A Brief History of the "Charrette Project"
and Its Basic Rationale

Le Chevalier de la Charrette (Lancelot)

The Old French narrative romance, Le Chevalier de la Charrette (Lancelot), composed by Chrétien de Troyes around 1180, tells the tale of Lancelot and his love for King Arthur's wife Guenevere. This seminal text was recast into the Old French Prose Lancelot in the 13th century, the primary source for Mallory's Morte d'Arthur, which in turn has been the source of modern retellings of the Arthurian legend (in English), including Tennyson's Idyls of the King, White's Once and Future King, and Renault's popular novels.

Modern edited versions of the romance contain approximately 7100 verses. These editions are prepared from as many as eight manuscripts dating from the 13th century:

MS A = Chantilly, Musée Condé 472
MS C = ("Guiot"), Paris, Bibliothèque Nationale de France, fonds français 794
MS E = Escorial, Real Monasterio de San Lorenzo M.iii.21
MS F = Paris, Bibliothèque Nationale de France, f. fr. 1450
MS G = Princeton, Firestone Library, Garrett 125
MS I = Paris, Bibliothèque de l'Institut de France 6138 (formerly 4676)
MS T = Paris, Bibliothèque Nationale de France, f. fr. 12560
MS V = Vatican, Biblioteca Vaticana, Regina 1725

Goals and Purposes

During the late 1980s, while the late Alfred Foulet and I were preparing our critical edition of Chrétien de Troyes's Le Chevalier de la Charrette (Lancelot), published in 1989 by Bordas (Paris) in the Classiques Garnier series, it became clear to me that working with a word processor like WordPerfect was greatly superior to typing (and retyping), as well as "whiting out" errors. The note-taking program known as Tornado Notes also proved useful as a means of recording ideas and significant variants as these emerged, the way they tend to do, in ad hoc fashion. By the time our edition, with Introduction, Bibliography, Notes, Rejected Readings, and various Tables, was ready for the printer, it occurred to me that "computerization" might well offer a highly practical means for setting down in transcribed form an entire manuscript tradition like that of the eight "complete" and/or "fragmentary" manuscripts constituting the surviving Old French tradition of the Charrette. Moreover, it would technically be relatively easy to enrich these diplomatic transcriptions with commentary, e.g., markers pointing out various poetico-rhetorical devices characteristic of Chrétien de Troyes's style (rich rhyme, chiasmus, adnominatio, etc.). Non-ASCII scribal characters--e.g., barred "p," abbreviated "Que" or "Qui," nasalized vowels--could be rendered by predetermined ASCII codes. Word boundaries would replicate those of the various scribes.

Since from the start our work coincided with the gradual imposition of Standard General Markup Language (SGML), alongside the specifications included in the internationally sponsored Text Encoding Initiative, as the norm for electronically recorded text, and since I believed it was important that what we were doing would not be rendered quickly obsolete by being tied too closely to any specific commercial word processing program, we decided that our transcriptions would be from the start couched in SGML-TEI format, despite the ugly--even virtual unreadability--of this format. We would also include an electronic (and more readable) ASCII version of the Foulet-Uitti edition of Chrétien's poem. Foulet-Uitti line numberings would also be employed in the case of each manuscript transcription so as to permit quick comparison among the various versions, despite their discrepancies.

Work on the transcriptions was begun in 1990. A small group of Princeton University graduate students in Romance Languages and Literatures interested in paleographical issues and in furthering their computing skills began to transcribe several of the manuscripts in WordPefect 5.1 (DOS) or Word for Macintosh format. Their work has been (and has continued to be) closely supervised by Professor Gina Greco, of Portland State University (Oregon). While participating in practically all the various aspects of the Project, Professor Greco has assumed most of the responsibility for the accuracy of these crucial transcriptions.

Although our applications for funding from the National Endowment for the Humanities proved to be unsuccessful, we were able to persevere thanks to the generosity of the Princeton University Committee for Research in the Humanities and Social Sciences, to that of several funding awards provided by Portland State University, and above all to the Alfred Foulet Publications Fund (established by a bequest made by Alfred Foulet in 1987 in order to facilitate Princeton University Romance Languages and Literatures research and publication in medieval language and literature).

I should like to record our gratitude to Professors J. Lionel Gossman and François Rigolot, chairmen of Princeton's Department of Romance Languages and Literatures, as well as to the administrative officers of Portland State University for their consistent moral and material support.

Lack of a major, over-arching Project grant has slowed our progress somewhat, however we have been able to move ahead over the years. Professor Greco has been able to spend several weeks each summer in Princeton, and has recently spent time at the Bibliothèque Nationale in Paris proof-reading our transcriptions copy; I have travelled to the necessary European libraries (Paris, Chantilly, Vatican and El Escorial) in order to secure digitizable (and quite legible) copies of the necessary manuscripts, as well as to inform interested colleagues (especially in France, Spain and Italy) of what we have been doing. We have been in a position to provide small stipends to our graduate student assistants.

From its very inception the Charrette Project has benefited from the enthusiastic support of Princeton University's Computing and Information Technology center (CIT), largely thanks to the sections headed by Jacqueline Brown. Humanities Computing specialist--now in Systems Support--Toby Paff, a trained linguist and a thoughtful expert in information retrieval and analysis, has helped us in ways too numerous to mention; he has accepted to be an Associate Editor of our Project, along with Dr. G.J. Murphy of the Rutgers-Princeton Center for Electronic Texts in the Humanities (CETH). Meanwhile, his successor as Humanities Computing specialist, Dr. Peter Batke, has taken on the task of general computing adviser to the Project, with a special interest in hypertext, graphics and in PERL-based textual analysis. Ph.D. candidate Peter Shoemaker completes the roster of Associate Editors, with a general responsibility for SGML-TEI and for HTML presentation, as well as helping in all of the Project's dimensions. (The Charrette Project has been developing closer links to the above-mentioned CETH organization in the persons of Dr. Murphy and its present Director, Susan Hockey; we hope that these links will be further strengthened.)

In early 1991 a specimen, or "demo," of our Project was launched on the World Wide Web (URL: http://www.princeton.edu/~lancelot). This specimen displayed a number of proof-read transcribed text files from various manuscripts, including a complete version of the important Charrette fragment preserved in MS Princeton University Library Garrett 125 (G), with digitized colored images of each manuscript folio accompanying the corresponding transcriptions. (By 1991 it had become clear that it was technically feasible to include with our SGML-TEI textual transcriptions of the manuscript tradition--i.e., with the central core of our database--folio-identified and line-numbered images of what our transcriptions represented.) Including these with the transcriptions the Project would allow textualists and paleographers to control our readings against the "originals," and modify the former according to their own lights. (Our procedure might well also contribute to preserve the manuscripts themelves from the effects of repeated handling by scholars and students--from otherwise inevitable wear and tear.) A computerized "cross-indexing" program was also included in our specimen, which permits immediate comparison of individual manuscript readings (e.g., v. 15 in MS G and its counterpart in, say, MSS E, T, and/or C).

In addition to the above our "demo" also provided a brief Introduction to Chrétien's romance, short biographies of the members of our editorial team, keys to the ASCII codes utilized in transcribing non-ASCII scribal characters, and two very tentative search and analysis programs (subsequently discarded). By Spring 1997 we will have placed on line proof-read SGML-TEI transcriptions of the entire textual corpus, with the exception of the extremely fragmentary MS Institut de France 6138 (formerly 4676) (I), as well as digitized and line-numbered images of the relevant manuscripts.

We have adopted the SPIRES format for our database and, for our present rather experimental purposes, we are using the highly restricted PAT program as a search engine.

Interest so far in our WWW "Charrette Project" specimen has been encouraging. "Hits" on our "demo" have come from a great variety of places, and it has been "hot-keyed" by a large number of other WWW sites. Commentary on our progress has been favorable, and has focused upon the Project's scholarly rigor and thoroughness. By late 1994 we decided to contact various French colleagues--specialists in medieval literature, in Old French textuality, manuscript curators, historians of French, and, in general, scholars interested in French medieval matters and computerization.

Contact with French Colleagues and the École Normale Supérieure (Fontenay/Saint Cloud) Colloquium (May 1995)

In late May 1995, thanks to the good offices of Professor Christiane Marchello-Nizia, of the É.N.S.-Fontenay/Saint-Cloud, an afternoon colloquium took place at Fontenay-aux-Roses. In addition to myself, Peter Shoemaker, Layla Ahsan Roesler and Daniel Solovay (from Princeton), those attending included (among others) Christiane Marchello-Nizia (Chair); Professors Gabriel Bianciotto and René Pellen, of the Centre d'Études Supérieures de Civilisation Médiévale (Poitiers); Professor Simone Monsonégo, of the Université de Nancy II (Unité de Recherche sur le Français médiéval) and of the Institut National de la Langue Française (INaLF); Professor Geneviève Hasenohr, of the Institut de Recherche et d'Histoire des Textes (Section Romane), with her colleague, Claude Rabel, of the Orléans branch of the Institut; Professor Jérôme Baschet, of the École des Hautes Études en Sciences Sociales; Professor Bernard Cerquiglini, Director of the U.F.R. de linguistique at the Université de Paris VII-Denis Diderot; Professor Claude Cazalé Bérard, of the Université de Lille 3 (Italian); and Professor Fernande Dupuis, of the Université du Québec (Montréal). Several of the afore-named presented papers on electronic projects upon which they were engaged, all of them of very considerable interest to us, while I was able to explain to the entire group the nature and progress to date of our Charrette Project. Discussion at that meeting and later correspondence proved to be encouraging also. It was clear to most of us that a longer and more detailed prise de contact involving French and American specialists was very much in order. The date 23-27 March 1997 was set for a full-fledged colloquium at Princeton University.

In addition to those named above, I also met with a number of colleagues who were unable to attend the Fontenay meeting: Professor Emeritus Jacques Chaurand, a specialist in French dialectology and language history (Université de Reims); Dr. Frédéric Vergne, director of the Chantilly Museum library and a specialist in ORACLE-based data organization; and Professor Michel Zink, who occupies the Chair of Medieval French Literatures at the Collège de France. Finally, this summer I met at some length with Professor Robert Martin, head of the INaLF office in Paris and director of the groundbreaking lexicological project devoted to Middle French, Le Dictionnaire du moyen français.

After open and (sometimes) quite lengthy discussions with these French colleagues, I am persuaded that our Project is of particular significance to such French institutions as INaLF, the Poitiers Centre d'Études Supérieures de Civilisation Médiévale, the C.N.R.S.-sponsored Institut de Recherches et d'Histoire des Textes, as well as to various on-going French research projects, for example, the above-named Dictionnaire du moyen français, the Université de Paris VII-based project titled Mutation de l'édition induite par le livre électronique (directed by Professor B. Cerquiglini and Jean-Louis Lebrave, of the C.N.R.S.), and the historical-image research undertaken by Professors Jean-Claude Schmitt, Jérôme Baschet, and others at the École des Hautes Études. It also seems to me that our Project's collaboration with French colleagues will serve to build bridges between them and such on-going United States-based activities as CETH and ARTFL. Scholarship on both sides of the Atlantic stands to gain from collaboration, in many ways--some surely unforeseen at the present juncture.

Let me add that the pedagogical implications of our experience should not go unmentioned. To continue to train students in Old French paleography and textual criticism without encouraging them to take full advantage of present electronic technologies--technologies with which the younger generations of today are far more comfortable than many of their elders--would, quite simply, be unconscionable.

What, it may be asked, does The Charrette Project have to offer at his particular point in time?

First of all, our Project provides a substantial linguistic and literary database: the complete surviving manuscript tradition of one of the major works--a corpus of some thirty to thirty-five thousand verse lines--by Chrétien de Troyes, who is arguably medieval France's most important narrative poet and who both lived and wrote during the "classical period" of literary Old French. Moreoever, unlike modern editions which utilize manuscripts in order to "recover," and print, the lost (presumed) holograph of its "author," our database presents an authentically medieval textual reality: Chrétien's "text" as it really existed within the dynamics of an Old French literary production extending from about 1215 to the close of the 13th century. We have preserved the manuscripts' punctuation (where it exists), and this might well offer clues as to how the texts were read out loud--a clearer grasp of their oral dimension; and, as stated above, we have done our best to replicate word separations as the diverse manuscripts have given them, which should furnish data as to the nature of the word, or lexical item, in Old French. Our poetico-rhetorical commentary tends to prove that a manuscript like the late-13th-century MS G is often more "faithful" to the 12th-century author's poetics than the much touted, and earlier, Guiot text (MS C). Interesting also for the medieval vernacular notion of authorship is the significantly higher incidence of hapaxes in the sections said to have been composed (with Chrétien's approval) by Geoffroi de Leigni when compared with that portion of the romance attributed to Chrétien himself. The Project database also provides a veritable laboratory for the study of the relationship between the (largely) francien-based Old French scripta and its actualization in dialect-colored (and largely oral?) manuscript redactions (especially Old Picard and champenois).

The Charrette Project has eliminated the need for a simple binarism of text vs. variant(s)--a requirement, it would appear, for modern printed editions. On the contrary, even in the case of so careful and poetically self-conscious an author as Chrétien de Troyes it seems that textuality comprises "variant readings." Thus, the copyist Guiot, whose interests (as displayed in the much praised MS Bibliothèque Nationale fonds français 794) were largely historiographical (and not poetic), did not perhaps so much "betray" Chrétien the artist as exploit what, in Chrétien (along with Wace and the matière de Rome), specifically served his main concerns. In a wider sense, then, a textual tradition like that of the Charrette represents the surviving corpus of intertextual relationships orbiting around the core of Chrétien's roman--the work, composed of approximately 7100 verses in rimes plates--and this fact has a number of important poetico-literary repercussions, including the undeniable reality that the Chevalier de la Charrette remained poetically relevant to a late-13th-century Picard-speaking audience, presumably an audience not in all ways similar to that of Marie de Champagne's court in 1180. (Other, clearly related intertextual dimensions include those connecting the Lancelot to such other rhymed narratives as the various Tristan et Iseut redactions, the romances pertaining to the matière antique, and even to Chrétien's own oeuvre: Yvain and the Conte du Graal, not to mention the prose redactions of the Vulgate cycle.)

Present and Future Possibilities

The Charrette Project suggests various possibilities for continued research in the present moment and for the future. Thus, work confined to the Chevalier de la Charrette is, itself, far from "complete," and, as time goes on, will certainly require further refinements and improvements; indeed, one suspects, "completeness" is beside the point, and, here also, the Project recalls the medieval notion of textual tradition which, though of course finite, can never be declared "complete." However, the core database--transcriptions and images--will remain to all intents and purposes constant, subject principally to the correction of errors and to the improvement of the images with the passage of time.

Still our Project possesses a certain model value. One would hope that the other works making up Chrétien's oeuvre romanesque, along with at least several other 12th-century "courtly" narratives like those of Marie de France, the Floire et Blanchefleur corpus, several romans antiques, and certain epigonal romances like the Perceval Continuations, Fergus and the Bel Inconnu could eventually be digitized and "imaged" according to our Charrette model--perhaps, also a representative selection from the Roman de la Rose tradition.

Were such a corpus to be established, particularly with the addition of several saints' lives, two or three chansons de geste, material from the Roman de Renard and from the fabliaux, sufficient documentation would be at hand for serious computerized study of the Old French literary language and scripta, in their relationship to the spoken language--indeed, of what I should be tempted to label the "Old French vernacular." One recalls Lucien Foulet's trailblazing masterpiece, Petite Syntaxte de l'ancien français (1923)--the first work to consider Old French as a language in its own right--which was based on documentation much more slender than that proposed above, and for which the resources offered by computerization were of course not available.

Ever since the days of early nineteenth-century Romantic scholarship--i.e., for close to two centuries now--debate has raged among Romanists as to the "orality" or the "written nature" of medieval French literature, particularly in respect to certain genres like the chansons de geste. At times these debates have been acerbic--one suspects, because of the "either/or" character of the thinking expressed. For the Central European Romantics and their early French disciples (like Gaston Paris), the Old French chansons de geste were to be considered as authentic expressions of the "voice of the people"--often of its Germanic stratum--articulated, as it happened, in Romance form. Thus, the epic songs were improvised virtually on the field of the battles which they celebrated, and later underwent revision and lengthening in order to assume narrative shape. Only the later, revised songs came to be written down, yet they are the product of original, oral poems or, to use Gaston Paris's term, epic cantilènes created by anonymous bards. During the early decades of the 20th century this view was challenged by Joseph Bédier, who argued that a 12th-century chanson de geste in Old French was exactly that: a 12th-century French poetic artifact, corresponding to the new late-11th-century fervor for massive pilgrimages across Europe, especially to Santiago de Compostela, and to the implementation of Pope Urban II's call for Latin crusading activity in the Muslim-held Holy Land--an activity he described as gesta Dei per francos.

Except for a few rare, mostly French, scholars influenced by the pioneering work of Edmond Faral, the important rôle played by what deserves to be labelled "bookishness" in this whole literary process, which ought to include also hagiography (e.g., the La Vie de saint Alexis) and, even quite self-consciously, the mid-12th-century emergence of romance narrative, was somewhat downplayed. By and large scholars hailing from nations peripheral to France--the Spanish (don Ramón Menéndez Pidal, with his theory of poesía latente), the Swiss (with their tradition of interest in Romance linguistic geography, dialectology and lexicology), and, to a somewhat lesser degree, Belgians--like the Germans, have tended to look with favor on "orality." (One thinks of the late Jean Rychner and his work on epic formulas and on the fabliaux, as well as on the late Paul Zumthor's fascination with African modes of story-telling associated with griots--both of these distinguished scholars were Swiss.) However, it has been suggested that very few indispensable structural elements central to the Oxford Chanson de Roland are lacking in Einhard's Vita Karoli (Uitti 1993). Analagously, the Old French Vie de saint Alexis exists as the vernacular resolution of, and as a commentary on, the much older Vita sancti Alexii (to be found in close to intact form in the Bollandist repertory of Latin saints' lives). That the mid-12th-century Roman d'Énéas was based on Virgil's Æneid has, of course, never been denied, nor has the pervasive presence of Ovidian values and procedures in this work ever been ignored. Yet, only rather recently have we begun to understand the importance of truly 12th-century vernacular values permeating this great poem: courtoisie, for example, in the treatment it gives to Virgil's beautiful bellatrix, Camilla, in her one-on-one combat with the disgracefully discourteous Tarchon, and in its depiction of the very touching man-woman friendship--the first of its kind in Western literature, to my knowledge--between Camilla and Turnus. The "vernacular values" we associate with courtoisie necessarily have their counterparts in speech constructs. The Roman d'Énéas thus fuses, so to speak, seamlessly Classical bookishness (the possession of clergie) with the essentially vernacular values of that form of speech we call "courtly diction." Similarly, the Oxford Roland combines the history of "our emperor"--Charlemagne--as reported by Einhard with the historical needs and aspirations of late-11th-century Capetian France. The Roland articulates both the legitimacy of that douce France and its identity.

Consequently, in my view, "orality" and "writing," when opposed antithetically to one another, come very close to resembling a red herring; in significant ways this opposition flies in the face of literary and spoken historical fact. In a very profound sense, the spoken French of the 12th and 13th centuries--like the language of the vernacular scripta, which, itself, is manifested in manuscripts usually colored (teintés) by the dialect of those for whom it was written (and to whom it was read aloud) and/or of the copyist, i.e., "oral" features--is inextricably entwined with book-based usages and "colors."

The issues I have just mentioned bear looking into seriously. Programs allowing the comparison of indirect speech of given characters with examples of oratio recta could easily be devised, as could search programs identifying examples of complex verbal constructs, or of various sorts of inversion. One thinks of the written/oral discourse attributed to the narrator figure in poems like the Alexis where present and past tense often appear to be "mixed up" in ways startlingly similar to very colloquial story-telling, even in our own day. Also, one might query, to what degree, and how, is speech possibly determined by gender? By what means does an "illustrious vernacular" come into being over a time? Evidence tending to prove the existence of such a concept in the period that interests us may be found in a number of Chrétien de Troyes's successors who do not stint in their praise of his French language. Written and spoken discourse mutually authenticate one another. (One is reminded of Erich Auerbach's sensitive derivation of a new, Romance sublime style from the old Christian sermo humilis of Antiquity.) Also, computer-assisted analysis of specimens of high courtly diction in comparison with examples of the systematic reversals of such diction in the fabliaux would be highly instructive. After all, a fabliau like "La Fille qui ne voulait point entendre parler de foutre" specifically demonstrates the 13th-century consciousness of this humorous, and sometimes rather touching, play of reversal.

What is required in order to bring off a properly articulated understanding of the "Old French" which I have been attempting here briefly to describe is a philologically-based and imaginatively oriented team effort calling upon language historians--lexicologists, syntacticians, phonologists, dialectologists--and specialists in the styles and languages of Old French narrative and lyric poetry. I believe that a satisfactory preliminary grasp of the conditions of verse must antecede, both poetically and linguistically, the devising of interesting approaches to prose. I understand here "verse" and "prose" as being both of intrinsic and of documentary interest. Codicologists, art historians, and historians tout court of the French Middle Ages--I am thinking of colleagues working under the auspices of such important institutions as the Institut de Recherche et d'Histoire des Textes and the Centre d'Études Supérieures de Civilisation Médiévale, as well as the École des Hautes Études--should from the start be associated with the enterprise which, in my view, ought to be multinational in scope, but founded on a specific collaboration between scholars from the United States and from France, with centers at Princeton and in France, at Poitiers, at the Centre d'Études Supérieures de Civilisation Médiévale.

Within this much larger context, our Charrette Project constitutes but a small initial step--a point of departure for other--similar or quite different--efforts that one by one, but taken together, will fully integrate into the domain of French studies that beautiful, and foundational, treasure we call "Old French." For was not also the vernacular language spoken in the lands that gave us, in the 12th and 13th centuries, Gothic architecture, Scholastic philosophy, the university, the beginnings of the nation state, and so much else that remains central to our own ways of being at the very heart of these gifts? To claim otherwise would be quite unreasonable.

Scholarly Works Cited

Auerbach, Erich. Literatursprache und Publikum in der lateinischen Spätantike und im Mittelalter (Berne: Francke, 1958) [Eng. trans., Ralph Manheim, Literary Language and Its Public in Late Latin Antiquity and in the Middle Ages (Princeton: Princeton University Press, 1965)]

Bédier, Joseph. Les Légendes épiques. Recherches sur la formation des chansons de geste. 4 vols. (Paris: Champion, 1908-13 [3rd ed. 1926-29]).

Faral, Edmond. Recherches sur les sources latines des contes et romans courtois du moyen âge (Paris: Champion, 1913).

Foulet, Lucien. Petite syntaxe de l'ancien français. 3rd rev. ed. (Paris: Champion, 1930 [1st ed. 1923]).

Menéndez Pidal, Ramón. La "Chanson de Roland" y el neotradicionalismo (Madrid: Espasa Calpe, 1959) [French trans., I. Cluzel, La "Chanson de Roland" et la tradition épique des Francs (Paris: Picard, 1960)].

Rychner, Jean. La Chanson de geste: essai sur l'art épique des jongleurs (Geneva: Droz, 1955).

------. Contribution à l'étude des fabliaux. 2 vols. (Geneva: Droz, 1960).

Uitti, Karl D. "'Ço dit la geste': Reflections on the Poetic Restoration of History in The Song of Roland," in Rupert T. Pickens, ed., Studies in Honor of Hans-Erich Keller: Medieval French and Occitan Literature and Romance Linguistics (Kalamazoo, MI: Medieval Institute Publications, 1993), 1-27.

8 January 1997