Contents

Anglo-French and the AND
Computerisation of the AND
Readers' Guide
Introductory notes for on-line AND users

Anglo-Norman or, more accurately, Anglo-French, is the form of French used in Britain between 1066 and the middle of the fifteenth century. The term ‘Anglo-Norman’ harks back to the time when the language was regarded as being the regional dialect of the Norman invaders who came across the Channel with William the Conqueror. Yet when account is taken of the heterogeneous composition of William’s army, which included many men from different regions of France, together with the fact that over the following three centuries the language must have been used in Britain by all manner of people from dissimilar ethnic backgrounds and whose linguistic competence, to judge by the writings which have survived, may be readily seen to have varied from a native mastery of French down to an elementary acquaintance with it, the generic term ‘Anglo-French’, the French of England, perhaps reflects the reality of the situation better than the more restrictive ‘Anglo-Norman’.

The title of this second edition of the Dictionary preserves the old name purely in order to maintain continuity with the first edition, which adopted ‘Anglo-Norman’ as being the term in current use in academic circles at the time in the later nineteen-forties when the idea of a glossary of the medieval French of Britain was first mooted. In this regard, it is perhaps worth noting that the dictionary of Robert Kelham going back to 1779, too early to be affected by the Neo-Grammarians who influenced the thinking behind the original Anglo-Norman dictionary project, was entitled A Dictionary of the Norman and Old French Language, a form of words indicating an awareness even at that time of a broader geographical base than ‘Norman’ alone.

Although the title of the new Dictionary remains the same as before, its contents are very different from those of the earlier work.2 In the first edition a significant change in both the depth and the spread of coverage is noticeable from fascicle 5 onwards (P-Z), in effect dividing the Dictionary into two dissimilar halves. This was made possible by the introduction of the important fichier gathered over many years by J.P. Collas, containing material from a wide range of both literary and non-literary texts, and also by the Dictionary of Law French compiled by Elsie Shanks but never published, the first serious attempt to record what is still the basic lexis of present-day English law. These additions transformed especially the non-literary side of the Dictionary, which up to that point was represented by only a small selection of texts outside the literary canon. This change of emphasis is now being extended in the new edition to the first half of the alphabet.

The work of Collas and Shanks will be carried on, adding to their texts numerous others from a variety of non-literary registers, whilst at the same time supplementing the literary register by drawing on texts absent from the first edition. The works used to promote this aim, in particular those outside the literary range, come predominantly from the fourteenth and early fifteenth centuries, a time which in the past has traditionally been considered by scholars working in the field of medieval French to be inferior in linguistic terms to the early period following the Conquest. These texts are usually in prose and they often fail conspicuously to observe the ‘rules’ of phonology as laid down by philologists in the Neo-Grammarian mould in their search for the normative form of medieval French supposed at that time to lie at the root of the modern language.

Recent scholarship has, however, substantially rejected many of the underlying assumptions which this approach implied. These later texts have also been depreciated in the past in terms of their cultural value, dealing less with the epics and romances of chivalry than with records of one kind or another. As a result of these additions, the new version of the Dictionary promises to be more than three times the size of its predecessor.

More important than the size of the new work, however, is the expansion in the range of source material referred to above. Whilst the narrow textual basis of the original Dictionary as first envisaged was progressively widened from one fascicle to the next, the new version covers a more extensive range of source material from the start and draws on a much greater number of texts within all the registers concerned in order to present a richer semantic picture of the language. In view of the qualitative and quantitative difference between the two editions, it may be appropriate to sketch in this preface the various areas of vocabulary from which the Dictionary material is now being drawn, setting them in the context of the society which produced them.

1 The first three sections of this essay are substantially identical to the corresponding introductory parts of the printed edition. For that reason, references to the detailed presentation of entries refer to the conventions of that edition, not to the sometimes different practices employed in AND1. Since this electronic edition supplements the completed AND2 entries with the as yet unrevised portions of AND1, readers are asked to bear that distinction in mind. More particulars of the differences between AND1 and AND2 on-line entries are given in the fourth section of this piece.
2 See also D.A. Trotter, ‘L'Avenir de la lexicographie anglo-normande: vers une refonte de l’Anglo-Norman Dictionary?’, Revue de Linguistique Romane 64 (2000), 391–405.

The basic difficulty with Anglo-French is that it was one of the three languages of post-Conquest medieval England, whose relationship to each other changed imperceptibly but inexorably over nearly four centuries. Anglo-Latin gradually lost ground to Anglo-French in its role as the official language of record at both national and local level, whilst Middle English emerged over time from being a predominantly spoken language to take over from the two others in the fifteenth century as the acknowledged national language, both spoken and written. This simple summary statement, however, hides a complex linguistic interplay brought about by the continuously evolving social situation in Britain and on the continent for many decades after the Conquest.

In the first place, the role of Middle English for over two centuries after 1066 cannot be determined with any precision in the absence of an adequate body of surviving recorded evidence before the fourteenth century, although it was there all the time in the background as the spoken vernacular of the majority of the population, despite many of them using French and/or Latin in their writings. Secondly, Anglo-French was not merely the language of the conquerors, destined to decline and eventually wither as the French component of the population dwindled from one generation to the next and was gradually absorbed into an anglophone society. If this process of absorption had been governed by ethnicity alone, it would not have taken over three centuries to take effect, but it was checked by a variety of outside factors.

The enduring links between France and the offshore island were not broken by the loss of Normandy in 1204, nor did they persist only in the form of military incursions, although Edward IV was born in France during the Hundred Years’ War, and even in the seventeenth century King James I was still called ‘King of Great Britain, France and Ireland’ on the opening page of what is known as the ‘King James Bible’. More important than the military contacts was the fact that French civilization did not stop at Calais or Dieppe, but was carried over into England on parchment and by word of mouth, as is demonstrated by translations of numerous Biblical works into Anglo-French, the production of a range of botanical and medical texts based on European works and the presence in France for long periods of cosmopolitan scholars from England like Adam of Petit Pont, John of Garland and Alexander Nequam, whilst, on the other hand, as late as the fourteenth century, the French chroniclers Jehan le Bel and Froissart were certainly not ignorant of the corresponding lettered class in England. At the same time thriving trade links with the countries on the mainland of Europe were similarly conducted in French, the medieval language of commerce in western Europe. It was not for nothing that the English Crown maintained a ‘Staple’ in Calais. French is even found in documents concerning the loans made to English kings by Italian merchants in the thirteenth and fourteenth centuries.3 Within the confines of England the various aspects of the machinery of government, at both national and municipal level, religious as well as secular, functioned in Britain largely through the medium of French until the fifteenth century.

4 ‘The Privy Seal Clerks in the Early Fifteenth Century’, in The Study of Medieval Records. Essays in honour of Kathleen Major, eds D.A. Bullough and R.L. Storey ( Oxford, 1971), 260–81.
5 In Standardising English: Essays in the History of Language, ed. J.B.Trahern Jr., Tennessee Studies in Literature 31 ( Knoxville, 1989), 37.
6York Memorandum Book (1376-1419), ed. M. Sellers, Surtees Society vol. 120 ( Durham, 1911), 13.
7The Little Red Book of Bristol, ed. F.B. Bickley, 2 vols ( London, 1900), I, 55.

In the area of jurisprudence, starting with the Laws of William the Conqueror, then proceeding to Magna Carta and on to the Statutes of the Realm and the Year Books, the extensive records of the legal system, also in Anglo-French and Anglo-Latin, call in their turn for similarly comprehensive (and ideally, multilingual) lexicographical coverage. To these must be added the lengthy treatises expounding in Anglo-French the procedures and jurisprudence of the law itself, texts such as Fet Asaver, Britton, the Mirror of Justices and the Court Baron. The contents of all these works offer to the lexicographer a whole register of specialised vocabulary that survives to this day in the courts of law in Britain under the guise of English, and which renders modern legal English so opaque to the majority of native speakers of English. Until the appearance of the Anglo-Norman Dictionary, students of the law had little help in acquiring an adequate understanding of the medieval part of their chosen subject. In his Manual of Law French (second edition 1990), a glossary, not a dictionary, J.H. Baker lists the early works from the later seventeenth century onwards which deal in one way or another with Law French (pp.7-10 and 24-32). Many of them are of little practical use and he concludes that: ‘The Anglo-Norman Text Society has at last succeeded where all the other attempts failed’ (p.10). The new edition of the AND can only reinforce that opinion. Whilst the legal volumes published year after year by the Selden Society for over more than a century now and those in the earlier Rolls Series have an English translation facing the Anglo-French, this does not apply universally, some texts leaving the student to his or her own devices with no guidance. In any case, the Selden Society and especially Rolls Series translations are often simply transpositions into forms of English of the Anglo-French originals, which offer in reality little help to the reader:8

the translation into English of Anglo-Norman texts of law on the one hand has long been a necessity and on the other has always tended to take a peculiar form […] the method of translation […] consists essentially in rendering into English the morphological and syntactical elements of the original, while leaving intact any and every term capable of definition or interpretation. The result is sheer jargon.

Yet not even this rudimentary assistance is always supplied. For instance, the later thirteenth-century Fet Assaver referred to above treats in great detail of all the different kinds of plea that can be made in the courts of law, providing also variant readings from different manuscripts of the text, but makes no concessions to the reader in the form of help with the languages used – legal Anglo-French and Anglo-Latin.

Not only the law of the nation as an entity was set down extensively, though not exclusively, in French, but so was the customary law applying both to large towns such as Southampton or Winchester and small ones like Romney or Winchelsea, indicating that a working knowledge of French for daily business dealings was far from being the preserve of the ruling elite. Indeed, legal matters between comparatively modest individuals on different sides of the Channel were often settled in French. In 1357 a Flemish ship laden with sea-coal and other goods belonging to a merchant of Amiens was seized in the harbour of Romney and its master and crew arrested on the order of the Warden. The mandate for the arrest and the confirmation of the execution of the order by the Bailiff and Barons of Romney were both in French. In 1358 the Flemish captain of a vessel named ‘Hardbolle’ based at Sluys drew up or had drawn up a receipt in French for money from a certain James Creoord of Romney in respect of sea-coal, slipstones, cables and anchors provided to him, together with an acquittance to this James Creoord and J. Heithe renouncing all manner of legal action against them. In the same year a dispute over a kiddle net (or kettle net) between persons in the parishes of St. Nicholas of Romney and St. Marychurch was argued out in French, although both sides in the dispute were English. The greater attention accorded in the second edition of the Dictionary to the legal register as a whole will be of help to all who deal with the different aspects of English law.

Eminent individuals kept their records in Anglo-French up into the late Middle Ages. The Register of John of Gaunt for the years 1371-75 and 1379-83, filling four volumes, was written for the most part9 in Anglo-French, as were the inventories of the possessions of John, Duke of Bedford and Regent of France, made between 1389 and 1435. Many of the objects listed in these inventories were works of art, their names or descriptions lying outside the general registers of French vocabulary, as for example: ‘Item, une basse coupe d’argent dorré […] esmailé au fons de fleures de ne me oblies mie’ (C20)10 (= ‘forget-me-nots’). Their counterparts in the world of business also kept their records in Anglo-French up into the fifteenth century, as may be seen in documents belonging to the Livery Companies in London such as the Merchant Taylors, Goldsmiths, Barbers and Scriveners, the specialist areas of vocabulary which they provide contributing to the enrichment of the second edition of the Dictionary. In 1418, when the Brewers appointed a new clerk to keep their records, his first entry was in Anglo-French, but when in 1422 it was decided to switch to English in order to be in line with other companies, the decision was, somewhat perversely, recorded in Latin.11 The Drapers were still using Anglo-French for their records as late as 1434, the first accounts in English dating from the end of 1440, although it must be conceded that the ‘French’ entries, being set down for an exclusively English readership, would have been incomprehensible across the Channel (like the fleures de ne me oblies mie referred to above), having abundant recourse to English endings attached to French words, or to English terminology, sometimes tricked out with more or less French endings, sometimes left in an unadorned English state. Amongst many such entries to be found as a regular feature in this kind of record are the following taken at random from the Drapers’ accounts: Item pour le Davbar pour parchettyng de le kychon (1430: p.322) in which Davbar is a form of dauber (‘plasterer’), parchettyng represents the Old French parjeter/porjeter (‘to plaster’) with the English ending yng,12 and le kychon bears little resemblance to the French cuisine from which it is derived. Similarly, pour le takyngdowne de lez draperz Steyne (1434: p.325), pour iij whell barwes (‘for 3 wheel-barrows’, 1425: p.301) and pour amendin dun vell whell barwe (‘for mending an old wheel-barrow’ 1425: p.302) pay only lip-service to French. Englishmen are thinking in English, but custom dictates that their records should be set down in official French, even though their command of its lexis is inadequate to the task. Such apparent disregard of the modern accepted linguistic boundaries is an indication of the multilingualism that was a feature of medieval Britain and which lexicographers must deal with.13 It is to be regretted that the records left by some of the mercantile companies (and guilds) have in the past been translated into English for publication without the original Anglo-French or Latin, and so need to be re-edited and presented in their original language before they can be used for lexicological/lexicographical purposes.

Likewise, the records of goods coming into and leaving Britain were in large measure kept in Anglo-French until well into the fifteenth century, French being the generally accepted language of trade (and the language often used for names of ships). The Port Books of Southampton reveal a thriving maritime commerce in the medieval period, at both the national and international level, especially with the Mediterranean countries, and the names of the goods carried provide a rich French vocabulary well outside the literary norm, a vocabulary which, like those of the administrative and legal registers, has been taken over into English. The Port of London must have been no less busy than Southampton, with trilingual records relating to the Thames running into the later fifteenth century. Additionally, however, in the case of London a mass of evidence, much of it still unpublished, confirms the extent of its trading links and the use of Anglo-French from quite a different angle.

Throughout the fourteenth century successive mayors of London produced a stream of correspondence in defence of the interests of their fellow-citizens in commercial matters which was addressed to anyone who threatened those interests, whether in Britain or across the Channel and the North Sea, all in Anglo-French. Nor was this type of correspondence in French confined to the influential traders of the capital city, but extended also to the small ports in Kent, as has been demonstrated above. The goods brought into Britain from abroad included not only everyday items such as coal or timber from the Baltic and wax from Poland, but a considerable number of more exotic products from the Mediterranean area, which may be seen to be put to practical use in two Anglo-French collections of culinary recipes dating from the early fourteenth century, followed by others in the next century. French ‘cuisine’ based on imports from distant lands was ‘on the menu’ (a French term again) in the houses of the rich, both lay and ecclesiastic, in medieval England.

9 This innocuous formulation (as often) hides a multitude of possibilities. Many, perhaps most, non-literary texts from medieval England are multilingual, either because texts in different languages are found in the same documents, or because various forms of code-switching and language-mixing are found within texts. See D.A. Trotter (ed.), Multilingualism in Later Medieval Britain ( Cambridge, 2000).
10 1428–32; PRO E154/1/39.
11 R.W. Chambers & Marjorie Daunt, A Book of London English 1384-1425 ( Oxford, 1931), 138–139.
12 With the prefix form par-, and with the sense of ‘to plaster’, the verb is attested only in Anglo-French (and British Latin); cf. D.A. Trotter, ‘L’anglo-français au Pays de Galles’, Revue de Linguistique romane 58 (1994), 461–487, and DEAF J article jeter (Stephen Dörr).
13 See the studies in D.A. Trotter (ed.), Multilingualism in Later Medieval Britain ( Cambridge, 2000).

In the private domain, the educated members of society exchanged letters in Anglo-French in abundance throughout this period. In the early years of the fifteenth century John Barton, originally from Cheshire, but with a background of scholarship in Paris, composed a detailed grammar of French to enable his contemporaries to communicate with people in France: Pour ceo que les bones gens du Roiaume d’Engleterre sont enbrasez a sçavoir lire et escrire, entendre et parler droit français afin qu’ils puissent entrecomuner bonement ové lour voisins.14 A great many letters must have been lost, but hundreds are still in existence on all manner of subjects. W.W. Shirley published two volumes of largely semi-official letters from royalty in the Rolls Series in 1862 and 1866, with some of them in French, but far more were edited in the twentieth century. For his Paris thesis, F.J. Tanquerey brought to light 164 letters written in Anglo-French between 1265 and 1399 by high-ranking officers of state, mayors, senior clerics and occasionally foreign dignitaries or even ordinary citizens, their subject-matter ranging from a request to be released from prison to complaints about the irresponsible behaviour of nuns, a business report to the Barons of the Exchequer from the manager of a royal mine in the Tamar valley, west of Tavistock, about the problems of water in the mine, and even an invitation to officiate at a funeral. Tanquerey went on to edit the Anglo-French letters sent by Edward I to the yeoman in charge of his hawks, in which the king shows both a keen interest and practical knowledge. The letters of Edward Prince of Wales written in French were published by H. Johnstone in 1931. Finally, a decade later, Dominica Legge edited a further 400 or so letters of widely varying kinds from Charles VI of France, Gaston de Foix, Charles III of Navarre, Isabella of Bavaria, Philippa of Portugal, John, Duke of Brittany, and Joan, Duchess of Brittany, the Duke of Milan and many other notables, down to a certain Nicholas Bargeman and two citizens writing in support of a candidate for the post of gaoler at Newgate Prison. All these letters in French date from the end of the fourteenth and beginning of the fifteenth century and had been brought together in just one manuscript in All Souls.

Letter-writing is the most personal form of written communication, and so is probably nearer to the spoken language than most other writings apart from conversation manuals. Tangible evidence regarding the use of French in speech in medieval Britain is naturally a rare commodity, so that the proof of oral testimony provided by Michael Richter for the early fourteenth century is especially valuable. In his study of miracles claimed to have been worked by a bishop of Hereford he notes that, in the investigation of nine witnesses carried out on behalf of the papacy, ‘only one gave his evidence in English, four of them did so in French, two, apparently lay people, spoke partly in French, partly in Latin, and the two clerics […] gave their evidence in Latin’.15 In the same year, 1307, evidence for another miracle was given in Hereford by people from Swansea, three of them from Swansea Castle, four burghers from the town and a priest. All those from the castle spoke in vulgari Gallico, as did the priest, whilst one of the burghers used French and the other three English.16 The evidence presented at enquiries of such importance cannot have been unintelligible, and nor does the use of Anglo-French by a large proportion of the witnesses appear to have been regarded as unusual, so it must be concluded that the language was currently used in spoken as well as written form, and not solely by a small minority of the rich or influential.

14 T. Städtler, Zu den Anfängen der französischen Grammatiksprache, Beihefte zur Zeitschrift für romanische Philologie, 223 (Tübingen, 1998), 128.
15 Michael Richter, ‘Collecting miracles along the Anglo-Welsh border in the early fourteenth century’, Multilingualism in Later Medieval Britain, 53–61 (p. 57).
16 Ibid., 58–59.

As was mentioned earlier, the part played by Anglo-French in its role of a widely-used language of record is of prime importance also in the ecclesiastical history of medieval Britain. After 1066 the Conqueror’s assumption of control over England extended to the spiritual and the secular sphere alike, so that the English leaders of the Church as well as the great landowners were ousted to make room for Frenchmen. Just as Norman castles steadily straddled the country following the Conquest, so the great cathedrals began to appear. The Norman cathedral in the strategic city of York was built only a few years after the Conquest along with others not only in the south in important centres of population and trade like Canterbury, Winchester and Exeter, but also in the east where Lincoln had the most extensive diocese in the country, and stretching from Gloucester further north to Hereford and to the key northern cities such as Chester and Durham, their imposing presence dwarfing any existing Saxon churches that were not demolished and proclaiming the dominance of the new order in religion just as the castles announced the new regime in government.

The castles and the cathedrals were linked by the fact that the modern separation between Church and State, between spiritual bishop and secular lord, often did not apply in post-Conquest Britain, where literacy and education, the levers of power, were the prerogative of the Church, so that high-ranking clerics might hold large estates from the king, incurring thereby the feudal duty of service, and could be called upon to occupy the great offices of state. Becket illustrates the dual role of Chancellor and archbishop in the twelfth century, secular and ecclesiastical posts being occupied again nearly two centuries later by John Thoresby who was Archbishop of York from 1352 to 1373 and Chancellor from 1349 to 1356, followed in turn by the hapless Sudbury, Archbishop of Canterbury and Chancellor, who suffered the same fate as Becket, losing his head in the Peasants’ Revolt. Earlier in the century Walter Stapeldon, bishop of Exeter and later treasurer of England, had similarly been decapitated in 1326. The military command exercised by the Bishop of Durham in the border region of northern England as a bulwark against the Scots by reason of his feudal tenure of its extensive territory was no less important than his episcopal role, a fact made clear by the contents of his fourteenth-century register in which the personal concerns of the secular lord and warrior loom larger than any solicitude of the good shepherd for the welfare of his flock. Similar registers of other senior ecclesiastics have survived from the Exeter area. In the same way that Latin and French were both used in recording the secular history of medieval Britain, they were also the linguistic vehicles of its ecclesiastical history. A relic of this medieval world survives today in the presence of archbishops and bishops on the benches of the House of Lords.

The teaching mission of the Church was also carried out after the Conquest in Latin and French, at least insofar as the literate section of the population was concerned, although English must be presumed to have been used on a large scale at the level of such elementary instruction as would be given to unlettered parishioners, especially in the countryside. Brief mention was made above of religion being one of the enduring links between France and England after the loss of Normandy. In greater detail, the early part of the twelfth century saw the appearance in England of the Oxford Psalter in French, termed the Libri Psalmorum versio antiqua gallica, and the Cambridge or Eadwine Psalter with the Latin glossed into French on the right-hand side of the page. A little later came the four books of Kings in French, Li Quatre Livre des Reis. Whether or not the appearance of insular French in these sacred Biblical texts at this time indicates an inability on the part of some of the clergy to understand the Latin, it certainly points to the acceptance of the vernacular on a par with the traditional language of learning. In all probability, the pre-Conquest tradition of using Anglo-Saxon as a religious language will have facilitated this development. The phenomenon may, however, also be seen across the Channel, where the link between the Church in France and its counterpart in Britain in this early period after the Conquest is illustrated by the Psalter Commentary composed in the language of north-eastern France for Laurette d’Alsace, the first section of which was completed in 1163–64, with the two remaining sections following before the turn of the century. Within a few years this immensely detailed and learned exposition in French of all one hundred and fifty Psalms, which would occupy at least four volumes were it to be published in its entirety, must have crossed the North Sea to Durham, where its earlier sections were copied in an insular form of the language.17 Eventually, the whole of the commentary was available in Durham and large sections of it were to be found also in Hereford Cathedral, Oxford and London. This movement towards the vernacular continued, considerable portions of the Bible, from both the Old and the New Testaments, appearing in Anglo-French form in the fourteenth century, when the Acts of the Apostles were also translated in two different versions in Anglo-French.

The amount of Anglo-French found in medieval didactic texts of a religious nature intended presumably for either the religious or the educated laity is very considerable. Not only is there an abundance of saints’ lives, many of them in verse, but Guischart de Beauliu composed a long sermon of nearly 2,000 verses in Anglo-French in the closing years of the twelfth century, an example that would be followed later by other writers of sermons. Before the middle of the thirteenth century Robert Grosseteste, Bishop of Lincoln, was writing his Injunccio penitenti gallice,Oraciuncula post prandium gallice,Le Mariage des IX filles du diable, Confessioun and his substantial Chasteau d’Amour. A few years later Jean de Howden produced his long allegorical treatment of the Passion of Christ, Le Rossignol; in 1268 the Lumere as lais was composed, setting out in almost 14,000 verses a detailed intellectual exposition of the tenets of the faith, the lais in the title being in no way the unlettered amongst the faithful, but simply those not in holy orders.

At the end of the century the Minorite Nicolas Bozon wrote not only a number of saints’ lives and sermons in French verse, but also a series of Contes Moralisés in which animals and birds are used to illustrate human failings, and the difficult Char d’Orgueil, where the different parts of the cart represent sinful traits such as arrogance, anger, malice, avarice and the like. In 1354 Henry of Lancaster chose Anglo-French as the language for his devotional treatise Le Livre de Seyntz Medicines in preference to his native English, clearly in anticipation of a readership versed in French, and before the turn of the century William de Wadington produced the Manuel des pechez, almost as long as the Lumere as lais, dealing yet again with the sins of the world, but without the personal stance adopted by Henry of Lancaster.

Additionally, since the sacred texts themselves admitted of translation from Latin into a vernacular, it is only to be expected that the regulations governing the lives of the religious who lived in accordance with the teaching of those texts might also be set out in the vernacular. For instance, two Anglo-French versions of the Ancrene Riwle have been published, one in verse, the other in prose, and various Anglo-French rules have been published in recent years by Tony Hunt, namely the Rule of St Augustine, rules for the priories of St Mary de Pré and Sopwell, and ‘An Anglo-Norman Treatise on Female Religious’, from the late thirteenth or early fourteenth centuries.

In the course of the thirteenth century, Middle English began to join Latin and Anglo-French in religious texts, creating a trilingual situation. The Ancrene Riwle is extant from that period in all three languages, although the relationship of the different versions to each other is not universally agreed. Towards the end of the Middle Ages, the trilingual De Quatuordecim Partibus Beatitudinis was assembled, its Latin translated into both Anglo-French and Middle English. This is paralleled about the same time by another trilingual religious text found in manuscript Latin 188 of Magdalen College, Oxford, where each line contains the three languages in the order Middle English, Latin, Anglo-French, the one set below the other, but, significantly, with the Anglo-French letters twice the size of those in Latin or Middle English.

For the educated in the medieval period, a preoccupation with the world of the spirit did not preclude an interest in the physical world in which they lived, both past and present. Knowledge of the natural world was set down in Anglo-French from the earliest times after the Conquest. In 1119 (or even 1113) Philippe de Thaon compiled his Anglo-French treatise on the computus from medieval Latin sources of both continental and insular origin, including Bede. The work is still extant in five manuscripts, a testimony to its perceived educational value. Works on the calendar in Anglo-French from the thirteenth century have been published by H.J. Chaytor (Calendar) and Tony Hunt (RAUF ANTS), whilst a short treatise on arithmetic from the fourteenth century was printed by L.C. Karpinski and C.N. Staubach (Algorism).18

A more general compendium of knowledge about the earth is to be found in the early thirteenth-century La Petite Philosophie, based largely on the Imago Mundi. The title of the Anglo-French work has little to do with the modern meaning of ‘philosophy’, being more in the nature of general knowledge. The traditional four elements, together with the earth, sky, sun, stars, points of the compass, hot and cold zones, winds, storms, descriptions of foreign lands with their legends and strange animals – these are the topics that make up the substance of the book. In two places (vv. 252-4 and 359-66) it is stated that the earth is round and not flat,19 moving and not fixed. This ties in with the Mappa Mundi, the Hereford World Map of around 1300, in respect of the shape of the earth, its geography and the animals living on it. The writing on the map is in Latin and Anglo-French. Almost two centuries earlier, Philippe de Thaon, the author of the Comput, had gone on to compose a Bestiaire dealing with the animal kingdom. The technical French vocabulary pertaining to the hunting of animals is detailed in the early fourteenth century in the Vénerie de Twiti, written in England by the huntsman of Edward II, which covers the hare, the boar, the deer, the fox and the wolf. This was translated into Middle English. The existence of earlier letters from Edward I on the subject of hunting with hawks has been mentioned above, and Henry of Lancaster deals with the tactics of fox-hunting in his devotional treatise Le Livre de Seyntz Medicines in 1354.

18 Abbreviated titles are those used in AND: see the List of Texts.
19 See Stephen Dörr, ‘La terre est un globe’, Revue de Linguistique romane, 66 (2002), 209–213.

The post-Conquest inhabitants of Britain were as interested in their past as in the present. By about 1140 Gaimar had written his long translacion of the Estoire des Engleis in Anglo-French verse, going back to the Saxon period. From around the same time there comes a Description of England, concentrating again on the Saxon period, but in particular on the shires and towns, including Wales and dealing with the struggle between Welsh and Normans after the Conquest. Jordan Fantosme’s Chronicle of 1175 dealt in detail with the events of his time, thus carrying on the story after Gaimar. This was followed early in the next century by another translation, called a Brut, which starts with the Trojans and moves across Europe through France to Britain. The reign of Edward I from 1272 to 1307 is recorded in great detail in Langtoft’s Chronicle, and the subsequent history of Britain in the fourteenth century was set out in the Anonimalle Chronicle from St. Mary’s Abbey in York, compiled late in the century and covering the years 1307-1381. The early part of that century is also covered by Friar Nicholas Trivet who wrote for the king’s daughter a lengthy history in Anglo-French starting with the creation of the world and running right up to his own times.

All this kind of Anglo-French material reflecting the civilization of medieval England in general needs to be represented in the new Dictionary. In addition, recent developments in the important field of medical terminology have transformed a technical register that could be represented only very inadequately in the first edition. Although remedies for bodily ills form part of the Secré de Secrez written in the second half of the thirteenth century by the author of the Lumere as Lais referred to earlier, the text is written from the standpoint of the general philosophy of human behaviour rather than as a specific subject in its own right. The original text is thought to go back to the sixth century and is supposed to be the words of Aristotle teaching Alexander the Great how to order his life. In terms of specifically medical texts a fragment of a thirteenth-century work written in the French of England was published as early as 1929, but it remained isolated until Tony Hunt turned his attention to editing a whole series of manuscripts which had remained untouched for centuries. In recent years his publication of the Chirurgia of Roger Frugard from the mid-thirteenth century in Anglo-French has linked Britain with the practice of surgery current on the continent at that time, and his editing of a wealth of medical receipts, some of them translations of lengthy authoritative continental works, others lists of simpler remedies of a more popular kind, has similarly transformed the understanding of Anglo-French medicine.

The heavy dependence of medicine on plant remedies in the medieval period means that these publications have advanced current knowledge of botany at the same time. Here again there is a strong link with the Church, the herb gardens attached to ecclesiastical houses being the source of many of the plants used by the religious in their ministry of healing. It is remarkable that a gap of almost a century separates the publication in 1887 of the Alphita, a trilingual ‘medico-botanical glossary’, based on an alphabetical list of plant names in Latin, with glosses in Anglo-French and Middle English, from the texts now being incorporated into the new Dictionary. Many of these medical texts share a characteristic which is being increasingly recognised now that the French writings of the later medieval period are no longer dismissed on account of their wayward orthography: like the religious texts referred to above, they are often multilingual in form, moving, seemingly with no difficulty, and certainly with no sense of unease, between Latin, French and English.

Since the innate ability to use Anglo-French inevitably became less common as the decades passed and the proportion of native French speakers became smaller in relation to the population of the country as a whole, recourse was had to didactic works intended to help anglophones attain to a working knowledge of French. The taking-over of the apparatus of the state by senior members of William’s entourage once the Conquest was complete made it clear that it was in the interest of the English to have a grasp of French if they wished to participate in the running of their country. As early as the second half of the eleventh century there is evidence of glosses in Latin, Old English and French. It has been shown that around the same time teachers were using Aelfric’s teaching of Latin grammar through the medium of English as a basis for teaching French: ‘his exposition of the parts of speech in Latin was sufficiently practical to attract the attention of several teachers who used his Latin text to elaborate the Anglo-Norman equivalents of the various paradigms, especially verb conjugations’.20 A number of manuscripts provide evidence that the vocabulary of grammar was well developed in Anglo-French by the end of the eleventh century. From that time forward, French was taught in England in a variety of ways to meet the requirements of different sections of the population. In the thirteenth century Walter of Bibbesworth produced his Tretiz, providing his patroness with the vocabulary necessary to run an estate in French, dealing with the parts of the body, of a cart or plough, items of clothing, the names of animals, plants, trees and flowers, the basic techniques of agriculture, house-building, brewing and so on. Copies and adaptations of this work are to be found up to the beginning of the fifteenth century.

On a different level, for an audience of prospective clerks or administrators in the broad sense, mainly young men based in town or city rather than on a country estate, a number of grammatical texts appeared, the later ones being detailed and complex, whilst the necessary vocabulary associated with these professions was taught from the fourteenth century onwards by the dictatores centred on Oxford who wrote specimen letters in Latin and French for their students to imitate. How much or how little spoken English was involved in these exercises must remain a matter of conjecture. A third strand in this teaching of French came with the Manieres de Langage from the later fourteenth and early fifteenth centuries. These consist of imaginary conversations between the English traveller in France and the people he meets in the course of his journey, offering a more colloquial, informal and, on occasion, vulgar vocabulary. All this didactic activity illustrates a widespread desire on the part of different sections of the upper stratum of the English population, even towards the end of the medieval period, to be able to use French in their daily lives.

Another link between French and English is provided by the great monastic houses, although it cannot be quantified in the present state of knowledge owing to the lack of available documentary evidence. As the thirteenth century advanced, religious orders from the continent, predominantly from la fille aînée de l’Église, established themselves up and down the land. To quote just one example: monks from Cîteaux, south of Dijon, spread out from their first house at Rievaulx across North Yorkshire, steadily building from one generation to the next a string of abbeys, of which Fountains is the jewel. The construction of these monasteries must have had a linguistic dimension, whether or not this can now be demonstrated in black and white. The extraction of the great quantities of stone necessary to build an abbey church of such a size and height as that at Fountains, with its nine side-chapels and extensive outbuildings, together with the fabrication of the tools necessary for the mining operation, the subsequent transportation of this mass of stone to the site, the work of the labourers and skilled masons, the splendid carved decoration now visible by modern techniques at the top of the building, not to mention the straightening of the little River Skell to meet the various demands of the monastery and create fish-ponds, with a fulling-mill coming later, all these operations would call for a large work-force possessed of a wide range of technical expertise.

The assembling of such a force must have outstripped both the available man-power of the monks themselves, even when reinforced by a body of lay brothers, and also their practical abilities. It would be reasonable to assume that both native labour and foreign professional skills were very probably b(r)ought in, French architects being associated with other ecclesiastical buildings in England during that period, as, for example, William of Sens who was responsible for rebuilding Canterbury Cathedral after the disastrous fire of 1271. At this distance in time it is impossible to make any informed judgement regarding the use of English and French during the process of construction, and it might be argued that the monks would have become ‘anglicised’ by constant contact with the inhabitants of the region. However, the rule of celibacy meant that new blood from abroad would have to be introduced on a regular basis if what were closely-knit monastic communities were not to die out. Ecclesiastical architectural terms such as ‘aisle, ‘chancel,’ ‘choir, ‘corbel’, etc., and place-names such as ‘Grange-over-Sands’ and the Cistercians’ ‘grange’ at Conistone on the Wharfe under Kilnsey Crag, show both the influence of French and also the semantic shift which often occurs when a ‘borrowed’ French term is taken over into English: the ‘granges’ in English were and are outlying estates, not just barns (cf. French dongeon and English ‘dungeon’ referred to later).

This link between French in England and the medieval monastic foundations was further strengthened when Cistercian houses such as Fountains, Rievaulx and Tintern became the leading exporters of wool across the Channel. Tintern owned its own ships, and access to the sea for the Yorkshire abbeys was facilitated by the fact that the French abbey of Meaux held the land where the river Hull fed into the Humber estuary. At a time when French was the language of trade, it would de rash to assume that it was not extensively used in this lucrative commercial activity which paid for the great Cistercian houses. An indication of this is provided by the fact that the finished articles that came back across the Channel carried the names of the French towns which had processed the wool, Middle English ‘chalons’ (Châlons-sur-Marne) or ‘arras’, terms apparently unknown to the French themselves, as was ‘blanket’, another product of France.

One of the most striking advances in our knowledge of the extent of Anglo-French in medieval Britain, and which is reflected in the new Dictionary, stems from the publication in 1991 of the three-volume Teaching and Learning Latin in Thirteenth-Century England by Tony Hunt. This work has many hundreds of glosses in Anglo-French and Middle English used by a wide variety of scribes to translate Latin texts by medieval authors such as John of Garland, Adam of Petit Pont and Alexander Nequam. The lexicographical value of the glosses (there are 3,154 citations of them in A–E of the second edition of the Dictionary) is greatly enhanced by their being the work of a large number of scribes writing in different parts of Britain and at different times, so that the range of possible vernacular equivalents for a particular Latin term is increased.

The value of the work lies not only in its undoubted important contribution to our knowledge of Anglo-French, but also in the light it sheds on the relationship between French and English. In many cases the different scribes use not only different terminology but different languages to translate the Latin, thus facilitating a comparison between the French and English terms used. In the vast majority of cases the glosses are straightforward, but it is by no means rare to find the scribe’s anglice followed by one or more French words.

For instance, in addition to glosses such as cicuta: gallice humbeloc, anglice herbe beneyt (ii 160), where the vernacular languages are simply reversed, or pompa: anglice boban (i 38), where boban is definitely French, or the form tricuspis: anglice treble pointe (ii 16), where both treble and pointe are French, sometimes the entry is more interesting. The Latin upapa is correctly glossed in English and French as ‘lapuhing, wype, vanel’, hupapa as ‘gallice vanele, anglice lepwinkel’(=lapwing), but then the shortened form pupa is glossed first as ‘gallice venele’, which means a ‘ginnel’ or ‘alley’ (both English words coming from French), instead of the correct ‘vanele’ (=lapwing), but then a second entry pupa is glossed as: ‘anglice puppe, gallice pupie’ (ii 20).

The medieval French poupee is attested in Godefroy’s Dictionnaire de l’ancienne langue française from the second half of the thirteenth century (6.351c and 10.391a), probably no earlier than the Anglo-French gloss here or the one in Cultura Neolatina 39 (1979), ‘puppa -pe fictio pannorum quam faciunt puelle s. puppe gallice’ (p.34), but no forms of it are attested in the OED until 1486, over two centuries later. It is highly unlikely that such a popular children’s word would have been ‘lost’ for this length of time and would then have suddenly resurfaced. The MED has a quotation for a plural ‘popettis’ from ‘a1500 (1413): ‘[…] as children make popettis for to play with whil thei be yong’, but this form is a diminutive, which would suggest the existence of an earlier standard form such as ‘pop(p)e’. The TLL attestation, then, is at least a century and a half earlier than the evidence in the dictionaries of English. Since later Latin has forms pupa and puppa in the sense of ‘doll’ found in Varro and Petronius, it would be reasonable to suggest that the forthcoming ‘P’ fascicle of the Dictionary of Medieval Latin from British Sources may well reveal the presence of the term in Britain much earlier than the Middle Ages. This is just one example amongst many which demonstrates that Teaching and Learning Latin needs to be studied in close detail by lexicographers working in English and Latin as well as in Anglo-French.

Such linguistic mixing is not confined to this one text, however. It was widespread from the twelfth century onwards, as may be seen in the DMLBS, where page after page provides evidence of French and English terms being ‘latinised’, an inevitable consequence of the inability of Latin to reflect the changed world that had developed over centuries after Classical Latin ceased to be a vernacular. Once the DMLBS is complete it will be possible for Anglicists to alter the dates of their first attestations of many English words hidden in a Latin disguise. However, whilst it is usually easy to recognise the artificial Latin, as in: […] mansionem […] in manum civitate reseisire et reassumere,21 or the English intruder in: que nul homme […] ne melte ascun metal […],22 it is sometimes difficult to say whether a particular form is to be regarded as French or English.

For instance, the linguistic status of estret(e) in the following diverse examples could well lead the lexicographer into error: en la haut estrete en la vyle de B.; deux hautz estretz […] & une place de gast ( l. degast[ee]) entre les deux estretz ; graunt noumbre des comunes […] furent esteauntz en les estretes pur luy vere; une autre schoppe de une chaundeler […] enmy le rwe del dit estrete; ou en les hautes estrés, ou en venelles.23 The meaning of estrete (‘street’) in these French contexts is not in doubt, and its forms do not identify it at first glance as being out of place in a French context, but it is unknown to the dictionaries of both medieval French and English. If it were found in only one instance, it might be possible to dismiss it as a random error made by an ignorant scribe, but the above examples come from a variety of texts, so it must be accepted as an Anglo-French term made from Middle English.

21 Chaucer Life-Records, eds Martin M. Crow and Clair C. Olson ( Oxford, 1996), 144.
22York Memorandum Book, ed. Maud Sellers, Surtees Society vol. 120 ( Durham, 1911) & 125 (1914), I, 94.
23 Respectively: Year Books of the reign of King Edward I, Rolls Series, 21-22, 55; Rotuli Parliamentorum, Record Commission (London, 1767-77), II, 84; The Anonimalle Chronicle 1333-81, ed. V.H. Galbraith, Publications of the University of Manchester, History Series 45 (Manchester, 1927), 41; ibid., 141; York Memorandum Book, I, 164.

Returning to the wider perspective of Anglo-French and Middle English in general, the absence to date of any published work based on the Middle English element in Teaching and Learning Latin leads to a consideration of the influence of Anglo-French on the overall lexis of literary Middle English. It is generally accepted that Chaucer makes extensive use of what are termed ‘French borrowings’, a practice that as a rule is put down to his French connections. However, he would be unlikely to discourage his compatriots from reading his work by using terms unfamiliar to them, so it must be assumed that his vocabulary, with its considerable French content, formed part of the lexis of Middle English in the later years of the fourteenth century. To see that such a mixed language was not peculiar to Chaucer alone, it suffices to look at the opening of Langland’s Piers Plowman, that most English of texts. In the first fifty lines the following ‘French’ words are found: seson, habite, heremite(s), banke, merveilouse, toure, dongeon, maner, wastours, glotonye, destruyeth, apparailed, contenaunce, disgised, prayers, penance, streyte (=estreit), ancres, selles, coveiten, likerous, plese, chosen, cheven (=French chever, cf. English ‘achieve’), mynstralles, iapers (i.e. japers), iangelers (i.e. jangelers), feynen (=French feindre), fantasies, foles, precheth, preve (=prove). If all the terms of French origin were to be removed from the works of these two writers, it would be very difficult to replace them by purely ‘English’ words without altering the sense of the texts: the Anglo-French terminology had been absorbed into the fabric of English, not ‘borrowed’. In other words, an important part of the legacy of the Conqueror was linguistic, the transformation of the vocabulary of English.

In the authoritative dictionaries of English, even in cases where the contribution of French to the lexis of modern English has been recognised, any mention of a French etymology for a word usually refers to the continental variety. The proportion of words said to derive from Anglo-French has up to the present been very small. Now that the new Anglo-Norman Dictionary is becoming not only available, but its contents electronically searchable on-line, many of the current etymologies given in the dictionaries of English will need to be altered to show a derivation from insular French. This is more than merely a change of label: it means that the Anglicist will be able to follow the history of many English words through the French used on both sides of the Channel and note any changes of meaning that came about in the process. It will be possible to show either a semantic continuity or a semantic divergence.

For example, the semantic development of forain/forein in both continental and insular French will be able to be charted in detail, showing a very large measure of convergence between the senses in the two forms of French during the medieval period, with Middle English absorbing the senses found in Anglo-French and so having a semantic profile similar to both varieties of French. Whilst Middle English passed on to the modern language most of its semantic content relating to ‘forein’, the situation in France itself is quite different. Successive dictionaries from the beginning of the seventeenth century onwards show the semantic range of forain diminishing until there is now little in common between the shrunken forain, its use restricted for all practical purposes to the vocabulary of the fairground, and its other former senses being largely replaced by étrange(r), and the widely-used English ‘foreign’. Anglo-French confirms that in this particular case the divergence between modern French and modern English has come about only after the medieval period.

On the other hand, a comparison of the modern French donjon with the English ‘dungeon’ indicates that here the parting of the ways was already developing in the medieval period. In the late thirteenth century Langtoft’s Chronicle states that ly roi descendist en un bas dongoun (ii 434), although at the end of the next century John of Gaunt ( gaunt 1, gaunt 2) was still using the word in the continental sense of ‘keep’ (like Chaucer around the same time). The Manuel des Pechez from the same period says that a rogue ought to be en dungun, which could be taken as meaning confined to the ‘keep’, or, more probably, down in the bowels of the keep – the ‘dungeon’, so the meaning of the word is to be determined by the context. The MED shows that, despite being attributed to ‘OF’ (i.e. continental medieval French), its ‘dongoun’ takes its current meaning from Anglo-French, because, whilst it attests the continental meaning of ‘keep’ up to the end of the fourteenth century, its quotation containing the modern English sense: ‘Now taken is Roberd & brouht unto prison, At Corue his kastelle sperd depe in a dongeon’ is dated ‘?a1400 (a1338)’. This sense could not have come from the continent, where it is not attested. The term mot(t)e shows a similar development. In continental French it developed from meaning ‘a clod (of earth)’ to a ‘mound’ from the twelfth century onwards, but only in Anglo-French did it take on the modern sense of ‘moat’, ‘defensive ditch’. Once again, the MED shows that the source of the modern sense is Anglo-French by providing two quotations dated ‘c1400 (a1376)’ and ‘c1400 (c1378)’ with the unambiguous meaning ‘defensive ditch, moat,’ unknown on the continent: ‘þe mot [var. moot, mote] is of mercy þe Maner al aboute, And alle þe wallis ben of wyt’ and ‘Conscience comaunded þo al crystene to delve And make a muche mote’.

All this new material means that the second edition of the Anglo-Norman Dictionary has a very different orientation from that of its predecessor. It is intended that its Anglo-French content should be viewed not as an isolated, defective imitation of francien, the imagined standard French of Paris, but as a language of civilization in its own right. Janus-like, on the one hand it illustrates the role of Anglo-French as an integral part of the civilization of Western Europe, and on the other hand it is capable of contributing substantially to the history of Middle English. It is hoped that it will become a standard work of reference for Anglicists and Latinists as well as for those working in the field of medieval French, either at home or abroad.

Yet there still remains much Anglo-French material which has not yet been exploited. The techniques and organization necessary for successful husbandry were set down in Anglo-French in about 1285 and published in 1971 by D. Oschinsky in Walter of Henley and other Treatises on Estate Management and Accounting, but similar detailed information in other technical areas such as the building of houses, churches or ships and the procedures used to make cloth, metal implements or beer is still lacking. In this respect the resources of the Public Record Office and the great libraries are far from having been exhausted. When, in 1952, Salzman wrote A Documentary History of Building in England down to 1540, he indicated in footnotes, sometimes ten or more to a page, the many unpublished manuscripts from which he had extracted his material, often in the form of a single word, a practice which meant that such isolated words could not be used for dictionary purposes without their supporting context. Moreover, it is not always clear from just the one word that he quotes whether the document from which it was taken was drawn up in Latin, French or English. In the intervening half-century since the publication of the book no attempt has been made to bring the material used by Salzman into the public domain and make it available to lexicographers working in Medieval Latin, Anglo-French or Middle English. Other cases in point include the Psalter Commentary referred to earlier and the extensive Account Rolls of Durham Abbey, of which only a small proportion were printed by Canon Fowler for the Surtees Society at the beginning of the twentieth century. Even this fraction of the documents yields a rich lexicological harvest, an indication of the potential value of the complete set of rolls.

Again, letters from the end of the thirteenth century preserved in the Chapter at Canterbury still await publication. Concerning the returns of London Guilds 1388-9, of which some sixty contain French, either alone or with Latin or English, Chambers and Daunt write that ‘None of the Latin or French returns has been published […]’ (p.275). The editor of The Little Red Book of Bristol, in which most of the documents are in French, states that ‘matters relating to foreign trade were regulated by a Gild of Merchants, the removal of whose early records from Bristol in the 17 th century, and the subsequent failure to discover them, cannot be too greatly deplored’ (Introduction, p.x). In the Introduction to his edition of The Black Book of Southampton (p.v.) A.B. Wallis Chapman writes that: ‘These ordinances and memoranda are for the most part written in French’, but he prints only one or two of them, as against scores of pages of Latin, thus completely distorting the linguistic balance of the documents. It is hoped that the appearance of the new Anglo-Norman Dictionary will lead to some at least of these silent witnesses to the use of Anglo-French in medieval England being brought into the public domain and thus made available to further enhance the lexicographical effort.

The computerisation of the AND began during the latter stages of the first edition, edited by William Rothwell, Louise W. Stone and T.B.W. Reid, with the assistance of a changing band of helpers over the years.24 If the Dictionary has made huge strides in the quantity and range of its coverage of the Anglo-French language, the changes in the working methods by which it has been compiled have been scarcely less significant. The advent of affordable personal computing as Fascicle 5 (P-Q) of the AND was in preparation had two major consequences for the first edition of the Dictionary which have been carried over – in much amplified form – into the second. Increasingly powerful word-processing systems greatly facilitated the drafting and revision of entries, eventually leading to the more fine-grained and user-friendly structural and layout conventions used in the new edition. A few years later came the ability to perform computerised analysis on a gradually accumulating corpus of digitised source texts. This made it possible for the first time for the editors to identify, and investigate systematically, lexical patterns which might otherwise have gone unnoticed. A third step-change has taken place since 2001 with the move to XML as the unifying data format for both Dictionary and corpus, now increasingly interlinked and mutually searchable on this website.25

24 The title page of AND1 reads as follows: ‘Edited by William Rothwell (1963–1992), †Louise W. Stone (1947–1973), †T.B.W. Reid (1973–1981) with the assistance of Dafydd Evans (F–Q), Stewart Gregory (R–Z), D.A.Trotter (R–Z), †Paul Staniforth (R–S)’. The preface printed in the final fascicle (7) outlines the history of the compilation of AND1.
25 See www.anglo-norman.net. For a useful survey of some of the issues from another (TLF) perspective, see Robert Martin, ‘Dictionnaire informatisé et traitement automatique de l’information lexicale’, in Peter Blumenthal et al. (eds), Lexikalische Analyse romanischer Sprachen (Tübingen, 1996), 63–72.

The great enrichment of source material brought about by the Collas bequest and the new non-literary source texts being edited and published throughout the 1980s meant that a more nuanced picture of the Anglo-Norman lexis was available and thus needed to be accommodated in the Dictionary. For more complex entries, the required level of semantic structuring was fast approaching the limits of what was reliably achievable by conventional means. In 1986, when prendre was being redrafted for the nth time by WR using the traditional tools of typewriter, knife, cutting board and pot of glue, it was realised that a more modern method of working was needed – and one had fortunately become available. The purchase of an Amstrad PCW (the UK’s first mass-market word-processor, which ran Locoscript under the CP/M operating system) allowed prendre, subsequent entries in P and the whole of Q, R and S to be drafted and revised in a newly flexible way, and, almost as importantly, to be printed out (albeit slowly) with fully apparent formatting (boldface, italics, superscript etc.) for enhanced ease and accuracy of proofreading. When the time came to move on to IBM PC-compatible machines, several generations of which were provided for the use of the editors by the MHRA, the PCW data were converted to an early format of Microsoft Word for MS-DOS thanks to a program devised by Michael Beddow, at that time Head of German at the University of Leeds. Electronic mail gradually became the routine medium of exchange between the editors working in different parts of the UK, with drafts of entries, then whole letters, moving backwards and forwards as file attachments, before eventually being sent in the same form to the printer.

Shortly after the French Department in Leeds purchased an HP Scanjet with Optical Character Recognition software, Michael Beddow drew the editors’ attention to John Bradley and Lidio Presutti’s new TACT text retrieval package (1990), published by the Centre for Computing in the Humanities at the University of Toronto. It was immediately apparent that the ability to scan and digitise printed editions of A-N texts and then subject their lexis to systematic analysis using TACT (version 1.2, then 2.0) offered a major enhancement to the editorial toolkit. Whereas in the past source texts had generally been combed once only, in linear fashion, and slips produced to capture any items perceived to be of lexicographical interest – an inherently uncertain process, and one whose vagaries were becoming increasingly apparent from lacunae in coverage in AND1 – computerised concordancing opened up the prospect of being able to revisit a text at any time and in relation to any word, to test a hypothesis or search for an example of usage. The computer, moreover, proved markedly more reliable than human editors. Equally important, TACT made it possible for the first time to examine collocations systematically, and therefore to identify locutions and establish their range of meanings more reliably, on the basis of frequency. Work therefore began, initially in an ad-hoc manner, on the construction of an electronic corpus for the use of editors working on stand-alone PCs, the first fruits of which were fed into the drafts of the final Fascicle of AND1 (T-Z).

The earliest corpus component was W. Rothwell’s edition of Bibbesworth’s Tretiz, which had been originated using the Amstrad PCW and so was already available in digital form. Gradually other texts were digitised and added to the corpus, increasing not just its size, but also its coverage of different lexical domains. Corpus construction continued beyond the completion of AND1 in 1992 since the MHRA had already taken the decision to encourage a revised edition of the Dictionary, providing financial and other support for several years thereafter. A number of individuals and projects contributed greatly to this enterprise by making electronic versions of their own editions available to the editors, prominent among them Pierre Nobel (Poème anglo-normand sur l’Ancien Testament), Tony Hunt (Anglo-Norman Medicine, Popular Medicine), and Chris Given-Wilson, director of the Leverhulme-funded project to re-edit the Rotuli Parliamentorum.26 By offering to the Dictionary the complete electronic text of the original edition that had been re-keyed for his project, he opened up to systematic scrutiny the whole of this vital text. In the late 1990s, TACT’s place in the project was taken by the more user-friendly Concordance package (currently in version 3.2) authored by Dr Rob Watt at the University of Dundee.27

Since April 2002, development of the AND and of its associated corpus has been supported by a series of grants from the UK’s Arts and Humanities Research Council. In 2002–2004, the award of a first Resource Enhancement Grant funded Phase 1 of the creation of the Anglo-Norman On-Line Hub, in which letters A–E of the draft revised Dictionary, converted to TEI-compliant XML, were mounted on the Web alongside a corpus of digitised source texts (including previously unpublished editions)28 and both republished and new academic articles on Anglo-Norman.29 Michael Beddow, now in the official capacity of Technical Consultant to the project, devised and implemented the innovative data-handling techniques which continue to put the Hub at the forefront of international developments in Humanities Computing in this area. A full-time Technical Support Officer, Siân Pilborough, was appointed to the corpus side of the project. A Major Research Grant for 2003–2007 is funding the continuation of the AND revision, specifically F–H, and has allowed the project to engage two full-time postdoctoral researchers, Virginie Derrien and Geert De Wilde, in addition to making new calls on the technical expertise of Michael Beddow. A second Resource Enhancement grant in 2004–2007 for Phase 2 of the Hub has further extended the remit of the Technical Consultant, at the same time allowing Siân Pilborough’s post to be continued and Tom Richens to be appointed as a second, full-time TSO. From the start of Phase 1 the joint project was divided between Aberystwyth (Dictionary) and Swansea (corpus); the Dictionary continues in Aberystwyth under the editorial direction of David Trotter, and the corpus is being taken forward on a full-time basis by the team at Swansea University led by Andrew Rothwell. WR, now also based in Swansea and holding Honorary Professorships at both Universities, continues to provide linguistic expertise, assistance and general guidance to the project, and remains its spiritus rector.

26 PROME (The Parliament Rolls of Medieval England), a new edition with translation of the Rotuli Parliamentorum, the official records of the English parliament from 1278 to 1503.
27 www.concordancesoftware.co.uk/
28 The first of these is an edition of Femina by W. Rothwell, to replace the inaccessible and inadequate version published by Wright nearly a hundred years ago.
29 The Hub published in 2004 its first new article, ‘ Ignorant scribe and learned editor: Patterns of textual error in editions of Anglo-French texts’, by W. Rothwell.

In addition to its role as the platform for delivery of the on-line version of the AND and for future work on the revision of the Dictionary, the Anglo-Norman On-Line Hub ( www.anglo-norman.net) is conceived as an expanding, fully-integrated and searchable resource for scholarly research into the Anglo-Norman language, the society which used it and the culture to which it gave rise. The dataset is held in TEI-compliant XML on a number of mirroring servers in the UK and abroad, and the storage and retrieval software used is open source, to ensure the longevity and sustainability of the data. The on-line corpus is currently in a phase of rapid expansion and already contains a large body of works from a variety of registers. Permission has kindly been granted by the Anglo-Norman Text Society for the digitisation and mounting on the Hub of all of its published editions, which is underway. This new material is being made available to the editorial team at Aberystwyth as they advance the revision of the Dictionary. An additional benefit of structural XML markup of the Dictionary is that the many thousands of citations contained in both first edition and the revision are themselves able to be searched for new lexical information, making the AND into its own mini-corpus. An on-line citation bank has recently been developed by Michael Beddow to allow the editors and other interested users to retrieve from AND citations the complete list of any word-forms of interest to them, together with the number of occurrences; each record can then be expanded to give the full text of the relevant citations, with a hyperlink leading automatically to the Dictionary entries in which they occur. It will soon also be possible to incorporate new citation material into the bank as it is discovered, ready for it to be used in future editorial work, thereby replacing paper fiches (on which such gleanings have still tended to be recorded) by a system that makes them instantly available on-line while also avoiding the need for subsequent rekeying. This is now, therefore, a truly corpus-based lexicographical enterprise, with the added advantage that its corpus is not only freely accessible to the wider research and teaching community, it is also increasingly interlinked with the on-line Dictionary itself. Thus it will soon be possible to click on a word in the corpus to look it up automatically in the Dictionary, a facility that is already available when perusing citations in the Dictionary itself, just as it is already possible (with texts whose digitisation on the site is already complete) to go from a Dictionary citation directly to the full text of the source at the appropriate location. Once other cognate dictionaries move over to XML, they will also be able to access AND2 data via their own interface, and vice-versa from the AND.

Before work began on revising the AND, the non-electronic part of the first edition (A-P) was scanned and digitised using the Kurzweil KDEM machine in the Humanities Computing Unit at King’s College London, through the good offices of MHRA Honorary Treasurer, Roy Wisbey. In addition to forming the starting point of the revision, the electronic files thus produced have now been checked for scanning errors against the print version, converted to the same TEI-conformant XML as is used for the AND2 material, and added to the on-line version of the Dictionary, so that it provides A-Z coverage of Anglo-Norman in an evolving hybrid of the first and second editions of the Dictionary. Ongoing editorial work (currently the revision of letters F-H) is now carried out entirely in XML, from which standard word-processor output can be generated for printing as needed. The process of initial conversion to XML of AND2 A-E, originally drafted in Microsoft Word for Windows, and the ongoing conversion of the later parts of AND1, has in itself led to the discovery of various formatting and presentational inconsistencies which inevitably accumulate when such complex data are being worked on repeatedly and by different editors over long periods of time, but which can be almost impossible to identify by visual inspection alone. The discipline of structural description and validation of the data enforced by XML has led to a more rigorous construction of many entries in the new edition. The form and formatting of sigla have also been regularised and they are no longer retyped manually, but selected from a menu linked to the canonical List of Texts, also held in XML. The numbering of senses and generation of summary boxes for the more complex articles are handled entirely automatically, which removes at a stroke a major potential source of error and allows the proof-reading task to focus largely on linguistic rather than presentational issues. Thus the migration of AND2 data from word processor format to structurally-constrained XML, initially a requirement for on-line delivery, has also brought with it a significant increase in the accuracy of the mechanics of the present printed edition.30

The choice of headword is largely determined by frequency (with the proviso that there is no scientific method of determining this with reliability), but the aim has also been to ensure that related words are adjacent. Square brackets indicate reconstructed or hypothetical forms, in some cases (e.g.: [aidement]) to ensure that words are not detached from their ‘family’ (in this case, aid- not eid-). Verbs are an exception to the general rule that square brackets are always used for hypothetical forms: unattested (when obvious) infinitive forms are used even if only finite forms have been encountered. Normal (round) brackets surrounding a headword form indicate that the editors have serious reservations about the reality of the form so listed.

Articles indicate first the part of speech (v.a., s., pron., etc.), then supply a gloss (italicised) and a quotation or quotations (in roman) illustrating that sense. Each quotation is followed by an abbreviated textual reference which is decoded in the List of Texts. Double clicking the mouse pointer on this reference will produce a window giving the List of Texts entry for the item concerned, together with a link that allows all citations from this source to be viewed. Where available, a link will also be given to the DEAF entry for the same source. The List of Texts also in many cases supplies an approximate date.

In the case of (e.g.) verbs, separate paragraphs deal with (e.g.) v.a., v.n., v.impers., etc.
Locutions follow main entries in a separate paragraph.

Entries are subdivided by numbered main senses and sub-senses are divided by the use of the symbol ♦. Numbered senses may be regarded as semantically distinct (though related) whereas sub-senses are gradations within one sense.

In the case of the more complex entries, the body of the article is preceded by a‘summary box’ of the senses which reproduces the glosses, and retains the main sense numbering, while replacing the ♦ symbol (for sub-senses) by a semi-colon. This is intended to allow readers more rapidly to see the organisation of the longer entries. Each listed gloss is a hyperlink to the portion of the entry concerned

The List of Texts provides the key to the bibliography of sources used. It decodes the short form used for each source text: on-line users can access the full List of Texts entry for any cited source by double-clicking on the siglum following the actual citation in the entry body, which will also provide a direct link to the corresponding DEAF entry (if present). In addition, it provides (where possible and appropriate) a date for the source, and the corresponding text siglum used in the Dictionnaire étymologique de l’ancien français (DEAF), whose system is now widely used as the industry standard, and which gives access to the critical bibliography which may now be consulted as DEAFBiblEl at the following internet address: http://www.deaf-page.de/. In this on-Line version of the AND, each List of Texts item for which there is a DEAF entry is accompanied by a hyperlink leading directly to the corresponding DEAF data, as well as a link which will show all citations from the source concerned that are currently to be found in AND entries.

Dates cannot always be provided (in the case for example of sources which are in effect compilations extending over many years, and sometimes over several centuries) and are often therefore necessarily imprecise. They are designed solely to assist readers in very approximately dating attestations and they rely heavily on other authorities. In particular, the editors gratefully acknowledge the use which they have made (in supplying these dates) of both the DEAF, and of Ruth J. Dean, Anglo-Norman Literature: A Guide to Texts and Manuscripts ( London, 1999). In some cases, on-line readers will find additional assistance in dating a citation (or an explanation of why it cannot be reliably dated at all) if they follow the link to the corresponding DEAF entry in the List of Texts.