1The papers in this special issue developed out of research presented at two workshops held in the English Department at the University of Zürich in January and July 2023. These events came about as part of a collaboration between the editors of this volume, and Annina Seiler and Olga Timofeeva of the University of Zürich, which was funded by UKRI (U.K. Research and Innovation) and designed to facilitate conversations amongst U.K. and Swiss-based researchers about our shared interests in the vocabulary of medieval English and its multilingual contexts. The editors are immensely grateful to Prof. Timofeeva, Dr Seiler and their colleagues for co-organizing and hosting the workshops, and for all their kindness and hospitality. We would like to express our sincere thanks to all those who attended and contributed to these events, in person and online; and to the editors of Lexis for publishing the outcomes of this work here.

2More broadly, this collaboration of course reflects the converging interests of participants in several recent and ongoing research projects tracing multilingualism and its effects in medieval Britain. These include the Bilingual thesaurus of everyday life in medieval England project (BTh) and its off-shoots (currently ‘The semantics of word borrowing in late medieval English’; Louise Sylvester, Richard Ingham, Kathryn Allan & Megan Tiddeman; funded by the Leverhulme Trust), ‘Gersum: The Scandinavian influence on English vocabulary’ (funded by the U.K.’s Arts and Humanities Research Council (AHRC), with a volume edited by Dance, Pons-Sanz & Schorn [Forthc.]), and now Olga Timofeeva & Annina Seiler’s project ‘Waxing and waning words: Lexical variation and change in Middle English’ (funded by the Swiss National Science Foundation). Many of these same research interests are moreover represented in the activities of the ‘Medieval English (c. 600–1500) in a multilingual context’ (MEMC) network (funded by the AHRC) and its collection of essays (Pons-Sanz & Sylvester [2023]). In their turn, these networks and projects are of course hugely indebted to the findings and the inspiration of recent milestones in lexicographical resources and related work on lexical borrowing and semantics, such as the Historical thesaurus of English (HTE), the World loanword database (Haspelmath & Tadmor [2009a and 2009b]), Durkin ([2014]) and the ongoing revisions to the Oxford English dictionary (OED) and other major historical dictionary enterprises.

3The seven papers presented in this special issue are all concerned with ways in which exploring the multilingual cultures of medieval England enriches our understanding of the history of words, helping us find out more about the make-up of vocabulary in English and other languages, the uses and interrelationships of lexical items and how these changed over the centuries. Some papers focus on particular periods of medieval English, with individual pieces dealing with Old English and early Middle English texts, and some also include findings for more recent periods and textual material probably produced outside Britain. They employ a variety of approaches and techniques, ranging from surveys of large lexicographical datasets that analyse their results via cutting-edge statistical and visualization tools, to fine-grained readings of specific medieval texts and authors, their sources, intellectual traditions, scribal contexts and stylistic nuances. Their key findings have to do with the history of particular semantic domains and their lexicalization, including semantic hierarchies, polysemy and word formation; with lexical borrowing and its relationships with innovation and obsolescence; and with sociolinguistic contexts, cultural traditions and the evidence for multilingual communities at home and abroad. The first five papers all treat the effects of these multilingual contexts on the vocabulary of specific conceptual areas, across one or more broad periods of English; and the sixth and seventh narrow the focus to case studies of individual manuscript copies of medieval texts.

4In the opening paper, Trends in the development of vocabulary for emotion and cognition in English: A millennial perspective, Kateryna Krykoniuk & Sara M. Pons-Sanz take an appropriately long view of lexical development across the centuries, from Old English to the present day. Their focus here is on two particular semantic domains, emotion and cognition, the histories of whose lexicalizations in English have never been compared in depth before. To achieve this, Krykoniuk & Pons-Sanz examine a sample of 1414 nouns drawn from HTE categories 02.01 (mental capacity) and 02.04 (emotion), marked up with etymological information from OED. They identify and visualize trends over time using clustering techniques (Principal Component Analysis and Correspondence Analysis) to help determine sets of nouns with similar properties. Their analysis shows that each historical period may be associated with developments in certain semantic subfields, and with the presence of particular word-formation processes and/or of items derived from one or more source languages. In the cognition field, for instance, the Old English period shows a clustering of words for knowledge, whereas in Middle English there are more items relating broadly to the subfield of memory; when it comes to emotion, on the other hand, Middle English clusters significantly with terms for fear, compassion, pleasure and suffering. A larger proportion of items that lexicalized emotion in the Old English period seem to have become obsolete compared to those in cognition; but the word-formation processes evidenced are more consistent across time in the emotion domain than they are for cognition, even if in both cases changes are visible. Krykoniuk & Pons-Sanz’s dataset supports the cross-linguistic findings generated by the samples in WOLD, namely, that cognition has the higher borrowing rates overall of these two domains. (This is an interesting result, not least in the light of the work later in this issue by Sylvester & Tiddeman on anger, which they identify as an especially highly lexicalized domain with plentiful borrowing in the Middle English period.) Breaking down the data by source language, cognition has a more significant presence of words derived from Latin, and also from Greek, whereas originally French and Scandinavian items play a greater role in emotion; but in both domains there are sub-fields which buck these overall trends, and some where native roots were and remained dominant (notably love). Zooming in on the Middle English data then allows one to explore the finer details behind these headline findings. The development of the memory sub-field, for example, is associated especially with the later Middle English period; and in both main domains, it is in later Middle English when the contribution from Romance languages first becomes really significant, with roots derived specifically from French becoming productive in English word-formation processes remarkably quickly following their introduction to English, something that bears out Ingham’s ([2018], and below) important conclusions about the sociolinguistic situation in England from the thirteenth century onwards.

5Richard Ingham’s own paper, Loanwords and polysemy: An investigation of specialized domain lexis in Middle English, is the first of three pieces in this volume that build on the data of BTh and its associated projects in order to explore the contexts and effects of lexical borrowing in semantic domains in the Middle English period. Here, Ingham investigates the polysemy of the French loanwords in Middle English found in three occupational domains, with significant results. As he points out, in (post‑)colonial contact situations in the modern period, and/or where borrowing from a “foreign” language is concerned, the tendency is for words to be loaned with just one meaning, illustrating the effects of “cultural borrowing”. But Ingham’s work (starting with Ingham [2021]) indicates that French words very often entered medieval English attached to several of their different senses in the source language, and he argues that this reflects the borrowing of French lexis into English by bilingual speakers from the higher strata of society who routinely employed both languages (with French probably acquired in the middle childhood years). The present paper takes this research further, to consider the characteristics of French loanwords in semantic domains where their borrowing may have been motivated by technical innovation. In such situations we might have expected a lesser degree of polysemy, with words being borrowed only in specialized senses germane to specific, technical concepts; but Ingham finds that, in fact, French loans in these domains in Middle English are only slightly less polysemous than those in other domains. Using BTh data in conjunction with the semantic category structures of HTE, and comparing the definitions of borrowed words and their etyma in the major historical dictionaries, Ingham investigates polysemy in French-only borrowings in Middle English in the BTh domains manufacture, farming and building. He finds that these loanwords continue two or more of the senses of their polysemous French sources in 70 cases out of 111; and moreover, that 15 of the remaining 41 are first attested in continental French citations dated to ca. 1350 or later, that is, to a period when a bilingual situation in England conducive to polysemous borrowing can no longer be assumed.

6The third paper, Louise Sylvester & Megan Tiddeman’s Lexicalization, polysemy and loanwords in anger: A comparison with non-affective domains in Middle English, presents some fruits of the ongoing ‘The semantics of word borrowing in late medieval English’ project (2022-). The BTh and its earlier off-shoot ‘Technical language and semantic shift in late medieval English’ (2017-2020) explored the vocabulary of nine everyday occupational domains (including those treated by Ingham in the foregoing article). Here, Sylvester & Tiddeman compare the findings for those domains with those for a new, tenth semantic field from the affective sphere: anger. The material analysed is divided into two datasets: the main one contains 5,818 Middle English terms, classified into 2,400 HTE categories and sub-categories, and arranged according to their place in a semantic hierarchy; the second element of the study then digs into a subset of these words, viz. 100 selected French and/or Latin loans, whose meanings in Middle English are compared in detail with those recorded in the historical dictionaries of their source languages. Sylvester & Tiddeman are able to show that, in some important respects, the terms and senses grouped under anger behave in ways closely comparable to the characteristics identified for the nine occupational domains. A key finding of the earlier studies, that loanwords arriving in the Middle English period only rarely went on to oust native terms with the same meanings, remains true in this affective domain. Likewise, the ratio of polysemy in Middle English loanwords from French and/or Latin to the polysemy of their etyma in the source languages is very similar in both anger and the occupational domains, and the same is true of levels of semantic overlap (the proportion of new senses unique to Middle English is roughly equal in anger and the other domains). However, a key point of difference for anger lies in the sheer number of lexical items recorded per sense: anger is far and away the most heavily lexicalized domain in Sylvester & Tiddeman’s corpus; especially remarkable are five “super-lexicalized” senses (each containing more than twenty co-hyponyms), and it is these that are particularly well populated with items of non-native origin, especially loans from French. Individual words in the anger domain, of whatever etymological origin, are also more likely than those in the rest of the dataset to be polysemous, including the most polysemous native and non-native words in the whole corpus (ME wratthe and greven). When it comes to the long-term outcomes of synonymy, Sylvester & Tiddeman demonstrate that this greater polysemy is most likely to result in the retention of the first item recorded in English with each sense, with incoming synonyms tending to drop out of use before Present-Day English; this finding supports Weinreich’s [1968] hypothesis that the lexis of affective domains will tend to show higher levels of redundancy, and be regarded by users as more frequently in need of renewal. It is also noteworthy that, amongst loanwords, senses newly developed in Middle English are markedly more likely to arise by metaphorical extension in anger than are new senses found in other domains.

7Gloria Mambelli & Johanna Vogelsanger’s paper, The church and the manor: Assessing and comparing the effects of language contact on two Middle English lexical domains, also focuses on words in specific semantic domains in Middle English, in this case comparing the lexicalization of two broad areas of meaning relating to important but contrasting medieval social environments, namely, “the church” and “the manor”. Both contexts were characterized by multilingualism, but in different ways and to different degrees: the upper echelons of ecclesiastical life would have been predominantly francophone and Latinate, and, while estate administration required some practical uses of literacy and competence in French, the world of agricultural labour and its tools and techniques would have been dominated by monolingual English speakers. In their study of these domains, Mambelli & Vogelsanger focus on nouns denoting locations and objects; their dataset for church is drawn from HTE (260 items) and for manor from the BTh and the manorial vocabulary in Mambelli’s doctoral research (348 items), with etymological information following OED and MED. For each domain, and for the terms denoting locations and objects within each, the article investigates source languages (combining French and/or Latin into a single group), semantic specificity (following the Category Level system devised by Sylvester, Tiddeman & Ingham [2020]) and survival rates. Amongst the very interesting findings are that the church domain contains significantly more French/Latin loans than manor (150 versus 72), confirming the central role played by French and Latin users in the ecclesiastical sphere (with words apparently borrowed directly from French being the lion’s share); and that, even though the Romance influence on the Middle English vocabulary of agrarian life is far from negligible, the proportion of native-derived lexis increases at the hyponymic levels for manor (especially in the objects sub-set). The bulk of lexical material clusters at the hypernymic level in both domains (i.e. there is the least variety amongst words denoting the most specific, technical concepts), but there is a noteworthy exception when it comes to the objects sub-set of manor vocabulary, indicating (at least if we take the whole Middle English period and its dialects all together) a relatively wide pool of terms available for the instruments and tools used on manorial estates (something that mirrors the distribution of material in the farming domain analysed in Sylvester, Tiddeman & Ingham [2020]). In addition, Mambelli & Vogelsanger’s analysis of obsolescence rates for their data confirms the findings of other recent studies (see also Sylvester & Tiddeman’s paper above), namely, that – contrary to common expectations – the arrival of loanwords does not seem to have caused significant losses amongst native lexical items; neither is there any major variation in survival rates across words deriving from different source languages in the two domains.

8The next paper, Social ties and negotiation of lexical norms in Old English: The vocabularies of vices, presents another study of lexical choices in a specific conceptual area, this time focusing on the rich evidence for the bilingual intellectual culture of the Old English period. Olga Timofeeva & Christine Wallis are our guides to the extraordinary world of words denoting types of sinful behaviour in Old English prose, and to the contexts, methods and resources for studying them. Their aim here is to explore how these items differed historically, regionally and in relation to the varied textual and cultural traditions that mark pre-Conquest writings. None of the Old English items that feature in this dataset are loanwords per se, but they were variously generated and re-generated as translations of Latin terms and may often have been coined on the basis of Latin models. As Timofeeva & Wallis illustrate, the lexicon of sin in this period is remarkably “polysemous and fuzzy”, showing considerable diversity across recorded Old English literature, even though the cataloguing and definition of the (usually eight) capital sins depended ultimately on the same influential Carolingian Latin sources. Although some of the terms (especially those denoting greed, gluttony and sadness) tend to be stable over time and space, others (particularly words for pride and sloth) varied significantly from text to text, even across the output of single authors, and none has survived to Present-Day English. Beginning with the principal Latin sources and their doctrinal contexts, Timofeeva & Wallis analyse the key Old English passages deriving from them, including the translations of the Capitula Theodulfi; Ælfric’s detailed explications and thesaurus-like categorizations of sins, their synonyms and sub-types in his Catholic homilies and Lives of saints; Wulfstan’s developing terminology in several of his works; Byrhtferth of Ramsey’s treatments in his Enchiridion; and a range of glosses, glossaries and confessional texts, the latter especially notable for the gamut of sins that they catalogue. Comparing these lexical traditions, Timofeeva & Wallis suggest that Ælfric and Wulfstan may both have adjusted their sin vocabularies as a result of their correspondence, but that Wulfstan also has specific elements in common with Byrhtferth which may show influence from Old English penitential literature, something that is most apparent in Wulfstan’s case in his homilies. Moreover, vernacular terms for sin also need to be understood in dialogue with the meanings attested for the same lexical roots in secular contexts (especially interesting here is the “Winchester word” modignes, which Wulfstan may have adopted from Ælfric before discarding it in favour of the less potentially ambiguous ofermodignes). Other notable developments include the creation of sets of words for the capital sins ending in OE -nes, perhaps designed to create equivalence with their Latin lemmata in terms of morphological structure and grammatical gender.

9Next, in Revising Layamon: The Otho scribe and his French additions, Max Fincher takes us forward slightly to the early Middle English period, to investigate one particular manuscript as a context for the occurrence of lexical borrowings. The scribal dialect in question is that of the revision of Layamon’s Brut in the late thirteenth-century London, British Library, MS Cotton Otho C.xiii. Fincher compares this version of the text to what we assume is a copy closer to the authorial original in London, British Library, MS Cotton Caligula A.ix, and asks what their lexical differences can tell us about borrowing, obsolescence and the relationship between word choice and literary-stylistic tradition. Of the two extant versions of this important early Middle English verse chronicle, the Otho redaction has long been recognized as lexically less conservative, and in particular as containing a larger range and number of loanwords from French. Fincher focuses here on those cases where the Otho text has a French loan but the “original” Brut in the Caligula manuscript does not, collecting the evidence for all these items (58 in total) in an appendix. In the first of two main sections, he analyses the words that have been replaced, and which have often been argued to belong to that layer of the original Brut’s vocabulary that has been described as archaic, archaizing or archaistic. He finds that a significant number of these terms are not attested in English later than the Brut (both versions), either at all or in the specific senses found in the Brut. This suggests that the Otho reviser replaced these items for the purposes of “updating” or clarifying what for them was old-fashioned or idiosyncratic usage, although it is interesting that they did not always do so consistently, occasionally being willing to preserve instances of this more “difficult” lexis, whether as a result of deliberate retention or a lack of diligence. Moreover, well over half of the replaced items (37) are not obviously archaic at all. In his second section, Fincher turns his attention to the kinds of prosodic linkage, viz., alliteration and rhyme, found in the Brut, and examines the relationship between Otho’s French-derived substitutions and the occurrence of these effects. Significantly, he shows that the presence of rhyme (which the Otho reviser is in general more likely to introduce) is never disrupted by these lexical changes; but that Otho’s replacements seem to be made without (much, if any) concern for how they might impact the operation of alliteration. This finding suggests that, however well integrated these French loans were into the Otho redactor’s active lexical repertoire, they do not appear to have been interested in using them to supplement the fund of alliterative synonyms available to the poetic tradition(s) of the Brut (in contrast with how Romance borrowings came to be deployed later in the Middle English period, in the verse of the “Alliterative Revival”).

10Finally, Christine Wallis, Annina Seiler & Heather Pagan investigate Multilingual glossing and translanguaging in John of Garland’s Dictionarius: The case of Bruges, Public Library, MS 536. Their focus here is another individual manuscript copy of a text, one written around the same time as the Otho Brut, but which survives (and may well have been produced) outside England and which contains material in Middle English and French as well as Latin. This mixture is fascinatingly emblematic of the complex, day-to-day encounters with multilingualism that characterized the international scholarly and ecclesiastical life of this period, including the communities of writers and readers at the University of Paris and the wider sets of interconnected groups in Flanders and in England who may have been the first users of this book. John of Garland was an Englishman who taught in Paris and Toulouse in the thirteenth century, and who produced his Dictionarius as a pedagogic resource to expose his students to Latin words for the (extraordinary array of) objects they might encounter on a walk through Paris. The copy of John’s text in Bruges 536 seems to have been designed to include glosses, and some 800 have been added, in Latin, French and Middle English. Wallis, Seiler & Pagan present here the first ever edition of these important glosses. They also provide a detailed linguistic analysis of the distribution of the glosses by language and by the different hands that inserted them, to explore how this copy of the Dictionarius was made and may have been used by its original group(s) of teachers, readers and language learners. The glosses are found mainly towards the beginning of the text, and cluster in chapters containing technical vocabulary. It is the Latin glosses that show the most varied functions: they are attached to lemmata belonging to all the main grammatical categories, and provide information about morphology and syntax, as well as lexical and explanatory glossing. The French annotations are the most numerous (more than 500), and are mainly lexical glosses on nouns. A dialectal analysis indicates the presence of features associated mainly with Anglo-Norman orthography. Another crucial indicator of the English connections of users of this manuscript is of course the presence of 58 glosses in Middle English. These provide some fascinating and hitherto unknown evidence for medieval English vocabulary, including three plausible antedatings of OED and MED attestations; while orthographically there are again Anglo-Norman features, suggesting that those responsible were perhaps more used to writing Anglo-Norman than Middle English. The English annotations were added by several hands, and in many cases they appear in double glosses with French terms. While some of the English interpretamenta are shared with glosses in other copies of the Dictionarius, this is not always the case, implying that the glosses in some sections (such as chapters 65-66) were added by bilingual speakers “as spontaneous reader reactions”. Interestingly, the glossators seem to have preferred French terms that are etymologically unrelated to their Latin headwords, often using words with Germanic origins; and moreover, a high proportion of the French glosses (more than 200) are attested as borrowings into Middle English before 1350. As Wallis, Seiler & Pagan remark, this may suggest a deliberate choice of words that “blur the boundaries between English and French”, and hence the adoption of a teaching approach that resembles the phenomenon now known as “translanguaging”.

11Together, the approaches and findings of these seven papers reflect the state of the art in a range of aspects of research into English lexical history, and point the way forward to future opportunities for deeper and more refined insights into the rich variety of contexts and effects of medieval multilingualism.