William Rothwell (1998)

It has long been a cardinal tenet of medieval English scholarship that later Middle English owes the richness of its lexis to large-scale importations of French terminology from the Continent, especially in the fourteenth and fifteenth centuries. However, although the massive influence of French on the lexis of English from medieval times onwards is undeniable and has been repeatedly noted and analyzed by Anglicists, lexicographers and linguisticians over a period of many decades, the attribution of this flood of new vocabulary to Englishmen returning from France laden with linguistic baggage either in their heads, having been picked up in the course of conversations, or in their bags, acquired from books, has always been based on a convenient assumption rather than on concrete evidence derived from a close study of the medieval French used on each side of the Channel. The reasoning behind the assumption would appear to be that, if a new word is found in a Middle English text and a similar, or not too dissimilar, word is recorded for Continental Old French, then the French word must have been introduced into England from France as a consequence of the activities of the traveller. Cross-Channel traffic of a commercial, diplomatic, educational and also of a military kind was certainly not lacking in the later Middle Ages, so it may seem not unreasonable to link such contacts with the enrichment of English at that time and take it for granted that the French ‘borrowings’ found in such abundance in the Middle English lexicon are taken from ‘Old French’. Yet a closer look in their own backyard, so to speak, might have been sufficient to raise substantial doubts on this score in the minds of Middle English specialists, had they been in a position to draw on an adequate knowledge of Anglo-French.

Given the circumstances obtaining nearly a hundred years ago, it was to a certain extent understandable that scholars at that time should treat the French vocabulary found in their Middle English texts as being introduced piecemeal from mainland France over a period of some three centuries. Back in the days of Jespersen, writing on the very threshold of the twentieth century, and even of Baugh in the mid thirties, the Middle English Dictionary^{1 p145} and the Anglo-Norman Dictionary did not exist, the Altfranzösisches Wörterbuch and the Französisches Etymologisches Wörterbuch were no more than embryonic dreams for the future that would grow only slowly, and so the most that even Baugh could expect by way of large-scale reference works on the linguistic state of medieval England and France as late as 1935, when he wrote his article on ‘The Chronology of French Loan-Words in English’,² were the recently published Oxford English Dictionary and Godefroy’s late nineteenth-century Dictionnaire de l’ancienne Langue française. However, this excuse is no longer valid. Over the last forty years the steady publication of the voluminous MED has profoundly altered the whole landscape of English etymology, whilst the appearance of the first edition of the AND has made possible at least a preliminary understanding of the presence in medieval England of a French that often differs considerably both in form and, more importantly, in semantic content from what were until recently accepted as the norms of standard Central French. In years to come the picture will be greatly refined and clarified: on the Continent the long-term revision of the FEW and the on-going publication of the new Dictionnaire étymologique de l’ancien français in Heidelberg will doubtless throw much new light on the overall history of the lexis of medieval French; in England, the forthcoming second edition of the AND, although not pretending to provide a historical survey of the Anglo-French lexis, with first attestations given for all entries, will inevitably greatly increase the knowledge available. An eventual third edition of the Oxford English Dictionary, one that would for the first time take full account of all the etymological knowledge accumulated from many quarters since the days of James Murray, would greatly facilitate a more informed approach. Yet, even without waiting for future developments, enough evidence has been readily accessible for a long time now to invalidate some of the views still being currently propounded by Anglicists and linguisticians regarding the relationship of English and French in medieval England and to offer more reliable alternatives.

At the most elementary level, any observant traveller who passes to and fro between England and France has for many decades been given a clear reminder on every journey that the current wisdom regarding the provenance of the French element in English must be highly suspect. At Victoria and Waterloo in London large boards indicate ‘Arrivals’ and ‘Departures’, whilst at the Gare du Nord and Saint Lazare in Paris similar information is given under Arrivées and Départs: the relationship of the two sets of words is close, but not close enough to warrant the facile assumption of a direct borrowing from French to English. In fact, no assumption of any kind is necessary, because the etymological path taken by the English terms runs unbroken from Anglo-French to modern English, without any detour via Paris. Arivail, in various spellings, is attested for Anglo-French from the fourteenth century (AND1), but is not recorded for Continental French at all, so that Chaucer and Gower, credited with the introduction of this along with many other items of the new vocabulary from France at the end of the fourteenth century, must have been acquainted with ^p146 the term through their knowledge of the French of London, rather than the Parisian variety.³ Similarly, departure in its English form and with its English meaning is not recorded for Parisian French, but is found in two fifteenth-century legal texts in Anglo-French, used as it would be in modern English. ⁴ Godefroy (2.511a) lists departeure, but can provide only one isolated and undated example using the form desparteure, with the meaning ‘death’. Tobler-Lommatzsch simply refer to Godefroy, without even giving an example. Moreover, the earliest Middle English attestation of the word in 1441 antedates the Anglo-French. ‘Departure’ is in line with a number of other similar terms that developed semantically on English soil. In later Anglo-French records the suffix -ure is frequently used to denote an act (faisure ‘doing’, ‘making’, escripture ‘the act of writing’ as well as the document written, etc.), a grammatical feature linking -ure and ‘-ing’ that will be referred to again very briefly later in this article, one that will have to be treated at length elsewhere. The traveller might also reflect with some satisfaction on the fact that he is linked only morphologically to the travailleur, and that his journey may perhaps be undertaken with no thought for the passage of time associated with journée.

If he were versed in medieval languages, our traveller might also pause to wonder why he is required to go through the French-looking ‘Customs’ on the English side of the Channel, whilst in France he must pass through the douane. ‘Custom’ in the sense of a ‘customary payment’, a ‘duty’ paid to an authority – local (for markets, etc.), or royal (for the import of goods, etc.) – is found in British Latin only a few years after the Conquest (DMLBS sub custuma) and is widely used in Anglo-French in a variety of senses from the thirteenth century onwards, being recorded with the meaning ‘customs duty’ from early in the fourteenth century and eventually being applied to the organization charged with the collection of the dues, and also to the building in which the duty is paid. The term was doubtless brought over to England as part of the linguistic baggage of the Conqueror and has remained in use ever since, extending its semantic field as the needs of society developed and passing into Middle English in the fourteenth century (AND sub custume), with the compound ‘custom-house’ appearing late in the fifteenth. Incidentally, by the middle of the thirteenth century Anglo-French had derived from custume a verb custumer meaning ‘to pay customs duty’,⁵a coinage that would point towards a date well before the fourteenth century for the introduction of the sense ‘customs duty’ for the noun. From 1372, instead of ‘Customs duty’ the French have paid their droits de douane (cf. Spanish and Italian ^p147 aduana, dogana). The interesting point here is not so much the Eastern origin of douane, adopted in the later fourteenth century and replacing a similarly Eastern fondique in the sense of ‘bazar, warehouse’, hence ‘customs-house’,⁶ thus indicating the Mediterranean focus of a great deal of French trade at that time, so much as the differing perspective from which the payment is envisaged. The English ‘duty’ and ‘dues’ – purely Anglo-French formations which are not found on the Continent, although their connection with deveir/devoir is obvious⁷ – imply that the payment is viewed from the perspective of the payer – an obligation to pay what is due; the French droit views the matter from the perspective of the receiver – his right to the payment. That this is no mere flight of fancy is shown by the fact that, whilst Anglo-French and Continental French have many senses of dreit/droit in common, the meaning ‘duty payable’ is not attested for the Insular variety.⁸ This is just another demonstration of the elementary fact that recognizing the presence of a particular form in a language is of minimal significance or usefulness without a knowledge of its accompanying semantic content and development.

The fundamental problem in the study of English etymology is that, whilst the presence of French in medieval England is generally acknowledged, it is rare to find an Anglicist or linguistician familiar at first hand with actual Anglo-French material, as distinct from being acquainted with the views expressed on that material by scholars of the past who are still regarded as enduring authorities on the subject. For example, despite all the massive new dictionary evidence and all that has been written on and around the subject in recent years, the fourth edition of the influential History of the English Language by A. C. Baugh and T. Cable (London, 1994), widely used in universities on both sides of the Atlantic, continues to perpetuate a view of the situation of French in medieval England derived unmistakably from K. Lambley’s The Teaching and Cultivation of the French Language in England (Manchester, 1920) and J. Vising’s Anglo-Norman Language and Literature (London, 1923). Both the tenor of the text on p.137 and all the supporting footnotes are firmly rooted in the early nineteen-twenties.

This uncritical dependence on the views of past scholars leads not only to the acceptance and dissemination of opinions that have been seriously challenged in recent years, but also to the formulation of theories that plainly run counter to common sense. By subscribing to what appear to be definitive pronouncements made by ‘experts’ in a field outside his purview, without testing them against the available evidence, the Anglicist or linguistician is free to develop any theory based on no more than second-hand statements. For instance, although expressed in the terminology of modern linguistics, the views of ^p148 C-J. N. Bailey and K. Maroldt in their article ‘The French Lineage of English’⁹ are founded on information about the characteristics of Anglo-Norman set down by Vising three quarters of a century ago and adopted grosso modo by writers on Anglo-Norman ever since, without the sobering corrective of first-hand documentary evidence. Bailey and Maroldt attempt to explain what they term ‘creolization’ in medieval England as follows: ‘French creolization seems to have proceeded in two steps – a major creolization, before 1200, and a minor one (mainly with Central French) that involved massive borrowing during the thirteenth and fourteenth centuries. The second creolization was from ( sic) French! ( sic)‘ (p.27). The italics and exclamation mark in the last sentence can only mean that the writers believe that the first ‘creolization’ was into French, that, in other words, the character of the French brought into England at the time of the Conquest was transformed into a ‘creole’ by the impact of English, whilst, to their apparent surprise, in the thirteenth and fourteenth centuries a second ‘creolization’, running this time in the opposite direction, did the same for Middle English by an influx of French terms imported from mainland France. The linguistic terminology used here would need to be accurately defined and its validity as applied to the situation of medieval England convincingly illustrated by documentary proof on a large scale before any credence could be attached to such claims. Yet, writing nearly twenty years after Bailey and Maroldt, and dealing only with the linguistic state of England at the close of the twelfth century, not the later period, Juliette Dor not only espouses the idea of creolization at this time, but adds to it pidginisation: ‘Ce n’est pas du bilinguisme, mais une simplification du français et de ses dialectes, dont on a réduit la syntaxe, les terminaisons, les genres et accords grammaticaux: … ils ( sc. the population of England) étaient dans la situation type de pigdinisation et de créolisation’.¹⁰ The correctness of these assertions regarding pigdinisation and créolisation will be examined later, but it is highly important to note at this point that the existence of a standard French – le français, different from ses dialectes, is already being postulated for a period even before the beginning of the thirteenth century.¹¹

No concrete textual evidence of any kind is offered by these scholars in support of their claims, but it is not difficult to hazard a guess at the provenance of their general picture of the development of French in medieval England. The separation into a ‘Norman’ period covering the twelfth and early thirteenth centuries, followed by a ‘Central French’ period goes back to Vising’s book Anglo-Norman Language and Literature, first published in 1923 and reprinted without alteration as late as 1970. This, together with Miss Pope’s manual of 1934, From Latin to modern French, with its similar views and similar history of being reprinted much later (1952, 1958, 1973) without significant changes to its ^p149 intellectual stance or content, has been the unquestioned authority to which Anglicists and linguisticians have turned for their knowledge of Anglo-French all through the last seventy years. Vising’s division turns up again, not surprisingly, in the otherwise very sensible and balanced account of the growth and character of the Middle English lexis in the Cambridge History of the English Language,¹² where an attempt is made to distinguish by differences in spelling the earlier – Anglo-Norman – ‘borrowings’ from the later Central French ones: ‘French influence came from two separate dialects of French: firstly, from Norman, both as spoken and written language, and later, as an artificially acquired literary language, from the French of the Ile de France’ (p.426). This contention too will be examined in more detail later.

Moreover, as long as they remain untrammelled by any burden of first-hand textual evidence, there is nothing to prevent writers from propounding the exact opposite of these views. Running directly counter to the statements expressed above is the opinion of S.G. Thomason and T. Kaufman in their book Language Contact, Creolization and Genetic Linguistics (Berkeley: University of California Press, 1988), where, like Bailey and Maroldt before them, they do not encumber the reader’s mind with any form of textual proof, but simply assert that ‘1. There were never many speakers of French in England. 2. They began giving up French by 1235 at the latest. 3. There is no reason to suppose that any large proportion of native English learned French between 1066 and 1250; after that point they had no reason to do so’ (p.308). They go on to state baldly that: ‘it can in no way be considered reasonable to suppose that any of the conditions for pidginization, creolization, or language mixture existed between English and French in the Middle Ages’ (p.309). Even though these conclusions are diametrically opposed to those of the writers mentioned above, the divisions in time adopted by the two sets of authors are broadly similar, with a first period running up to about 1235 and being markedly different in character from the following two centuries.

To get at the truth in this welter of unsupported and contradictory assertions it is only necessary to look at the bare facts of the linguistic situation in medieval England as indicated by history, bringing to bear on that history such textual evidence as is available. It is generally claimed that the French of England came originally from ‘Norman, both as spoken and written language’, as quoted above from the Cambridge History of the English Language. Yet with all the spoken language of the eleventh century gone for ever and no substantial body of French texts to hand from that period which can be unequivocally located in Normandy, this simple, self-evident statement may not be quite as straightforward as it appears at first sight. It is beyond dispute that the invasion force that took possession of England in 1066 was not a purely Norman army: not only did it contain many Bretons, who spoke their own non-Latin language, but also a large contingent from farther up the French coast – Picardy and beyond. At a time when the dialects of northern France were by general consent heavily fragmented, the French of these non-Normans would not have corresponded exactly to that of the men from, say, Dieppe or Coutances. Furthermore, the social aspect of this dialectal fragmentation makes it is highly unlikely that, even within the Norman section of William’s army, nobles, high-ranking ecclesiastics and peasants turned soldiers would have shared a common speech form. In England today, after well over a century of universal compulsory education, half a century of nationwide radio and decades of terrestrial and now satellite television, with road and rail providing rapid communications between widely separated areas, it is still possible in many cases to distinguish by their speech – lexically and syntactically, as well as phonetically – men from regions as near together as Cheshire and West Yorkshire, North Yorkshire and Tyneside or Essex and Sussex, to pick out the farm labourer from the landowner, the shop-floor worker from the manager, the private soldier from the colonel. Since no one was on the beach at Pevensey with a tape-recorder in 1066 to greet the disembarking invaders, the nature of the speech they used must remain a matter for conjecture, but it is most unlikely that it could be adequately described in terms of any all-embracing formula such as ‘Norman dialect’. This most elementary fact is at the root of the whole question of the languages of medieval England, yet rarely finds so much a passing mention in the learned disquisitions on the subject. The whole of the spoken language output of Conquest and post-Conquest times, both English and French, is irretreivably lost: all that can legitimately be used in evidence is the – by comparison – minute amount of ink on parchment that has been preserved, however great may be the ingenuity of the philologists’ attempted reconstructions. Moreover, not only is this written evidence no more than a drop in the ocean when set against the sheer bulk of the collective daily speech of untold thousands of people over many generations, it is also highly unrepresentative, being exclusively the product of the scribal class. To assume, as phonological studies invariably do – either by statement or, more usually, by implication – that sound values for vowels and consonants elaborated by nineteenth- or twentieth-century philologists from the rhymes of medieval texts composed or copied by a tiny minority of literate people can be said to reflect accurately the spoken language of the whole population of an extensive area at any one period is no more than wishful thinking. It may well be the best that can be done at this distance in time, but this ought not to blind us to the practical realities of the situation. To put the matter in modern terms: it would be the height of folly to use the diction of BBC announcers in the 1930’s and 1940’s, together with the written language of ‘The Times’ newspaper in those years, and claim that the result gave a correct overall picture of English at that time, either spoken or written, either in London or in the country at large, although the quantity of data so produced would greatly exceed that available for study from surviving records of the French of medieval England. To make matters worse, in the case of Anglo-French the very modest amount of writing that has come down to us from the early period does not start to appear in any quantity until more than half a century after the Conquest, so there is quite literally scarcely a shred of hard contemporary evidence from England on which to base reconstructions. Even when written evidence does start to become available for study in England, often it cannot be stated ^p150 with any certainty whether the writers and scribes were native French speakers who had settled in England or Englishmen and English women who had learned French as a foreign language, either at home or abroad, to a high standard of competence or only enough to enable them to copy out texts already drafted;¹³ whether the native French speakers amongst them came originally from Normandy or from other regions; whether they had travelled far from their birthplace to acquire their education and professional skills, perhaps modifying their French as a result of prolonged contact with another dialectal form; whether they were conforming in any noticeable degree to some kind of code of scribal practice. All these points are no more than the most basic common sense, yet they need to be stated in the face of all-embracing certainties built on nothing more than empty hypotheses. They serve to explain why editors have often remarked on the presence in Anglo-Norman texts of dialectal features from areas of France well outside Normandy, why some of these texts are highly praised by their editors, both English and French, for their ‘correctness’ and compared favourably with similar works coming from mainland France at around the same period, whilst others are full of ‘errors’, and why there is still hesitation as to whether certain texts are of Insular or Continental provenance. The widespread anonymity and potential mobility of the medieval scribal class militate against any dogmatic pronouncements regarding the detailed nature of the earliest French used in England.¹⁴ Frankwalt Möhren put his finger on the key factor in this problem in his recent review of the Anglo-Norman Dictionary: ‘Man muss in Erinnerung rufen, dass fast jedes agn. Wort auch ein kontinentalfr. Wort ist. Der AND belegt so manches Wort, das wegen mangelnder Überlieferung oder lückenhafter Lexicographie erst wieder als mod. Dialektwort auftaucht, und er gibt viele Bedeutungen, die bisher noch niemand erfasst hat, obwohl sie existieren’.¹⁵

These cautionary observations apply not only to what has been termed the ‘Norman’ period of ‘creolization’ or ‘borrowing’ (better, ‘adoption’) from one language to the other in medieval England, but also to the later period in the thirteenth century, when the influence is supposed to come from ‘Central French’ as a result of the large-scale immigration of courtiers and their officials following the marriage of Henry III to Eleanor of Provence in 1236. Here again, there are questions to be raised. In fact, Vising himself signalled very clearly the need for scepticism on this score when he wrote that: ‘The king now lavished his favour on the queen’s Provençal and Savoy compatriots’ (p.10), but the ^p152 implications of this important statement have been completely ignored for decades. There is no proof that the first language of all these ‘Provençal and Savoy compatriots’ who were brought over to England by Queen Eleanor was even a form of the langue d’oil (Northern French), as distinct from a variety of the langue d’oc (Provençal) or even of the lingua del sì (Italian). If indeed by some strange chance they were all native speakers of some form of northern French, whether this French was of a uniform kind must be very doubtful in view of the dialectal conditions referred to above. Even in the highly unlikely event of their using an approximately common form of speech, whether this corresponded to what modern philologists are pleased to call ‘Central French’, despite the lack of any evidence for its existence at that time,¹⁶ is even more doubtful. Yet these assumptions must have been routinely taken on trust as copper-bottomed facts by Anglicists and linguisticians who have subscribed over the years to the idea of a wholesale enrichment of the Middle English lexis from Central French sources set in train by the immigrations during the later thirteenth century. ¹⁷ Moreover, whatever the speech of the immigrants might have been, only by the passing of any new terminology through the filter of the scribal class and its entering the records of one or other of the three languages of medieval England could any permanent enrichment of Middle English have taken place.

Whilst much has been written about the provenance of ‘French’ terminology found in Middle English, up to now no reference has been made to the reception of these ‘borrowings’ in England. Along with so many other assumptions it has simply been taken for granted that the immigrants used a ‘new’ word or phrase and that somehow this was taken up into Middle English. Yet, for ‘borrowing’ to have taken place on such a large scale would require a most unusual set of circumstances. For example, on their arrival in England Eleanor and her followers, all using their totally unproven uniform ‘Central French’, must have found themselves in the midst of thousands of users of French – spoken or written – in the literate classes who, according to this version of events, did not know the hundreds of everyday French words which, thanks to the new immigrants, would eventually find their way into Middle English. So these benighted people, equipped only with their ‘Norman French’ from a bygone age, must have been using a language unfit to meet the normal requirements of daily intercourse on account of its manifold lexical deficiencies – let alone its crippling syntactical ones if Dor’s pidginisation is to be believed – and must have been ^p153 waiting expectantly for it to be supplemented and brought up to date by the new vocabulary. Yet there is ample documentary proof that these same benighted people had been travelling back and forth across the Channel in number ever since 1066 – and even before, when Edward the Confessor was resident across the water; that many of them had spent years in France, often in the centres of learning, rubbing shoulders with the cultural elite on the Continent; that the writers amongst them had produced a wide-ranging literature in French that would stand comparison with that written on the mainland and which gives not the slightest hint of a paucity of lexical resources;¹⁸ that the many officials of all ranks, with the help of the growing class of scribes, had been successfully administering all aspects of their country largely through the medium of what can only have been, if we are to give credit to this theory, their grossly defective French, somehow creating in the process a legal and administrative register far richer than anything so far recorded for mainland France. It is difficult to imagine such people waiting impatiently for Eleanor to bring them her linguistic bounty and so enable them to partake fully, if belatedly, of the benefits of French civilization.¹⁹

As was mentioned above with reference to these later ‘borrowings’, The Cambridge History of the English Language switches the emphasis from the spoken to the written language, claiming that the lexical enrichment came from ‘an artificially acquired literary language, from the French of the Ile de France’ (p.426). Yet this idea too is not without its difficulties. If the flood of new vocabulary is not to be ascribed to a few outstanding personalities wielding massive influence over their contemporaries, it postulates a constant stream of Englishmen travelling to Paris over decades in order to learn Parisian French and then return to England with their bags filled with sheets of new words or expensive manuscripts written in the language of the French capital.

Although speculation of the kind examined above was understandable in the past, because the knowledge necessary to form a more informed picture of the growth of the English language in the later medieval period was not available, as was explained earlier, in present-day circumstances a more productive course of action is waiting to be adopted. Rather than rushing blindly to heap credit on Eleonor and her entourage for the enrichment of Middle English through the medium of the spoken word or looking around for written Continental sources that might be enlisted to explain the phenomenon, it would be more profitable to take all the terms supposed to have been introduced into Middle English from about the second half of the thirteenth century onwards and, by using all the modern dictionaries of English, French and Latin, determine when and from where they came into written English, not forgetting their subsequent diffusion and semantic history. Even though work on the new dictionaries – Middle English Dictionary, Anglo-Norman Dictionary and Dictionary of ^p154 Medieval Latin from British Sources – is far from complete, a close study of the published sections of these works will suffice to transform current views of the growth of the Middle English lexicon during the later medieval period. For example, it can be demonstrated that the distinction between the ‘Norman’ spellings of words taken into Middle English during the early period and the ‘Central French’ spellings attached to words adopted later, is far from being as clear-cut as it is made out to be in the reference books – for instance in the Cambridge History of the English Language pp.430-431. This alleged difference in spelling is regarded as a key element in determining the origin of ‘borrowings’ from French. The source quoted for this distinction is Fraser Mackenzie’s Les Relations de l’Angleterre et de la France d’après le Vocabulaire, that goes back to 1939, even though much more up-to-date evidence for most of the words involved can now be seen in the MED. If we add to the MED the information now available from Anglo-Norman sources in the AND, it becomes clear that there was no uniformity of spelling at any particular period, either in the original French – Continental or Insular – or in the Middle English into which the words were eventually adopted. Whilst full treatment of the dictionary evidence now available would demand a long article to itself, the point may be made here simply by looking at the treatment of the Latin ‘c’ followed by the vowels ‘a’ or ‘o’ in a few early Anglo-French texts taken from different registers – the administrative/legal, the literary and the glossarial. According to theory, there ought to be visible a clear distinction in time and origin between forms beginning with c- (claimed to be Norman in origin) and those with ch- (said to come from Central French). In the Leis Willelme,²⁰ the earliest body of law in French, written in the first half of the twelfth century and as ‘Norman’ and ‘official’ a text as it is possible to find, the scribe uses the ‘Central French’ forms chaceur, chalenjurs, chambre, charn, chatel, chemins, cheof, cheval(s), chevestres and chose, some of them several times; in the first volume of the Borough Customs,²¹ also from the twelfth century, may be seen chascun (p. 23), chascune (p.177), chiet, chaist (p.47), etc.; in the literary field, Gaimar’s Estoire des Engleis (1135-40) ²² uses forms of chair (vv. 212, 225, 690, 2660 etc.), chalces (v.4259), forms of chaleir (vv.2085, 3150, 3886, etc.), chalengier (vv.2389,4137, 5966), champaine (v. 2823), etc. in the same text as cacier (vv.870, 1829, 2222), campiun (v. 2978), cangier (v. 447), etc.; the Anglo-Norman writer of Li Quatre Livre des Reis (c.1170), ²³ a text rich in ‘Anglo-Norman’ spellings of all kinds, nevertheless uses chaldes, champiuns (p.4), chescune (p.7), chevel (p.8), chose, chai, champ (p.10) etc. alongside cancelant (p.5), kachevels (p.11), etc.; the Roman de toute Chevalerie²⁴ of about the same date contains numerous forms of chair (vv.363, 364, 620, 776, 1196, etc.), forms of chaleir (vv.809, 1739, 2493, etc.), chalenger (vv. ^p155 2320, 3558, 3943, etc.). On the other hand, recent evidence from an extensive range of multilingual glosses written in widely separated areas of the British Isles and covering mainly, but not exclusively, the thirteenth century ²⁵ ought, in theory, to show a clear predominance of ch-forms, reflecting the shift in emphasis from the ‘Norman’ to the ‘Central French’. However, abundant testimony from this less elevated type of record confirms that both c-and ch-were in use side by side from the earliest times. In Aelfric’s Glossary, the manuscript of which ‘was written in the second half of the eleventh century’ (vol.I, p.24), we find chevilla, chacius, chavasuriz, chao alongside campiun, canne, cardin (=’garden’), caz (=’chalk’), cisel. In the Indexes to the whole book, which gather together all the vernacular forms used by the many scribes who contributed to the rich harvest of glosses over generations, forms in ca-occupy three double-columned pages and cha-forms something over four pages, hardly a clear endorsement of the theory of separation by date. This rough list of examples of the c-/ch-alternation found in Insular French texts of the eleventh to thirteenth centuries could easily be multiplied many times over, but enough examples have been adduced from different types of record to show that it is not possible to postulate a clear separation between an early period when c-was used and a later period from the thirteenth-century onwards when, under the influence of a Central French – which, it must be remembered, remains no more than a hypothesis at that date – ch-took its place. As was mentioned above, writers and scribes in England did not live in total isolation, fixed in one spot from cradle to grave, cut off from all contacts and influences outside their place of residence.

More importantly, perhaps, than this acceptance on the part of Anglicists and linguisticians of the views of a past generation of Anglo-Norman specialists as regards the division of the French language in medieval England into two sharply differentiated periods is a preoccupation with the form of French words taken into Middle English at the expense of their meaning, morphology taking precedence over semantics. After all, a word is nothing more than an empty shell until it is imbued with meaning: a language such as Middle English or Medieval French cannot be adequately learned from treatises on phonology and morphology, but only by the assiduous reading of texts and the absorption of syntax and semantics. In ‘The French Lineage of English’, Bailey and Maroldt write that: ‘The only question of substance is whether Old French was creolized with Anglo-Saxon (…), whether Anglo-Saxon was creolized with Old French, or whether the mixture was of so thorough-going a nature that it makes little sense even to pose the question at all’ (pp. 22-23). They attempt to solve this strange problem of their own making by counting the main entries in the Shorter Oxford Dictionary – not the full OED, nor the MED – and come up with the finding that 28.3% ‘are of French provenience ‘ (p.31). What they really mean is that this percentage of words has a French form – nothing more. In order to ^p156 demonstrate the importance of this French element they quote a number of what they call ‘basic concrete terms that come from French’. Their list contains ‘cattle’, but no mention of ‘chattel’ (the c/ch alternation here cannot be used as a marker of date of entry in either Anglo-Norman or Middle English); ‘person’ without any mention of the peculiarly Anglo-Norman development to ‘parson’ (again completely undifferentiated in spelling in both languages); ‘catch’, with no reference to ‘chase’. There is no sign that the writers realize (another French word that the English have adopted, altered semantically and have seen re-emerge in present-day France with its new English sense) that their counting process is no more than a sterile exercise, since they lack any understanding of the historical process by which words of ultimately French origin became part of the lexis of English as a result of the myriad daily contacts between Anglo-French and Middle English in the minds and under the pens of a whole literate class in England between the Conquest and the end of the fourteenth century. Each of the three words ‘cattle’, ‘person’ and ‘catch’ is one element of a pair of doublets and owes its existence to this interplay of English and French in medieval England, not to continental French. The continental chatel/catel – both spellings were in widespread use – never had the meaning ‘cattle’ and has long since disappeared from standard French altogether; the ‘parson’ and his ‘parsonage’ were formed in Anglo-French, without any neat and tidy differentiation in spelling from the forms with e, and have been taken from Anglo-French into Middle and then modern English (see AND), never having existed in France itself; again, ‘catch’ illustrates the creative activity of Middle English with regard to its inherited Anglo-French component, being a semantic extension of ‘chase’, a sense found only in Middle English from the early thirteenth century (see MED). As with ‘parson’ and ‘parsonage’, spelling is of little help in determining the sense of this pair of doublets. Medieval French (on both sides of the Channel) and Middle English alike display a most cavalier attitude towards orthographical ‘correctness’. The list drawn up by Bailey and Maroldt contains also the Anglo-French ‘chisel’ (cf. French ciseaux) , ‘cherry’ (cf. French cerise), where, in theory, the initial ch ought to indicate a Central French form, one that, however, is lacking in Central French, and ‘flour’ (Old French flour/flur, but modern French farine), clearly a case of Anglo-French going its own semantic way irrespective of continental developments. ²⁶ ‘Close’ is also present in the list, but without any indication that it has senses in English as both adjective and noun (e.g. ‘a close friend’, ‘a close call’, ‘close-fisted’, ‘cathedral close’, etc.) not found in France in either medieval or modern times. In the same semantic area they might have added ‘privy’, which as adjective and noun has senses that are specifically Anglo-French and English, even though its origin is unmistakably French (e.g. ‘to be privy to a secret’, ‘a Privy Counsellor’, ‘Privy Seal’, ‘a privy’) and ‘foreign’, which is linked semantically to the noun ‘privy’ in ^p157 chambre foreine or maison foreine in Insular French, ²⁷ but which developed there senses as adjective and noun that bring it into the semantic field containing estrange and alien.²⁸ Middle English took up chambre foreine from about 1290 (MED sub forein 6) and can have ‘borrowed’ or ‘adopted’ it only from Anglo-French. The modern English ‘foreign’ and the modern French forain, its meaning now altered, probably by its proximity to foire, have only their form to remind us of a bygone common origin. The same applies to the Anglo-French chambre coi and English ‘coy’, yet another of the terms for a latrine or privy. The OED (sub coy sb.3) gives French as the source of the word on the strength of an entry in Cotgrave, but his fosse coye is not recorded for medieval Continental French by either Godefroy or Tobler-Lommatzsch, whilst chambre coie occurs in 1354 in the work of Henry of Lancaster.²⁹ There is room here for a thorough semantic study of the interlocking and changing relationships as all these and other similar terms develop through the centuries in Continental French, Insular French and Middle English.

Whilst it is ‘nice’ in the modern English sense to have numbers of words counted and lists like the one put together by these linguisticians, it is also ‘nice’ in the Middle English sense of the word to leave their enquiries at such a superficial and rudimentary stage. ³⁰ The answer to the question of ‘creolization’ which they posed may be sought in the number of these faux amis – the hundreds of words that have the same form in French and English, but differ semantically. The base language is English, with French terms being taken into it in quantity, but many of them coming from Anglo-French, not the continental variety, then developing senses unknown to the parent language – Continental French. ³¹ The term ‘creolization’, whatever definitions of it may be given in the manuals of linguistics, suggests the corruption of a highly developed language of civilization by close contact with a less sophisticated one used by a more primitive people. When linked with pidginisation, as in the article by Dor, the ^p158 disparaging³² intention is unmistakable. As such it is grossly misplaced when applied to the mixing of English and French on English soil during the Middle Ages. Neither Old and Middle English on the one hand, with a centuries-old history of high achievement in literature and the arts associated with it, nor, on the other hand, French, whether it be the Continental variety chosen by the Florentine Brunetto Latini in 1262 to be the vehicle of his Tresor because it was the most widely used and most pleasing language (after Latin, he was careful to add) or the insular form, with its own wide-ranging literature of the imagination and also of science, law and medicine, can be dismissed as undeveloped languages in any sense of the term. The mixing that has produced modern English took place between two highly sophisticated languages of culture as used by a literate elite. It is absurd to regard Chaucer’s language as the product of ‘creolization’; it would be equally absurd to pretend that its incomparably rich texture was shared by the whole population of medieval England.

First-hand contemporary evidence of the nature of the cultural contacts between Latin, English and French in England after the Conquest is provided as early as c.1135-40 by Gaimar’s lengthy Estoire des Engleis, ³³ written for Constance, the wife of a wealthy Lincolnshire landowner, Ralf FitzGilbert. Both Latin and Old English sources went into the making of this highly regarded Anglo-French text: ³⁴ nowhere is there the remotest suggestion of either ‘creolization’ or ‘pidginisation’. Somewhat later the anonymous writer of the Roman de Waldef tells of the existence of a written history of the English kings that was very popular until the Conquest, and begins his work by explaining the reason for his translation from the English:

Quant li Norman la terre pristrent Les granz estoires puis remistrent Qui des Engleis estoient fetes, Qui des aucuns ierent treites, Pur la gent qui diverserunt E les languages si changerunt. vv.39-4435

On two occasions in the following verses he refers to the numerous translations made into French of this historical material from the Old English past (vv. 45, 54) and twice mentions his own translation from the English (vv.79, 86). There ^p159 is no hint in either Gaimar or the Waldef writer of any disparagement or condescension towards English. Long before the close of the twelfth century we have to be prepared to deal with ‘the trilingualism that was the unique cultural characteristic of twelfth-century England’.³⁶ This trilingualism is the hall-mark and the product of the scribal class, and it is to this class that we must turn for reliable knowledge about the linguistic state of medieval England, not to modern phonologists or linguisticians.

The sad necessity for this warning is easily demonstrated. A decade later than the essay by Bailey and Maroldt, Dekeyser’s ‘Romance loans in Middle English: a re-assessment’³⁷ does not even list Anglo-French or Anglo-Latin amongst the languages that contributed to the enlargement of the Middle English lexis: only Old English, Old French, Medieval Latin and Latin-French – whatever that may be – are deemed necessary. Reference is always made to ‘Old French’ as though it were immaterial whether the ‘borrowing’ came from the French of London or of Paris. Presumably, in the writer’s mind French is French – un point, c’est tout. Similarly, neither Anglo-French nor Anglo-Latin figures in the Cambridge History of the English Language amongst the languages which have contributed to suffixation in English: ‘French and Franco-Latin are, however, by far the most prolific sources of foreign derivational suffixes’ (p.449). Why English should go out of its way to seek on the Continent suffixes currently in use in Anglo-French and Anglo-Latin on its own doorstep is not easy to understand.³⁸ In the newly published Cambridge Encyclopedia of the English Language, edited by David Crystal (CUP, 1995) , it is claimed that: ‘After the Norman Conquest, the influx of words from the continent of Europe, especially French, doubled the size of the lexicon …'(p.126). Yet, earlier in the same work we read: ‘The main influence on English was, of course, French – strictly, Norman French, the language introduced to Britain by the invader’ (p.30). Confusion reigns supreme. The reality of the situation is that Anglo-French, not Continental French, was for centuries one of the two languages of record as used in government, the law, commerce and education in medieval England³⁹ as well as of a wide-ranging literature. Insular French evolved in parallel and in constant contact with Middle English on the soil of England; it was not some sort of foreign decoration lightly superimposed on the native idiom. The officials of all ranks and their clerks who drafted and copied records all day in ^p160 Latin and French were in large measure English and moved freely from one language to another according to the nature of their work and the company in which they found themselves.

In recent years the reality of this mixing, summarily dismissed by Thomason and Kaufman without so much as a glance at the abundant evidence available, has been confirmed from a quite different quarter. Attention has been increasingly directed towards the study of personal names found in medieval records as a possible means of determining the extent to which French and English were mixed in the lower strata of society in medieval England, the illiterate classes that remained without a voice for centuries. The pioneering work of Cecily Clark in this area⁴¹ has now been taken up by D. Postles,⁴¹ who has studied the names found in the taxation rolls for a number of English counties between 1290 and 1332. Amongst his conclusions drawn from a considerable body of first-hand evidence are the following statements: ‘C’est donc au cours du XIIIe siècle que les noms français vont commencer à s’étendre chez les paysans’ (p.8); ‘Ces statistiques montrent sans aucun doute que les noms français étaient employés dans une large mesure depuis la noblesse et les groupes sociaux les plus élevés jusqu’aux paysans de condition moyenne’ (p.10-11); ‘ces statistiques démontrent que les noms français se retrouvaient dans la paysannerie rurale tant de condition libre que de condition servile’ (p.12). Well aware that the names he finds in the taxation rolls were recorded for posterity not by their bearers themselves but by the scribal class referred to earlier in this article, thus raising possible doubts as to their use in the wider community, Postles rightly stresses the significance of the presence of nicknames made up of two or even three French parts – Portebrefe, Queordelyon, Duredent, Blancpayn, Jeovousdy (p.13). His argument is strengthened by the use of similar evidence taken from Coventry at roughly the same period, where forms such as le Fourner, le Peschur, Orfevre , le Bucher etc. are found in number, also composite nicknames such as Peissable, Grauntpé, Barbedaveral, Croulebois, etc. (p.12).⁴² Even before the twelfth century was out similar evidence is found in the Records of the Borough of Leicester, where in 1196 on the first Merchant Gild Roll reference is made to ‘Ric. filius Rogeri sopere’, ‘Will. homo Walteri le mercer’, ‘Stein le muner’, ‘Ric. filius Roger palmer’ and many other similar English and French names.⁴³ The fact ^p161 that the English and French names are set in the Latin framework demanded by convention would seem to indicate that they are not creations of the scribe, but genuine names by which the persons concerned were known in the town. In this connection a significant entry is found in the records dealing with the canonization of Thomas Cantilupe, bishop of Hereford, in 1307, where a Walterus Fraunceys gives his testimony in English.⁴⁴ The sum of all this evidence would point towards a mixing of both peoples and languages from the later twelfth century onwards, but, once again, the evidence comes to us through the medium of the scribal class, always very much a minority in the population. On occasion, it is difficult to state with certainty whether we are dealing with French or English in the records drafted by these scribes. In 1280, the justiciar of West Wales inspected the castle at Aberystwyth and reported that: ‘… pur ceo ke le fundement de la tour est assis trop pres le fosse du chastel, dunt le fundement crachet de jour en jour pur le graunt carck del mayre.’⁴⁵ Whereas the French authorities have no attestation for craquer before 1546, ‘to crack’ was in use in England by the end of the first millenium, so it would not be outside the bounds of credibility to claim that we have here an English word slotted into a French sentence, a practice that was to increase in frequency with the passage of time.

In this mixing of Anglo-French and Middle English that took place at all levels and from an early period there are two discrete processes at work, not just one indiscriminate muddle. English words are increasingly found in what are ostensibly French texts during the later period,⁴⁶ but this disappeared with the demise of French in England: the permanently productive part of the process was the passing into Middle English of large numbers of Anglo-French terms, leading to the enrichment of English. This enrichment was at once a lexical and a morphological/syntactical one. On the simplest level, words were taken into English from Anglo-French with their form and meaning unchanged: for instance, ambage(s), claimed variously by Middle English specialists and dictionaries to have been introduced by Chaucer from France or from Italy, was in use in Anglo-French from the late twelfth century, nearly two hundred years before it is attested in Continental French (AND). Moreover, the word had been part of the Anglo-Latin lexicon for even longer (DMLBS sub ambages), so the claim that it was a ‘borrowing’ from the Continent is hardly persuasive. It must not ^p162 be forgotten that Chaucer the administrator was in contact with both Anglo-French and Anglo-Latin for many years in the course of his varied daily work, as is shown by the wealth of documents gathered together in the Chaucer Life Records.

In other cases, the form of a French word was taken into Middle English, but its meaning reflects the French of London rather than that of Paris. To give no more than one example of a widespread phenomenon that has been treated in some depth elsewhere, ⁴⁷ the French donjon and the English ‘dungeon’ are obviously related by form, but, equally obviously, separated by meaning. The donjon stands high in the centre of the castle, whilst the ‘dungeon’ is a dank cell in the bowels of the earth. The Continental donjon has not changed its meaning in the last eight hundred years or so, but the Anglo-French donjun can be clearly shown to move in the thirteenth century towards the modern English sense. That it is originally a structure high above the ground is shown by a quotation from towards the end of the twelfth century:

la reine … A tant devale le dungun⁴⁸

In the mid-thirteenth century it is used by Grosseteste in the continental sense of ‘keep’, the strongest part of the castle:

Les treis bailles du chastel Ki … defendent le dongon⁴⁹

This meaning is still to be found as late as 1354 in the devotional treatise of Henry of Lancaster:

la tour par dedeinz, q’est en my lieu de castel, qe homme appelle dounggeon ⁵⁰

However, before the end of the thirteenth century there is clear evidence of the basic sense of strength and security being expressed by dungun without any concomitant meaning of elevation being involved:

il est bricun, E (var. El) clostre dust estre, ou en dungun ^{51 p163}

The crucial element in the modern English definition of ‘dungeon’, that of being below ground, is present by the first half of the fourteenth century:

ly roi descendist en un bas dongoun⁵²

From the evidence at present available, therefore, it would appear that in the later thirteenth and the fourteenth centuries the modern English sense of ‘dungeon’ was being developed in the French of England, whilst the old, continental sense was still in use. The MED puts the earliest use of this modern English sense in an English text somewhat later – ‘?a1400 (a1338)’.

Alongside this lexical enrichment went a morpho-syntactical one. At the beginning of this article it was shown how the English ‘arrival’ and ‘departure’ are apparently French words, yet are not found in mainland France. These are not isolated examples, but illustrate another facet of the intimate mixing of French and English in medieval England. In ‘arrival’ and ‘departure’ both stem and suffix are French, but they are combined to create words not used by the French themselves. French was apparently regarded by the literate classes not as it is today, a discrete, self-contained language entirely separate from English, but rather as part of the common stock of linguistic material available for use in records either in the form of complete lexical items or as components that could be combined with English elements. This phenomenon becomes visible with the increased number of surviving records from the late thirteenth century onwards. For example, in 1383 the sum of twelve pence is paid for canvas used in ‘calfattyngge’ (i.e. ‘waterproofing’) a boat on the Thames: ⁵³the semantic element of this composite word is French, ⁵⁴ but the suffix denoting the carrying out of an operation is English. In 1406 the same Thames records show that beer was bought and given to the workers ‘circa le launchyng del shoute in Thamis’: again, the verb la(u)ncer had been in use in Anglo-French with reference to putting boats into water from the early thirteenth century at the latest, but the suffix is again English. ⁵⁵ Later in the fifteenth century reference is made in these records to ‘hewyng & apparellyng’, also ‘werkyng in apparelling’ stones called ‘bridge asshelar’ or ‘brigassheler’ ( ibid. p.48). The Anglo-French elements here are the verb apparailler (‘to make ready, prepare’, etc.) and the noun aiseler (Modern English ‘ashlar’), neither of which needed to be sought across the ^p164 Channel, having been in active use in the French of England for centuries.⁵⁶ Similar mixed forms occur in abundance in the Merchant Taylors’ archive in the early fifteenth century. In 1404 there is mention of an account ‘pur le peinture de les novelles torches et pur le refresshing et floresshing de les veiles torches’,⁵⁷ where the Anglo-French refresscher (one of the forms of what is usually listed under refre(s)chir) and florir (florisser is as yet unattested) have been given the English suffix -ing, whilst in peinture the same sense of an action being carried out is rendered by the French suffix – ure. The same grammatical procedure occurs again when fees are paid for the ‘doublinge del chartre’.⁵⁸ At around the same time the compilers of the records of the Grocers’ Company paid for ‘asselyng de lez pois’ ‘authentication of the weights’.⁵⁹ Again at this crucial period comes the first attestation of ‘costiveness’, the English suffix – ness being added to the Anglo-French costivé. The bonding may, however, be made the other way round, a French ending being grafted on to an English stem, as when the terms fosser ou pyler are applied to the foundations of a wharf in 1389, the English ‘pile’ being made into a verb by the addition of the French infinitive ending – er.⁶⁰ This morpho-syntactical mixing goes deeper than the straightforward adoption of lexical items and needs to be treated separately at some length elsewhere.

From the textual evidence provided above it is clear that the question of the enrichment of the Middle English lexicon by the introduction of French terminology needs to be tackled in the first instance by the study in depth of the interplay of the three languages that went to make up the trilingual civilization of medieval England: most of the answers of concern to modern English will be found on English soil, not in France. Perhaps Chaucer himself may be called upon to provide the final illustration of the centuries-old mixing of Anglo-French and English in medieval England. In the prologue to his tale the knight says:

‘I have, God woot, a large feeld to ere, And wayke been the oxen in my plough’61

^p165 It is generally accepted that ‘to ere’ (‘to plough’) in this context is a survival from Old English, but, without casting any doubt on this attribution, it is worth mentioning that Anglo-French was using a similar verb with the same meaning from the twelfth century:

Les teres laborent e erent E richement les cultiverent62

The scribal class of medieval England, responsible in large measure for the enrichment of later Middle English, was in varying degrees a trilingual one. All three languages of medieval England need to be studied as making up a unitary linguistic situation, especially in the later period when all three are often found in one sentence, rather than approached individually without reference from one to the other.

W. ROTHWELL

---