Corps de l’article

1. Introduction

Translation is a privileged site for the introduction of lexical innovations in a given language and therefore offers equally fruitful opportunities, retrospectively, for their historical study. With respect to scientific or otherwise technical terminologies, this role has been richly described, notably within Romance studies tracing terminological and epistemological genealogies between classical and vernacular texts or between vernaculars (Lépinette 1998; Bertrand et al. 2007; Duris 2008; Ducos 2013 and this author’s whole oeuvre; Thibault and Lo Vecchio 2020). Within etymological lexicography, it is common practice to acknowledge the exogenous origin of words whose earliest attested use is found in a translation (for French, TLF, FEW, DMF; for Italian, LEI; for English, OED, and generally see Durkin 2014: 306-307). More generally, in recent years increasing linguistic scholarship has begun to operationalize the study of translation as a site of language contact (McLaughlin 2011; Baumgarten and Özçetin 2008; Steiner 2008; Malamatidou 2016), a development which has helped highlight the prominent role translators and their productions have played in the elaboration of modern standardized languages (Delisle and Woodsworth 2012: ch. 2).

The concrete methodological tool that is translation analysis deserves to be more explicitly recognized for its utility to lexicology, particularly due to the vast potential for updating word histories in the digital era (Maconi 2016; Shapiro 2018; Wright 2023: 4-5). With new electronic resources making it easier than ever to follow cultural and linguistic flows across texts in multiple languages, it is not an exaggeration to state that our historical knowledge of any semantic field or concept could be significantly renewed through in-depth examination of the textual record – more specifically the translational record, as it indicates where cross-linguistic contact informs language use. Perhaps surprisingly, a translation-centered approach is useful not just for scientific or technical vocabularies, but equally so in a semantic field as pragmatically marked (“taboo”) as queerness – taken here in a broad transhistorical sense to refer to the margins of sexuality and gender, however those may be defined at a particular time and place. Many basic historical terms used to name queer people and practices are, in fact, highly internationalized, and traceable within a complex interlinguistic dialogue in various languages.

This paper presents case studies of nineteenth-century internationalisms in five modern languages (French, Italian, Spanish, English, and German): it will particularly focus on the semantic calque of sodomy as ‘bestiality’ and on derivative constructions in the sapphism and uranism series; other related series (inversion, homosexuality) are briefly addressed.[1] The goal is to illustrate some ways in which the lexical-discursive construction of queerness has operated cross-linguistically, via translation, with a focus on late nineteenth-century innovations built upon morphemic stock common to the European languages. This paper is an attempt to explicitly articulate the translational analytical approach that underpinned my specialized dictionary of historical LGBTQ-related internationalisms. In the introduction to that work, I noted:

Les textes traduits sont la forme la plus ancienne de ce processus de transfert culturel et leur étude s’est avérée très fructueuse dans notre travail: dans une large mesure, cette étude pourrait se résumer à une démarche traductologique qui confirme l’importance pour la lexicologie historique d’une comparaison rigoureuse d’un texte avec ses traductions.

Lo Vecchio 2020: 4

The importance of a translational approach to historical comparative lexicology was not assumed a priori but emerged empirically during the course of my investigation. With so many telling early examples found in translated texts, it became impossible to ignore the singular role of translation in the construction of an internationalized lexis. This proved to be the case throughout the modern era: from biblical terminology introduced via translations of the Vulgate or other ecclesiastical texts in Latin (abomination, contre nature, nefando), to medieval vernacular developments with long-term relevance (bugger, bardache), or early modern re-creations of classical Greek forms (tribade, pederast) bearing little conceptual resemblance to the ancient models on which they were based, but representing instead translinguistic creations in the modern languages through the mediation of Renaissance Latin (see Lo Vecchio 2020, 2021). The role of translation remains strong throughout the twentieth century. For popular in-group transfers such as gay and queer, translations may not necessarily have been the main source for activist speakers’ adoption of the terms, but in the broadest sense they establish points of cultural contact between speakers (and writers) and their respective linguistic communities (Lo Vecchio forthcoming).

These points of contact in the written record are where we can witness the very process of cultural and linguistic transfer, and the origins of category convergence, as encoded in lexical terms. Limiting the investigation to one semantic field, we will see how such points of contact help map out flows of ideas across cultures, sometimes with multiple, concealed intertextual layers, through several languages, between an original source and the ultimate translated text.

1.1. A theoretical reconsideration: from lexical borrowing to colingualism

With their complex translinguistic origins and diffusion, lexical internationalisms question established models of linguistic borrowing; a reconsideration of the category seems therefore appropriate here (see thorough treatments on internationalisms in Braun et al. 1990, 2003). Justifiably, it has become routine to acknowledge the inadequacy of both the terminology and the theoretical foundations of “loanwords” or lexical “borrowing” (emprunts, préstamos, prestiti, Lehnwörter, etc.) as belonging to an outdated structuralist view of language. In recent years, the translanguaging paradigm has been particularly useful to emphasize the inherently political and ideological nature of categorizing any set of language practices in relation to a named language (see a synthesis in Otheguy and García 2024). Where more traditional philological-etymological research (e.g., Wright 2023: 6-8) and proponents of newer translanguaging approaches to lexis (e.g., Ahn et al. 2023) have begun to align, however, is in the observation that, when examined in the actual discursive situations of creation and use (rather than abstracted out for attribution to a particular named language’s lexicon), translingual neologisms cannot simply be described as moving, as it were, from one particular code to another. Instead, they reflect the multilingual reality of language users (i.e., speakers, signers, writers) whose lexical repertoires may be cognitively integrated.

Yet internationalisms reveal the limits of the translanguaging framework, too, as they do not merely result from the workings of one integrated cognitive apparatus at the individual level, but from mass-level, mediatized, social processes of lexicalization over time. Another way to put it is that common lexical items possess shared, cumulative social meaning that must be explained by drawing without hesitation on the “social ontology” (Otheguy and García 2024) of a given language, as perceived by its users. Language users’ metalinguistic consciousness, which calls upon extralinguistic encyclopedic knowledge of institutionalized languages as learned in situated contexts, thus tempers the premises of the translanguaging framework: language users are indeed aware of the sociopolitical dimensions of their languages. Frequent metadiscursive statements – especially in the presence of neologizing practices, translingual or not – attest to metalinguistic or epilinguistic awareness, on the part of users, of the existence of distinct codes (Achard 1998 [1993]: 48-49, 114-117). It might be said that this metalinguistic consciousness is what makes languages real.

This paper attempts to recover Renée Balibar’s notion of colingualism (colinguisme, see Balibar 1985; 1993), which insists on the historicized co-construction of languages, as normative institutions, via writing and translation. The latter are forms of “communication réfléchie en langues,” as opposed to the “communion intuitive du langage” (Balibar 1993: 8). Specifically with respect to French, “toute approche du français doit le situer à l’intérieur du système qu’il constitue avec le latin et les autres langues européennes codifiées” (Branca-Rosoff 2001: 1; see also Branca-Rosoff and Wionet 1998: 7). Colingualism (colinguisme) is contrasted with plurilingualisms (plurilinguismes), which in turn refer to the translingual practices omnipresent across the world: “Distinguons clairement le colinguisme, qui définit et réunit les langues, et les plurilinguismes, qui juxtaposent ou mêlent les langages” (Balibar 1993: 7). As pertains to the historical lexicon, plurilingual practices do not necessarily lead to major system-level linguistic changes, with results attested broadly among many users over time and, in the most extreme cases, being sanctioned by dictionaries. Some relation to linguistic normativity seems required in order to explain widespread lexical developments, notably via mediatization: this traditionally includes print sources, but obviously extends nowadays to audiovisual and digital media (Wright 2023: 6, passim).

Balibar’s theory of colingualism may have been largely overlooked in linguistics (but cf. Achard 1998 [1993]; Branca-Rosoff and Wionet 1998; Branca-Rosoff 2001) precisely for the importance it places on writing – and, more pertinently, on translation – in the institutionalization of languages as sociopolitical constructs in particular historical circumstances. That is: the historicized elaboration of langues as opposed to the human universal of langage, a distinction lost in English (Balibar 1993: 7-8). As developed by Balibar, colingualism does not apply universally to all language traditions everywhere or throughout time, but to the particular situation in the Judeo-Christian West – which potentially represents a heuristic limitation (see Branca-Rosoff 2001 for contextualization and critiques). Pared down to its essential insight – that colingualism and plurilingualisms oppose each other in their relation to institutionalized norms – it seems broadly applicable. According to this framework, translanguaging practices can be treated on their own terms, as elements of plurilingualism, without necessarily assigning each lexical item in a speaker’s sequence of discourse to a particular language. Yet, because all discourse occurs within socioculturally situated contexts, there must be some relation to historicized language norms, which for some lexical fields have been co-constructed across societies in a colingual context.

Balibar’s plurilingual–colingual continuum (or nexus) naturally invites comparison with Bourdieu’s notion of the linguistic marketplace (Bourdieu 1982; and see Boutet 2016 [2010]), in which language practices reflect, and enact, the complexly imbricated multiplicity of extralinguistic power relations. Plurilingual practices coexist with institutionalized colingual structures in any form of discourse that can be situated vis-à-vis a standardized language – especially written discourse, which is inherently inscribed within normative standardized conventions.

Internationalized lexis presents a unique case in point (Balibar 1993; Branca-Rosoff 2001: 20-22; see applications in Branca-Rosoff and Guilhaumou 1998 on socialisme and Balibar 1998 on communication). From the perspective of colingualism, internationally convergent vocabulary does not result from isolated “contact” between languages within multilingual communities – that is, through plurilingualism – but from a sustained multiplicity of translational linguistic flows over time, crucially mediatized in such a way as to achieve a critical mass. The internationalizing nineteenth-century lexis of queer sexuality and gender offers an instructive look at how the specifically colingual construction not just of lexical items but of concepts themselves operated within mediatized discourse across culturally proximate societies.

1.2. Methodological considerations

The importance of a comparative approach for any kind of internationalized lexis thus cannot be overstated. Balibar’s framework can be tied to Schmitt’s (also overlooked) notion of “euromorphology” comprising “euromorphemes” (1996; 2000); compare also the notions of “interlexeme” and “intermorphemes” as described in Braun et al. (1990, 2003). Because European languages share etymological roots and combining forms, studies of lexically convergent vocabulary must look past the morphological radix alone and situate the emergence and evolution of derivative or composed lexemes within the interdependent “cultural superstructure” (Überbau, Kultursuperstrat) in which concepts are coded linguistically across languages (Schmitt 1996).

In a lexicological study – as opposed to discursive analysis of individual translational choices made for each token within a text – the focus is on how the lexical types found in specific translations relate to the larger internal lexical structure of a language – including, but not limited to, instances when these types are first documented in their earliest attested or otherwise pivotal uses (attestations charnières, Glessgen 1993); these are data points of fundamental importance (and in need of contextualization) for any historical dictionary. The approach is inherently diachronic, in that it begins with specific lexicalized types with conventional forms and meanings, then goes back in time to locate and analyze key texts that may have been instrumental in their introduction or diffusion.

The analyses presented here do not focus on randomly chosen texts, but on targeted sources with a demonstrably strong value in the diffusion of the selected queer-related terms. Such key texts have been identified through intertextual cross-references within and across languages, including explicit contemporary metalinguistic discourse on the role of certain authors or texts. Crucially, the analysis results from a global cross-linguistic reading that also takes into account the absence of the lexical types in the textual record, including the main electronic corpora. Thus, rather than starting out with individual texts and asking questions about how particular translators chose to interpret their messages, my lexicological approach starts out with lexemes and asks how particular translations may help us situate them vis-à-vis the internal lexical structure of each language.

Lexical transfers between languages (“borrowings”) could all to some extent be considered a kind of translation – obviously so for semantic calques or mirroring phraseology, but also for lexemic loans, as these may represent a communicational choice not to translate. Here, however, I refer to translation in a narrower, concrete sense, as an actual practice taking multiple forms. Without any suggestion of exhaustivity or universal applicability, we can distinguish between four general types of translation practices among those described in the case studies below. Distinguishing between them is helpful to contextualize the multiple vectors of lexical introduction and integration at work in the corpus studied.

  • Monographic translation of individual texts. Lexical items may be introduced as part of fully translated monographic texts, whether books, articles, or other individual texts: this is translation in its prototypical sense. A coinage may first occur in a work that was not translated from the original text but via an intermediary language (see Pięta et al. 2022), whether or not this practice is indicated in the published translation. French has played a famous role as an intermediary throughout the modern era for many languages, and more particularly here for Spanish.

  • Reported translation or partial translation of individual items. In some cases, only individual passages or phrases deemed of relevance themselves are transferred, “reported” as translations in texts in other languages. The most prominent scenario is in book reviews, notably scholarly ones, which have a long tradition, but other kinds of specialized or popular press may report partially translated text and thus introduce lexical innovations.

  • Bilingual/multilingual dictionaries. Sometimes lexical items are created via morphological calque in bilingual dictionaries, in imitation of a lexeme that already exists in another language but has not yet been lexicalized, nor is even attested, in the target language. It is hard to determine whether a lexicographer was responsible for creating such dictionary words via morphemic translation, but if these significantly precede first attested uses outside of the lexicography, that may serve as a clue.

  • Translational context. Lexical innovations are often introduced in a translational context, without the precise site of occurrence necessarily involving translation of an underlying lexemic model from a given source language. Here two scenarios are discussed: 1) textual annotations or marginal commentary involving expansions over the source text, a well-known situation in philology; 2) unattributed, piecemeal, at times plagiarized (at least by contemporary standards) adaptations of a text, frequently encountered as part of synthesizing encyclopedic treatments. Another translational context, not addressed here, involves multilingual publications where translation is central to their purpose, even if each individual item is not itself a translation (Lo Vecchio forthcoming).

These categories together help to chart out the early trajectory of exogenous or translingual lexical innovations in a way that no strictly synchronic account could do.

New electronic resources make it easier than ever to follow linguistic flows across texts in multiple languages. In the field of historical lexicology, corpora should be used as supplementary tools, rather than being the object of study itself (as is often the case in corpus linguistics). Highly developed or tagged corpora (such as Frantext, CORDE, the British National Corpus, etc.), for all their utility in linguistics, tend to yield the least interesting results in comparative research, precisely because the sample is relatively small and contains few translated texts. The “raw” data available in less mediated textual databases – such as Gallica, the Biblioteca Digital Hispánica, Google Books, the Internet Archive, or Hathi Trust – requires much more patient and careful examination, involving repetitive and creative search queries (Maconi 2016; Wright 2023), but is absolutely indispensable at this point in any etymological study (understood in its fullest modern sense: thus étymologie-histoire, rather than narrow étymologie-origine, in Baldinger 1959’s well-known terminology; see Buchi 2016).

Such corpus methods ought to be deployed alongside exhaustive readings of historical and etymological lexicographical sources (including specialized glossaries) as well as targeted consultation of secondary sources that may in turn help identify primary sources for closer scrutiny. An onomasiological approach, which involves consolidating a manageable number of searchable keywords and starting out from certain core texts, is surely the most effective way of proceeding, as opposed, for example, to the old dictionary method of starting with a particular letter of the alphabet (Lo Vecchio 2021: 99-100; Harm 2023: 108-109). Focusing on a single concept facilitates a global, contextualized examination of the record: neither artificially assigning value to isolated occurrences, nor assuming the textual record has nothing of value to tell, but making reasoned hypotheses on the empirical basis of the philological evidence taken transversally as a whole. Where evidence is too scant to make surefooted claims, this is acknowledged; where later textual discoveries come to light, early etymological hypotheses may need to be revised – all etymologies being indeed but hypotheses (Durkin 2014: 13-15).

2. Nineteenth-century internationalisms built on common lexical stock: generalities

The second half of the nineteenth century is when the role of cross-linguistic intertextualities in queer-related naming most conspicuously emerges from the philological record. Not just new lexical categories but actually social ones were then being created, with discursive practices both generating and being generated by their sociocultural context.

Most strikingly, 1860s Germany is where we find the clearest early evidence of proto-queer activist movements in which self-conscious neological acts were central to the project, with assertions of linguistic self-determination occurring together with a desire for social change. Two major lexical families – uranism and homosexuality – were then created or first attested within this activist context, though by the end of the century they would be coopted by the international medical sphere (albeit based on pseudoscientific assumptions), a process that can be read transversally in the written archive. While technical in appearance due to their learned word formation, widespread lexical items such as homosexuality or uranism thus did not follow a prototypical determinologization process – a fact strikingly indicative of the complexity of the linguistic marketplace. Queer denominative subjectivity evidently shifts over time, yet in no era is a queer (or any) lexis purely in-group or purely out-group; the complex power relations and social hierarchies are reflected in who uses it, how, in what situations, including in the transfer from one language to another.

Word formation based on learned morphemes of Greek and Latin origin, a device common to all major European languages, thus facilitated the translation and transfer of lexical items – in the superficially invisible way common to any morphological calque, yet rendered visible upon painstaking examination of the textual record (again see Schmitt 1996, 2000). A distinction needs to be made between lexical base and innovation: that is, between the common morpholexical stock available within European languages, and the actualized discursive combination of these, by agentive speakers or writers, via derivation and compounding. More or less historically prototypical meanings may be coded into morphemic bases, but in combination these are subject to semantic variation and evolution, depending, crucially, on the point of view of the people combining them. The cultural meaning attributed to, say, the ancient poetess Sappho and her fellow Lesbians has evolved considerably over time. While today it may seem a given that Sappho and the ancient Lesbian women represent the prototypical ‘homosexual’ woman, historically many other associations were attached to the inhabitants of Lesbos. Lexical items formed upon the bases sa(p)ph- or lesb- could variously describe a female literary genius, a heterosexual woman passionate in love or in lust, or phenomena in other fields like music, poetry, wine, or architecture. It is impossible to neatly identify when, where, or in what language the now conventionalized ‘homosexual’ view of Sappho and the Lesbians eclipsed the others, because the colingual processes occurred gradually over centuries and across societies. The history of individual morphological formations created upon these bases may, however, be traced more accurately.

This leads to a discussion of “morphological metaphysics,” to adapt Rainer’s (2012) useful expression. Because European languages share a common stock of classical-origin morphemes, a particular combination of these could potentially be coined in any language. But this does not mean, first of all, that the actual morphological realizations do occur spontaneously in each language independently; nor, secondly, that identical formations would carry identical meaning, since all morphemic combinations are only partially compositional at best. As Buchi (2016: 345) notes: “etymologically speaking, there is no such thing as a transparent derivative!” This is where the internal morpholexical structure of a language meets real-world pragmatic actualization in discourse, drawing simultaneously on both individual cognitive and intersubjective cultural factors: after all, a given speaker’s encyclopedic knowledge of the world is by necessity socially conditioned and grounded in a particular communicative context. There are many telling examples in this field where identical morphemic constructions carry very different, if not nearly antonymic, meanings, due to the specific sociopragmatic circumstances of their interpretation. English heterosexuality and Italian eterosessualismo, for instance, referred at times to pathological heterosexual behavior. bisexual, in several languages, could formerly denote pairings of one man and one woman, thus ‘heterosexual’ in a way. Early French precursors such as homéosexuel precede the later equivalent German homo innovations, with the same denotation, yet they had no lasting influence, being confined to obscure sources never subject to international dialogue and diffusion. It is only through inherently historicized social and colingual processes of pragmatic-semantic encoding – of lexicalization – that particular morphemic formations become widely adopted, with conventionalized meanings, in various languages.

As I aim to show in the three case studies below, it is legitimate to see some of the queer-related lexis of the period being actively formed within translational practice, the textual medium forming the foundation upon which (later or concurrent) oral uses were based. Of all such innovations, inversion stands out as being a veritable international transcreation. The process is too complex to lay out here in detail (see Lo Vecchio 2020: 303-336; and compare Bauer 2009; Beccalossi 2012). Yet its history is traceable in the rich metalinguistic and translational discourse that transformed the German conträre Sexualempfindung (‘contrary sexual feelings’), mostly via Italian and French intermediation, into the recognizable sexual inversion form, so successful in late nineteenth-century Europe that the Latinate form later displaced the earlier German compound as well. The all-important homosexuality series, also too multifaceted to be fully addressed here (see Lo Vecchio 2020: 337-394), follows the same general patterns as will be observed below, with the specificity of sparking popular metalinguistic debates, notably in France, about adopting the German-origin barbarism homosexualité – which was nevertheless, morphologically speaking, perfectly French. Following the transfer of homo and hetero forms from key German texts to other languages, we also observe a profusion of neological transcreations at the earliest stage – some of them obscure, like Italian omogeneo, eterogeneo, omoestesia, or Spanish homoerasta, homoestesia, all of them worth closer study – before later consolidation around the established forms still in use today, a situation that only attests to the dynamic cognitive process that is translation.

One methodological aspect worth emphasizing for this period is the “véritable terrain d’expérimentation” for neology represented by the study of terminology à la source, that is, going essentially in the opposite direction of etymology (Buchi 1994: 98). Instead of starting with a given lexical item and working backwards to explain its origin, when we know the source of an innovation – as is the case with uranism and inversion, and to a lesser degree homosexuality and several others – we can work forward from the point of lexicogenesis and then follow the pathways of diffusion and redistribution.

3. Sodomy as ‘bestiality’

According to a broad ecclesiastical view starting in medieval times, and later also observed in juridical and medical contexts, sodomy was a catch-all category of nonprocreative acts, including both homo- and heterosexual oral or anal sex, masturbation, and bestiality (sex between humans and animals) – a technical view that should not, however, be conflated with the more prototypical meaning of sodomy referring to anal sex between men, as sustained predominant usage and dictionaries in the modern languages make clear. In Western European forensic medicine in the second half of the nineteenth century, though, sodomy became restricted to the meaning of ‘bestiality’, to the exclusion of the other meanings including its most prototypical one, in English, Italian, French, and Spanish. This sudden, even incongruous sense development originated as a semantic calque of German Sodomie, introduced in translations of monographs and in scholarly book reviews, especially that of a major work by forensic pathologist Johann Casper.

While German shares a common biblical etymon with other European languages, as well as an early sense history, by the mid-nineteenth century the German form Sodomie had consolidated a semantic restriction to ‘bestiality’. One explanatory hypothesis could be the form’s relatively weakly rooted semantism in German, where Sodomie is far less frequently attested throughout the modern era (and even still today) than its cognates in other languages. Other lexemes like Ketzerei (‘buggery’) or Knabenliebe (‘boy-love’ thus ‘pederasty’) among others appear to do the same pragmatic-semantic work to refer to proscribed same-sex activity, thus leaving the less semantically entrenched Sodomie more susceptible to such a shift. Whatever drove the shift, ‘bestiality’ remains the prototypical meaning in German, as long attested in modern dictionaries.

The semantic transfer to other European languages can be traced in translations starting in the 1850s. An 1854 German medical text translated into English the following year serves as one key early example:

The English adapts and condenses the German, eliminating autonymic markers that added metalinguistic emphasis to the terminology in the original, via both letter spacing (the common German typographic practice of Sperren) and synonym reduplication of Sodomie and Thierliebe (‘animal-love’) with conjunction oder ‘or’. The result in translation (1a) is a candid and uncommented instance of English sodomy, a surely understandable, yet peripheral, use of the term in comparison with its then more prototypical meanings (‘homosexual act’, ‘male-male anal penetration’).

The translations of Casper’s Practisches Handbuch der gerichtlichen Medicin (1858) into Italian (1859), French (1862), and English (1864) are the most important vectors of semantic transfer. This was due both to the influence of Casper’s work, which was constantly cited in the scientific literature, and to the textual economy of the translation which, by calquing a metalinguistic commentary originally valid for the German, applied it to other languages and thus transcreated a new, restricted meaning in the target tongues. In the short “Sodomie” section of his medico-legal handbook, Casper briefly addresses the question of bestiality, noting that it has little relevance for forensic physicians. He states that the strict or “narrower” (enger) sense of Sodomie refers to sex acts between people and animals.

All translators here reproduce the metalinguistic commentary to apply the restricted meaning to the target language, following the German source. The Italian version (2a) expands the original (as is quite common in such materials at that time), even repeating the metalinguistic commentary, “nello stretto senso/significato della parola.” The French (2b) goes so far as to consider this the “true” (vrai) sense of the term.

Judged from the point of view of linguistic accuracy, all these translations (2a, 2b, 2c) could justifiably be considered rather poor, even incorrect, since examination of the documentation in English and the Romance languages, and especially their lexicography, shows that this restricted sense is not lexicalized in any significant way. A word-by-word translation of a metalinguistic commentary rarely leads to a meaningful statement in the target language, and usually needs some sort of adaptation (if not deletion), even if such statements may at times roughly coincide cross-linguistically. Yet in each case here paronymic resemblance with the cognate form prevailed over idiomaticity – a common situation known to any proofreader of translations between related languages that have semantically divergent cognate terms. Leaving aside any judgment of accuracy or elegance, we must simply acknowledge that the translators’ choices played a role in explicitly establishing a neological restriction to sodomy as ‘bestiality’, setting the tone for later authors through the century and into the next.

In addition to the instances above in monographic translations, an 1863 compte rendu of the French Casper translation provides an example of a reported translation of this semantic calque, with a periphrastic gloss for sodomie specifying its restricted use:

The semantic restriction is also observed in Spanish. This is a particularly revealing case because it typifies the cultural and linguistic flow of information among European languages in this field at that time. As in many other examples from the time period, the earliest known transfer to Spanish occurs via French. The novel meaning is used in the fourth edition of Pedro Mata’s Tratado de medicina y cirugía legal (1866-1867), altering the treatment offered in the prior edition (1857) to take account of Casper’s opus.

This is an example of a semantic neologism introduced in a translational context, as defined above. Neither example (4) nor any similar use of sodomía in Mata’s text is translated from Casper; no attribution to Casper or anyone else is directly made when the word is used with this sense. Yet Casper was Mata’s source, as is clear from the many citations throughout the volume. Tellingly, subtle intertextual traces emerge in the “Sodomía” section that, in context, can reasonably only be seen as originating in Casper’s Handbuch. For example, in echo of the statement in example (2) above, Mata remarks that sodomy is more common in the countryside than in the cities: “Esa abominacion es mas frecuente en los campos que en las ciudades” (Mata 1866: 476). What is more, we know that Mata did not consult the original German – he says so early on (“no hemos visto el original”) – but Baillière’s translation into French (Mata 1866: 23). So, without there being any explicit translation in the Spanish text with respect to sodomía as ‘bestiality’, the new meaning occurs within the translational context of Mata’s Tratado. The lexical innovation thus presented in Spanish is built upon an internationalized foundation ultimately traceable to the German, via French, yet with no direct signposting of this layered history.

For the Spanish as for the Italian, French, and English languages, whether the strict ‘bestiality’ meaning of sodomy was ever widespread or even known outside of specialist legal medicine circles is beside the point. Among a particular community – a kind of speech community, as it were (or a “community of practice,” as applied in Wright 2023) – sodomy had a certain currency with this restricted sense, due to the influence of German science and, more particularly, to the translations that introduced this novel sense to readers of other languages. Methodologically, this example underscores the importance of a comparatist method in historical lexicology. Studied individually in the languages where it is attested and, in some cases, recorded in historical dictionaries, the origin of the abrupt semantic development is opaque, if not impossible to identify; only by contextualizing the lexical item transversally with related European languages can its origin be found, then followed forward through intertextual links.

4. Sapphism

As a fairly typical case, the origins of the sapph- series cannot be neatly identified, as it is part of a word family (conceptually tied to lesbian) that already had a long but ambiguous history in relation to female (homo)sexuality. Earlier adjectival forms adapted from Latin Hellenisms – French saphique, Italian saffico, Spanish sáfico, English sapphic, German sapphisch – were sporadically, at times euphemistically, coded with (potential) ‘homosexual’ meanings, including in progressively lexicalized collocations such as amour saphique, sapphic love, sapphische Liebe, etc. Because the queer meanings are related to wider, shared understandings of Sapho and her Lesbian consoeurs, it would not be possible, or meaningful, to singularly attribute such uses to the influence of one source language. All the same, a global, contextualized cross-linguistic reading allows us to establish a relative chronology for each representative item in the lexical family. Early, isolated attested uses of English sapphism and sapphist in the late 1700s predate corresponding forms in other languages; while the philological record prior to the mid-1800s is too lacunary to make firm judgments, the discursive and extralinguistic context suggests that these were Gallicisms (Lo Vecchio 2020: 223-224). From the mid-1800s, translations make it possible to trace the sapphism/sapphist pair (in its modern use) more clearly to the French.

In German and Italian, the earliest known attestations are found in bilingual lexicographic sources, two decades before the appearance of other occurrences. This suggests that these lexemes began as translated “dictionary words” that had no prior lexicalized status in each language but were instead the creations of lexicographers who calqued the morphological construction from the French. The German Sap(p)hismus treatment appears first, in Adolphe Peschier’s update to Mozin’s dictionary (1842-1846).

Peschier’s source was Barré (1838-1842), a recent supplement to the dictionary of the Académie Française: Peschier’s definition is largely copied from Barré. In the earlier French-German entry (ex. 5), the clue that this was a morphological calque from the French is the spelling, with only one -p-. The simplified French spelling saph- was by then generalized, in comparison to German or English, for example, where the more etymological sapph- forms were still dominant (and still are). The form is then corrected to German spelling norms in the German-French entry four years later (ex. 6) as Sapphismus.

Italian saffismo appears to have been transcreated by morphological calque in the same way, in the 1855 Nouvel Alberti dictionary (French-Italian), also drawing on Barré (1838-1842), with a similar definition that is clearly derived from it.

No saffismo entry is found in the 1859 Italian-French volume (Nuovo Alberti), however, which supports the hypothesis that the lexeme then had no lexicalized status in Italian but was merely translated from the French so as to provide an equivalent morphemic structure.

A pivotal period in the process of lexical transfer from French saphisme to other languages begins in the 1880s, following the work of French doctor Louis Martineau, a ubiquitous authority in the Western medical literature of the time. The restricted technical meaning this author gave to saphisme ‘(oral) clitoral stimulation between women’ spread via reported translations of his texts in scholarly reviews. Curiously, Martineau initially attributed this innovation to the American gynecologist Marion Sims: “le saphisme, suivant l’expression par laquelle Marion Sims a caractérisé les habitudes des femmes lesbiennes contemporaines de Sapho” (Martineau 1877: 193). I have not located this use in a cursory examination of Sims’s published texts, but the archives of this notorious physician could be a source for future revelations about the lexical history of the term in English. Similar texts by Martineau circulated in multiple publications, but here is one of the earliest versions, followed by translations (some of which were based on slightly different original versions). The most important aspect is the metadiscursive introduction of the new terminology, which syntactically is almost identical across languages:

This is just a sampling of such reported translations (themselves circulating alongside monographic translations); many other examples may be found, all attributing the discussion to Martineau and using his terminology of sapphism. Again, the influence of the French can be further observed, in the spelling of Italian safismo and English saphism in examples (8a/b): these forms imitate the spelling of the French, going against historical orthographic conventions and what would become normalized as Italian saffismo and English sapphism. The German passage cited in example (8c) does not reproduce the spelling with one -p-, but the Saphismus spelling is found elsewhere in the same review, in imitation of the French. Overall, formal variation at this early stage is indicative of the novelty of the lexical item in Italian, English, and German.

The more general meaning of sapphism as ‘female homosexuality’ would be the one to spread outside of the medical field. This sense is actually also attested in Martineau’s texts (along with several other shades of meaning), as a result of a metonymic shift in which a specific sex act becomes semantically contiguous with the woman who practices it – a reductive if not uncommon framing within queer-related lexis. However, the limited number of attestations prevents making easy inferences about the spread of this general meaning via the translation of Martineau’s oeuvre. As for the analogical term sapphist, it is also used in Martineau’s oeuvre, but later, and as such was not transferred in the earliest translations into other languages. Yet it follows a similar pattern: in all cases studied for this period, an underlying French textual model can be more or less explicitly identified (see Lo Vecchio 2020: 197-229). As we will see below with the uranism/uranist pair, formations such as sapphism/sapphist cannot be taken for granted: the fact that they could theoretically have been coined in any language does not make it less significant that all of the earliest forms empirically observed in the Italian, German, and English textual records are modeled on the French.

The Spanish transfer follows the typical route at the time, with key Martineau-informed attestations appearing slightly later and intermediated via French. Yet Spanish safismo had actually already been introduced in the 1874 Spanish translation of a popular French novel. We can interpret this as evidence of the translator’s awareness of the internationalizing “scientific” terminology of the time; it also indicates a gap in the textual record (meaning that this whole discussion could potentially be further antedated), since this is clearly not an isolated use, as revealed by a metalinguistic commentary. Safismo does not appear in the literary text itself, but in the translational context of a pathologizing footnote added by the translator.

In this particular case, there is no underlying lexemic model for Spanish safismo in the original French novel. Nor did the translator, Amancio Peratoner (1874), use the terminology of safismo in his cited text on the “DANGERS OF LOVE” (some lesb- formations are used there, though). Both texts more generally reflect knowledge of the growing body of European (pseudo)science in the field, by translating key authors selectively or otherwise calling upon them intertextually.

5. Uranism / Uranist

The uranism series is an excellent example of the fertile experimental ground presented by a terminology studied à la source (Buchi 1994: 98). Not only is the moment of innovation quite precisely identifiable in the record – Karl Heinrich Ulrichs’s 1860s series of pamphlets, Forschungen über das Räthsel der mannmännlichen Liebe (‘Research on the mystery of male-male love’), along with some published personal correspondence – but it involves dozens of lexical creations forged by a prolific neologue who very consciously drew on classical sources to produce a novel vocabulary adapted to the moment, although only some of these coinages would find further resonance in translation (see Beachy 2010 for an overview of Ulrichs’s role; Matzner 2015a/b on classical influences).[3]

Far from representing some mechanical application of the “meaning” attributed to Aphrodite Urania – a pure and celestial love as described in Plato’s Symposium – Ulrichs’s creations are deeply original, idiosyncratic even, with a wealth of real-time commentaries and metalinguistic discourse articulating their basis. For this reason, a closer lexicological reading refutes the ad hoc argument by Bauer (2003: 399), who attempted to explain the fact that Ulrichs’s works were not translated in their time by reasoning that they derived from a “humanist tradition […] providing a common source of cultural concepts that need not be translated among languages,” thus positing their “a-translatability.” Leaving aside the obvious ideological factors that might have prevented their translation, we observe that Ulrichs’s terminology was in fact very much translated, although not in the more traditional form of monographic translation. Likewise, Kaylor’s argument (2006: xii) that the English uranian series was an independent innovation unrelated to Ulrichs’s neologisms could only result from an incomplete analysis, as it fails to account for the internationalized environment informing its emergence (see also Beachy 2010: 805-806 n. 16).

As the production of an openly, militantly queer person who saw themself as a woman trapped in a man’s body, the uranism series also presents considerable value in a sociopragmatic and ideological perspective. Despite its activist origins, Ulrichs’s terminology would later be coopted by science and spread largely in that field (see examples 10-14 below) – although not exclusively, since it was also used by authors such as John Addington Symonds (Lo Vecchio 2020: 298) and Edward Carpenter (Rowbotham 2008: chs. 10-11), among others. The interplay of in-group and out-group usage, as partially as it emerges from the documentation, reflects an uneasy social dynamic at a pivotal moment in queer history. As such, it highlights the importance of a nuanced discussion around queer naming. It is crucial not to conceive of queer people at the time as mere victims of a discursive reality over which they had no control, but instead to acknowledge the complex ways in which social hierarchies are reflected in lexical use (see Eribon 2012 [1999], 426-428; Boutet 2016 [2010], 87-89; McConnell-Ginet 2020: 8-9; Turton 2020: 225-228).

The common established forms in this series in all languages under study neatly condense Ulrichs’s vast terminology into an -ism/-ist pair: uranism and uranist. However, no such analogical construction is found in Ulrichs’s writings. There, German Uranismus is predominantly used to refer to the abstract concept, while initially Uranier and later Urning refer to the person. By comparing translations, we learn, first, that the -ism/-ist formations can in no way be assumed to arise spontaneously and independently in each language; in fact, various urning(-) transfers were initially put forward, in line with the original. More specifically, the historicized roots of the -ism/-ist pair can be traced to French translations, which again were influential in the process of lexical diffusion. If it seems incongruous to argue that the -ism/-ist pair in other languages was due primarily to the French innovation in the 1890s when Uranismus had been in use for three decades in German, this case reflects indeed the highly internationalized context in which lexical transfers were occurring in this semantic field. In any case, if we empirically follow the intertextual trail of translations, rather than projecting current metalexical intuitions onto the past, this is the hypothesis we arrive at.

While Ulrichs coined the terminology of uranism in the 1860s, two other German-speaking authors were instrumental in its spread to other languages, mostly in the late 1880s and early 1890s: Richard von Krafft-Ebing and Albert Moll. Krafft-Ebing, in particular, was of the utmost importance for the diffusion of the homosexual lexicon, especially via the various editions of his main work, Psychopathia Sexualis. Where Ulrichs had used several abstract noun forms – first and foremost Uranismus, later Urningsliebe and Urningthum – Krafft-Ebing favored the form Urningthum, starting from the first edition of Psychopathia Sexualis (1886). (Uranismus there appears only in the titles of cited works.) This would prove significant in terms of its translation, because the earliest relevant translations were modeling their terminology not on Uranismus, but on Urningthum. Among the first translations of Krafft-Ebing’s work to appear was the Italian by Sterz and Waldhart (1889), where the original Urning was adapted into urningo and Urningthum into urningismo (Krafft-Ebing 1887/1889). Similarly, Chaddock’s English translation (1892) adapts the Urning and Urningthum found in Krafft-Ebing’s main text as English urning and urningism. Significantly, though, English uranism does also first appear in Chaddock’s 1892 translation, not in the main text but in a quotation by Albert Freiherrn von Schrenck-Notzing cited in the translator’s preface. Here, the immediate model for English uranism is indeed German Uranismus, following Schrenck-Notzing’s (1892) text, which was absent from Krafft-Ebing’s original work.

In 1893 was published Pactet and Romme’s French translation of Albert Moll’s Die conträre Sexualempfindung, where Moll makes very extensive use of Ulrichs’s terminology, with hundreds of tokens to be found. Unlike Krafft-Ebing and now explicitly following Ulrichs, Moll favored Uranismus as the abstract noun and Urning to name the person (Moll 1891). In translation, then, Pactet and Romme modeled French uranisme after Moll’s German Uranismus by morpheme substitution and simultaneously, by analogy, coined uraniste – terms presented throughout, other than on the initial references, without autonymic marking, presenting a fluent semiotic integration into French (Moll 1891/1893). Of course, the translation of these terms cannot be seen somehow as a morphological necessity but as resulting from dynamic cognitive creation; the Italian and English formations heretofore show other ways translators chose to adapt the terminology. The French was not inherently superior to the earlier terminologies, but it was crucially based on two learned morphemes of classical origin (as opposed to the Germanized or properly Germanic combining elements urn-, -ing-, -tum[4]) that could be easily transferred to other languages, as would prove to be the case.

The same year, Paul Blocq adopted Pactet and Romme’s terminology in several publications. Most significant was a brief overview in the Gazette hebdomadaire de médecine et de chirurgie synthesizing the most important works in this field (including the French translation of Moll), an article that would in turn be translated/adapted into both Italian and English – the site of the earliest attested uses of the -ism/-ist pair in both of those languages. Translation is multiply layered here, with passages from Moll’s monographic translation being reported, themselves then translated further into Italian and English. Here the Moll translation is explicitly given as the source of uranisme:

The articles also use uranist several times in the discussion of Moll’s work, although without specifically attributing the term to him. As there is no metadiscursive treatment, the innovation appears semiotically integrated already upon these earliest attestations. For example:

By the time the French translation of Krafft-Ebing’s Psychopathia Sexualis appeared in 1895, the -ism/-ist terminology had been further consolidated, including in that work (Krafft-Ebing 1893/1895). That was also the year of Ulrichs’s death and the year that, so far as I have been able to ascertain, Uranist first shows up in German in a scholarly review of a work by French author Marc-André Raffalovich. This German instance is not visibly modeled on any particular lexical item in the text under review, although the reviewer was evidently familiar with the now internationalized terminology (Appelius 1895: 350).

Such superposed cross-cultural flows are also vividly apparent when examining the introduction of this terminology into Spanish, both Peninsular and Latin American. The earliest known transfer occurs in an 1895 translation from an Italian text by Scipio Sighele, who describes the terminology and attributes it to Ulrichs. The Italian adaptation urningo, based on Ulrichs’s original coinage, is rendered as such into Spanish (with obvious morphological adaptation for the plural).

The feminized noun urninga is also coined in the Spanish text (p. 148: “la urninga mujer”), where the Italian had feminized urningo adjectivally as urningo femmina. I have not encountered later traces of these isolated coinages in Spanish, but they represent important points of cultural contact – especially as they appeared in a publication by the prestigious España Moderna house (Yeves Andrés 2002: 102). It should further be noted that Sighele’s treatment, here applied to both men and women, already departs conceptually from Ulrichs’s original work, where Urning specifically referred to (birth-assigned) males. (Yet Ulrichs also later revised their thinking, applying specific terminology to women, notably Urningin, with the German feminine ending -in.) This is important to keep in mind, because the many cross-linguistic transfers in this field over time were constantly accompanied by conceptual reimaginings, so that strictly originalist, literalist interpretations of what, say, uranism (or homosexuality or inversion…) denotatively “meant” should be viewed with extreme suspicion.

Directly following this is a Spanish version (1896) of Albert Moll’s work, which I have just argued was primarily responsible, through translation, for the spread of the -ism/-ist pair. In this case, indeed, the Spanish does not follow the original German, but the influential French translation. No explicit indication is given in the Spanish edition that the French was the intermediary source, but it is evident through close comparison of the texts, most conspicuously their titles: where the original German was titled Die conträre Sexualempfindung (‘Contrary sexual-feelings’), the French reads Les perversions de l’instinct génital. Étude sur l’inversion sexuelle, a title subsequently calqued in Spanish as Las perversiones del instinto genital. Estudio sobre la inversión sexual. Ulrichs’s terminology is explicitly introduced and employed throughout.

Beyond the obvious calquing from the French in this particular passage, the terminology employed all throughout the text is evidence that the Spanish version followed Pactet and Romme’s translation, employing uranista, after the French, uraniste, where Moll’s German uses Ulrichs’s original coinage, Urning.

Early attestations of both urning and uranista in the Latin American Spanish context similarly reveal concealed intertextual layering. The following example from a Chilean forensic medicine handbook is actually an adapted translation from a French work by Thoinot.

Puga Borne thus condenses the original, only obliquely acknowledging Thoinot’s relevance to this discussion in a mention lower down on the page (“diremos con Thoinot”). Yet no direct attribution is made to him as the original author of these lines, putting this treatment somewhere between actual translation in terms of its syntax and, because the latter can only be discerned through attentive inspection, some kind of translational rewriting. By modern standards, this would be seen as plagiarism, yet here it pervades the whole work – as was the case in the above-mentioned manuals by Peratoner (sapphism) and Mata (sodomy), and in many other similar publications in various languages.

By studying multiple layers and types of translated text from source to target, we are able, as accurately as possible, to follow the early trajectory of what would ultimately become common established forms in European languages. Which, of course, is not to suggest that the earlier urning(-) (or other) formations immediately or entirely disappeared; some formal variation continues to be attested here as would be the case for any innovation, exogenous or internal. The influential role of French – central in fin de siècle European medicine, as in many other fields at the time – and more particularly the timing were both doubtless decisive. Due to the sheer volume of texts published on the subject, it is legitimate to consider the 1890s as a kind of turning point in the discursive construction of (homo)sexualities in some Western societies. However, caution is warranted so as not to exaggerate this ahistorically, as if all lexical creation pertaining to sexuality before and after this period were somehow irrelevant and subject to different discursive forces, or as if all sectors of all (Western) societies shifted their discursive practices at once.

6. Conclusion

The case studies above present examples from one particular semantic field, yet the translational approach partially described here can be applied to help map out the lexical history of other terminological contexts in which internationalized language contact, broadly defined, has played a role. The overarching point is that a lexicological approach to translation analysis focuses neither singularly on individual translation practices in discourse, nor on notional translation of concepts cross-culturally, but on how the evidence observed within certain material translations relates to, or even itself explains, the items in the historically structured lexicon of a language. In the most general sense, translations provide concrete chronologies and geographies of cultural contact fundamental to the description of any internationalized lexis. In many cases, as with sodomy as ‘bestiality’, sapphism, uranism/uranist, inversion among others, the origin of the lexical innovation can legitimately be seen as occurring through translational practice.

Methodologically, translations are where researchers of historical, contact-induced lexical change would do well to turn first. Theoretically, Balibar’s notion of colingualism offers a compelling alternative to traditional lexical borrowing and translanguaging as a way of explaining cross-linguistic lexical and conceptual convergence. Such convergence emerges not as a result of plurilingual processes but as part of sustained interdependent discourse in the translational written record. The more striking point may lastly be that this co-construction of lexical categories is not restricted to scientific or technical terminologies, but also extends to fields of major social relevance.