1The research reported here investigates the relationship between lexical borrowing and polysemy in Middle English, taking a polysemous word as usually defined to be an item with two or more distinct but related senses sharing an etymology. As is well-known, the Middle English period saw extensive borrowing from medieval French. The issue to be addressed is the extent to which those loanwords were generally adopted in English in a single sense, or whether lexemes polysemous in French were borrowed into Middle English with source-language polysemy. This is a distinction with important implications for the nature of the borrowing process taking place in England at this time. Lexical borrowing has been common in modern-day colonial / post-colonial scenarios, where the relevant languages are usually in a diglossic relationship (Fishman [1967]), the colonists’ language being High and the indigenous language(s) being Low. Lexical borrowing may also occur where the source language is a foreign language, i.e., the borrowing language speakers do not inhabit a geographical area controlled by speakers of the source language. In both contexts, according to Winter-Froemel [2014: 73], “loanwords are normally borrowed in one meaning only. This means that S[ource] L[anguage] polysemies (and homonymies) normally disappear in processes of borrowing”.
2The position taken by Winter-Froemel is amply borne out in colonial scenarios for borrowing. In former British colonies in Africa, such as Kenya and Tanzania, monosemous loans are very commonly reported. Swahili has borrowed English party as phathi, but only in the sense of ‘political party’, and the English verb pass as pasi, only in the sense of ‘to pass an exam’ (Lusekelo [2013]). Kikuyu has borrowed English note only as nooti ‘bank note’, and in Urdu, press has been borrowed only in the sense of ‘to iron’ (Aslam [2020]). In former French colonies in the Maghreb, such as Tunisia, Arabic has borrowed Modern French demande (‘request’) only in the sense of ‘application’, and orange only in the colour, not the fruit, sense (Sayahi [2014]). Similarly, in dialects spoken in Carpatho-Ukraine, formerly part of the Austro-Hungarian Empire in which German was the High language, lexemes polysemous in German are typically used in a single sense, e.g., Zug (‘draught’), not e.g., ‘train’, ‘chess move’, and ‘trait’ (Hvozdiak & Zymomrya [2019]). Where a colonists’ language borrows from an indigenous language, monosemy again appears to be the case: Calude, Miller & Pagel [2020] found that out of a corpus of 113 Maori lexemes borrowed into New Zealand English, 105 showed no polysemy, e.g., of the three senses of Maori pāua ‘abalone’ (a kind of mollusc), ‘fishing bait’, and ‘hoof’ – only the first has been borrowed by New Zealand English.
- 1 Loanword polysemy is to be distinguished from loanshift, where a native word takes on an additional (...)
3When the languages are not associated in a former colonial context, what might be called ‘international’ borrowing occurs. Languages enjoying prestige in certain domains are likely to function as source languages, e.g., English in sport, French in food. Again, monosemous borrowing appears to be the normal case. In Slovak and German, the English loan star denotes only an outstanding artistic or sports person widely known in public (Gallová [2009]). Sports domain loans from English are likewise monosemous in Indonesian (Nurweni [2013]). In English, the French loanwords croissant and éclair have only the culinary sense, not those of ‘half-moon’ and ‘flash (of lightning)’, respectively1.
4English borrowing from French during the later medieval period was on a very sizable scale, with estimates tending towards a figure of around 10,000 words (Kastovsky 2006). Most such loans are first attested after 1250 or so (Durkin [2014]). Mainstream history of English textbooks, e.g., Baugh & Cable [2002], usually discuss French loans in terms of domains in which the Norman ruling elite would have left their stamp, such as fine food, dress, the law, religion, and cultural pursuits. It would therefore be natural to consider the French-origin vocabulary of Middle English to represent mainly cultural borrowing, in which case it would be predicted to have displayed rather little polysemy. However, this view has recently been challenged by Ingham [2021], who uncovered a strong tendency for French loans in the Middle English period to be polysemous. Over three-quarters of 51 French-origin polysemes sampled, and attested in insular French, were found in the Middle English dictionary (MED) to have at least two of the senses of the respective Old French etymon. Ingham [2021] explained this finding on the basis that the language contact setting in later medieval England was quite distinct from those giving rise to cultural loans in recent times. It involved educated elite bilingualism in English and French, whereby the usual senses of a polysemous French item borrowed into English would have been known to bilingual speakers. This differed from the context typical of nineteenth- and twentieth-century colonial contexts, in which monolingual speakers of a local language borrowed a lexeme from a colonizing elite that tended to be fluent speakers only of the dominant colonial language. It also differed from the contemporary context in another way: nowadays a word travels from one language area to another via international communication, as with sports and IT terminology. In the medieval period, international communication was far more limited.
5Motivation for borrowing, as discussed by Matras [2009], can take the form illustrated above of prestige loans, when bilingual speakers use an item from a language associated with prestigious contexts, and of cognitive pressure. This last Matras describes as pressure on a bilingual speaking one language to relax ‘inhibition’ of the other language. Much psycholinguistic research has shown that both of a bilingual’s languages can be activated in speech processing, resulting in cross-linguistic interaction. For Ingham [2021], the borrowing process involving French loanwords took place in a context of individual bilingualism; loans would have been adopted from one of the bilingual’s languages into the other language, i.e., they would have been borrowed via the bilingual lexicon (see, e.g., Schreuder & Weltens [1993]; Kroll & De Groot [2005]; Grosjean & Li [2013]). Heine & Kuteva [2005: 100] pointed to a process of polysemy copying whereby, under contact influence, a native word takes on an additional sense or senses replicating the semantics of its source language equivalent. The semantics of a French loanword would in like fashion have been copied when borrowed into English, accounting for the prevalence of loan polysemy in the medieval context. The context for borrowing at this time was predominantly not one where French words passed from a language used by a socially higher culture to a language used by the hierarchically inferior one, but one where both languages were used by the bilingual higher strata of society. The concept of prestige loan is therefore ill-suited to the sociolinguistic environment within which loans took place in the Middle English period. Bilingual speakers would have had more than one sense of a polysemous word available to them in the languages they knew, and could therefore borrow more than one French sense in their English.
- 2 I would like to thank an anonymous reviewer for raising this point.
6The existence of an educated bilingual stratum of society does not by itself explain, however, why French loans in Middle English were most often polysemous, while loans into indigenous languages in contact with modern English in (post-)colonial settings have strongly tended not to be2. In numerous colonial regimes, particularly in the later years of their existence, a stratum of local society became bilingual in their native variety and in the colonists’ language, allowing them to participate in administration and other aspects of public life. Yet as illustrated above English-origin loanwords in local languages tend not to be polysemous. Two distinctions between these scenarios are nevertheless potentially relevant. The medieval English contact setting differed from (post-)colonial ones in that the settler and indigenous population strands (Schneider [2007]) had by the 13th century become intertwined by marriage, whereas in English-speaking postcolonial settings there was very limited intermarriage between white settlers and indigenous peoples. Another difference, perhaps more significant linguistically, is that, in Britain’s African colonies especially, English was usually an instructed foreign language (Schneider [2007]). In medieval England, early quasi-naturalistic acquisition of French may have facilitated a level of competence in the language that gave rise to the psycholinguistic priming effects discussed in Ingham [2021], whereas it seems plausible that later learners of English do not experience such effects. This must remain a matter of speculation on our part pending further enquiry into the L2 learner mental lexicon.
- 3 Or American terminology such as hood, trunk, and fender.
7The borrowing scenario proposed by Ingham [2021] did not, in any case, take into account what happened in language domains where borrowing may have been strongly motivated by technical innovation. According to Prins [1941: 285], “[m]any of the French terms introduced into Middle English were introduced as technical terms with very limited sphere of application, because they were used as such in Anglo-Norman”. It is known that French-speaking master builders were imported into medieval England, e.g., those employed for building cathedrals and other large buildings (Corrigan [2019]). In domains such as this, technical terminology may have been borrowed in only the sense appropriate to the one in which French speakers worked and came into contact with their English-speaking workforce. It is an everyday observation that languages borrow items as monosemes for a technical purpose, e.g., a technician using English as a second or foreign language may know English terms for a motor vehicle such as transmission, ignition, bonnet, boot and radiator3, without knowing other senses of these items. Ingham [2021] took only a general sample of Middle English loans from French, without considering the situation in more technical walks of life. A question arising was therefore whether in such areas loanwords were borrowed only with senses related to a particular specialization; those borrowing them would not have known other senses for them outside their own occupational domain.
8That issue will be pursued here using the resources of the Bilingual Thesaurus of Everyday Life in Medieval England. This makes available language related to medieval occupational domains where some of the terminology was specialized in one sense for a particular technical concept, similarly to the Modern English motor-vehicle lexemes mentioned above. On that basis, Middle English loans from French in specific occupational domains might be expected to show less polysemy than was found in the general sample used by Ingham [2021]. They might have gained only the sense appropriate to a particular domain, while the other sense or senses of the lexemes remained unborrowed. This would have resembled the modern-day cultural loan scenarios in which loans are not adopted within a bilingual speech community, but by borrowing from a colonial or foreign language, and polysemy does not normally arise. The alternative hypothesis is that Middle English loan lexis of particular occupational domains were just as polysemous as with loan vocabulary generally, reflecting the practice of a bilingual speech community that tended to replicate in English the polysemy of its French source lexemes. As with the modern-day motor vehicle analogy mentioned above, one sense might have related to a particular domain, while other senses had more general application. The research reported here investigated which was the most common outcome of borrowing from French in contexts favourable to technical specialization.
9French in England initially had the status of a language spoken by a mostly French-speaking army of occupation, and by settlers following in their train. However, the ethnic divide between Saxons and Normans appears to have weakened considerably by the end of the twelfth century (Short [1980]). Bilingualism spread during this period and later, though probably applying only to strata of society favoured by birth or by education. The long-lasting use of French in aristocratic and higher ecclesiastical strata is attested in surviving documentary texts, but also in those created by merchants, lawyers and civic administrators. Though French in England was unquestionably a language of the socially privileged, privilege could derive from formal education; it did not necessarily mean aristocratic lineage. The clerks that worked in various literate occupations, and the clergy who also functioned as teachers, could be of fairly modest origin. As part of their education, they would normally have acquired French, though not in the form of an instructed foreign language. Contemporary testimony of medieval schooling practices tells us that up to the time of the Black Death French was a vehicle language in grammar schools through which Latin was learnt (Orme [1973]). Such a pedagogic procedure would hardly have worked if a good grounding in French was not acquired first. This was argued by Ingham [2012] and [2015] to have occurred in pupils’ first school, known as a schola cantūs, or song school, in a naturalistic ‘immersion’ context, providing them with the communicative competence in French that enabled them to learn Latin successfully in the grammar school classroom.
- 4 In accordance with the practice of the Anglo-Norman dictionary and the Anglo-Norman Text Society, t (...)
- 5 Insular morphology sometimes appears to depart from continental French when the loss of a phonologi (...)
10Evidence from linguistic features of insular French support that scenario, in which knowledge of the language would have been generally transmitted via child second language acquisition, in the middle childhood years around age 5. Meisel [2009] found that morphosyntactic competence in a second language is reliably acquired from frequent exposure to the language up to age 6-7, but thereafter requires explicit instruction. However, to be acquired naturalistically a nativelike phonology needs exposure at an earlier age, by around 3 years. Now the morphosyntax of later Anglo-Norman (A-N)4 was still largely that of continental French (Wilshere [1993]; Ingham [2012])5, while its phonology was definitely not (Pope [1934]). If French was acquired naturalistically in the middle childhood years around age 5, these diverging linguistic outcomes receive a satisfactory explanation.
11Up to the mid-fourteenth century. educated speakers in England would thus have possessed, in addition to native speaker English, some competence in French, though to a level which would have varied considerably in the individuals concerned. In the case of clerks and others involved in administrative and other aspects of public life, it is likely to have been to a fairly proficient level, to judge from the plentiful documentary record surviving (Trotter [2003]; Rothwell [1993]; Lusignan [2011]).
12In short, the linguistic situation in England with respect to French and English should be seen as evolving over the two centuries after the Norman Conquest from the coexistence of a high-status and a low-status language, each mainly used by monolingual speakers, to one where both languages coexisted in the repertoire of educated bilingual individuals, i.e., from societal to individual bilingualism. The latter scenario is regarded as one that favours contact effects of a more intense nature than in less intense contact settings such as those illustrated in the previous section (cf. Trudgill [2011]). Importantly, the great majority of French loanwords first occur in English from the later thirteenth century onwards, after the spread of individual bilingualism among the educated community. Contact influence via bilingualism no doubt encouraged a tendency to borrow more than one sense a lexeme had in French, whereas contact influence without substantial bilingualism would have produced single-sense lexical borrowing. While that position, argued for by Ingham [2021], may be valid for the general vocabulary, it remains to be seen whether it holds good for specialized domains where technical lexis was borrowed. The present research sought to investigate this issue by drawing on the lexicographical resources described in the next section.
13Reliable semantic criteria for polysemy remain problematic, not least because they usually invoke speaker intuitions about relatedness of meaning (e.g., Lyons [1977]; Lehrer [1990]; Geeraerts [2001]), which are often fluid and dependent on individual judgment. Nevertheless, lexicographers continue to make such judgments, as they must if they are to create usable and consistently laid-out dictionary articles. Historical linguists researching polysemy using dictionary resources will inevitably depend on their judgments. Earlier states of language using the extensive lexicological resources now available electronically, both for English and for French, allow lexical searches for target items, but for the identification of particular senses possessed by a lexeme, one is dependent on lexicographical information. The research presented here made use of a number of lexicographical resources in order to investigate the extent to which the polysemy of a French source word was reflected in the item borrowed into Middle English. They are discussed in turn below.
- 6 Some adjustments to conceptual categories had to be made in BTh for the specificities of medieval d (...)
14The Bilingual thesaurus of everyday life in medieval England (BTh; Ingham & Sylvester [2017]) was established using the category and subcategory structure of the Historical thesaurus of the Oxford English dictionary (now, Historical thesaurus of English, HTE). This offers a conceptual classification, in a hierarchical series of categories, of the lexis of English from the Anglo-Saxon period to the present day. The data is arranged in numbered hierarchical categories, each consisting of a defining heading followed by chronological lists of all the words, with their dates of currency, used as synonyms or near-synonyms for the definition. The headwords in a conventional onomasiological dictionary are ordered alphabetically and represent lexemes for which appropriate definitions are provided. In HTE the headings structure the classification and determine the ordering of the hierarchies6.
15BTh consists of seven occupational domains, each of which was intended as a set of practices and conceptual spaces, articulated by a conventional code (cf. Bourdieu’s [1979] concept of “habitus”). Those practices reflected some degree of input from bilingual English-French speakers, and would have been particularly receptive to French-origin lexis in domains where supervisory control was exercised by higher-status French users such as master builders. The vocabulary was classified in terms of semantic role (Jackendoff [1972]), and the language used for the agents, processes and objects special to each domain was identified via reverse look-up searches of two electronic dictionary sources, MED for English and the Anglo-Norman dictionary (AND) for Anglo-Norman. Around 500 Middle English headwords were designated etymologically as exclusively French-origin items by MED.
16The Dictionnaire du moyen français (DMF) is a large and well-regarded online dictionary comprising the lexis of French in the fourteenth and fifteenth centuries, a very large part of which would have been in use since the Old French period, which is conventionally taken to have given way to Middle French around 1325-1350. A DMF entry most commonly displays first a basic or literal sense of a lexeme, followed by analogical senses and senses by extension, e.g., (translated):
bande:
‘a piece of diverse materials longer than broad’;
by analogy ‘piece of land’.
17The distinctions drawn in DMF between sense categories involve different kinds of opposition that depend on the item in question, as can be seen in the following labels (translated by the present author) for distinct senses, such as ‘literal’, ‘physical’, ‘figurative’, and ‘by analogy’, cf. the following entries translated and abridged here for illustrative purposes, and with the original French category label between brackets where provided:
gentil:
Of a person: ‘of noble birth…’;
by analogy (p. anal.) ‘With an idea of…quality’
liaison:
Literal (au propre) …. ‘bond serving to tie someone up’;
figurative (fig.): ‘that which confers a moral obligation’
réflexion:
Physical (domaine physique) … ‘phenomenon by which light is reflected’;
figurative (fig.): ‘ability of the mind to go back so as to give deeper thought to something’
saillie:
‘Action of going out’;
metonymic (p. métonym.) ‘passageway to the outside’.
In such cases the senses so distinguished are usually labelled ‘A’ ‘B’ and ‘C’, e.g., the entry for the item gentil:
gentil, adj.
A. -[D’une pers.]: ‘Of somebody’
1. De condition noble, bien né: ‘Of noble status, well-born’
2. [Idée d’agrément physique et moral]: ‘Sense of physical and moral attractiveness’
3. [D’un animal] Gracieux, élégant, racé: ’Graceful, elegant, thoroughbred’
B. -P. anal. [D’une chose]: ‘Of something’
1. [Avec une idée de noblesse, de perfection, de qualité]: ‘With a sense of nobility, perfection, quality’
2. [Avec une idée d’agrément]: ‘With a sense of attractiveness’
Senses differentiated by capital-letter A and B distinctions were taken to indicate polysemy.
18The DMF’s editorial principles seem not to have been uniform, however. In some cases capital Roman numerals are used to distinguish senses of a lexeme, e.g., s.v. investir, with no other level of sense distinctions:
investir:
I. - Investir qqn (de qqc.). Mettre qqn en possession de qqc. (‘put someone in possession of something’ )… ... ... ... ...
II. - Investir qqn / un lieu. Cerner, attaquer (‘surround, attack’)
Here, polysemy was taken to be indicated by the capital Roman numeral differentiation.
19Where multiple subsenses appear at different levels, but under the same capital letter or Roman numeral heading, this too was counted as polysemy, e.g.:
- 7 This entry has been simplified for expository purposes.
couvrir7:
A. Au propre (‘Literal’)
a) Placer sur, disposer sur (‘Place, arrange on’)
b) En partic. CONSTR. Édifier un toit au-dessus d’un bâtiment…(‘Esp. CONSTR. build a roof over a building…’)
c) Répandre sur (‘spread on’)
d) … Couvrir qqn ou qqc. pour le protéger ou le dissimuler (‘Cover someone or something to protect it or hide it’)
20In some cases, a distinction that was arguably grammatical rather than semantic produced an entry with different capital letter or Roman numeral entries for different constructions in which the verb appeared, i.e., transitive and intransitive uses of retirer (‘withdraw’). So as to avoid possibly subjective departures from the DMF’s practice, these cases were also counted as polysemous.
21Where two or more defining terms are placed at the same level in DMF, the item is taken to be monosemous, e.g.:
cordonnier:
Celui qui fabrique, répare, vend les chaussures, cordonnier (‘one who makes repairs sells shoes, cobbler’)
ponce:
Pierre dont on se sert pour polir, nettoyer, pierre ponce (‘Stone use to polish clean pumice stone’).
A ‘shoe-maker’ and a ‘shoe-seller’ might be seen as constituting different senses, but the format used by DMF was followed for reasons of consistency, and any such items were taken as having a new single sense in French. They therefore did not become candidates for loan polysemy in the present study.
22The Middle English dictionary was compiled over a period of many years with the purpose of providing definitions of lexemes used in later medieval English texts in their various senses, and explanations of their use in phrases. They can be fairly lengthy and elaborate, with numerous senses distinguished numerically. A fairly typical example of how MED handles polysemy is the entry for assai (use in phrases omitted):
- 8 This sense, evidently used in biblical writing to correspond to Latin gentiles, is not found in DMF(...)
assai:
1. (a) The testing or a test of quality (as of a metal or other object); assaying of metal; (b) of god ~, of good quality or make; (c) the testing (of a report or problem).
2. (a) A testing of character or personal traits (such as faithfulness, friendship, faith, fortitude); trial, ordeal; (b) character or personal qualities (as demonstrated by trial).
3. (a) A test of arms, combat; an attack or sally; (b) fig. a sally (as of the Devil or an enemy).
4. Ceremonial tasting (of food or drink); also, a taste or foretaste.
5.(a) Investigation, examination, inspection; observation, experience; also, power of observation, reason [quot. c1600]; (b) what is observed or observable; evidence, sign
6. (a) An attempt; striving, effort; (b) an enterprise; a feat.
The Middle English lexeme derived from OFr. gentil, receives similarly expansive treatment:
gentil:
1. (a) Of noble rank or birth, belonging to the gentry, noble; -- often implying character or manners befitting one of gentle birth; -- also used of Christ; sup. the gentileste, the noblest, the noblest men; (b) ~ blod, noble ancestry; also, persons of noble ancestry;
2. (a) Having the character or manners prescribed by the ideals of chivalry or Christianity; noble, kind, gracious, etc.; (b) courteous, polite, well-bred, charming; graceful, beautiful, handsome; (c) used ironically.
3. (a) Belonging to a person of rank; also, suitable to a nobleman; noble, gracious, refined, graceful, beautiful; (b) of animals, birds, bees, fish: of excellent breed or kind; superior, excellent, fine; (c) of things: excellent, superior; (d) of qualities, actions, words, behavior: noble, gracious, kind, generous;
4. Pagan, heathen8.
5. In surnames.
23Comparing the layout of MED gentil with the layout of its DMF counterpart shown above, we found that the latter makes an initial two-way distinction between gentil when used for an inanimate or abstract thing, and when applied to a person or animal. It then offers further subdivisions relating to breeding, high quality and attractiveness which cross-cut the animate / inanimate initial distinction. MED does not make an explicit sense distinction between gentil used for an animate and an abstract entity. It defines the word in sense 1 as relating to a person’s birth and social standing, and in sense 3 as relating to qualities seen as refined and superior. Sense 2 relates to notions of courteous, chivalrous conduct, hinted at in the DMF sense A2 by the use in the definition of the word moral. Although the layout of the respective entries in the two dictionaries appears somewhat different, the senses noted by DMF – superior breeding, excellent quality, and refinement – can also be found in the MED entry.
24The same can be said for ME coveren, compared with the simplified entry shown above for its French etymon couvrir. The MED entry for coveren recorded three of the French senses:
coveren:
1 (a) To cover (an object) by laying or spreading material over or about it…
2(a) To cover (sb., the body, its parts) with an article of clothing or armour…
7. To cover (a building) with roofing material
Here, the general sense A(a) of the original is referred to in MED sense 1(a), the construction sense A(b) of the original in MED sense 7, and part of sense A(d) – (’protect’) – in part of MED sense 2(a) – ‘armour’.
25It was felt that the DMF entries were more rigorously structured than their MED counterparts, which often gave the impression of listing uses rather than senses (Möhren [2000]). As will be described in the next section, DMF was therefore relied on methodologically for sense differentiation.
26The other dictionary used in the analysis procedure was AND, which—as noted below—served to check whether the senses available in French of a word borrowed in Middle English were attested in the insular variety. In general the senses found in AND matched fairly closely their counterparts in continental French has represented in DMF. As an example, we take the AND entry for gentil:
gentil:
1.noble, lofty (birth, rank etc.) ♦ (of a person) noble by virtue of position, character, or exploits ♦ (of persons) having the character appropriate to one of good birth ♦ (formu laic) perfect, noble ♦ (used in polite address) noble
2. beautiful, shapely
3 (of thing) splendid, fine, superb ♦ (of thing, food) excellent, fine, of good quality ♦ (of fragrance) delicate, subtle
4. (of language, action) kind, courteous
5. (of animal) domesticated, tame, quiet, easily managed
6. wealthy
Senses A1-3 and B1 of the DMF entry for gentil, providing for notions of noble rank and character, when used of a person, and of superior quality when applied to a thing, occurred also in Anglo-Norman. DMF sense B2, relating to attractiveness, is strongly implied by the terms ‘beautiful’, ‘splendid’ etc. used as translation words in AND senses 2 and 3. The sense of gentil found in references to animals differs somewhat, however: DMF alludes to the animal’s breeding whereas the AND sense refers to its manageability. (The MED entry picks up the continental but not the insular French sense.) The very substantial but not total overlap between these DMF and AND entries typifies what was observed in comparing the two dictionaries, not unexpectedly since Anglo-Norman was a dialect of medieval French.
27The procedure followed was to search the following BTh domains for French-only loans in Middle English: Manufacture; Farming; Building. The BTh search facility was used first to obtain all monolexemic items in each domain with French derivation (either continental or insular). Items whose etymology is attributed by MED also in part to Latin or another language were discarded. The French etymons of the remaining Middle English items were then looked up in DMF, and their senses were compared with those provided by MED. Monosemous French lexemes were classed separately.
28Senses relating to feudal rights, duties, taxes or other customs in operation in France but not to our knowledge in England were disregarded for the purpose of establishing polysemy. Polysemous French lexemes were categorized according to whether the corresponding loan in Middle English possessed two or more of its senses, or else a single sense only, i.e., no distinction was made between cases where two senses and cases where more than two senses were shared between etymon and Middle English lexeme. Polysemy arising from sense extensions only in Middle English was ignored. AND was consulted as to whether senses of Middle English lexemes used were also recorded for the French etymon in insular usage. If a sense was recorded only in insular French, the relevant lexeme was counted separately, as in Ingham [2021]. In all cases, the general principle followed with the analysis was that senses distinguished in DMF at separate levels, and thus warranting an analysis of polysemy, were searched for in the MED definitions without requiring them to be arranged in similar fashion to the presentation in DMF.
29The identification of separate senses was undertaken as discussed above. A number of examples will be provided next, showing how this was implemented in practice. In the case of a word such as plastre, matters were straightforward: both the dictionaries made it clear by using different level entries that the word possessed distinct senses, and the two senses afforded by DMF showed up as items 1 and 3 in the MED article for plastre:
plastre (DMF):
A. Matière blanche tirée du gypse, plâtre
B. P. méton. 1. Préparation médicinale à usage externe, emplâtre" ...
plastre (MED):
1. Med. (a) A plaster, poultice, or compress made from herbs, meal, or other substance, often applied on a cloth to the affected area, for a variety of medicinal purposes…
3. (a) Gypsum, sulfate of lime; calcined gypsum, plaster of Paris; … (b) a composition of lime, sand, etc., for plastering walls or other surfaces…
On the other hand, senses shown at separate levels by DMF could be presented in MED in a single-level entry. e.g.:
haras (DMF):
A. Ensemble d’étalons et de juments destinés à la reproduction ... ...
B. Lieu destiné à la reproduction ou au dressage des chevaux, haras
haras (MED)
1 (a) A place where horses are bred or kept, a stud; also, a collection of breeding horses.
30Haras was taken as polysemous, since both DMF’s ‘collection of horses’ sense and the ‘place’ sense were found among the MED definitions. Since different dictionaries were compiled along different principles, as discussed above, it was taken that an item was polysemous if the senses shown separately in DMF were distinguished in some way in MED, either by numerical or by small letter listing, as in the MED entry for plastre, or by a semi-colon, as in the MED entry for haras. Further examples of the procedure adopted were:
forgeur (DMF)
A. Celui qui forge (un objet en métal, des monnaies...)
B. Celui qui fabrique, qui crée qqc. (une chose abstr.)
forger (MED)
1. (a) One who works at a forge, a smith; (b) a carpenter, a wood carver; (c) an agent; (d) a creator; (e) an inventor of lies; …(f) a forger or counterfeiter of documents.
Both the metalworker and the creator sense, shown at different levels in DMF, are recorded in the MED entry, albeit listed in MED in a single level entry. That being so, ME forger was taken as polysemous.
31The case of Fr. giste posed a problem. In DMF it is considered to have a range of senses, including the following:
giste (DMF):
A. Endroit où l’on couche, où l’on réside temporairement ou habituellement… (‘place where one sleeps, where one lives temporarily or habitually’)
B. [Idée dominante d’horizontalité] (‘dominant sense of being horizontal’)
1. CHARPENT. Solive, poutrelle de plancher… (‘floorboard’)
In MED, both senses are present, but distributed across two headwords, giste, n.2. and giste, n.3:
giste, n.2 (MED)
1(a) A stopping or resting place, lodgings, hostel
giste, n.3 (MED)
1(a) One of the timbers supporting a floor, etc., a joist;
Whereas MED seems to treat them as homonyms, to judge from the different headword numbering, the French original considers giste was polysemous. The decision was taken to include ME giste, considering it polysemous, for reasons of consistency insofar as the analysis procedure followed the sense differentiation of DMF, not of MED, as in the cases of words such as haras and forger.
32Where only one of the senses of a polysemous French etymon was recorded in MED, the Middle English item was scored as monosemous, e.g.:
plastrier (DMF):
A. Ouvrier qui travaille avec le plâtre (‘worker working with plaster’)
B. [Valeur dépréciative] Homme grossier (celui qui ne sait que gâcher) (‘[derogatory force] Clumsyman (one who can only mess up)’)
plasterer (MED):
1. a) One who plasters the walls of buildings, a plasterer; (b) as surname.
Here the metaphorical sense of the French item is taken as not borrowed, leaving ME plasterer as monosemous. (Uses in names were ignored.) Likewise with Fr. plancher, ME plauncher:
plancher (DMF):
A. [Dans un bâtiment] (‘[in a building]’)
1. Assemblage de solives recouvert de planches séparant les étages d’une construction; face supérieure de cet assemblage
2. Étage (‘storey’)
3. Face inférieure d’un plancher, formant le plafond d’un appartement (‘underside of a floor forming the ceiling of an apartment’)
B. [Dans un moulin] Pan de charpente horizontal (‘horizontal part of timber frame’)
C. Planches formant le dessus d’une table (‘planks forming the top of a table’)
plauncher (MED)
1. (a) A plank, beam; also, planking, beams; (b) a wooden floor, floor made of planks.
This time several senses are offered for the French etymon, only one of which – ‘wooden floor, floor made of planks’ – is found in the MED entry. ME plauncher did develop a further sense – ‘plank, beam’ – but this is an extension not found in the French original, so the Middle English item was taken as monosemous.
33The following items were polysemous in Anglo-Norman with the same technical sense as in Middle English, but in DMF they lacked that sense (the second sense shown in 3(a-b) below):
carpenterie (AND):
1. woodwork, wooden part:
2. carpenter’s work:
roule (AND):
1. Rolled up material, small roll or tube
…
4. Pulley: rolle
These were originally taken as monosemous, but their polysemous status in Anglo-Norman could have arisen by polysemy copying from Middle English, and they were therefore omitted.
34Altogether, it was found that some 180 items from the three target domains were borrowed in Middle English, 69 of them monosemous in continental French and about which nothing further will be said here. The 111 items polysemous in continental French were first attested in the main in the post-1350 period, as can be seen in Table 1.
Table 1. First attestation dates of polysemous French-origin loanwords
|
pre-1300
|
1300-1349
|
1350-1399
|
1400-
|
First attestation date
|
15
|
28
|
29
|
39
|
35When this distribution is compared with first attestation dates of French-origin loans in Durkin [2014: 258], based on the third edition of the Oxford English dictionary (OED), it can be observed that, in our data, a rather higher proportion than in OED were first attested from 1400 onwards. This may be to do with the lack of source material relating to these occupational domains prior to that date. We return below to the issue of interpreting first attestation dates.
36Quantitative results of the polysemy analysis by occupational domain are shown in Table 2.
Table 2. Number of loanwords polysemous in French attested in MED
|
Building
|
Manufacturing
|
Farming
|
Total
|
Polysemous in French & ME
|
20
|
20
|
30
|
70 (63.1%)
|
Polysemous in French only
|
18
|
9
|
14
|
41 (36.9%)
|
Total
|
38
|
29
|
44
|
111
|
- 9 Lusignan [2011: 34] found that 86% of the 669 texts or manuscripts in Anglo-Norman referenced in th (...)
37Of the 41 polysemous French items borrowed into English as monosemous; however, it was noticed that in 15 cases the unborrowed sense(s) were first attested in continental French in citations dated to 1350 or later. The second half of the fourteenth century was a period when the bilingual context for borrowing loans from French showed signs of being severely weakened (Ingham [2011], [2012]), and by the early fifteenth century insular French was dropping out of use in nearly every aspect of public life in England.9 It would be unsound to assume, therefore, that contact with French remained rich enough to afford the same opportunities for new word senses to be borrowed in Middle English as had been the case prior to the Black Death.
38If the 15 items mentioned are accordingly removed, the numerical outcome is that 70 out of 96 French polysemes attested prior to 1350 were polysemous in Middle English in the same senses. The percentage figure then reaches 73%, not far short of the 77% level in the general sample of French polysemes obtained by Ingham [2021].
39Either way, it is clear that a reduction in polysemous borrowing as compared with Ingham [2021] is observable in specialized domains, as was predicted. That said, the difference is not particularly great, especially when only loans are counted whose French etymons were definitely polysemous before 1350. It must be acknowledged, then, that whether the general or the specialized loan lexis of Middle English is studied, a much higher proportion of polysemy replicating the French source item is found than would be the case in modern contexts for borrowing (see Section 1). French-origin words in the three large and representative occupational domains targeted in this research were overwhelmingly not borrowed in a single sense. They showed at least two meanings reflecting the source language, e.g. (a précis of the senses is given, rather than their full definition in MED):
(4) assai: to test, to attempt
attempren: to mix, to moderate
barre: long piece of wood, barrier
compassen: to shape, to go around
creste: top part of helmet, ridge of ploughed field
culter: ploughshare, knife
curtilage: garden, gardening
enhauncen: to raise, to become arrogant
forgen: to create, to falsify
punchoun: pointed tool, dagger
scantiloun: sample of cloth serving to show its quality, one of a pile of logs exemplifying their size
somer: beast of burden, load-bearing piece of timber construction
verrer: maker of glass items, someone who fits glass
vessel: ship, container
40Although these words all exhibit related pairs of meanings (they exhibit polysemy, not homonymy), the relationships seem to be quite idiosyncratic. Importantly, none seems to show regular polysemy whereby two senses arise thanks to a recurrent type of relationship (Apresjan [1974]; Pustejovsky [1995]), such as is seen in the following examples:
maser }:
|
material / product alternation
|
verre }
|
|
bef }:
|
animal / meat alternation
|
vel }
|
|
haras }:
|
location / collection of animals kept there
|
pultrie }
|
|
- 10 This hypothesis was informally taken up with linguistically trained native speakers of Japanese and (...)
41The lexical items in (4), by contrast, seem to possess extended senses where no similar regularity is evident, and for which it is unlikely that any given language would strongly tend to have one lexeme with both meanings. In the great majority of polysemous cases studied in this research, the likelihood that polysemy arose by sense extension within Middle English, independently of source language influence, does not seem compelling.10
42Before concluding, it is worth returning to the point made above (see Table 1) that from around 1350 the bilingual context for borrowing showed signs of being in terminal decline. In fact, the great majority of the French loanwords investigated were first attested from 1350 onwards, and many of them not till the 15th century. There might at first sight seem to be a contradiction here, but a timelag needs to be recognized between the period when loanwords were borrowed in spoken use, and the dates where they surface in extant written texts, a point which is hardly controversial. The time-period of regular contact between the languages, which gave the opportunities for borrowing, preceded the first textual attestations of a great many loans, probably quite substantially. First attestation dates, as in lexicology generally, are no more than a terminus ante quem.
43The research reported here aimed to identify semantic outcomes when lexemes polysemous in French with one sense relating to certain occupational domains were borrowed into Middle English. The procedure adopted was designed to obtain an estimate of how frequently the other senses were also borrowed. It was found that the great majority of the time Middle English loans from French shared at least two of the senses of the polysemous etymons. When senses of French etymons first recorded on the continent after 1350 were discounted, sense sharing between French and Middle English polysemes reached over 70%. The likelihood that sense extension within Middle English, mirroring French polysemy but independent of it, has been considered but provisionally rejected as most cases fell outside recognized types of regular polysemy. Future research making cross-linguistic comparisons of lexical polysemy, taking further our informal comparison referred to in footnote 9, may make this conclusion more firmly supported.
44Bearing in mind that the target vocabulary involved lexis with domain-specific senses, not the general vocabulary of the language, the prominence given to cultural borrowing as a mechanism for the creation of French loanwords in Middle English evidently needs to be reconsidered. Middle English loanwords from French were polysemous far too often for that account to suffice, given what we know from modern-day lexical borrowing about the rarity of polysemy borrowed with loanwords. It was found that dozens of lexical items passed from French into English denoting concepts where prestige could have played a considerable role. Yet they most often display polysemy, as with plastre, somer, coveren, jointure, assai, in the building domain; verre, verrer, attempren, temprure in manufacture; and graunge, curtilage and creste in the farming domain. In each case, the loanword was adopted with at least one sense outside the specialized domain in question, suggesting that the source word was used by a speech community that knew more than its technical sense.
45It is recognized that the impetus given to borrowing by the prestige of French was an important factor in the medieval period. However, both prestige and cognitive pressure borrowing mechanisms, according to Matras [2009], involve the agency of bilingual users of the source and recipient languages. While prestige would have been a motivation for bilinguals to draw on their French, it no doubt interacted with cognitive pressure, both facilitating the adoption of loanwords. It seems we should, as with Ingham [2021] for general vocabulary, attribute a very sizable influence to French-English bilinguals as agents of borrowing in specialized occupational domains as well. At least in the domain vocabulary examined in this research, monosemy was not, as in modern contexts, the norm with French-origin loan lexis at this time. Rather, French loanword adoption in Middle English on the whole took the same form as with French-origin words generally: the specialized sense sat alongside other senses in the vocabulary of the educated speech community.