objects’. Hanks and Taylor both draw attention to the fact that while people can technically use language however they want, if they wish to be understood by others, then the meanings they apply to words cannot be totally arbitrary but must conform to established norms of use. In light of this, Reference HanksHanks (2013: 91) suggests that it is the responsibility of a dictionary to identify these norms in its definitions, ‘encapsulat[ing] the shared beliefs about word meaning and word use on which writers and readers, speakers and hearers mutually rely’. However, Hanks reverts to the old binary by saying that the writing of such a definition should be a descriptive task, one that presents the ‘normal’ use of a word rather than prescribing some ‘normative’ standard (431).

It is not obvious that normality and normativity are separable in this way. Because dictionaries typically present a definition as though it were an abstract and invariant equation between word and meaning, the definition cannot but appear as a rule – one whose force depends on the authority of whoever issued the dictionary. Reference Taylor, Joseph and TaylorTaylor (1990: 22) wryly comments that the OED, the greatest exemplar of descriptivism in English lexicography, has itself become ‘by far the most authoritative […] normative influence on the behaviour of individual speakers and writers of English’. It is no solution to insist that people who treat the OED, or any ‘descriptive’ dictionary, as a regulator are misusing it, for instructing users not to read definitions prescriptively is a prescription in itself. Reference CameronCameron (1995: 8) notes the perils of positioning any linguistic work as scientifically objective: ‘Because science itself has authority in modern society, while at the same time the discourse of value remains a highly salient one for everyday talk about language, the absolute distinction between observing norms and enforcing them cannot be maintained in practice.’ Ultimately, ‘“description” and “prescription” turn out to be aspects of a single (and normative) activity: a struggle to control language by defining its nature’ (8). As such, rather than using descriptive or prescriptive to classify a dictionary, I think it is more helpful to use descriptivist or prescriptivist to denote the ideology of a dictionary-maker. Acknowledging the normativity of all lexicography is vital in a study that includes dictionaries compiled under a range of principles, ‘scientific’ and ‘prescientific’. Without downplaying the differences in lexicographers’ ideals, we should recognize the overlaps in their practice. A focus on linguistic norms also provides an opportunity to unite language ideology with queer notions of performativity.

Linguistic Normativity and Queer Performativity

In the thirty years since it was coined by Michael Reference WarnerWarner (1991), heteronormativity has remained a significant – if not the central – object of critique for many queer scholars. Yet while heteronormativity is a useful target for studies of present-day Western or Westernized societies, it is less suitable for analysing sexual concepts that predate the coining of heterosexuality in the late nineteenth century. Without wishing to re-entrench the period divides that Traub has urged us to cross, I take her warning that looking for heteronormativity in earlier periods risks imposing an anachronistic and monochromatic vision of heterosexuality over an array of disparate acts, desires, and identities founded on cross-sex relationships, obscuring that precisely how these concepts were understood and valued varied between times and places (Reference TraubTraub 2016: 14–15).

Queer deconstructions of heteronormativity owe a lot to Judith Butler’s early work on performativity, but Butler’s insights need not be restricted to contemporary expressions of identity. While Foucault was chiefly concerned with societal shifts in the production of sexuality – a concern that led him to present the rise of a minoritizing model of ‘the homosexual’ as a fait accompli, for which Sedgwick censured him – Butler is interested in how social categories are negotiated at the level of the individual. In Reference ButlerButler’s (1990: 33) famous assertion, the attribution of an identity to a person is not an act that is ever accomplished once and for all, but one that must be performed again and again: it is a ‘repeated stylization of the body, a set of repeated acts within a highly rigid regulatory frame that congeal over time to produce the appearance of substance, of a natural sort of being’. Butler addressed this remark to the production of gender rather than sexuality per se, but later argued that gender and sexuality are subject to overlapping forms of regulation (Reference Butler1993: 238). Both may be understood in relation to performativity.

The concept of performativity was initially set out by J. L. Reference AustinAustin (1962) to designate a class of statements that have no classical truth value. As opposed to describing an existing state of affairs, the utterance of a performative itself brings about a new state. Thus, a speaker in the right circumstances can make a wager simply by saying, ‘I bet … ’ or name a child by saying, ‘I christen you … ’ Butler extended Austin’s concept to suggest that identities are also performative: they are constituted by the act of their invocation. Gender and sexual labels do not objectively describe pre-existing biological realities; rather, they assign bodies to social categories. These assignments, through their constant repetition, create the semblance of a stable subjecthood. At the same time, the social categories themselves are coherent because they have been invoked on many other occasions to constitute many other subjects. Over the course of their lives, these subjects will be beholden to a set of cultural expectations about gender and sexuality which they will have to meet, at least in part, in order for their gendered and sexual selves to remain identifiable to others. If an identity fails to ‘cite’ any norms that are recognizable within its culture (or subculture), then it will be regarded as unintelligible.

Queer linguists have mostly drawn on performativity theory to analyse how language is used by speakers to create individual identities (see Reference Milani, Hall and BarrettMilani 2019). However, Rusty Reference Barrett, Campbell-Kibler, Podesva, Roberts and WongBarrett (2002: 28–29) observes that performatives constitute identities not just through ‘the assignment of individuals to particular categories’ (e.g. she’s a woman), but through ‘statements attributing particular social attributes or practices to a particular category’ (e.g. women are attracted to men). In the latter case, the performative does not impose a subjectivity onto a real body but instead constructs a generalized ‘imagined referent’ (29). Barrett does not suggest dictionaries as a source of such generalizations, but working from his argument, it is not difficult to interpret definitions of social categories as performative utterances. Indeed, one need only draw a line from Reference ButlerButler’s (1993: 232) proposal that the performance of an identity is the ‘citation of a norm’ to Reference Taylor, Joseph and TaylorTaylor’s (1990: 24) proposal that definitions are also ‘citations of norms’.

Due to their different objects of analysis, Taylor’s and Butler’s models are not entirely co-extensive. For one thing, they differ in their approaches to agency. Butler is sceptical of the role played by voluntarism in the construction of an individual subject. As the subject does not socially pre-exist its initial performative invocation, its agency only emerges from, and is dependent on, the constraints of that invocation. Reference ButlerButler (1993: 232) maintains that the performance of a personal identity is therefore ‘forcible’. Individuals cannot easily opt out of the identities by which they are recognized by other people – though they can in some measure subvert the norms through which those identities are produced.Footnote ⁸ Obviously, this issue does not arise in the same way for definitions of social categories. There is no point in asking after the autonomy of an imagined referent in a dictionary entry. We can, however, consider the agency of the dictionary-maker. In this case, Taylor’s perspective is more relevant. He observes that if we accept that all metalinguistic practice is to some extent normative, then that raises ‘fundamentally political questions of responsibility, power, authority and ideology’ (1990: 9–10). By whom, through what means, and to what ends is a particular norm enforced? To Taylor’s list may be added questions of historicity, as raised by Anne Reference CurzanCurzan (2014: 6–8): when did a norm emerge, how has it evolved, and has its underlying authority been in any way altered? Here, it is useful to turn again to Reference ButlerButler (1993: 232), who asserts that the citation of a norm always comes with a ‘complex historicity [that] is indissociable from relations of discipline, regulation, punishment’.

Admittedly, the historicity of the word norm itself is complicated. As Karma Reference LochrieLochrie (2005: 3) points out, it can mean both ‘that which is usual, in the sense of being most prevalent’, and ‘a rule or standard, a type that defines an ideal as well as deviations from that ideal’. The two are often conflated. Reference TraubTraub (2016: 153) voices a concern that if historians of sexuality use norm to mean ‘the typical, the usual, the commonplace’, this might lead us to make incorrect assumptions about which behaviours really were commonplace in earlier periods. Even so, I will retain the word to refer to models of behaviour, particularly those approved by dominant cultural institutions, whether or not the behaviours were actually prevalent. While Reference LochrieLochrie (2005: 3) notes that the OED has found no evidence of norm in English earlier than the nineteenth century, the dictionary also records instances of enormity (in the sense of a deviation from what is right or lawful) going all the way back to the fifteenth. And if there are explicit enormities, then there must be an implicit order of things from which they have strayed. So much can be shown by examples from either end of this book’s timespan.

In 1914, the first edition of the OED defined tribade as ‘A woman who practises unnatural vice with other women’. To read this definition as a performative is to accept that it does not describe a universal fact of language any more than it sets down a truth of nature or a moral absolute. Instead, the definition marks out a social category in relation to a body of norms governing naturalness and morality. For the definition to be coherent to the dictionary’s users, the norms it echoes must already be familiar to them. Whether they have encountered the word tribade before does not matter, but they need to be aware of wider cultural ideals that judge human acts and their actors to be either natural or unnatural, vicious or virtuous. The precise ways in which the OED’s tribade definition was interpreted by individual users in 1914 are lost to us, but we can get a sense of how the norms it cites may have been understood by certain people at the time. Four years after the definition was published, in the final throes of the First World War, England was momentarily distracted by a sensational court case: the actress Maud Allan accused a newspaper of libel for claiming that her performance of Oscar Wilde’s play Salome would incite ‘obscene and unnatural practices among women’ (Reference BillingBilling 1918: 461). Though talk of the trial was widespread, its details remained obscure to some. The socialite Lady Diana Manners, who was volunteering at a London hospital, wrote in a letter to her sweetheart Duff Cooper, ‘The nurses […] ask me all the time about the case and are totally ignorant of any significations [i.e. terms for sexuality between women]. They have a dim vision of Sodom and Gomorrah, which is built for them by the word vice, but even that is hazy’ (Reference CooperCooper 1983: 66). The letter is tinted by class condescension, but it seems reasonable that this passage does reflect the linguistic and sexual knowledge of certain women (and men) in wartime Britain. If any of them had somehow come across the OED’s entry for tribade, the biblical image conjured by the phrase ‘unnatural vice’ would have been enough to make the definition intelligible as a cautionary statement.

Clearly, a performative may be coherent without being informative. The definition fails to specify that the tribade’s vice is erotic. The OED was not unique in this respect: when it comes to explaining same-sex relations, lexicographers have historically favoured disapproval over detail. In 1604, Cawdrey’s Table Alphabeticall had defined sodomitrie as ‘when one man lyeth filthylie with another man’. This definition has nothing to do with identity – linguists’ usual area of concern when they deploy performativity – but it still constructs a category of proscribed acts as much as the OED’s tribade definition constructs a category of proscribed persons. The precise behaviour of tribades, or the manner in which two men can lie together ‘filthylie’, is apparently so sordid as to be unmentionable. Yet norms do not have to be explicit to be successful. In fact, reticence is another part of these definitions’ prohibitive force, demonstrating how ‘silence’ can be ‘rendered as pointed and performative as speech’ (Reference SedgwickSedgwick 1990: 4).

Sedgwick’s plotting of the space between silence and speech, ignorance and knowledge, marks out one way of reading the gaps in dictionaries. Reference SedgwickSedgwick (1990) argues that knowledge does not straightforwardly equal power, for ignorance can be just ‘as potent and as multiple a thing’ (4). Moreover, particular ‘ignorances […] are produced by and correspond to particular knowledges’ (8). This is a helpful guideline for any researcher wishing to understand the ideological function of lexicography as it intersects with sexuality. Dictionaries contain not only erotic knowledges but their complementary ignorances. They offer models of what the dictionary-user ought to know and not know about sex, which sexual practices are approved or prohibited by law or custom, and how the user should act (or be acted upon) as a proper erotic subject (or object). At the same time, because ignorance is multiple, the researcher must bear in mind that the empty spaces found in every dictionary arise for different reasons. An absent headword or vague definition may indicate, among other things, a withholding of information that the dictionary’s target users are deemed unfit to know; a suppression of material thought to be obscene, whether out of a lexicographer’s personal scruples or for fear of social or legal censure; an elision due to spatial constraints or generic conventions; an assumption that something is so transparent or well-known that it needs no explanation; or a genuine mistake or lack of learning on the part of the lexicographer. Examples of all these ignorances will occur over the following chapters, as the book considers how different knowledges were produced in dictionaries aimed at audiences whose presumed gender, age, education, occupation, and wealth varied over time and across texts.

Reading a Dictionary – Text, Practice, Genre

In selecting the dictionaries for analysis in this book, I have been greatly helped by the bibliographies painstakingly assembled by other scholars (Reference WheatleyWheatley 1865; Reference AlstonAlston 1966–2006; Reference Starnes and NoyesStarnes and Noyes 1991; Reference ColemanColeman 2004–2010; Reference CowieCowie 2009; Reference RussellRussell 2018; Reference McConchieMcConchie 2019). My starting point for data collection was to check each dictionary against a list of sexual (especially same-sex) terms known to have been in use around the time of its compilation, drawn up with the aid of the Historical Thesaurus of the Oxford English Dictionary (2010–) and augmented by my own reading of primary and secondary sources.Footnote ⁹ Relevant entries were then transcribed into a database. Many of the collected definitions are reproduced (with certain provisos) in Appendix II, which I hope will substantiate some of the broader claims I make in this book. It may also serve as a preliminary resource for other scholars wishing to study dictionaries’ role in the making of erotic knowledge and ignorance.

When charting the sexual and linguistic ideologies of a dictionary, the researcher must have an eye to more than word entries alone. For this, I have drawn on a number of analytical tools from linguistics and literary criticism. Many of these have already been applied to lexicography in areas other than sexuality. For example, in their work on colonial and nationalist attitudes in dictionaries, Antje Reference HornscheidtHornscheidt (2008) and Wenge Reference ChenChen (2019) have proposed ways of doing ‘critical lexicography’ that are informed by critical discourse analysis – in particular, CDA’s concern for uncovering how structures of power are bolstered or destabilized through language. Reference ChenChen (2019) adapts Norman Reference FaircloughFairclough’s (1992) three-dimensional model of discourse to argue that a dictionary should be read not just as a text but as a discursive practice and social practice. That is, the analyst ought to attend to the content encoded in the dictionary (text), the processes through which that content has been created, distributed, and used (discursive practice), and the socio-historical context in which those processes have taken place (social practice).

There are several sources of evidence for the historical production and circulation of a dictionary, the most obvious of which lie within and around the printed book. Reference McConchie, Tyrkkö, McConchie and TyrkköRoderick McConchie and Jukka Tyrkkö (2018), in their introduction to a volume on exploring lexicography through the concept of paratextuality (Reference GenetteGenette 1997), point to the value of reading the liminal elements that surround a dictionary. These include both the paratexts internal to the book (title page, preface, lists of contributors and subscribers, appendices, errata) and those external to it (announcements of the proposed work, advertisements of its publication, reviews). Often these present the best remaining clues to a lexicographer’s intentions and methods, as well as to their public reception. If researchers are unusually lucky, they may also have access to an archive that preserves a lexicographer’s correspondence or draft materials, allowing them to reconstruct something of how the dictionary was planned and revised. Similarly, private responses to a dictionary by individual users are sometimes recoverable from annotated copies or letters and diaries. Further testimony of the dictionary as a discursive practice can be found in its intertextuality (Reference KristevaKristeva 1980) – that is, its relations to other texts, including other dictionaries. The lexicography of earlier centuries was frequently appropriative, with compilers borrowing freely from each other’s publications (with or without acknowledgement). The ways in which lexicographers altered the material they copied offer insight into both their perception of their peers’ work and their methodological or ideological divergence from it. As such, it is useful to think of intertextuality not as a set of static relations but as a process of entextualization (Reference Baumann and BriggsBaumann and Briggs 1990: 72–78): pieces of a text are decontextualized from their prior site of use and recontextualized into a new text, where their content may be endorsed, disputed, or rewritten to serve a new purpose.

As the above suggests, intertextuality also provides evidence of the dictionary as a social practice – its interaction with wider society. Outside of the dictionary, information on the milieu in which a lexicographer lived and worked can be gleaned from contemporaneous sources as well as historical scholarship. Yet the researcher should look within the dictionary too to see how it integrates allusions to texts from domains beyond lexicography: literature, law, religion, medicine, and so on. These allusions testify to the role that dictionaries have played as broadcasters of cultural discourses. Quotations from other works, or citations of the names of prominent authors and scholars, are among the most overt ways that a dictionary can reinforce the norms of the courts, the church, the medical professions, and other influential bodies – though alternatively these human institutions may be masked behind appeals to more abstract authorities, such as the laws of God, nature, and science. As with copying from other dictionaries, the selection and citation of non-lexicographical texts is never neutral. Reference JohnsonJohnson (1755 I: B2^v) ruefully observed that when a quotation is clipped in order to fit it into an entry, ‘the general tendency of the sentence may be changed: the divine may desert his tenets, or the philosopher his system’. Aside from practical constraints, distortions may be motivated by ideological concerns.

As my emphasis on the intertextuality and paratextuality of specific dictionaries indicates, the present book favours a qualitative approach to criticism. This will, however, be supplemented with quantitative findings in cases where notable discursive patterns across a number of dictionaries can be identified. Such patterns tend to be most apparent in texts that belong to the same genre. Not only do generic conventions inform the cultural domains that a dictionary invokes, but they influence its wordlist, size, style, scholarly pretensions, and target audience. The lexicography of sex in particular presents the researcher with a wide range of potential genres for investigation. The ones I have chosen to explore have been determined by my own interests. Because this book focusses on same-sex intimacy, it will have nothing to say of the clichés of cross-sex courtship as they were satirized in Georgian dictionaries of love, or about country terms for wedding customs in Victorian dialect dictionaries, or on any number of other subjects. There may be times when the reader is disappointed by what I have overlooked, but I have tried to survey a diverse set of genres: dictionaries of hard and general words, law, medicine, classical myth, women’s biography, etymology, cant, and slang. Some of these can be further divided into subgenres, such as medical or classical dictionaries aimed at a popular audience of women and men versus those intended for an elite usership of (presumptively) male professionals. As Lindsay Rose Reference RussellRussell (2018: 16) points out, genre and gender are ‘interlocking systems of sociality’ in the history of dictionaries. Both condition the erotic content that is made available in a given lexicon. This is especially apparent in dictionaries’ treatment of sexuality between women: silently passed over in most general works, its vocabulary must instead be sought in specialist texts. As such, when tracing the intertextual relations between dictionary genres – as between dictionaries and other text types – it is necessary to ask both what is let through and what is held back at the border.

Mapping this Book

The ordering of the following chapters is principally thematic. Chapters 1 to 4 overlap in time, each following a specific parcel of cultural discourses as it moved across dictionaries from the early sixteenth to the late nineteenth century (though the chapters vary in the weight they give to the periods in this span). Dictionary genres have been divided up among the chapters according to the discourses that seem most salient in them. Hard-word and general dictionaries are an exception: discussion of them threads through the whole book, serving as a means of comparison between the knowledges that lexicographers thought were suitable for laypeople and those they reserved for specialists.

Chapter 1 introduces some of the dominant institutional structures through which sexuality in Britain was interpreted. It surveys the relationship between lexicography and the law – human, divine, and natural – as they converged around buggery and sodomy, perhaps the two words most familiarly associated with same-sex intercourse in pre-1900 English. I attempt to defamiliarize them by comparing the diverse explanations given to them across hard-word and general dictionaries, law lexicons, and legal treatises. At the same time, I propose that the semantic instability that dictionaries attributed to these words spilled over into definitions of copulation and other terms for ‘conventional’ sexual intercourse, unsettling their underlying androcentrism and cross-sex normativity. Chapter 2 argues that lexicography was implicated not only in the upholding of law but in the making of a national imaginary – one to which same-sex intimacy was cast as inherently alien. Dictionaries positioned buggery, sodomy, and pederasty as ethically and ethnically remote, vices practised by ancient pagans and modern heretics. However, the moderns’ veneration of ancient Greece and Rome would complicate how classical myths of same-sex desire were retold by lexicographers. Above all, the contradictory portraits of Sappho given in general, classical, and women’s biographical dictionaries reveal the tensions intrinsic to the project of culture-making.

Chapter 3 presents another side to the dictionary as gatekeeper, showing how lexicographers could police the borders of English not just by distancing same-sex practices from English society but by symbolically disbarring words for those practices from the English language. The chapter contrasts dictionaries of ‘proper’ English with lexicons of criminal cant and fashionable slang to trace the limits of the speakable in polite and impolite society. Yet what was unutterable for the layperson was not necessarily so for the professional. Chapter 4 surveys the ways in which transgressive sexualities were openly anatomized in medical discourse, and how the tribade in particular, erased from most other dictionary genres, became an object of enduring fixation in specialist medical lexicons. That works aimed at physicians were able to imagine women’s sexualized bodies with such candour prompts wider questions about how the accessibility of dictionaries to different readerships was socially and materially conditioned.

The chronology of Chapters 1 to 4 ends around 1884, when the first part of the Oxford English Dictionary was published. The OED forms the focus of Chapter 5, which combines a study of the printed dictionary with an exploration of the draft entries, letters, and memoranda that went into making its first edition and 1933 Supplement. I ask what the archive can tell us about the spaces between the compilers’ scientific principles and their actual practice when it came to handling sexual vocabulary – a vocabulary that was growing in size alongside the OED, as the rise of sexology brought new erotic taxonomies to public attention in Britain. Nonetheless, the pathologization of same-sex desire had to compete with the writings of apologists and activists who rejected diagnosis, reclaimed taxonomic labels, or coined their own identity terms. I will survey the extent to which these dissenting voices were represented in the OED, and the challenges they posed to the universal truths of language and of sexuality that scientific lexicographers and sexologists sought to record.

The concluding chapter will look ahead to the legacy of descriptivist ideology for present-day dictionaries, and the limitations inherent in any project whose aim is to document a language as it is commonly and widely used. The advent of corpus-based lexicography – and, more recently, the rise of crowdsourced online lexicons – may seem to hold out the promise of a truly democratic dictionary. However, I argue that ‘majority rule’ or ‘free-for-all’ approaches to lexicography cannot avoid marginalizing speakers whose erotic practices or identities, and the language by which they express them, diverge from dominant norms. Here, as in the rest of this book, I hope to encourage readers to see dictionaries past and present not just as scholarly aids but as texts deserving – and requiring – critical study on their own. The sexual history of English cannot be written without them.