The notion that the intelligibility of a language might sometimes be impaired by the presence of two or more terms having the same or very similar spelling and/or pronunciation was expounded by the linguistic geographers around the beginning of the twentieth century – the concept of the ‘homonymic clash’. The linguistic geographers were working on modern dialectal material, but their idea of the homonymic clash has been applied on occasion to historical linguistics also.² The aim of the present article is to draw attention to the skilful use made of homonyms and near-homonyms as a major aid in the teaching of French some six and a half centuries before the time of the linguistic geographers, and to demonstrate how the work of a medieval Englishman can sometimes add to and even correct the work of modern lexicographers.

The example quoted above is one of some fifty sets of homonyms or near-homonyms used by Walter de Bibbesworth in the eleven hundred or so verses of his Tretiz composed to teach French to anglophone landowners around the middle of the thirteenth century in order to help them manage their estates ‘in French’. A printed version of the text has been available for nearly a hundred and fifty years now,³ and a critical edition was produced some sixty years ago,⁴ yet, whilst quotations from the work have been used – although not systematically – by Godefroy, Tobler-Lommatzsch and the Anglo-Norman Dictionary, little attempt has so far been made to treat in detail the considerable non-literary vocabulary in the text or, in particular, to tackle the homonyms which were such an important aspect of ^p258 Bibbesworth’s teaching method.⁵ It would be impossible to treat all the homonyms in the space of a single article, but a variety of examples will be examined in order to show their different levels of complexity and the diverse areas of vocabulary from which they are drawn, thus providing an indication of the range of Bibbesworth’s acquired knowledge of French,⁶ as well as comparing his material with that given in the current dictionaries.

The play on words in the verses given above opens by bringing together the reflexes of the Latin regina and rana, which co-existed in Medieval French as reine until near the end of the fifteenth century.⁷ Then for rei Bibbesworth uses the homonymic clash between regemrei and retem > rei which was resolved in favour of the royal word by the loss of rei meaning ‘net’ and its replacement by filet after the end of the medieval period.⁸ He further demonstrates his command of medieval French by introducing desrei, a negative noun in frequent use, the opposite of arrai ‘order’, ‘arrangement’. The whole play on words is rounded off by the near-homonyms reoun ‘furrow’ and reume ‘realm’. For a thirteenth-century Englishman to be able to bring together such semantically disparate terms in French, without the benefit of dictionaries, and use them as a teaching tool implies a very thorough knowledge of his acquired second language.

Many of the homonyms used in the Tretiz present no difficulty, and so may be dealt with by brief mention of a small varied selection. Temple with its religious and anatomical meanings is obvious , poun as ‘peacock’ and ‘pawn’ in chess , littere in the senses of ‘bedding (of straw)’, ‘straw’ and ‘bed/couch in which a person is carried’ , arable as noun ‘maple’ and as adjective ‘arable’ – these and other such pairs call for no comment. Houce (M.E. holintre), (modern French houx) and houce (M.E. tabart) ‘tabard’, ‘coat’ likewise form a straightforward pair, only a minor change of spelling being involved. The same applies to near-homonyms such as greve (‘parting in the hair’) and grive (‘fieldfare’) , rupie ‘maldrop’ (i.e. ‘drop on the end of the nose’) and rubie (‘ruby’) , to the group made up of livre (‘book’ and ‘pound’), levre (‘lip’) and lievre (‘hare’) ,⁹ to the bale group¹⁰ and to a number of others. Although many of the words brought together in this way are not true homonyms according to the nineteenth-century ‘laws’ governing the historical development of French, this argument cannot be adduced to dismiss Bibbesworth’s groups as being no more than insular vagaries, the product of ignorance, since the dictionaries show that spelling in medieval French in general was varied in the extreme. This may be seen, for example, in Godefroy’s continental forms hos, hus, hours and ous for the modern French houx (sub hous 9.771b) mentioned above, showing that Bibbesworth’s houce is far from outlandish. Further evidence of this important point will emerge as more complicated groups of homonyms are examined (see Note 43). The ^p259 dictionaries of Godefroy to a limited extent and Tobler-Lommatzsch as a matter of general principle create a false impression of tidy conformity to ‘rule’ by not listing in their lemmas all the forms found in the material quoted in the articles below the lemmas. Only a close examination of the entries themselves will reveal the variety of forms actually used in medieval French.

In gathering together a large number of homonymic groups in French Bibbesworth displays a quite remarkable range of active vocabulary for an Englishman and, more importantly, a familiarity with non-literary registers that were only rarely written down. Take, for instance, the apparently simple juxtaposition of vanele/wanele (‘lapwing’) and venelle (‘alley’, ‘lane’).¹¹ Godefroy (sub vanneau, 10.829b) is able to provide just one example of this word before the sixteenth century, taken from the Roman de la Violette and written only two or three decades before Bibbesworth. Similarly, to bring together ventrer (‘belly-band’) and ventrere (‘midwife’) seems obvious enough,¹² but Bibbesworth’s use of ventrer as applied to horses antedates Godefroy’s earliest attestation (10.842c) by well over half a century, whilst T-L (11.211) can quote only this Bibbesworth example. A study of the vocabulary of the Tretiz as a whole, including its many variants in the surviving manuscript versions, would doubtless reveal other similar lexicographical information.

Bibbesworth’s range of knowledge is shown again when he sets down no less than four meanings for rai – ‘spoke’, ‘sunbeam’, ‘ray (fish)’ and ‘striped cloth’.¹³ Any suggestion that only his insular inability to appreciate the concept of gender enables him to include in this list of masculine nouns the reflex of the Latin feminine raia ‘ray-fish’ is refuted by Godefroy’s masculine forms rais/raiz in this sense from continental sources (6.558c; cf. raie 10.475a). Incidentally, Godefroy’s 2. rai (6.558c), treated as a separate word and incorrectly glossed ‘rayure’ instead of ‘étoffe rayée’, ought to be included under his 1. rai (6.558c). His sole example of this sense is taken from an Anglo-French statute of Richard II, well after Bibbesworth’s time, so the present example from the Tretiz ought to take chronological precedence. Also, Bibbesworth might usefully have added to his list of homonyms the rei forms mentioned at the start of this article (‘king’, ‘net’), since they are not always separated in spelling from the rai group.

A more unusual set of homonyms is formed from two verbs and a noun: Jeo oy toner (M.E. thonner) ; veir il toune (M.E. thondres) , Dunt la cerveise empire en toune (M.E. toune) . Ore me suffrez, mon pee toune (M.E. slepeth) .¹⁴ The common French verb tonner ‘to thunder’ needs no explanation, but the presence of the Middle English noun ‘toune’ in a French phrase shows how closely the two vernaculars interacted in medieval England.¹⁵ The third element – mon pee toune – is an example of the Anglo-French fondness for aphetic verb-forms. The infinitive is given by ^p260 Godefroy as entomir (3.267c-268a), but his list of variants includes entonnir. Bibbesworth omits the prefix in this instance in order to make his set of homonyms and elsewhere uses the past participle of the verb as an adjective, but with a prefix not listed by Godefroy: ‘Pur le destreit de l’yver Dunt avez la mein si estomie (M.E. comeled) Ki pur le freit ne purrez mie De deis fere la capinole’.¹⁶ This use of the ‘wrong’ prefix es- instead of en- with this verb is found in both T -L (3.1406) and the FEW (17.384a), and so cannot be attributed to insular ignorance of ‘correct’ French. It is just another example amongst thousands proving that, although the forms of Old French prefixes are in the main clearly derived from Latin, the use to which they are put in Old French cannot be automatically inferred from their role in Latin.¹⁷

The Tretiz, however, contains more interesting cases such as fusil, given as ‘steel, fire-iron’ and ‘spindle’. The difficulty here is that Bibbesworth gives three senses (‘spindle for spinning’, ‘mill-spindle’ and ‘steel for making a spark’) for the two objects represented by the modern French fuseau and fusil: De treys choses seert le fusil (M.E. spinnel): Le fil est filé du fusil, E le fu de cailloun (M.E. flint) fert le fusil (M.E. virhirne), E le blé est molu par le fusil (M.E. mulne spinel) . The fact that the one spelling fusil is used for all these meanings cannot, however, be interpreted as a mark of Bibbesworth’s English ignorance of the ‘correct’ French forms. As was mentioned above, continental medieval French did not separate its meanings neatly by spelling any more than insular French did, and Bibbesworth is supported here by Godefroy, who gives fuseil as ‘fuseau’ (4.185c), fuisel (plus ten other variant spellings, including fusil) as ‘petit instrument qui sert à tordre et à enrouler le fil, le fuseau moderne’ (4.177b-c) and fusil a second time as ‘pierre à fusil’ under foisil (4.45c). Both Bibbesworth and Godefroy show here yet again the untidy reality of medieval French orthography as opposed to the false tidiness found in the manuals of historical French grammar.

Later in the Tretiz, when listing the parts of a cart, the mention of axles leads Bibbesworth’s mind to the hub of the wheel – moail – and its close resemblance to mouwel, the yolk of the egg.¹⁸, moal (M.E. nave), moel (M.E. yoolk) (C) 11va, moyaus (M.E. naves), moyal (M.E. nave), mouhel (M.E. yelke) (T) 130v, etc. The quotations given by Godefroy (sub 2. moiel 5.357a, moieuf 5.360b and moieul 10.162b) and by Tobler-Lommatzsch (sub moiuel ‘Eidotter’ 6.171 and moiuel ‘Radnabe’ 6.172) show how interchangeable were the forms used for these two meanings. Given that the ‘hub’ and the ‘yolk’ are at the centre of the wheel and of the egg respectively, familiarity with Bibbesworth’s text might have encouraged Godefroy and T-L to treat their material as having a common origin, as is done in the AND sub muel, rather than make separate entries for each sense.

In another case Bibbesworth and the Middle English glosses to his text may be shown to be not only correct, but also to provide new information to the dictionaries. After pairing the noun lesche (M.E. sziver; i.e. ‘sliver’) ^p261 and the verbal form lesche (M.E. liketh) , he goes on: ‘Ore donez (sc. the pan) a chael a flater (M.E. lappen) Ki lesche la rosé (M.E. dewe) de l’herber. Eschuez flatour ki seet flater (M.E. losenge) E les genz espeluper (M.E. glonden)‘ . The accuracy of flater (‘to lap’) in this quotation is confirmed by two examples in Bozon’s Anglo-French Contes Moralisés referring to a fox lapping milk.¹⁹ In the main body of his dictionary Godefroy lists flater only in the sense of ‘to throw’ (4.26-27), but in the Complément (9.625b), after an example juxtaposing flater and losengier, he gives a quotation from Ronsard containing the phrase flata sa joue, which he translates as ‘caresser’. This stroking action is that of the whelp in Bibbesworth’s Tretiz lapping/licking clean the pan and of the fox in Bozon’s Conte lapping up milk. Although the dictionaries attest the abstract, moral sense of flatter centuries earlier than the concrete stroking by hand or tongue, this insular evidence shows that the semantic development was as normal – from concrete to abstract, not vice versa.

A more unusual addition to the dictionaries of medieval French could be made from Bibbesworth’s near-homonyms varole, verole and virole.²⁰ Virole and verole are, as in modern French, a ‘ferrule’ and ‘pox’ respectively, but varole is unknown. The French text explains that the varole grows on cabbages, and it is glossed in Middle English as ‘wratte or wort worm’ and ‘a grene worte worm’, so we must be dealing with the caterpillar of the cabbage-white butterfly.²¹ Godefroy has forms veret (8.188b) and verot (8.200b), which he glosses ‘petit ver’. In support of his single attestation of verot he refers to the same meaning ‘worm’ being used for the term in the modern dialects of the Centre and Normandy. It seems, therefore, that Bibbesworth’s varole may be a genuine French term, rather than merely an invention, as might at first sight be suspected. In fact, he could have added to this series another near-homonym found elsewhere in his Tretiz – virel or virol: ‘ E de un conoil (M.E. rocche, i.e. ‘distaff’) vous purveez, Mes le fusil ja ne ubliez, E le virel (M.E. werve, ed. werne) ki a ceo suffist ‘ . The virel is the ‘wharve’, the whorl or fly-wheel of the spindle, both the French form and the Middle English gloss being supported by similar forms in other manuscripts.²², also meaning ‘whorl (of spindle)’ (see AND sub vertoile) Although Godefroy does not record virol with this sense, in his second quotation under virelle (8.257a) the plural form virelles is glossed vaguely as ‘Sorte d’ustensile’, and the context would suggest a whisk for mixing mustard. This quotation has no connection with his first example meaning ‘ferrule’, and the two should be given separate entries. The utensil for mixing mustard is not far removed from Bibbesworth’s ‘whorl’, but the Tretiz antedates Godefroy’s quotation by close on two centuries. This not only shows Bibbesworth’s remarkable command of French but, more importantly, is an indication of the knowledge about the real medieval world outside the pages of romance waiting to be gleaned from unglamorous glossarial material, nominalia and the like.

^p262 Similarly, Bibbesworth’s juxtaposition of the verb form jaroile ‘quacks’ and the noun garoile ‘trap’²³ could enrich the dictionaries of both Godefroy and Tobler-Lommatzsch. Godefroy’s earliest quotation for jargoillier (4.636b) comes from the 14th century, whilst Bibbesworth was writing around the middle of the 13th; moreover, Godefroy’s glosses ‘gazouiller, murmurer’ are hardly adequate for the harsh quacking of the duck. Tobler-Lommatzsch (4.1586) offer no information on the subject, merely referring the reader to Godefroy . As far as the noun is concerned, Godefroy gives no meaning for his single quotation under garoil (4.236c) taken from an Oxford manuscript, whilst Tobler-Lommatzsch (4.195) translate the term as ‘Banner’, accepting Vising’s interpretation in Studia Neophilologica 15 (1942-43) p. 203. Vising is following the late Femina,²⁴ but at least four earlier manuscripts of Bibbesworth’s text gloss the word as ‘trap’, which would also fit the Godefroy quotation,²⁵ whilst ‘banner’ definitely would not.

Bibbesworth’s play on the form louche as noun and verb can be made to provide information to help fill out a strangely thin entry for lisser in the dictionaries:’E mettez vostre veille rouche (M.E. hivve) Desuz vos pos, noun pas la louche (M.E. ladil). Priez Jonet ki ta coyfe i leuche (M.E. szhike) De un lucchere (M.E. szhikinston) suiz la louche (M.E. ladil)‘.²⁶ The other manuscripts, as might be expected, show variations at this difficult point, with ‘Johnny’ becoming female or a ‘lady’ and the sleek-stone being not ‘under the ladle’ but, more plausibly, ‘on (top of) the chest’ (la huche).²⁷ The general consensus, however, is for texts and glosses to support the homonymity of the noun l(o)uche ‘ladle’ and the verb l(e)ucher ‘to smooth’. Godefroy (10.87b) has no attestation of lisser before 1564, whilst T-L (5.741 sub luschier, lischier) can only offer its readers this Bibbesworth evidence, and then only from one manuscript.

Another of Bibbesworth’s pairs of near-homonyms serves not to add to Godefroy , but to correct one of his entries: Mes war ki la chouue (M.E. the co) Ne touche vostre jouwe (M.E. cheke), (‘But take care that the chough/crow does not touch your cheek’) .²⁸ The Middle English glosses ‘co’ and ‘crow’ contradict Godefroy’s translation of choe (1.126b) by ‘chouette’. Not only are the Bibbesworth glosses supported by other insular material,²⁹ but three of Godefroy’s own continental quotations associate the bird with the colour ‘black’, a difficult combination if choe means ‘chouette’.³⁰

A more difficult problem arises when the Bibbesworth manuscripts use an unusual term kakenole in the section devoted to the parts of the body, with Middle English glosses in several of them translating the word as ‘the rime/reme of the brain’.³¹ The homonym kakenole two verses below is not glossed in the base manuscript at this point, but is translated later in the text as ‘a cake of spices’ . In other manuscripts this kakenole ‘cake’ is replaced by forms of brachole.³² The link between the ‘cake’ form and the ^p263 “/>anatomical kakenole in the base manuscript would seem to be one of texture, the spongy matter at the back of the brain being not unlike that of a cake (cf. English ‘sponge-cake’ and Godefroy sub espongne 3.546a). Bibbesworth’s mention of this homonymic pair makes it possible to draw together a number of apparently random strands quite unconnected with his Tretiz. Under chachevel (2.30a) Godefroy gives variants kachevel and quaquevel, with examples from the Anglo-French Li Quatre Livre des Reis and the fourteenth-century Évreux manuscript of the Dyaloge Saint Gregore, translating the word as ‘crâne’. Elsewhere, under kakenole (4.681b), he simply quotes the Wright version of Bibbesworth, with a question-mark instead of a translation, whilst under caquerole (1.781b) he gives late quotations from Oudin and Rabelais with the gloss ‘coquille, ou limaçon de mer’. Clearly, he has not made the link between his chachevel, kakenole and caquerole. Given that it is often impossible to distinguish between a scribe’s n and his u/v, and in view of the general interchangeability of c-, ch-, k- and q(u)- in medieval French, together with the fact that all these terms are similar in their general sense, it appears likely that we are dealing here with variants of a single word. Paul Meyer came across cakenole in an Anglo-French Manière de Langage of 1415³³ and saw that it might be the same as caquerole, but did not bring chachevel into the equation. Tobler-Lommatzsch (5.5) under kakenole give the glosses ‘Hinterkopf, Schädel’ and ‘Kuchen’, without exploring the matter further. Pulling all the available evidence together, it may be postulated that medieval French, both continental and insular, had a term kakenole or, more probably, kakevole (cf. DMLBS caccabus ‘metal pot, cauldron’) in several spellings, meaning ‘(back of the) skull’ or ‘the spongy matter at the back of the brain’.

In a different area of language comes the complicated coin group. From the ‘quince’ listed amongst the trees in the orchard – li coigner (M.E. quinse tre) – the play on homonyms leads us into axes, wedges, stocks and money: Les sires (M.E. lordes) font son neif (M.E. bondman) coigner (M.E. to sitte) En ces ceps (M.E. his stokis) pour chastier (M.E. chastying). Et pernez le coigne (M.E. the axe), abatez (M.E. fell) le coyner (M.E. the quinstre) Et copoe (M.E. hewe) de coigne (M.E. axe) un coigner (M.E. a wegge). E coingner (M.E. coigner) est il ensement Que fait la monee (M.E. mone) d’argent (M.E. of silver).³⁴ The starting-point for all this is the homonymity of the French reflexes of Latin cotoneum (Old French coin, modern French coing, ‘quince’), and cuneus (Old and modern French coin, etc., ‘wedge’, ‘corner’, etc.). The spelling of the term for ‘quince’ is now slightly altered in modern French, but the homophones remain unchanged. The reflex of cotoneum underwent no semantic development and so is of little interest here, but a close study of the Bibbesworth text at this point could have helped Godefroy and Tobler-Lommatzsch to present a fuller and more coherent picture of the development of the Latin cuneus in French.

^p264 In his original entry for coin (2.173b) Godefroy gives only one quotation, and this carries the later derived sense of ‘droit de frapper la monnaie’; in the Complément he first glosses 1. coin as ‘corps solide terminé en angle; extrémité d’un corps solide terminé en angle; pointe, sommet’ (9.120c-121b). The senses of in his illustrative quotations under this heading are a mixture of ‘corner’ and ‘point’, with the exception of two late examples which clearly belong not here but under his second heading ‘T[erme] de monnaie, morceau de fer trempé et gravé qui sert à marquer les monnaies et les médailles’.³⁵ Elsewhere he gives coigne as a separate lemma, although it is obviously a form of coin, glossing it as ‘le coin dont on frappe la monnaie’ (2.172a). Moreover, the example he gives for this sense is insular, not continental, and means ‘currency’, not ‘die, stamp’.³⁶ His other quotation in this entry is also Anglo-French and confirms the sense of the first one.³⁷ Here the French of medieval England may be seen to be extending the meaning of coin further than the continental variety.

The French of the Bibbesworth manuscripts again diverges from the continental norm in its repeated use of coin glossed as ‘axe’. Godefroy makes a clear distinction between coin and coigniee (9.120b), which alone is glossed as ‘sorte de hache …’. At first sight it might appear that Bibbesworth and his copiers are using coin idiosyncratically, simply in order to create a further example of homonymity, but the DMLBS gives the following meanings (with examples) for cuneus: ‘wedge, chisel, stamp (for coining), coining iron, mint, coin, coinage; (partly conf[used] w[ith] I conus) corner’. The presence of ‘chisel’ here is reinforced by the glosses in TLL, showing that coin could mean ‘axe’ (as in Bibbesworth), securis being glossed by coyn (II, 105), cuin (II, 109), cuyn (II, 90), and secures by coyns (II, 43), alongside numerous examples of the expected coin(i)é, etc. This means that since cuneus could have the sense of a cutting tool (‘chisel’) in British Latin and since the British glossators used coin(g) etc. to render securis ‘axe’, the Bibbesworth texts and glosses must be considered correct for Anglo-French.

From the root coin we come to the derivative coi(g)ner as verb and noun. The Bibbesworth manuscripts use the verb coigner to indicate ‘putting (in the stocks)’, so providing examples that could strengthen Godefroy’s quotations with this sense that are taken from just one author. Coi(g)ner as noun is found as ‘coiner’, ‘minter’, a normal formation along the lines of bof/bover, charrette/charrett(i)er, charue/charruer, etc., but one that is not attested in connection with money in mainland France. The Middle English glosses ‘coiner’ and ‘menetere’ (i.e. ‘minter’) leave no doubt as to its correctness.³⁸ The word is also found in the sense of ‘wedge’, alongside coignee.³⁹ This whole development of the semantic field based on cuneus in medieval England shows, firstly, that Anglo-French needs to be taken into account in any etymological or semantic study of medieval French and, secondly, that it is not enough simply to compare Anglo-French with the ^p265 French of the mainland: Anglo-French and Anglo-Latin together formed the linguistic tool of the literate classes. Those who used the one were very likely to use also the other in their professional dealings, many administrative and business records being drawn up in a mixture of all three languages, Latin, Anglo-French and Middle English, each of which affected the other two.⁴⁰

Not only Godefroy , but even the prestigious FEW has something to learn from the neglected efforts of a thirteenth-century Englishman to teach French to his compatriots. A number of the Bibbesworth manuscripts list a group of near-homonyms that are not found together elsewhere. They distinguish between naer (‘to swim’), noer (‘to drown’), nager (‘to sail/row’) and negger (‘to snow’). By the end of the medieval period nager would have replaced naer in the sense of ‘to swim’ and naviguer in turn would have taken over the meaning ‘to sail’ from nager.⁴¹ This change is explained in the FEW as follows: ‘Im mfr. zeit fängt man an, nager für ‘schwimmen’ zu gebrauchen, weil im fr. die vertreter von *NOTARE und NODARE homonym werden’ (7.65). This explanation is hard to swallow, because the postulated homonymic clash on which it is based had apparently been easily tolerated for centuries, ever since Latin intervocalic d and t began to be confused in French, yet, so we are asked to believe, at some time in the fourteenth century it was suddenly regarded as an impediment to comprehension. The FEW does not, unfortunately, provide any indication of the precise circumstances in which a hearer would wrongly imagine that his interlocutor was talking about swimming when in reality he was talking about tying. There is indeed a homonymic clash here, but not the one claimed by the FEW. Bibbesworth hit the mark over seven centuries ago: the genuine clash is between the reflexes of natare and necare (given in the Tretiz as naer and noer), with nodare (> nouer) not even entering the equation. The fourteenth-century manuscripts of Bibbesworth help their readers to distinguish the various forms by separating naer and noer along etymological lines, but the form with a is confined to the Tretiz: in French as used in the real world, both insular and continental, noer is found with the meaning ‘to swim’ as well as ‘to drown’. This state of affairs is also reflected in the fifteenth-century Femina: ‘Ensi avoms noer, nager & neger (M.E. Also we havyth to drowne, swymme, rowe & snowe)‘ (p. 55). With four Middle English glosses for only three Anglo-French verbs, noer here is intended to cover the senses of both swimming and drowning. Immediately below this the writer uses the form with a, but again the homonymity is complete for the senses of ‘to swim’ and ‘to drown’: ‘En mear naee le pisson, En mear est naee mult prodom (M.E. In zee swymmeth the fyshh, In zee ys drowned many goudman)‘.

A close scrutiny of entries in the dictionaries of medieval French reveals ample proof of this claim that the real homonymic clash is between the reflexes of natare and necare. Nowhere in Godefroy’s lengthy entry under ^p266 noer, nouer, nuer, noier ‘nager’ (5.511a-b) or the even lengthier one in Tobler-Lommatzsch noer, naer ‘schwimmen’ (6.696-99) is there a single instance of this verb being used where the context of water is not immediately obvious or where there is any possibility of the verb being interpreted as ‘to tie, knot’. Conversely, the entries for noer ‘attacher, fixer, en attachant un noeud’ (G.10.205b-c) and noer ‘knüpfen’ (T-L 6.699-701) contain nothing that could be misconstrued as referring to water. On the other hand, in T-L’s entry for noiier, niier (necare) (6.710-13) several quotations use forms which are not given in the lemma as being products of necare: e.g. Ne la puet mer naier (711.10-11); El le feist ardoir ou en la mer naier (711.13-14); salive d’omme jeun naie le basilique (711.45-6); tote l’ost d’iaue naiot (713.40). None of these carries the sense of natare, despite the spelling. The rescue of Moses from the bull-rushes by Pharoah’s daughter is expressed as follows: ele l’aveit gari de neger, et trait de l’aigue ou il esteit gitez, par le comandement au rei Pharaon, qui aveit comandé que tuit li enfant malle … fussent nee (6.711.4-8). In this quotation the meaning of neger and the past participle nee, both derived from necare in its restricted sense of ‘to kill by drowning’, is confirmed by the Biblical text (Exodus I, v. 22).

Another of T-L’s quotations and one from the AND, both given as ‘to swim’, reveal just how difficult it can be on occasion to say with certainty which meaning is intended – ‘to swim’ or ‘to drown’. In a passage from the Roman de Renart the wolf has been chased into deep mud by the peasants: ‘Le vilein trovent en la boe Grant et parfonde si qu’il noe; Fors l’en ont tret a moult grant paine” (T-L sub noer ‘schwimmen’ 6.697, 24-26). The hunted wolf was surely drowning in the deep mud, not swimming in it. Less clearly, the Vie de Saint Gregoire tells how in a great flood ‘Les hautes tours veissez neier E les vielles meisons noer’ (AND sub noer¹ ‘to swim’); this probably means that the old buildings were drowned/submerged (neier necare) and only the tall towers were ‘floating’ (noer natare) above the water. However, it is the context, with the two verbs being opposed, not any phonetic or orthographical feature, that decides the meaning. The product of necare in French, then, whether spelt as naier, neer or noer, is liable to be confused with that of natare whether spelt as naer, neier or noer. That this jumble of forms is not just the blind result of a phonetic accident is shown by other quotations in T-L: noer ferai (Latin: natare faciam) tute nuit mun lit, de mes lermes mun lit aruserai (699.15-17). This is given under noer, naer (natare), but a quotation with the similar sense of ‘to flood’ is found under noiier, niier: sa grant glotenie En avarice con ( l. çon) cuer nie (711.36-7). The semantic blurring which allows this apparent confusion is to be seen already in Classical Latin, where Lewis and Short give as one of their glosses under nato ‘to swim or overflow with any thing, to be overflowed’, their examples coming from writers of the stature of Cicero and Virgil. In modern English a similar semantic closeness may be seen in ^p267 phrases such as ‘the fish and chips were swimming in vinegar’ and ‘the cathedral was flooded with light’. Too much reliance on phonetic factors alone as forces affecting the semantics of a language can be dangerous.

One of the most difficult sets of near-homonyms in the Tretiz is based on the axle of a cart. In the base manuscript are juxtaposed essel, assel and ascel.⁴² The glosses for the first and third of these terms – ‘axetre’ and ‘armeholle’ – show that we are dealing with the modern French essieu and aisselle. Other manuscripts support this, even if their spellings are sometimes sharply different.⁴³ The second of Bibbesworth’s three terms – ascel, with its Middle English gloss ‘clout’ – means ‘plate’, usually of wood or, as in the present context, of iron, fixed on the axle-tree to prevent wear (see OED sub clout I.2). The base manuscript does not make the passage any easier to understand by using esseaus for ‘axles’ and also for ‘plates’: ‘Le chartil lyoms sur les esseaus; En les moyeaus sunt les esseaus’ , but (C) makes the situation admirably clear: ‘Le chartyl (M.E. cartbody) lyoms sur le assel (M.E. cartclout) Et en le moal (M.E. the nave; i.e. ‘hub’) gist le essel (M.E. axetreo)‘ (f.11vb). (O) replaces the Middle English ‘cartclout’ by ‘lide gate’ (f. 338rb), a lidgate being literally a ‘swing-gate’, but here presumably signifying a non-rigid plate that would move with the movements of the cart. Notwithstanding the lack of uniformity in spelling, the reality of Bibbesworth’s three terms can be verified by reference to Godefroy as well as by the Middle English glosses: 1) aissel ‘essieu’ (1.199c), 2) aisselle ‘toute sorte d’ais ou de planche, aisseau … ‘ (1.199c-200b), aissil ‘ais, petite planche en forme de tuile …’ (1.201a-b) and 2. aisseau ‘petit ais ou planche très mince …’ (8 Comp. 64b), 3) aisselle ‘mod. aisselle’ (8 Comp. 64b-c). The spellings used in Godefroy’s entries for these words are as varied as those found in the Bibbesworth manuscripts, so that the work of the English teacher of French in this passage cannot be dismissed on either orthographical or semantic grounds.

However, Bibbesworth is not content to leave matters there: he proceeds to introduce another set of terms close to the essel group, but which are more difficult to explain coherently. His section on the vocabulary of the cart puts him in mind of the horse that pulls it and of its collar:’ Les couls dé chivaus portent esceles (M.E. hambrowes) E colers du quir en lur osseles (M.E. homes)‘ ,⁴⁴ and he returns to this theme later when dealing with the making of a fire: Si des osceles (M.E. hambors) du chival Facez asteles (M.E. szhides), vos frés mal .⁴⁵ On the face of it, we have once again three near-homonyms – esceles, osceles or osseles, and asteles. The last of these, asteles, is the easiest to deal with, being well attested in both insular and continental French in the sense of ‘stick’, ‘chip of wood’, a meaning confirmed by its Middle English gloss (see OED sub shide). The glosses ‘hambors’ and ‘homes’ mean, respectively, ‘horse-collar’ and ‘the two curved pieces which go to make up a horse-collar’ (see ^p268 OED sub hambargh and hame). The fact, however, that the base manuscript contains first esceles and then osceles with the gloss ‘hambrowes/hambors’, whilst osseles is also glossed as ‘homes’ (i.e. ‘hames’), suggests an understandable case of synecdoche, a semantic confusion between the whole of the horse-collar and its constituent parts. The variant manuscript readings at this point confirm this suspicion, the gloss ‘hambargh’ (under several spellings, of course) being applied to esceles (C), oceles (C), hosceles (C), hoceles (A),⁴⁶ and oeles ( l. oceles) (T), whilst ‘hames’ is given as the Middle English gloss for esteles (A), hesteles (T), osseles (C), osteoles (O). This apparent confusion is only to be expected, since it would be unrealistic to imagine that all those who later copied Bibbesworth’s text would be fully conversant with all the technical vocabulary it contains, especially since they would also have to face the difficulty of distinguishing between the very similar groups sc and st in their manuscript sources. It looks, therefore, as though an orthographic confusion could well be added here to the semantic one. Yet it would be unwise to dismiss all the forms as being perhaps no more than the product of Bibbesworth’s imperfect knowledge of French. No less than four of the manuscripts would suggest otherwise. The Oxford text gives a perfectly coherent reading: Et les colers de quir (M.E. leather) od lour ostooles (M.E. coddes)(f.338va). ‘Cod’ is a cushion or padding (OED sub cod²), so the meaning would be: ‘And the leather (horse-)collars with their padding/padded hames’. The readings in (A), (T) and (C) are less coherent, but all have a significant extra word bourle, which in one instance is glossed as ‘flockes’ i.e. ‘padding’ (used in the horse-collar).⁴⁷ The correctness of this word in the context of the horse-collar is supported by Godefroy’s bourel, borrel, boureau, ‘le collier, et en général tout le harnois d’une bête de somme” (1.704b-c) and by T-L’s ‘Kummet’ (i.e. ‘horse-collar’) for borrel (1.1076). There is definitely a genuine term for ‘horse-collar’, ‘hame’ amongst the various forms in the Bibbesworth manuscripts, so the next step is to determine which of them is to be regarded as the ‘correct’ one, as far as any notion of correctness is at all applicable to medieval French.

Under 1. eschielle (3.387c-88b), surrounded by many examples carrying the broad sense of ‘ladder’, Godefroy puts an incongruous set of four continental quotations from the thirteenth and fourteenth centuries which, as he says, mean ‘pillory’ (3.388b). The characteristic feature of the pillory is that the neck is encased in a collar not unlike the horse-collar,⁴⁸ so it would appear at least possible that this is the same word as Bibbesworth’s plural esceles, all the more so since the forms in Godefroy’s four examples have the spelling eschelle and esquele. Yet, bearing in mind the numerous forms with st rather than sc found in the Bibbesworth manuscripts, and the difficulty of distinguishing between them palaeographically, one is inevitably reminded of the modern French for ‘hame’ – attelle.⁴⁹ Attelle is the modern form of the medieval astele, the form we started with, so a possible ^p269 solution to the problem raised by Bibbesworth’s play on the near-homonyms esceles, osceles/osseles and asteles is that the first two are in reality variant spellings of the same word and represent part of the semantic field of astele, the third element. Under different spellings Bibbesworth appears to be using astele in its basic meaning of ‘stick’, etc. and also in its derived sense of the ‘wooden hame’ forming part of the horse-collar. Evidence from both continental and insular sources is available to support this suggestion. Although Godefroy gives ‘attelle, mod. v. astele‘ (10.233b), under astele itself (1.456c-457b) he provides no example of the modern sense of attelle ‘horse-collar’; he does have relevant evidence for this, however, under the diminutives astellet, without giving a translation,⁵⁰ and under astellette, translated as ‘partie du collier des chevaux, à laquelle les traits sont attachés’ (1.457c).⁵¹ These diminutive forms must indicate the existence of a similar sense for the standard form astele, even if this remained unknown to Godefroy . For Anglo-French TLL provides a significant gloss: ‘arquillas: hames, estels’ from a Worcester Cathedral manuscript of the thirteenth century (II, 154),⁵² whilst T-L gives ‘Kummet’ (i.e. ‘horse-collar’) for estele (3.1372). The suggestion is, then, that Bibbesworth’s three near-homonyms – escelles, osseles/osceles and asteles – are all forms of one basic term usually found as asteles and having the root meaning of ‘stick’, ‘chip of wood’, but also the derived sense of ‘stick used to make the wooden part of the horse-collar’ (i.e. ‘hame’) and so ‘horse-collar’ itself. The different spellings reflect principally the lack of a clear distinction between sc and st, which is a regular feature of many manuscripts, in addition to the normal random orthographical variations, including initial a- and e-, found all over medieval French. This attempted explanation of a tangled word-complex may well be defective or erroneous, but it would at least indicate that the Bibbesworth text at this point cannot be simply disregarded as no more than the invention of an Englishman out of his depth in the finer points of medieval French.

On occasion, the Tretiz can be shown to be wrong, as when a distinction is made between pail (M.E. chaf) and paille (M.E. stre; i.e. ‘straw’) ; nor is Bibbesworth’s text always as clear as could be wished. For example, his distinction between le char, la char and eschar poses a problem. His first statement separates the three clearly and simply:’Vos avez la char (M.E. fleyx) e le char (M.E. wayn), Mes cuntregardom de eschar (M.E. scorn)‘ . The difficulty comes a little later, when he attempts to explain the different meanings by setting the words in context:’ Eschar par folour hom revilist; La char par hidour en hom fremist (M.E. quakes). Jeo vi la char (M.E. fleyz) seer en char (M.E. wayn) E de la char fere eschar (M.E. scorn)‘ .⁵³ This last sentence is intelligible only if la char is taken in its theological sense of ‘someone’, ‘a human being’, but the current dictionaries of medieval French have no record of this, although the DMBLS lists ‘body, bodily existence’ amongst ^p270 the meanings of caro. The different manuscripts of the text offer no real help here, but the late Femina glosses char as ‘lady’, evidently taking the word to mean ‘human flesh’, i.e. ‘human being’, ‘person’.

In a few places the Tretiz appears to be both unclear and incorrect, as when Bibbesworth attempts to explain the series of near-homonyms espandre, espaundre, espendre, pandre and espandere.⁵⁴ Most of these are used correctly, but, leaving on one side the unrealistic separation of espandre and espaundre, only the second of which is glossed (twice, with the Middle English ‘scheden’ and ‘schedes’ giving the correct and unequivocal meaning ‘to shed, spill’), difficulties arise with pandre and espandere. As regards pandre, the text states that the fish paunde when it is caught live in the net. Other manuscripts have the same verb here, and the Middle English gloss ‘flakerers’ (i.e. ‘flaps about’; see OED sub flacker) in the base manuscript is supported by ‘flikeres’ in (O) and ‘bat’ in (C). Clearly, the copyists agree as to the sense of the word, but it does not seem to be attested elsewhere. Espandere raises a different question. The Middle English gloss ‘spele’ and the form espeau (present indicative, third person singular) given a few verses later would point to the attested infinitive espeudre (see AND sub espeleir), but this would not fit into Bibbesworth’s run of near-homonyms, nor would the emendation of espandere to espaudere which immediately springs to mind on account of the frequent identity of scribal n and u. This group of near-homonyms calls for further study.

By no means all the interesting homonyms in Bibbesworth’s Tretiz have been treated in the present article, nor is it claimed that the explanations offered here for some of the more difficult ones necessarily provide a definitive solution to the problems raised. What is important is to provide material to stimulate a comprehensive study of the non-literary registers of Medieval French that Bibbesworth attempts to teach and that were so essential for his contemporaries and successors in their daily living. The French found in the many saints’ lives or romances of chivalry could not possibly be adequate for the running of an estate in medieval England. The Tretiz, with its practical French vocabulary and its Middle English glosses, comes into the same category as the mass of glossarial works involving Latin, Anglo-French and Middle English brought into the public domain in recent years by Tony Hunt. The presence of Middle English in this material adds an extra dimension to the study of medieval French, one that can often provide the key to comprehension when dealing with difficult terminology, and one that is available only in conjunction with Anglo-French. Far from being no more than an awkward appendage to medieval French that may conveniently and thankfully be ignored on the grounds that it is unreliable and those who produced it incompetent, the whole corpus of Anglo-French didactic writing could make a major contribution to the next generation of dictionaries of both English and French.

W. ROTHWELL

---