The New Philology
Introduction: Philology in a Manuscript Culture
By Stephen G. Nichols
Thoughts on the Discipline
In medieval studies, philology is the matrix out of which all else springs. So we scarcely need to justify the choice of philology as a topic for the special forum to which Speculum, in a historic move, has opened its pages. On the other hand, if philology is so central to our discipline, why should one postulate a "new" philology, however ironically? While each contributor answers this question in a different, though complementary, way, the consensus seems to be that medieval philology has been marginalized by contemporary cognitive methodologies, on the one side, while within the discipline itself, a very limited and by now grossly anachronistic conception of it remains far too current. This version, formulated under the impulse of political nationalism and scientific positivism during the second half of the nineteenth century, continues to circumscribe the "discipline" of medieval studies. The forum presented here undertakes to explore and interrogate presuppositions underlying current philological practices.
What is "new" in our enterprise might better be called "renewal," renovatio in the twelfth-century sense. On the one hand, it is a desire to return to the medieval origins of philology, to its roots in a manuscript culture where, as Bernard Cerquiglini remarks, "medieval writing does not produce variants; it is variance."1 On the other hand, a rethinking of philology should seek to minimize the isolation between medieval studies and other contemporary movements in cognitive methodologies, such as linguistics, anthropology, modern history, cultural studies, and so on, by reminding us that philology was once among the most theoretically avant-garde disciplines (cf. Vico, Ampere, Michelet, Dilthey, Vossler).
Medievalists are frequently viewed by modernist colleagues as hostile or indifferent to contemporary theory. In such strictures, philology often figures both in the attack and in the defense: the modernists oppose theory to philology; the medievalists cite philology as a sufficiency that either precludes the need for theory or renders modern theories anachronistic in a medieval
1 "Or 1 Venture medi£va!e ne produit pas de variantes, elle est variance. La recriture incessante a laquelle est soumise la textualke" m£dievale, 1'appropriation joyeuse dont elle est l'objet, nous invitent a f'aire une hypothese forte: la variante n'est jamais ponctuelle." Bernard Cerquiglini,
Eloge de la variante: Histoire critique de la philologie (Paris, 1989), p. 111.
SPECULUM 65 (1990)
1
2 The New Philology
context. This split between modernist and medieval studies could be felt already in 1948, when René Wellek suggested that "philology" should be dropped from the lexicon of literary studies.2 It was open to misunderstanding, Wellek argued, because it had come to signify too broad a domain of applicability:
Historically, it has been used to include not only all literary and linguistic studies, but studies of all products of the human mind. Though its greatest vogue was in nineteenth-century Germany, it still survives in the titles of such reviews as Romance Philology, Modern Philology, Philological Quarterly, and Studies in Philology. [Philip August] Boeckh, who wrote a fundamental Encyklopadie und Melhodologie der philolo-gischen Wissenschaften (1877, but based on lectures partly dating back to 1809), defined "philology" as "the knowledge of the known" and hence the study of languages and literatures, arts and politics, religion and social customs. (P. 38)
Even while Wellek composed its obituary, Leo Spitzer, Erich Auerbach, and Ernst Robert Curtius were at the summit of their careers, practicing philology that ranged from Spitzer's etymologically based stylistic studies, through Auerbach's efforts to see in linguistic expression a profile of historic moments, to Curtius's insistence that poetic form, thanks to the power of transhistorical typologies, asserted the complex unity of European culture.
The philology of all three of these masters was grounded in texts, but edited texts, rational products of philological endeavor. This was consistent with Auerbach's conviction that philology grew out of specific Renaissance technological and intellectual movements: humanism, the Reformation, and the invention of the printing press. Humanism and the Reformation needed to collect and edit manuscripts from the ancient world to better articulate principles of moral philosophy and theology, while the printing press permitted the fixing and dissemination of sources and the new principles predicated on them.
The need to go back to the sources, a need felt as much by the Humanists as by the reformers (many Humanists were among the chief proponents of the Reformation), led to the founding of pkilology. And the invention of the printing press also contributed toward this end; many printers were also distinguished Humanists, and some of them were strong adherents of the Reformation. It was at this time and under these circumstances that the collecting and editing of manuscripts . . . became necessary and developed with complete spontaneity. In addition to their scholarly work, which involved editing, composing works on the grammar and style of Latin and of their own mother tongues, on lexicography and on archaeology, these Humanist philologists accomplished an important task of popularization: they were translators of the great works of antiquity.3
By its origins, in Auerbach's view, philology represented a technological scholarship made possible by a print culture. It joined forces with the mechanical press in a movement away from the multiplicity and variance of a manuscript culture, thereby rejecting, at the same time, the representation
2 René Wellek, Theory of Literature, 3rd ed. (New York, 1956), p. 38.
5 Erich Auerbach, Introduction to Romance Languages and Literatures (New York, 1961), p. 147.
Introduction 3
of the past which went along with medieval manuscript culture: adaptation or translatio, the continual rewriting of past works in a variety of versions, a practice which made even the copying of medieval works an adventure in supplementation rather than faithful imitation. In its place, the philology inherited by Auerbach's generation installed a preoccupation with scholarly exactitude based on edited and printed texts. The high calling of philology sought a fixed text as transparent as possible, one that would provide the vehicle for scholarly endeavor but, once the work of editing accomplished, not the focus of inquiry. It required, in short, a printed text.
We see this dramatically in Leo Spitzer. Most often working at the micro-textual level with discrete words, phrases, or expressions in order to open up insights into a poet's style, Spitzer based his philological analyses, which frequently took into account variant readings, on edited texts in published editions, rather than on manuscripts where the variants could be viewed in context. Spitzer's article "Parelh paria chez Marcabru (ou l'origine de la pastorela)" offers a case in point.4 He makes a claim for an entirely new reading of this famous pastourelle in a dazzling display of philological erudition where the poem and its variants are conceived solely in terms of print, rather than manuscript, culture. The crux on which Spitzer's whole claim for Marcabru's poetic talent turns, the creation of a neologism, parelh-paria, "avec trait d'union," as he insists (p. 419), could not possibly be found, or even conceived, in the six (of seven extant) manuscripts that preserve this lesson. Compound words identified by hyphenation are conventions of a print culture, not a manuscript culture where writing is dictation and reading, oral (as Suzanne Fleischman points out in her article below). Finally, it is notoriously difficult to determine fully variant lessons from the critical apparatus of many editions which are — as in the case of the poem Spitzer deals with — necessarily incomplete. Editors of the "old" philological persuasion sought to limit variation, not reproduce it, as Fleischman makes clear. See also in this respect "Modernitě textuaire" in Cerquiglini's Eloge de la variante.
The medieval artifact, for Spitzer, was the edited text, or, preferably, edited texts; literary language could only be adequately described by multiple examples from many texts, which then permitted him to identify the invariant signaling a universal or to demonstrate the normative deviation signaling stylistic originality which would reveal the mark of a superior poet. He made this point in an anecdote contrasting himself to the famous positivist Fustel de Coulanges. Spitzer recalled that Fustel de Coulanges insistently asked his students when they made a historical statement: "Avez-vous un texte?" Spitzer's own response held that "the student in historical semantics must ask: 'Have you many texts?,' for only with a great number of them is one enabled to visualize their ever-recurrent pattern."5
Philology was system, model, for Spitzer, as it was for Auerbach in a somewhat different sense. Auerbach, too, conceived of the text as singular,
* Spitzer's article first appeared in Romania 73 (1952), 78-82, and was reprinted in Romanixhe Literaturstudien, 1936-1956 (Tubingen, 1959), pp. 418-21.
s Leo Spitzer, Classical and Christian Ideas of World Harmony (Balumore, 1963), p. 1.
4 The New Philology
almost transparent, a vehicle for the higher ideals expressed by its language. In his search for realism — which for him was perfected in its nineteenth-century French version6 — it is not the materiality of texts (their physical historical presence) that interests him so much as their ideality: But an even stronger limitation than that in terms of class results for the realism of the courtly romance from its legendary, fairy-tale atmosphere. It is this which makes all the colorful and vivid pictures of contemporary reality seem, as it were, to have sprung from the ground: the ground of legend and fairy-tale, so that . . . they are entirely without any basis in political reality. The geographical, economic and social conditions on which they depend are never explained. They descend directly from fairy-talc and adventure.'
Auerbach's position here is a more sophisticated, but not qualitatively different, conception of the Middle Ages viewed from the scientific (specifically, philological) perspective voiced by Gaston Paris in his inaugural lecture at the College de France in 1866:
Le moyen age est une epoque essentiellement poetique. J'entends par la que tout y est spontane, primesautier, imprevu: les hommes d'alors ne font pas a la reflexion la meme part que nous; ils ne s'observent pas, ils vivent naivement, comme les enfants, chez lesquels la vie reflechie que developpe la civilisation n'a pas etouffe encore la libre expansion de la vitality naturelle. Ils n'ont ni dans le monde physique ni dans le monde social cette id6e de regularite prevue que nous a donne la raison.8 Gaston Paris's sentiments accord with his conception of literature as "being no more in sum than one of the aspects of the life of a people." Before undertaking to study literary history, in consequence, "one must understand what the people are who produced it, ponder the influences they underwent, what milieu traversed, and the phases of their development before the hour when their literary history began."9 Neither here, nor later, when he outlines the tools that gave philology its scientific status — etymology and "the inflexible laws which governed the evolution of Latin sounds for twenty centuries"1" — does he talk about the material specificity of medieval texts, the manuscript matrix, or the way they interacted with the social formation they describe.
6 Cf. Linda Hutcheon, A Theory of Parody (New York, 1985): "Auerbach . . . wrote Mimesis (1942-45) from the implied stance that nineteenth-century French realism is the only true realism, and therefore thai all attempts before Stendhal are but imperfect steps en route, and that any after are a sign of its decline." The last remark corroborates Bernard Cerquiglini's notion that the philology put in place in the latter part of the nineteenth century was one based on decadence {Eloge de la variante, pp. 82-85).
'Erich Auerbach, Mimesis: The Representation of Reality in Western Literature, translated from the 1946 Swiss edition in German by Willard Trask (New York, 1957), p. 116.
s Gaston Paris, La poisie du moyen age: Lemons et lectures, 2nd ed. (Paris, 1887), p. 9. We should note that although this was indeed Gaston Paris's inaugural lecture at the College de France, it was prematurely so. His father, Paulin Paris, designated Gaston to replace him in 1866 but reclaimed the chair for two more years in 1867. Gaston Paris again replaced his father from
1869 to 1872, when the elder Paris definitely retired, leaving the chair vacant for Gaston. Ibid., p. 240.
9 Ibid., p. 43.
10 Ibid., p. 251.
Introduction 5
This is so in part because Paris's conception of philology does not use a language model in conceiving of the way in which the beliefs of the Middle Ages shaped its institutions and textual artifacts. Instead, he posits a pre-anthropological model, the people (le peuple), whom he views as having spontaneously "sung" or expressed themselves in poetic form devoid, as Howard Bloch reminds us in his contribution, of subterfuge, complexity, or opacity. As a spontaneous, collective emanation it is "true," that is, unsophisticated by the conventions of literary theory:
La litterature fut 1'image de cette vie. Elle en a la liberte, la variete, la franchise. Elle n'est pas, comme la notre, suveillee par des lois, ni retenue par les prejuges ou les convenances, ni dirigee par des exemples classiques; rien ne l'empeche de dire pleinement et entierement ce qu'elle veut dire. Aussi est-elle vraie avant tout, el c'est la son grand merite. Sans se preoccuper des regies, des theories, des questions de forme, elle exprime simplement ce qui s'agitait dans les ames; elle donne une voix, souvent peu nette et peu forte, mais fidele, aux sentiments, aux idees de tous. Ce n'est pas une litterature de livres, destinee a occuper quelques instants dans 1'attention des lecteurs, qui d'ailleurs n'en sont pas dupes et ne lui accordent qu'une faible partie de leur ame: c'est une poesie toute vivante et exteYieure, a laquelle chacun croit et que chacun pourrait avoir faite, qui se chante el qui park, au soleil, dans les rues, dans les places, au milieu des batailles, sur les routes qui menent aux p£le-rinages ou aux foires, sur les navires qui emportent les croises, dans les eglises ou sous leur porche, dans les chateaux, dans les assemblies brillantes, aux festins des rois, aux repas des auberges." (emphasis added except for first instance)
Gaston Paris merits attention. Stripped of its romantic lyricism, his description articulates a poetry of enonce based on a spoken, rather than a written and codified, language that defines a contextual social formation as important for its meaning as the discursive intertextuality (semiotic codes) we know also to have been an important constituent of this poetry. Suzanne Fleischman, Howard Bloch, and Gabrielle Spiegel, in their contributions, develop different aspects of Paris's insights. His theory of representation, not unlike Auerbach's own, grounds itself in mimesis, rather than semiosis, on direct, rather than mediated, imitation.
What separated Gaston Paris's philology from that of the generation of Auerbach, Spitzer, and Curtius, and what deflected his insistence that twelfth-century Old French was not a poetry of the book was Joseph Berber's espousal of the unique manuscript approach to text editing first adumbrated in the introduction to his SATF edition of Jean Renart's Le lai de I'ombre in 1913, and later revised as a definitive statement of his principles of text editing of Old French literature in 1928.12 B£dier's program struck directly at Lach-mannian principles of "scientific" editing by classifying manuscripts into a "genealogical tree" that would permit the editor to discover the manuscript(s) closest to the lost original, principles that had been introduced in France by
" Ibid., pp. 19-20.
12 Jean Renart, Le lai de I 'ombre, ed. Joseph Beclier (Paris, 1913). The famous article containing his "Reflexions sur l'art d'editer les anciens textes" appeared under the rather innocuous title "La tradition manuscrite du Lai de l'ombre," Romania 54 (1928), 161-96 and 321-56.
6 The New Philology
the young Gaston Paris.13 The new approach had the virtue of emphasizing an authentic medieval manuscript as opposed to a hybrid reconstruction, since Bedier believed in finding and using one best manuscript as the basis on which to edit a work (see his article championing the primacy of the Bodleian Library's Digby 24 as the best manuscript of the Chanson de Roland).14
Far from encouraging focus on the manuscript matrix, however, Bedier's insistence on a single manuscript had the effect of putting that manuscript into a relationship with its printed edition analogous to that of a unique manuscript of a modern printed book. Indeed, the analogy of the modern holograph manuscript is so real that Bedier even postulates a medieval poet, like Jean Renart, revising his work for a "second edition" on the occasion of copying a new manuscript for a patron. The seven existing manuscripts may thus be related to two versions of the poem, both coming directly from the poet's hand. Like the modern holograph manuscript, the medieval manuscript may thus be seen as linking us closely to the hand and mind of the author. As though reinforcing this technocentric view of medieval manuscript culture, Bedier uses terms — publication, tirer, etats — in speaking of this "second edition" that have strongly marked semantic associations with the lexicon of printing:
Je suppose ici que Jean Renart a d'abord lance son Lai de I'Ombre dans la circulation sous la forme d'un manuscrit pur de fautes, 0 . , . . Trois mois, six mois apres cette premiere publication de son ouvrage, Jean Renart I'a relu dans un manuscrit identique k Ol, et en a tire de sa main une copie nouvelle, Oa, pour 1'offrir & quelque patron ou pour la vendre a quelque jongleur. En recopiant, chemin faisant, me-content de son premier jet, il a refait certaines lecons. ... Si nous supposons ainsi que nos sept manuscrits peuvent representer . . . deux "etats" du texte tour a tour avoues par le poete, quoi de plus naturel en soi qu'une telle supposition? Pourquot les ecrivains anterieurs a I'invention de Vimprimerxe n'auraienl-ils pas fait ce que nous voyons fane a tous leurs confreies venus apres evx?ib (emphasis added)
Whereas Gaston Paris's preanthropological model for medieval literature made the collectivity the generative locus, Bedier, good modernist that he was, placed the author at the origin not only of the literature but, as we have seen, for as much of the manuscript tradition as he could. Both Paris and Bedier believed literature to be referential mimesis where reference begins with its origin. The origin of a work, in this view, marks not only the content but the transmission as well, finally influencing the philologist's work. Thus Bedier makes Jean Renart responsible for the "two versions" of Le lai de I'ombre he confects from the interesting manuscript variations. Surplus variations in the manuscript tradition cease to have critical interest because they
15 See Cerquiglini, Eloge de la variante, "Gaston Pans et les dinosaures."
14 Joseph Bedier, "De l'autorite du ms. d'Oxford pour i'etablissement du texte de la Chanson de Roland" Romania 41 (1912), 330-45.
15 Le lai de I'ombre, pp. xxxvii-xxxviii. Bedier concludes: "On peut done ires legitimement former l'hypothese que les quatrc-vingts lecons propres au manuscrit E seraient du jean Renart le plus pur et le plus authentique" (p. xxxix).
Introduction 7
simply reflect scribal errors. Where no author can be postulated for a work, as with the majority of the early Old French epics, then place and function generate an unknown poet. For the epic, this means anonymous clerics in shrines along the pilgrimage route, or in Bedier's by now immortal dictum: "Au commencement fut la route, la route jalonnee de sanctuaires."
Bedier did not lack for challengers, but his theory was so strong and such a counterthrust to the genealogical-tree approach to text editing that it gained enormous prestige during the years when Auerbach, Spitzer, and Curtius were beginning their careers. We can better understand their relative lack of concern for the material artifacts of medieval literature, the manuscript culture per se.
It is that manuscript culture that the "new" philology sets out to explore in a postmodern return to the origins of medieval studies. If one considers only the dimensions of the medieval illuminated manuscript, it is evident that philological practices that have treated the manuscript from the perspective of text and language alone have seriously neglected the important supplements that were part and parcel of medieval text production: visual images and annotation of various forms (rubrics, "captions," glosses, and interpolations).
The medieval folio was not raw material for text editors and art historians working separately. It contained the work of different artists or artisans — poet, scribe, illuminator, rubricator, commentator — who projected collective social attitudes as well as interartistic rivalries onto the parchment. The manuscript folio contains different systems of representation: poetic or narrative text, the highly individual and distinctive scribal hand(s) that inscribe that text, illuminated images, colored rubrications, and not infrequently glosses or commentaries in the margins or interpolated in the text. Each system is a unit independent of the others and yet calls attention to them; each tries to convey something about the other while to some extent substituting for it.
Sometimes we see graphic examples of the systemic rivalry, as, for example, in the case of decorated or historiated initials at the beginning of passages which are so ornate that it may be difficult to read the image as a letter. Similarly, the rubric — the annotations in red that comment on the text or provide captions to images — does not simply "explain" or describe what is to be found in the miniature or passage it introduces. Appropriating to itself the role of commentary or directed reading, the rubric focuses attention at specific moments, telling us what it is we are to see in the visual scene or laying out the narrative thrust of the verbal text.
The same kind of mimetic appropriations occurs in the relation between painted miniature, poetic text, and the copies of a manuscript. A miniature we admire as a work of art in its own right also represents a scene in the poetic narrative, now transposed from the verbal to the visual medium. On the other side, the poetic narrative offers luxuriant ekphrases that we recognize as poetry substituting for picture. Not infrequently, such ekphrastic passages form the basis for miniatures (such as the illuminations accompanying the portraits on the wall of Deduit's garden found in many manuscripts
8
The New Philology
of the Roman de la rose) which transpose the ekphrastic descriptions into a visual register where the artist seeks to outdo the poet.
The apparently straightforward act of copying manuscripts is not free from mimetic intervention, either. In the act of copying a text, the scribe supplants the original poet, often changing words or narrative order, suppressing or shortening some sections, while interpolating new material in others. As with the visual interpolations, the scribal reworkings may be the result of changing aesthetic tastes in the period between the original text production and the copying. Even in such cases, however, the scribe's "improvements" imply a sense of superior judgment or understanding vis-a-vis the original poet.
Recalling that almost all manuscripts postdate the life of the author by decades or even centuries, one recognizes the manuscript matrix as a place of radical contingencies: of chronology, of anachronism, of conflicting subjects, of representation. The multiple forms of representation on the manuscript page can often provoke rupture between perception and consciousness, so that what we actually perceive may differ markedly from what poet, artist, or artisan intended to express or from what the medieval audience expected to find. In other words, the manuscript space contains gaps through which the unconscious may be glimpsed.
The dynamic of the medieval manuscript matrix — I am talking here particularly of illuminated manuscripts — involves cognitive perception as two kinds of literacy: reading text and interpreting visual signs. This double literacy involves mimetic repetition to the extent that the visual art repredi-cates the poetic text. I am not suggesting that the visual art simply imitates the verbal, but that insofar as it illustrates a prior narrative, it opens itself up — along with the poetic narrative it reflects — to the psychic apparatus of repetition. In consequence, illuminated manuscripts double the potential for that rupture between perception and consciousness I spoke of above, and thus offer a dual route of penetration to the underside of consciousness.
What I am suggesting is simply that the manuscript matrix consists of gaps or interstices, in the form of interventions in the text made up of interpolations of visual and verbal insertions which may be conceived, in Jacques Lacan's terms, as "pulsations of the unconscious" by which the "subject reveals and conceals" itself.16 If the subject is divided by the effects of language in ordinary speech, then the doubling of perceptual fields in the manuscript matrix into verbal and visual forms produces conditions favoring an even greater split in the subject represented by the speaking voice(s). It is that division that has been insufficiently explored by medievalists working with illuminated manuscripts when they argue, as some have done, that a study of the visual components will simply help to confirm intentionality in the verbal texts.
I offer this vignette, necessarily incomplete, by way of illustrating the kinds of questions that can and should follow in one small sphere of interest if we turn our attention back to the manuscript culture of the Middle Ages. If we
lfiJacques Lacan, The Four Fundamental Concepts of Psycho-Analysis, ed. Jacques-Alain Miller, trans. Alan Sheridan (New York, 1981), p. 188.
Introduction 9
accept the multiple forms in which our artifacts have been transmitted, we may recognize that medieval culture did not simply live with diversity, it cultivated it. The "new" philology of the last decade or more reminds us that, as medievalists, we need to embrace the consequences of that diversity, not simply to live with it, but to situate it squarely within our methodology. That is what the contributions to this forum have tried to suggest.
The Contributions
The contributors do not represent a particular school or tendency. Indeed they speak with remarkable individuality, circumspection, and freshness. Each has a specific aspect of the discipline to explore and innovative techniques to suggest. The order of presentation traces a continuum from a clear description of the aims and methods of philology as it has been understood, with suggestions for innovation, to more radical approaches based on alliances with contemporary cognitive methodologies. What is "new" in the philology common to all the contributions may be found in their insistence that the language of texts be studied not simply as discursive phenomena but in the interaction of text language with the manuscript matrix and of both language and manuscript with the social context and networks they inscribe.
The five contributions make a cohesive continuum. The first three — by Siegfried Wenzel, Suzanne Fleischman, Howard Bloch — deal specifically with aspects of text editing and language analysis in medieval texts, philology in its basic sense, as it were. Siegfried Wenzel opens the forum with a contribution that bridges the space between philology as traditionally understood and new philology. He presents "a practitioner's concern with what philology is and with what continuing value for the study of medieval literature it has." The paper constitutes an excellent mise au point of the scholarship and aesthetics of text editing by a master of medieval Latin and Middle English. Suzanne Fleischman brings her combined strengths in the disciplines of linguistics and Romance philology to bear on the specificity of Old French language. Arguing that rigidly codified conceptions of language bequeathed by nineteenth-century philology have prevented scholars from understanding the degree to which textually preserved Old French reflects a spoken rather than a written language, Fleischman skillfully demonstrates how current work in linguistics can help scholars view Old French as a user-defined (spoken) system of language. Recognition of this fact should lead to an entirely different, less rule-constrained approach to text editing and language analysis in Old French literature.
The literary analysis consequent on Fleischman's findings may be seen in part in Howard Bloch's article. Bloch's contribution makes a natural fulcrum that complements and extends the issues raised in the first two articles, while shifting to the extended connotation of philology as language analysis, historical reference, and issues of the discipline itself addressed from the standpoint of their different specialties by Gabrielle Spiegel and Lee Patterson. Bloch joins Wenzel and Fleischman in considering specific aspects of philology focused on textual and linguistic study, extending issues they raise into
10
The New Philology
Old French lyrico-narrative poetry. At the same time, by recalling the broader conception of philology as cultural hermeneutics first proposed in the eighteenth century by Vico in his Scienza nuova, a book that profoundly influenced Auerbach, Bloch initiates the discussion continued in the last two papers. He is precise and insightful in defining what it is that the new philology seeks to define itself against.
In a bold and, for many, welcome move, Gabrielle Spiegel argues that new philology can provide an effective counter to the dehistoricized cognitive methodologies fostered by poststructuralist discourse analysis. As a prolegomenon to her presentation, Spiegel provides a valuable precis of recent methodologies hostile to history. In her quest "to restore history as an active agent in the social construction of meaning," she finds that New Historicism has not managed to rehabilitate history itself as distinct from textual construction about historical topics. Insisting that the text be distinguished from its historical context(s) by remembering that "texts represent situated uses of language," she argues that the sites of linguistic usage are "essentially local in origin and therefore possess a determinate social logic." Rejecting such totalizing concepts as "language" and "society," the historian can study the social logic of situated linguistic usage in discrete geographical or political units. She demonstrates effectively her program of relational reading between text and context by analyzing the literary and historical dialectic of genealogy in the Pseudo-Turpin chronicles as played out in the rivalry between the Flemish aristocracy and the Capetian monarchy.
Lee Patterson's paper makes a logical ending to the forum because it illustrates in summary issues touched upon in one way or another by the previous papers, but recast as an analysis of the status of medieval studies in the larger context of the profession. Like Spiegel, Patterson evokes contemporary theories, taking them as a challenge to medieval studies. As Siegfried Wenzel begins the forum by linking traditional and new philology, Patterson closes it with a challenge to rethink attitudes in and toward medieval studies. He carries the critique outside of medieval studies (whose part in its own marginalization within the academy he also shows) to expose the "master narrative" by which academic departments privilege Renaissance humanism and modernism to the exclusion of medieval studies. His analysis of some recent and highly regarded Renaissance scholarship reveals the complacency with which colleagues in that field have fostered the myths by which modernism — with the unwitting complicity of medievalists themselves — guards the gates of academe's medieval ghetto. Patterson offers strong medicine, but if the sense of absorption and isolation of medieval studies can be overcome, it will not be by placebos.
Stephen G. Nichols is Associate Dean of Arts and Sciences and Edmund J. Kahn Professor of Humanities at the University of Pennsylvania, Philadelphia, PA 19104.
Reflections on (New) Philology
By Siegfried Wenzel
As the following remarks are to reflect my own scholarly commitment and experience, I should begin by saying that they come from a medievalist who in his work is always conscious of dealing with the works of a past state of civilization. They also come from a historian of literature, who in contrast to political or economic historians makes written documents the subject of his study, and who in contrast to linguists looks at them as works of verbal art. And finally, they come from a professor of English literature who works closely with colleagues whose ultimate aim is the aesthetic analysis and appraisal of verbal objects, and of whom many today feel driven to ask deeper theoretical questions about, not only the nature of aesthetic values, but the nature of language itself and its function to communicate "meaning," about the relation between object and perceiver, about the psychological or political function of literature and of language, and so on. Working in such an environment naturally brings with it a certain amount of confrontation, which in turn has its occasional political side in discussions about such practical matters as departmental curricula and new appointments. Though I would not rank myself among philologists who have little use for literary criticism and theory, I can of course not help being conscious of pressures and of attitudes on the other side — however unvoiced they may remain — that would reject "philology." Hence, my remarks will reflect a practitioner's concern with what philology is and with what continuing value for the study of medieval literature it has. They will describe a position and are in no way intended as a Forschungsbericht; the few illustrations cited are no more than examples that have crossed my attention in recent months.
By tradition, "philology" can be taken in either a narrow or a broad sense. In the former it designates the academic discipline of studying or "scientifically" elucidating the basic, literal meaning of verbal documents. Originally, and primarily, this meant the study of older stages of particular languages, whose records, in contrast to contemporary utterances, required careful investigation and the application of a wide-ranging knowledge in order to make sense. Thus, philology dealt first with the classical languages, Greek and Latin, and with surviving documents of other "dead" languages, such as ancient Hebrew and other Near Eastern languages, and eventually the various branches of the Indo-European family. In a short time, philologists extended their studies to the older stages of modern European languages, of their national tongues, as well. To uncover their meaning, it was necessary to understand a variety of ancient writing systems and the phonetic value of individual symbols, whether these have continued in use or not. Once the hypothetical sounds hidden behind written signs had been identified, one could set out to deal with words and sentences and their meanings and
speculum 65 (1990)
11
12 The New Philology
eventually construct grammars, dictionaries, and comparative histories of individual languages. In the course of the last hundred years, this intellectual enterprise has become much refined and diversified and, more importantly, has given rise to a new academic discipline, linguistics, which studies language as its final object, not as a method to understand ancient texts.
In its wider sense, philology preserves the basic urge to understand a work of verbal and usually written communication, but in doing so it goes beyond the strict concentration on language and its aspects, to include whatever contextual information might help to elucidate a text: first of all its sources, but then also political and other aspects of history, biography where the writer is known, socioeconomic conditions (such as patronage or practical demand for books), the conditions and processes of writing or copying a document and of printing, the religion or Weltanschauung of the culture from which the text comes, the intended use of the text (its Sitz im Leben, to use the fashionable phrase), and much else. In this wider sense, I would think of philology not so much as an academic discipline with a clearly defined object and proper methods of investigation, but rather as an attitude. It is precisely what the etymology of the word declares, "love of the word": an appreciative attraction to verbal documents that seeks to understand their meaning, starting with the surface and penetrating to whatever depths are possible, but also alert to the fact that a given text comes from and is shaped by a specific time and place that usually is significantly different from that of the observer. It is this wider sense that I would apply to my own scholarly orientation and with which my remarks here are concerned.
Of the possible objections to philology in this sense as viable, fruitful, and even necessary, I exclude two that are not entirely germane to the present discussion. One concerns the supposed irrelevance of studying old texts (all the way up to this morning's newspaper, of course), dead languages, and the past in general; to answer this objection is properly the province of the historian. On the other hand, the objection that philology's traditional and central concern with language has developed into the much more "professional" discipline of linguistics is not entirely pertinent since, in my view as expressed above, the study of language, including the detailed knowledge of linguistic changes and so on, is not an end in itself but a means, even if a most important one, toward the understanding of texts.
There are, however, several pertinent objections that might be raised to the claim that philology continues to have an important heuristic value. One is that it is old hat: it was done, and perhaps done well, in the nineteenth and early twentieth centuries; but — quite apart from the fact that our modern intellectual quest has moved on beyond the typical concerns of philology — the major insights which philology could furnish have been furnished, and nothing new can be added. Second, philology is no longer relevant to the main objective of literary studies, to understand, appreciate, and evaluate works of literature as objects of art. It has been, and should indeed be, replaced by literary criticism or even literary theory. And third, philologists are often thought of as working in blissful ignorance of such notions as structuralism, the hermeneutic circle, or indeterminacy, which
Reflections 13
indicate an immensely more sophisticated awareness of the entire process of cognition and reflect deep uncertainties that set modern man very much apart from his nineteenth-century forebears. These are serious objections, and I think they need consideration.
To begin at the end: an awareness of the contemporary discussion of cognitive theory, based on such fundamental, even if by now banal, insights as that one's question will determine the answer one gets, or that an investigation of facts always implies interpretation, is certainly a most desirable asset in a good philologist. But I would maintain that such insights occur quite naturally once one begins to reflect on the process of one's philological work. In transcribing a late-medieval Latin manuscript, for instance, one is often faced with a text that is full of abbreviations and lacks a meaningful system of punctuation to guide the modern reader's comprehension. Thus the editor has to supply both expansion and punctuation, and in doing so must make choices at the most basic level. This can be done only if the editor already understands what the text says: one must know, not only Latin and paleography, but also — and I often think, far more importantly — the subject matter of the text and its special vocabulary. Thus, the letter p with a cross stroke through its descender can stand for per, par, or por, and one's choice will depend on the lexical meaning that is "required" in the reader's opinion, and occasionally on other factors such as (in vernacular texts) the dialect of the text, which itself might depend on the date and geographical provenance of the document. To be sure, in most cases these are minor details, yet the choice between aperuit ("he opened") and apparuit ("he appeared," often written with single p) can have major implications for a passage's meaning. The proverbial difficulty of reading minims, which wreak havoc with any clear distinction between the letters m, n, u, and i, is another case where one has to decide for oneself what the text one is trying to understand says. And the same goes for more linguistic matters. In order to determine the lexical meaning of individual words in many Middle English poems, for instance, the scholar has to rely on a combination of insights furnished by etymology and context. Given the absence of a written standard for Middle English, and hence the wide variation in spelling and in dialect forms, in questionable cases the reader nearly always has a choice of several word meanings before him, from which one will be chosen that "fits the context." And the same hermeneutic circle is even more threateningly present at the higher level of critical editing. The recent edition of Piers Plowman, for example, tries to reconstruct what the poet wrote or intended to write, and it does so on the basis of specific assumptions about the poet's dialect, individual vocabulary, and, especially, prosodic characteristics. Though not everyone has agreed with the critical method adopted by George Kane and E. Talbot Donaldson, I do not. believe anyone will deny that their work ranks as a monument of philology. And Piers is not an isolated case; the field of editing all major poetic texts in Middle English is in great turmoil, in which the hermeneutic circle plays a major role. Here the cognitive problem of modern science is indeed felt most keenly.
In fact, the principles for editing a medieval text continue to be subjected
14 The New Philology
to intense debate and probing theoretical questions. The classical notion that a critical edition will attempt to reconstruct the author's original or intended meaning has, to some scholars, become a rather Utopian idea — leading them, in the case of certain works, to produce parallel texts from two or more manuscripts instead. In the case of the Canterbury Tales, textual criticism has arrived at the point where even the earliest surviving manuscripts are felt to have undergone editing on the part of their scribes and therefore must be accepted as more or less intelligent responses by first-generation readers. Further, editors and critics in many areas now believe that any surviving text expresses less a single authorial intention than that of various "social" forces. Hence the ideal might be to furnish facsimile reproductions. Yet while photographic and xerographic processes provide a wonderful aid to scholarly work that must be undertaken far from the actual sources, they are of course no substitute for penetrating analysis, discussion of the state of the transmitted text, and ultimately an informed judgment based upon a multiplicity of considerations. The spectacle of a seasoned scholar presenting and defending solutions for a handful of textual cruces has become a standard ingredient of nearly any professional gathering of medievalists. Though hardly a new phenomenon, it tends to be conducted with ever increasing sophistication. If to an outsider such activity may suggest that Matthew Arnold's Grammarian was, regrettably, not buried after all, the pursuit of strange variant readings nonetheless still holds the prospect of unprecedented discoveries both in the preserved texts and in their history before and after they were written down.
Nor is the minute study of a given manuscript, together with alert attention to scribal processes, simply a continuation of methods over a century old. During the past two generations, codicology and paleography have become immensely refined disciplines and have produced new bodies of information that would rightly be the envy of our elders. In contrast to them, we can no longer consider a codex as a mere receptacle that happens to have preserved the text under investigation; instead, a modern editor will have to look at the manuscript "holistically," as a total unit about whose physical makeup, composition, and history he or she will want to know as much as possible. Modern paleographers have also given us a much clearer and more differentiated nomenclature of medieval scripts, especially for the later Middle Ages, and even if a great deal remains to be done and agreed upon in order to either correlate or differentiate between terminologies used in different countries, we possess a tool that makes a,significant part of our labors considerably less idiosyncratic or questionable than it was not too long ago. The introduction of such descriptive labels as "Angličana" or "Secretary" has fairly revolutionized a corner of our field, and the notion of a "hierarchy of scripts" that has come with them reflects a new awareness of differences in audience, market conditions, scriptoria, and purposes of the transcribed text that throw new light on the study of texts. The very recent concern with layout, with its sensitivity to the visible disposition of text on the page, among other things suggests a utilitarian and rather prosaic function for even the most artistic illustrations in some medieval codices. In this very specific area we have not
Reflections 15
only made important progress but are still far from having heard the last word. I suspect, for instance, that further interest in manuscript layout may lead to the somewhat embarrassing realization of how incomplete or limited even the most minute studies of this moment are; for example, in trying to find some serviceable reproductions of manuscript pages that would show the marginal annotations in sermon manuscripts designed to guide a medieval preacher or student to the main parts of the scholastic sermon, I came to realize that the plates in those magnificent modern catalogues of manuscripts, particularly of dated and datable manuscripts, commonly leave off the medieval margins altogether.
Yet these catalogues represent a magnificent advancement of philology, and they stand next to other projects that collect and order a vast amount of data and make them available to modern scholars. The new research into and publication of medieval library catalogues is a fine example of fresh work whose impact will be felt only in the years to come. It is joined by such projects as the Old English Dictionary, based on a complete listing and analysis of the vocabulary found in Old English texts, or the Middle English Dictionary, whose completion is now in sight. The two dictionaries are much more than simple updates of their late-nineteenth-century predecessors, which attempted to do little more than give the modern meanings of medieval vernacular words; in contrast, the new projects furnish a wealth of additional information that reveals important differences in geography, time, and — perhaps most importantly — literary contexts and registers and is therefore of immediate and invaluable help to literary historians and critics. Middle English dialects have of course always been in the center of English philology devoted to that period — often very painfully so for readers who are not particularly interested in the reflexes of Old English vowels. Yet it must be noted how important the precise knowledge of dialect variation and etymological derivation of Middle English words is for an accurate understanding of words, as I mentioned earlier, and for the aesthetic appreciation of our major Middle English authors. How questionable the information found in major authoritative works often is can be illustrated by an experience I had recently that involves line 510 of Patience:
Bitwene be stele and be stayre disserne 11031 cunen.
Two modern editors gloss stele as "the upright of a ladder" (thus yielding: "they could not distinguish between the upright of a ladder and the rung"); but I remembered seeing the English word in a slightly earlier Latin sermon with the clear meaning of "rung."1 A trip to the Oxford Engluh Dictionary yielded the following results. "Upright side of a ladder" is one of several meanings assigned to two different (modern) nouns, stale and steal, which are said to be "not clearly distinguishable" from each other. In each case this meaning is labeled as obsolete and is directly coupled with a more recent
1 In a comparison of the Blessed Trinity to a ladder: "Inter duo latera scala [read scale | est pe stele vel ronge coniungens duo latera simul. - . . Sed in ista scala sunt ix stadia, anglice steles, per que ascenditur ad celum." Oxford, Merton College MS 248, fol. 64vb.
16 The New Philology
meaning of "a rung." The supposedly older meaning of "upright of a ladder" is in each case supported by a single quotation; for steal the dictionary cites the line from Patience, and for stale it refers to the following sentence from Ancrene Riwle:
Vilitas et asperitas . . . beod }>e twa leaddre steolen be beoo up iriht to heouene, and bitweone beose steolen beod of alle gode beawes be tindes ifestned bi hwucche me climbed to be blisse of heouene.2
In all the other quotations for the composite meaning, each word unquestionably has the "later" meaning of "rung." Now, the passage in Ancrene Riwle itself refers to Saint Bernard, and its ladder image has in fact been traced to a Latin text which, quite properly, speaks of latera scalae and gradus (sides or uprights, and rungs).3 Thus, appeal to the translated source, here happily identifiable, does indeed confirm the meaning of stele that was established by etymology, though not in a strongly convincing manner. But this seemingly clear picture in Ancrene Riwle, which is already somewhat disturbed by the noun's "later" meaning, becomes further dimmed when one traces the dictionary meaning of the companion term to stele, namely, tindes. Here, the Oxford English Dictionary gives "rung" as the third meaning, labeled "obsolete" and "rare" and supported by but one citation (from Ancrene Riwle, of course), of a noun that basically designates "each of a series of projecting sharp points on some weapon or implement," including the two branches of a deer's horn or of a stream — or, one suspects, of a ladder, though the OED does not say so.4 Has the Ancrene Riwle author perhaps confused uprights and rungs? It would seem that his usage is, to say the least, very puzzling, and it certainly casts a strong shadow over the "earlier" meaning of stele that supposedly still exists in the line from Patience. The modern definition of the word in the late-fourteenth-century poem, derived as it is from the context, may make sense, but surely it rests on a very questionable history of the word and demonstrably violates contemporary medieval usage.
This admittedly very minor detail may easily provoke impatient literary critics to cry "Hyperprofessionalism!" But the case gains some critical significance when one notices that line 510 of Patience occurs in a passage that poses a major problem of sequence and was even thought to have been canceled by either poet or scribe.5 Of the various morals implied in this case, one surely points to the false sense of security that reading editions give, which gloss their "hard words" with only the best meaning suggested by the context and fail to discuss textual and lexical problems of this kind, thus
2 The English Text of the Ancrene Riwle: Ancrene Wisse, ed. J. R. R. Tolkien, EETS 249 (London, 1962), p. 181.
'See PL 184:460, identified by R. W. Chambers in Review of English Studies 1 (1925), 19. * The Oxford English Dictionary, ed. James A. H. Murray et al. (Oxford, 1933), under stale. Heal sb. 1, and tine sb.1
5 See Malcolm Andrew and Ronald Waldron, eds., The Poems of the Pearl Manuscript (London, 1978), p. 205, whose text and line numbering I have followed- A fuller discussion is given by William Vantuono in his more recent edition, The Pearl Poems: An Omnibus Edition (New York, 1984), who rejects the traditionally accepted meaning of stele (vol. 2, p. 233).
Reflections 17
doing potential damage to literary critics who, in graduate school, were never required to work their way through at least one critically edited medieval text with a glossary that cites etymologies.
The continuing collecting of primary material and evidence naturally extends far beyond the vocabulary of a past age. Among the many projects one could cite here is the "Index of Middle English Prose," for which a targe group of scholars is, at long last, systematically combing medieval manuscripts for Middle English prose texts, an area of study that until now has lacked a bibliographical guide. Another project of particular interest for Middle English literature is the equally systematic combing of local archives for references to plays and dramatic performances and other related matters. Both projects have already begun to yield very important and fresh information about the sociology of literature, whether this concerns the production of written documents, particularly in the fields of devotional and scientific prose writings, or the popularity and variety of playacting through the fifteenth and well into the sixteenth centuries. Here and in some other areas, from which I would of course not exclude my own field of Latin sermons, a great deal of new factual material is coming to light on all levels that pertain to the study of literary texts. This makes me think that the note of sad resignation voiced in a recent discussion of folklore might be somewhat premature: "The Ballad of Hind Horn' has been popular in recent times; but during the late Middle Ages? Nothing can be known with certitude."6
All these efforts to discover new information obviously have significant consequences for the work of interpretation, whether it deals with a single poem, such as "Maiden in the Moor Lay"; or with broader cultural phenomena, such as Chaucer's feminism or the political orientation of the Elizabethan stage; or even the highest reaches of semiology. To these and a host of other contemporary critical questions and approaches to literature, philology continues to serve as a handmaiden, furnishing the material basis on which they must stand. Handmaidens are proverbially humble and modest; and however fascinating, even all-consuming for its practitioner, the quest for an elusive etymology or textual variant may become, in the larger scheme of humanistic scholarship and the pursuit of the examined life it certainly has its limitations. At the same time, scholarship is not an absolute monarchy but a republic, in which the handmaiden, while doing her job of preparing the necessities of life — intelligible texts and tools for their understanding — will also remain constantly watchful and critical of the nobility. To order the disciplines devoted to the understanding of literary texts hierarchically, in the shape of a pyramid with paleography at the base and semiotics at the apex, is tempting but dangerous, because such a model allows the semiotician as well as the literary critic in the middle ranges to remain above and aloof from the concerns of philology. Not just an ancillary discipline, philology is an attitude
6 Bruce A. Rosenberg, "Medieval Popular Literature: Folkloric Sources," in Thomas J. Hef-fernan, ed.. The Popular Literature of Medieval England. Tennessee Studies in Literature 28 (Knoxville, 1985), pp. 61-84, esp. p. 64.
i 8 The New Philology
of respect for the datum, for the facts of the text and its contexts, which should be cultivated at all levels of our enterprise to understand and appraise.
Philology thus holds not only a material value, in that it provides the raw materials for understanding, but equally a disciplinary one, by continuously demanding that the intellectual systems built by interpreters or theoreticians be tested against and anchored in the realities of the subject matter. As one should not take the authoritative listings and lexical decisions of the Oxford English Dictionary automatically for granted, so one must not read a poem's words "bin fader was a bond man" as expressing the dawning social consciousness and revolutionary feelings of late-medieval lower classes, without careful examination of their manuscript context.7 Similarly, a philologist will and should remain acutely uncomfortable with the speculative leap that causes a modern critic, after noticing that Unwine — who is only cryptically mentioned in a handful of medieval English texts — may, plausibly, mean "son born beyond hope" and that, in the slim historical record, he is not listed as succeeding his father, to claim that this Germanic hero was a "type ... of rebellion against the father" like the biblical Absalom8
What, then, about "New Philology"? The attitude and orientation I have described and called philology has, to be sure, at various times occupied various positions on the stage of humanistic scholarship — now standing in the limelight, now being pushed into the wings. If "New Philology" means that it is currently moving a little more to the front, that is all to the better; the movement vindicates the faithfulness of those who have made its cultivation their main task. But as I have defined it, "love of the word" that seeks understanding is a lasting concern of the intellectual life and as such stands above the currents of fashion. This is not to denigrate the many -isms that strut for a while; not only do they play their part in the ongoing performance of intellectual exploration, but they occasionally refine and enrich the more basic work of philologists by developing new "optics," thus sharpening our sights and adding new dimensions of awareness. Yet respect for the facts, for the concrete realities of the text, is and must remain basic. Nor is such an orientation restricted to medieval studies, as the present controversy over editing James Joyce's Ulysses demonstrates. Whether old or new, it would seem that philology is very much alive and has a surprising amount of blood in it.
1 See Siegfried Wenzel, Preachers and Poets and the Early English Lyric (Princeton, 1986), pp. 8-
8 Lee W. Patterson, Negotiating the Past: The Historical Understanding of Medieval Literature (Madison, Wise, 1987), p. 222; for the historical record, see R. W. Chambers, Widsith; A Study in Old English Heroic Legend (Cambridge, Eng., 1912), p. 219. with its "must have been" and "presumably." If one must find biblical typology, John the Baptist would seem to be a much better candidate. I wish to state emphatically that in this and the preceding examples I have cited I intend no attacks on the respective scholars, whose work I happen to respect very highly
Siegfried Wenzel is Professor of English at the University of Pennsylvania, Philadelphia PA 19104. '
Philology, Linguistics, and the Discourse of the Medieval Text
By Suzanne Fleischman
Philology, as Stephen Nichols suggests in his introductory remarks, has come to be equated in the minds of many with a dessicated and dogmatic textual praxis which, through the minutious methodologies of paleography, historical grammar, and the textual criticism of "Monsieur Procuste, Philologue,"1 has reduced medieval literary "monuments" to the status of "documents."5 The Oxford Roland, in my initial philological encounter with it, was alternately a subtext for deciphering sound laws or a node in a tree diagram mapping the scriptural genesis of a legend. And there it ended.
Clearly, the study of medieval texts has progressed beyond the state of the philological art at the moment of this first confrontation with what I later came to appreciate as one of the major literary monuments of the European Middle Ages. But the "crisis of philology" to which this anecdote points is still very much at hand; indeed, it stands out all the more sharply as the intellectual climate of postmodernism challenges us to reexamine the premises and presuppositions of our traditional methodologies and disciplinary practices and to renovate or replace them if need be with alternatives which can make the old texts speak to us in ways more consonant with our modern, now postmodern, episteme. In short, we find ourselves at a crucial moment of charting new directions that will justify — for ourselves and notably for our students — continuing to "do philology" at all.
In the following pages I should like to offer some suggestions for directions of research which can potentially revitalize philology in one of its areas of
My thanks to Jonathan Beck for suggesting a number of improvements to this paper. By the "off-the-hook" topos, I absolve him of responsibility for any misconceived ideas or discursive infelicities that readers may encounter.
1 This epithet is one of several suggestive chapter titles in Bernard Cerquiglini's critical essay on manuscript variation, whose subtitle asserts its relevance here: F.loge de la variante: Histoire critique de la pkilologie (Paris, 1989).
2 This useful heuristic distinction was introduced by Paul Zumthor. "Monuments" are what we might alternatively call "literature": a body of discourses, or texts, which a society considers worthy of dissemination and preservation in essentially constant form and which are not perceived as a purely utilitarian use of language ("documents"). The advantage of this definition — which locates "the literary" in a community's reception of a text rather than in authorial intent or properties of the language itself — is that it circumvents the traditional linking of literature to writing and to belles-lettres. Without such a definition, we could not speak of "medieval literature," given that literature only established itself as an institution toward the end of the eighteenth century.
speculum 65 (1990) 19
20
The New Philology
central concern: the language of medieval texts. Specifically, 1 propose to survey recent linguistic contributions which I feel can enhance our understanding of medieval textual language beyond what can be gleaned from the historical grammars and manuals of the early vernaculars on which medievalists have traditionally relied. I will focus on Old French; however, my critiques of the available materials and my proposals for fruitful research agendas should be understood to have wider application.
1. "Text Languages" of the Middle Ages
in the beginning was the voice
On parle dans sa propre langue, on ecrit en langue etrangere.
— Jean-Paul Sartre
It is now commonly accepted that the European Middle Ages were "oral," insofar as writing was dictated and reading was carried out viva voce.3 The term for writing as a method of composition was dictate, whereas scribere generally referred only to the physical act of putting pen to parchment: these were different activities, carried out by different individuals. Legere, as late as the fifteenth century, normally entailed an oral articulation of the sounds being decoded.4 Even after the introduction of printing, a "literate" textuality
s The particular blend of orality with writing that characterizes the European Middle Ages is what Paul Zumthor, in Introduction a la pohie orale (Paris. 1983), calls "mixed orality": although the culture is in possession of a writing system, the influence of writing remains external, partial, or for some reason deferred.
4 The practice of reading aloud has generally been interpreted as a reluctance to abandon the communicative activity of speech even in the face of writing. Jeffrey Kittay, in "Utterance Unmoored: The Changing Interpretations of the Act of Writing in the European Middle Ages," Language in Society 17 (1986), 209-30, suggests that it may also have been necessitated by a deciphering problem that only vocalization could resolve. In order to understand manuscript writing — which often consisted merely of a series of letters with imprecise boundaries between words and sentences, given the inconsistent use of spacing, punctuation, and majuscules —■ the reader was obliged to sound it out. Compare the following word-level variants from two manuscripts of Marie de France's Lai de Lanval as cited by Bernard Cerquiglini, Jacqueline Cerquiglini, Christiane Marchello-Nizia, and Micfiele Perret-Mitiard, "L'objet 'ancien francais' et les conditions propres a sa description linguistique," in Methodes en grammaire francavse, ed. Jean Claude Chevalier and Maurice Gross (Paris, 1976), pp. 185-200 at p. 192:
MS H: net ensouvient    MS C: ne lensouvir.nl        "he doesn't remember" nel apele ne tapele "he doesn't call (her)"
del amur de lamur "of love"
un kit ot servi un que lout servi • "one who had served him."
As the authors observe, elided object pronouns and definite articles are regarded in H as enclitic, in C as proclitic.
Kittay's observation that "words as units do not exist on the [manuscript] page" (p. 215) is correct, but from this it does not necessarily follow that words "were created by the reader as the writing was vocalized." Evidence from linguistic field workers puzzling out the grammars of unwritten languages, together with synchronic and diachronic evidence of "false segmentations" ( petit Tom" for petit homme; "tes dicaments" calqued on medicaments; older I'endemain becoming modern [(e) lendemam, with agglutination of the article), suggests that our lay notion of "words" is inescapably tied to the grammar of written languages and to the spacing that writing inevitably
The Discourse of the Medieval Text
21
only gradually achieved the place it has today in cultures where most reading is silent. Hence it is not surprising to find "oral residue" in texts known to have been composed in writing and under circumstances of increasing literacy.5
Medieval vernacular documents confront the modern reader with idiosyncrasies and incoherences, with gaps in the text, with anomalies of grammar, script usage, and text structure. In the Old French corpus, for example, word boundaries and orthography are inconsistent; case marking is idiosyncratic and obviously no longer functional;6 tense usage seems to defy grammatical logic, with jarring alternations between the past and the present; clauses are often simply strung together, with little or no formal "connective tissue" (subordinators and coordinators) to articulate the junctures;7 narrative chronology can be illogical, showing major temporal gaps and prolepses as well as conspicuous repetitions of the same events.
Attempts to explain these "anomalies" have typically involved vague ideas about the "expressivity" of the early language or its "greater freedom and grammatical flexibility" as compared with the language of modern texts. However, a compelling case can be made for viewing the language of Old French texts as not yet a codified, written idiom; from the standpoint of its grammar and discourse structure, Old French is very much a spoken language, the communicative instrument of a fundamentally oral culture,
imposes on the stream of speech. What is unassailable in Kittay's argument is that "writing did not bring with it the ideographic status of the word" (p. 215).
5 The term "oral residue" was coined by Walter J. Ong, S.J., in "Oral Residue in Tudor Prose Style," PMLA 80 (1965), 145-54. Franz Bauml, in "Medieval Texts and the Two Theories of Oral Formulaic Composition: A Proposal for a Third Theory," New Literary History 16/1 (1984), 51-66, at pp. 41 ff., makes the case that certain marks of orality are "by their writtenness" converted into stylistic attributes of written texts. In written texts features of orality often have a "referential" function quite different from the "mechanical" (that is. pragmatic) one they had in their oral antecedents. Bauml illustrates this argument with the example of epic formulas. Formulas are a mechanism essential to the process of composition-in-performance as well as to the reception and retention of material in primary oral epics. They are also culturally essential references to the tradition they encapsulate and transmit. In the process of formulaic written composition, formulas no longer fulfill a crucial mechanical/pragmatic function but take on a referential function: they refer to a specific type of text (oral) and thus represent the convention which determines the composition of the written text.
6 In my book Tense and Narrativity: From Medieval Performance to Modem Fiction (Austin and London, 1990), I argue, apropos of tense-aspect morphology, that when morphology is available that is no longer being used for its basic grammatical function, languages — being efficient natural economies — will often "recycle" this morphology to do other kinds of work. It is evident from the Old French texts that as early as the eleventh century case morphology was already dysfunctional. On the recycling of flexional -s prior to its disappearance altogether, see Cerquiglini et al., "L'objet ancien francais,'" pp. 185-91.
' Even where interclausal connectives are present, their functions in the articulation of discourse are not always clear. Cerquiglini et al., in "L'objet 'ancien francais,'" observe a blurring of the distinction between coordination and subordination: the texts show que, which grammars call a "subordinator," to be at times interchangeable with "coordinativc" el and si, thereby calling into question the validity of the taxis distinction for Old French.
22 The New Philology
adapted — sometimes better, sometimes worse — to writing.8 It has been characterized as the "literary elaboration" of a language which, while it served as the vehicle for a vernacular poetic tradition already in place by the twelfth century, "had not yet become the object of a grammatical discourse."9
The oral mental habits of all languages that have not grammaticalized writing necessarily leave their mark on linguistic structure. Orality played a crucial role in shaping the grammar (in the linguist's sense) of medieval vernaculars and, consequently, the linguistic structure of our texts. Yet in the extensive philological literature devoted to the Old French corpus, there has been relatively little productive analysis of orality. Not that medievalists have ignored the oral context in which medieval texts functioned, but simply that the implications of this orality — how it influences syntax, text structure, inscription, and the production of meaning in texts10 — do not carry forward into critical analysis. Like Paul Zumthor, in his provocative essay "The Text and the Voice," I invoke orality here "not so much ... to insist on [its] importance ... in the transmission and . . . creation of medieval poetry, but rather to appreciate and gauge what this orality implies; not so much to evaluate the size of the 'oral part' in the corpus of extant texts as to integrate into my perception and my reading the properties thus explained" (my emphasis).11
Certain of the "idiosyncrasies" of grammar, notation, and text structure that I and Romance philologists working in Paris (B. Cerqutglini, C. Mar-chello-Nizia, and M. Perret) have investigated bear striking resemblance to disconcerting phenomena confronting modern linguists whose object of study
8 Roger Wright, in Late Latin and Early Romance, m Spain and Carolingian France, ARCA Classical and Medieval Texts, Papers, and Monographs 8 (Liverpool, 1982), links the "invention" of Romance writing to the development of medieval Latin speech: for those in France trained to recite Latin correctly (that is, according to the reformed pronunciation introduced by Alcuin, ca. 800), an oral performance in the vernacular could not be unambiguously scripted except by adapting the standard Latin orthography to the non-Lattnate phonetics of what was by then "French." Wright thus situates the origins of vernacular writing not in an attempt to record what the writer has heard others say, but what he wants others to say. In line with this view he argues that the two earliest documents in French — the Oaths of Strasbourg (842) and the Sequence of St. Eulaiia (ca. 880) — were written the way they were in order to prompt a desired oral performance; they were not, as traditionally assumed, attempts at a phonetic rendering of what was said.
9 Cerquiglini el al., "L'objet 'ancien franc.ais,'" p. 191. The terms "grammar" and "grammatical" are to be understood here not in the linguist's senst in which all languages have grammars, resident in the heads of their speakers, but in the normative sense of ban usage — a concept inseparable from writing.
10 Ethnomethodological research into oral versus literate strategies in discourse suggests that in literate traditions "the meaning is in the text," in the actual written words, while in oral situations "the meaning is in the context" and in the implications of communicative acts. See Jack Goody and Ian Watt, "The Consequences of Literacy," in Literacy in Traditional Societies, ed. Jack Goody (Cambridge, Eng., 1968), pp. 27-68; David R. Olson, "From Utterance to Text: The Bias of Language in Speech and Writing," Harvard Educational Review 47 (1977), 257-81; and Richard Bauman, Story, Performance, Event: Contextual Studies of Oral Narralive (Cambridge, Eng., 1986). See also n. 19 below.
11 Paul Zumthor, "The Text and the Voice," New Literary History 16/1 (1984), 67-92.
The Discourse of the Medieval Text 23
is likewise a spoken idiom. Consider the sentences below from thirteenth-century French prose:
(1) Et quant Aucassins Yentendi: "Ha, Dix.," fait il, . . . "sont cou mi anemi mortel?" (And when Aucassin heard this, he goes: "Oh God! . . . are these my mortal enemies?") Aucassin et Nicolette 10:16—18.
(2) Li empereres, cant il oi ensi parler le mesagier, si fu tous esbahis. . . . (The emperor, when he heard the messenger speak thus, [he]'2 was thoroughly confounded. . . .) Li conies dou roi Constant i'empereur?'
These sentences, which illustrate respectively the phenomena of tense switching and "left-dislocated" topicalization, are analogous to structures that occur commonly in spontaneous informal speech: "So when he heard the news, he goes: 'My God, 1 won the lottery!'" or "Mon frere, des qu'il a epouse sa femme, on ne le voit plus" ("My brother, ever since he got married, you never see him any more").
Regrettably, most modern descriptive grammars (normative or generative) have little if anything to say about such constructions of spontaneous speech. Transcribed on paper, spontaneous spoken language often strikes the literate observer as unstructured and incoherent — in a word, ungrammatical. Recently, however, these and analogous constructions observed in natural speech have begun to capture the attention of linguists who have sought to shift the focus of inquiry, particularly in the domain of syntax, away from a search for elegant descriptive formalisms of isolated (and often constructed) sentence types, toward an understanding of the "pragmatic" functions of discursive strategies used by real speakers in real contexts of communication." As a linguistically oriented philologist, I am convinced that many of the disconcerting properties of medieval vernacular texts — their extraordinary parataxis, mystery particles, conspicuous anaphora and repetitions, "proleptic" topicalizations, and jarring alternations of tenses, to cite but a few — can find more satisfying explanations if we first of all acknowledge the extent to which our texts structure information the way a spoken language does, and then proceed to the linguistic literature that explores the pragmatic underpinning of parallel phenomena in naturally occurring discourse.
12 Words enclosed in brackets do not occur in the Old French.
13 Ed. L, Moland and C. d'Hericault in Nouvelks franqmses en prose du Kllle siecle (Paris, 1856), p. 30.
14 Linguistic "pragmatics" focuses on contextualized uses of language, that is, on language viewed not as an abstract system detached from its users, but as a communicative instrument that responds to and is shaped by the pressures of actual situations of verbal interaction. It deals with speakers' "communicative competence," the knowledge that enables them to produce and comprehend utterances in relation to specific communicative purposes and specific speech contexts. Part of the task of pragmatics is to describe how larger utterances and verbal exchanges cohere internally, the kinds of communicative functions they perform, and the unspoken "rules" assumed to be in play when the language is used. Its domain of analysis is, therefore, ultimately discourse, not the sentence.
24
The New Philology
The Discourse of the Medieval Text
25
from the voice to the text
A aucun moment le linguiste ne doit perdre de vue qu'une langue "morte" tut est irr£mediablement etrangere.
— Christiane Marchello-Nizia15
Though from the standpoint of grammar and discourse structure the language of medieval documents is essentially a spoken language, it is at the same time exclusively a "text language."16 This raises the issue of the gap that separates the performance from the manuscript, the voice from the text.
We cannot assume that any extant manuscript represents a faithful representation of what was originally performed. This holds true for scripted performances (texts read aloud from a manuscript) as well as for texts composed in performance. As Zumthor observes, the fact that these texts have come down to us mediated by writing means that our efforts to approach their orality will be at best approximative;17 however the texts may have originally been composed and received, we confront them as fixed, written documents from which much of the paralinguistic and pragmatic information they contained as "storytelling events" has been lost.18 Add to this the alterations introduced in the process of scribal transmission and then again through modern editing, and the result is a text at a considerable remove from its oral origins. But should we therefore abandon all efforts to determine what features of our texts can be explained by orality? I think not.
For there are clear traces of the oral infrastructure that have survived the transformation from performance to manuscript. Some have even survived the transformation to printed editions, though more have been obliterated by editors' notions about what makes a "good" text.19 As Cerquiglini notes,
:5 "Question de methode," Romania 106 (1985), 481-92, at p. 484.
16 By this term i mean simply a dead language or "langue de corpus" (Marchello-Nizia, "Question de methode"), one for which there are no native speakers and all evidence derives from texts. This should not be confused with what Walter Ong refers to as a "textualized" language, such as Latin was throughout much of the Middle Ages; see "Orality, Literacy, and Medieval Textualization," New Literary History 16/1 (1984), 1-12. Given that Latin had ceased to be a vernacular around 500-700, no one who spoke it during the period under discussion (roughly the twelfth to fifteenth centuries) had learned it the way one normally acquires a native language; all who knew Latin had acquired it through the use of writing. As Ong notes (p. 7), everything that was spoken in Latin was measured against its written texts, for independent of these if had no existence.
We might view the relationship between Latin and Old French (or any other medieval vernacular) during the later Middle Ages as one of "complementary distribution." For Latin, the text controlled the voice; speech was modeled on a written idiom with an established tradition of grammar and rhetoric. For French, the voice controlled the text, improvising as best it could a functional ecriture.
17 In La po&sie et la voix dans la civilisation me'die'vak, Essais et Conferences, College de France (Paris, 1984).
13 Marchello-Nizia, "Question de methode," p. 485, states this position even more strongly: "II s'agit toujours d'une langue de registre ecril; quelque effort que I'on fasse, on ne peut identifier cette face cached de la langue medievale qu'6tait la pratique orale. . . ."
" Hildegard L. C. Tristram, Tense and Time in Early Irish Narration (Innsbruck, 1983), surveys a corpus of Middle Irish texts that preserve features of orality despite conspicuous evidence of written composition (see also n. 5, above). Among these features is tense switching. She suggests
the most valuable information linguists get from native speakers of a living language is the dynamic of the language (le mouvement langagier), the play of form and meaning; it is precisely this, he argues, that is offered to us by the manuscripts but typically lost once they are edited.20 If the task of philology, "the ancillary science," as Cerquiglini describes it (p. 110), is to establish the data for historico-comparadve linguistics and the other hermeneutic branches of medieval studies, then one appropriately "postmodern gesture" of the New Philology, as suggested by Stephen Nichols in his introductory remarks, is a return to the manuscripts, not merely as sources of editions, but as "the original texts."
writing in the plural
L'ecriture medievale ne produit pas des variantes, elle est variance.
— Bernard Cerquiglini2'
Cerquiglini's recent critical essay, Eloge de la variante, responds directly to this challenge, outlining the practice of a postmodern textual criticism in which "the text" is destabilized into the plurality of its variants.
The editor of a medieval text is typically confronted with manuscript variation. In such a situation, Cerquiglini insists, it cannot be decided — nor is it interesting to ascertain — which variant is closest to the elusive Urtext. The philologist's task should be comparison, not archaeology,22 since the latter reduces to singularity what acquires meaning precisely through plurality, through variation.23
Medieval literary aesthetics, Cerquiglini argues, was until the end of the thirteenth century founded on an "ecriture de la variance," which the prevailing methodology of textual criticism has served to camouflage. He points, first of all, to philology's traditionally atomistic focus on individual lexical or grammatical elements, on details, typically at the expense of the larger picture. The (more or less) critical treatment of variants has thus traditionally concentrated on the word. Yet Cerquiglini insists that "the variant is never punctual"; it is not at the word level that the mechanism of variance reveals
that the extent to which tense switching has been preserved in the manuscript form of a tale be considered an index of that manuscript's proximity to its oral origin. In my forthcoming book (7"ense and Narrativity) I draw similar conclusions concerning the manuscripts of Villehardouin's Conquest of Constantinople. 20 Cerquiglini, Eloge de la variante, p. 108. Ibid., p. 111.
22 However, the literary-theoretical context of Cerquiglini's reevaluation of the medieval variant is, in fact, "archaeological." It is an approach known as "literary genetics," which over the past decade has been enjoying some currency, particularly in France, as applied to modern literature. The thrust of this approach is a shift in emphasis away from "the text" as fixed and final product and onto the process of its coming into being; on additions, deletions, and alterations to the text, on versions sous rature, all of which are perceived to be as interesting, if not more so, than their ultimate reifkation into "the work."
23 Cerquiglini, Eloge de la variante, p. 67.
26 The New Philology
itself. It does so at the level of the sentence, perhaps, though its operation is most visible at the higher level of discourse.24
To make the point that little can be learned about the aesthetic of medieval writing from the traditional philological protocol of printing the "best" manuscript and noting variants individually at the bottom of the page, Cerquiglini compares variant readings in paragraph 70 of Villehardouin's Conquest of Constantinople (thirteenth century). The standard text is Edmond Faral's 1938 edition, based on MS O:
Icil1 Alexis sim prist son frere I'empereor", si° li traist les iaulz de la teste et sep fist empereorq en tel traison con vos avez oir. (This Alexis, [he]" took his brother the emperor; [he] pulled his eyes out of his head and made himself emperor through the treachery you have just heard about.)
The apparatus cites variants in MS B and variants common to MSS CDE. When pieced together, the text in CDE reads:
Cil prist son frere l'empereor et li traist les iaulz de la teste et se fist empereor en tel traison com oes. (He [lit. "the latter"] took his brother the emperor and pulled out his eyes from his head and made himself emperor through the treachery you're hearing about.)
We observe, first, a difference in tense usage (present oes versus perfect avez oi), which will not concern us for now. We observe also a syntactic contrast that corresponds to a basic difference in the way information is structured.26 The CDE version is syntactically a single sentence with the "demonstrative" cil (referring to Alexis) as subject of three coordinated predicates conjoined by et. By contrast, the O version contains two sentences, not formally conjoined; both begin with si (the second contains two predicates coordinated by et), and neither has an expressed subject. The particle si marks the left-hand boundary of sentences in which it occurs; its occurrence in O leads to an analysis of icil Alexis not as a grammatical "subject" but as a discourse "topic,"27 left-dislocated outside the boundary of the sentence.
24 Ibid., p. 111.
™ Although, as previously noted, words enclosed in square brackets do not occur in the Old French, my translation "This Alexis, he . . ." appropriately renders the topicalszing syntax of MS O.
26 Along similar lines, Daniel Poirion, in "Les paragraphes et le pre-texte de Viliehardouin," Langue francaise 40 (1978), 45-59, has shown that the paragraph divisions in Faral's edition do not reflect the disposition of the text in the manuscripts, which show variation in this regard. Faral's paragraphing follows that proposed by the nineteenth-century philologist Natalis de Wailly and adopted by most subsequent editors. From a comparison of the actual paragraph divisions of MSS A and B, representative of the two major families, Poirion concludes that modern editors have segmented Villehardouin's material according to a principle of chronology that was not the author's own organizational principle and that, moreover, camouflages his thinking and artistry. An edition of MS B is available, prepared by the CRAL research team at Nancy (1978), which adheres to the original punctuation and paragraphing of that manuscript.
2' "Subject" is a syntactic category, referring to the nominal argument in a sentence that governs agreement marking on the verb; "topic" is a category of information structure used to refer (I grossly simplify) to "what is being talked about" at a particular point in the discourse. Topic status normally correlates with information already "given" in the discourse or "recover-
The Discourse of the Medieval Text 27
The difference in discourse organization that emerges from a comparison of the manuscripts — subject-verb-object versus topic-focus — reveals itself only if we take a broader view of the variance that informs, indeed defines, medieval writing. As Cerquiglini puts it: "Medieval writing does not produce variants, it is variance. . . . Variance is its foremost characteristic: fluidity of discourse in its concrete alterity, the figure of a premodern writing, to which editing should give primary recognition."28
Cerquiglini's insistence on a return to the manuscripts as a step toward understanding the "premodern writing" of medieval France constitutes the first of a series of "postmodern gestures" in the refurbishing of philology. Once the data are established (not to minimize the effort involved in this undertaking), it still remains to interpret them. Radically simplifying, we can divide this hermeneutic process into two stages. The first, which I will concentrate on in the remainder of this article, involves deciphering the texts as linguistic documents: what the language means; how it works; how its grammar responds to demands of the communicative contexts in which it was used. This task has traditionally been assigned to the branch of philology known as historical grammar.29 The subsequent hermeneutic stage, discussion of which I leave to the other participants in this forum philologicum, involves investigating the texts as monuments — poetic and/or historical artifacts (these two categories do not exhaust the typology, nor are they discrete) of a premodern culture — and seeking out new ways to make them continue to speak to us.
2. Grammar, Pragmatics, and Medieval Textuality
grammar and postmodernism
Jjaamffijir is a becoming rather than a being.
— Paul J. Hopper30
Manuals of the early vernaculars have traditionally been written by grammarians or early language "specialists," not by linguists. They are, accord-
able" by the addressee, while what it contrasts with, "focus," correlates with "new" information about the topic which is introduced into the discourse at that point. Though topics do tend to appear as sentence subjects, the two categories must be kept distinct at the theoretical level. The claim has been made, notably for French (see Knud Lambrecht's forthcoming book, Information Structure and Sentence Form: The Pragmatics of Syntax in Spoken French [Cambridge, Eng.]) that the syntactic structure of informal speech seems to be determined less by syntactic typology (that is, the unmarked ordering of the principal syntactic constituents subject, verb, and object) than by basic, cognitively motivated principles of information structure, one of which is germane to our example from Viliehardouin: state your topic first; then predicate something about it. 28 Cerquiglini, Eloge de la variante, pp. Ill ff.
" The "historical grammar" rubric conflates two approaches to language whose methodologies are quite different: one has as its goal a synchronic description of an earlier elat de langue; the other, an analysis of the changes a language has undergone in the course of its diachrony (see Marchello-Nizia, "Question de methode"). Only the former approach is of concern here.
M "Discourse Analysis: Grammar and Critical Theory in the 1980s," Profession 88 (1988), 18-24, at p. 23.
28 The New Philology
ingly, even in recent instances,31 largely unreflective of developments in linguistic theory and methodology. The manuals still virtually all adhere to a structuralist "parts of speech" approach to grammar, which explains the overall emphasis on morphology at the expense of syntax; to the extent that syntax finds a place in this approach, it is predominantly sentence based. Yet over the past two or three decades many linguists have come to acknowledge that certain linguistic phenomena, syntactic phenomena in particular, can only be understood properly from the viewpoint of their functional motivation in the higher-level units of language referred to as discourses or texts.
Discourse analysis in linguistics has endeavored to explicate grammar in terms of contexts larger than a single sentence, by showing either that sentence-level phenomena have some kind of grounding in discourse or that the sentence itself is suspect as a core unit. By considering entire discourses as the core units of a data base, proponents of a text-based linguistics "argue, or assume, or hope that . . .the hitherto invisible problems of context raised by isolated sentences will be thematized."32 The move in linguistics toward discourse can, if taken a step further, have profound ramifications for a theory of grammar that are a fortiori germane to the project of formulating grammatical descriptions of "text languages."
Since the advent of structuralism we have become accustomed to thinking of grammar as a relatively stable structure or system which all adult members of a speech community possess in reasonably similar form, one in which forms — the recognized categories of the grammar — have meanings that hold constant independently of context. This view of grammar, which mainstream linguistics formalizes in terms of rules, underlies what little theorizing there has been about the grammatical description of medieval languages. Thus Marchello-Nizia writes: "For a dead language, a synchronic state can only be defined ... by the presence of a certain number of rules and consistent forms,"33 though she acknowledges the possibility of modifying the description, presumably through the inductive-deductive procedures linguists use to test and refine descriptive statements. The only difference would be that native speakers are replaced by texts.
In several recent papers Paul Hopper has proposed that we abandon this a priori notion of grammar, on which standard theories of linguistics are based, in favor of a postmodern view of grammar as "emergent," that is, not as a synchronically stable bedrock of form-meaning or form-function correlations, but as a set of linguistic transactions that are continually being negotiated in individual contexts of communication. As he states, "in place of forms with contexts ... we will have texts with forms, and contextuality will be replaced by textuality. Structure and grammar in general, instead of being seen as present a prion, will emerge out of quite concrete repetitions in
91 For example, William Kibler's Introduction to Old French, 2nd rev. ed. (New York, 1985), or Philippe Menard's Syntaxe de iancien francais, 3rd rev. ed. (Bordeaux, 1988). 32 Hopper, "Discourse Analysis," pp. 18 ff. 53 "Question de methode," p. 486.
The Discourse of the Medieval Text 29
discourse."34 The relationship Hopper posits between "structure" ("grammar" in the traditional sense) and "discourse" (speakers' use of forms in specific contexts — what is alternately referred to as "pragmatics") is a dialectical one: "structure, or regularity, comes out of discourse and is shaped by discourse as much as it shapes discourse in an on-going process."35 Structure, in this postmodern view of grammar, is thus always provisional and always deferred (that is, a continuous movement toward structure that is always only partially achieved), and, crucially, always "parasitic on texts."36
Medievalists deal with linguistic data that are constituted exclusively as texts. With no native speakers to appeal to, the texts are all we have. This situation prompted the Paris Romance linguists (Cerquiglini et al.) in 1976 to call for a "linguistics of the medieval text." Yet since that time little work has been done on the early stages of French along text-linguistic, or discourse-pragmatic, lines. The work that has been done, on which I propose to comment, might be construed as yet another postmodern gesture that may slow down, and indeed perhaps reverse, the eclipse of philology.
is there a speaker in this text?
C'cst dans et par le langage que 1'hommc se consume comme sujet. .. . Est "ego" qui dit "ego."
— F.mile Benveniste
Among the basic presuppositions of language as a code for communication is the notion, which may at first seem trivial, that every utterance, every discourse, has a speaker and an addressee and is produced in a specific extradiscursive context. This context includes the time and place of the speech event, the identity of the participants and their relations to one another, plus a variety of cultural or real-world knowledge which the participants presumably share. The utterances of a text are in this sense not decontextualized pieces of language; even the act of writing, which may sever them physically from their origin, does not ipso facto obliterate connections to a speaker, a context, and the locutionary act that produced them — in French, their enonciation.
Some of the most illuminating recent work on the language of Old French texts is based on problematics of enonciation. This approach is founded on. the proposition that the utterances (enonces) of a text inevitably contain traces of the locutionary activity (enonciation) that produced them, the context in which they were produced, and the subjectivity of the producer. A major research agenda of the Paris Romance linguists has involved identifying and interpreting these linguistic traces through which the act of speaking and the beliefs and attitudes of a speaker leave their imprint on the surface structure of the text.
M Hopper, "Discourse Analysis," p. 22.
35 Paul J. Hopper, "Emergent Grammar," in Proceedings of the Thirteenth Annual Meeting of the Berkeley Linguistics Society, ed. Jon Aske, Natasha Beery, Laura Michaels, and Hana Filip (Berkeley, 1987), pp. 139-57, at p. 142.
M Hopper, "Discourse Analysis," p. 22.
30 The New Philology
Cerquiglini looks into the pragmatics of the Old French "mystery particle" mar,?7 illustrated in the following examples from Roland:**
". . . Felun pai'en mar i vindrent as porz. / jo vos plevis, tuz sunt jugez a mort." ("Vile pagans, woe unto you that have come to this pass. I promise you, you are all destined to die," vv. 1057-58.)
Li quern Rollant ki ne I'otreit mie, / En ptez se drecet, si li vint cuntredire. / II dist
al rei: "Ja mar crerez Marsilie----" (Count Roland, who is not at all in agreement,
rises to his feet, [he] went to contradict him. He said to the king: "Woe unto you who will believe the words of Marsile," vv. 194-96.)
The gloss "woe unto you" does not adequately capture the function of mar, which Cerquiglini characterizes as a "thetic modalizer," a marker of subjective evaluation through which the speaker of a sentence casts a prophetically negative light on the outcome of the proposition contained in it, a negative outcome which the grammatical subject of the sentence (the affected participant) does or did not see. Thus, we might paraphrase the first sentence above as "Vile pagans, I hereby judge your decision to come to Roncevaux pass to have been an unwise move — contrary to your understanding of this act!" — a reading that elucidates the articulation between this sentence and the one immediately following ("I promise you, you are all destined to die"). The issue of discourse connectivity (cohesion) in Old French will be discussed further below.
Two other monographs from the 1980s target different aspects of the language of Old French texts from the pragmatic perspective of their fonctions mandatives. Marchello-Nizia offers an exhaustive semantico-syntactic analysis of one of the most elusive elements of Old French grammar, the mystery particle si.39 She interprets si as an "evidential" particle signaling the speech act of assertion; by means of this particle, the speaker vouches personally for the truth of the proposition that follows — whence the book's title Dire le vrai.w While I am convinced that the function of this particle relates to the
"This term is borrowed from Robert E. Longacre's article on discourse strategies in Amerindian languages: "'Mystery' Particles and Affixes," in Papers from the Twelfth Regional Meeting, Chicago Linguistic Society, April 23-25, 1976, ed. Salikoko S. Mufwene, Carol A. Walker, and Sanford B. Stcever (Chicago, 1976), pp. 468-79. He observes (p. 468) that "in analyzing the sentences of languages found in some parts of Ihc world, certain . . . affixes and sentential particles may continue to defy analysis even at a relatively advanced stage of research. Typically, the native speaker uses such affixes and particles with complete assurance but is unable to verbalize anything very concrete as to their meaning and function. The analyst is tempted here to resort to a rather crass theory of free variation to explain the presence of these elements. . . . Almost inevitably 'mystery' particles and affixes of this sort are found to have a function which relates to a unit larger than the sentence, i.e., to the paragraph and the discourse." Mutatis mutandis, his observations are particularly germane to the Old French sentential particles mar and si.
58 Bernard Cerquiglini, La parole medievale (Paris, 1981).
59 Christiane Marchello-Nizia, Dire le vrai: L'adverbe "si" en franfais medieval, Publications Romanes et Franchises 168 (Geneva, 1985).
40 Roman Jakobson, Shifters, Verbal Categories and the Russian Verb (Cambridge, Mass., 1957), reprinted in his Selected Writings (The Hague and Paris, 1971), 2:130-47, introduced the term "evidential" as a tentative label for a verbal category which indicates the source of the information
The Discourse of the Medieval Text 31
pragmatics of discourse, I have reservations about the evidential analysis, likewise about the claim that si indexes a speaking subject or speech-act context, as has been argued more compellingly for mar and for the locative "shifters" ci, ca, la, and iluec.4]
Michele Perret takes a pragmatic approach to a set of locative adverbs — ci, ca, la, and iluec — occurring in French prose texts from the fourteenth and fifteenth centuries.*2 The first two members of this set (ci, ca) translate roughly as "here," the last two (la, iluec) as "there." But the point of Perret's analysis is that to understand how these terms function as discourse articulators of Middle French prose, we must consider not only their meanings, as "signs" (the first term of her title) on the level of langue, but also their "mentions" (the second term of her title) in parole, in actual discourse, to carry out various operations of reference: to the extradiscursive context ("deictic" ca, la, iluec); to previously mentioned elements in the context, conceptualized as a linear space ("anaphoric" la, iluec);ri and, in the case of ci, to nothing other than itself, to its own occurrence in the text through the linguistic activity that produces it. Perret's monograph demonstrates the inadequacy of a purely syntactico-semantic analysis and the need to look beyond "grammar" (as traditionally understood) to the pragmatics of "discourse" — in this instance to the pragmatics of prose writing, which during the Middle French period evolved its own discursive protocols as distinct from those of earlier verse texts. Nor is it accidental, Perret observes, that the emergence of a grammar of prose coincides with the passage from oral to written transmission of the cultural legacy and with the expansion of didactic literature in the vernacular.44
Linguistic implications of the transition from verse composition to prose composition are also explored in Cerquiglini's La parole medievale. Part 1 of the book focuses on transformations in the presentation of speech (whence his title) that occur as one moves from twelfth-century verse romances to their thirteenth-century prosifkations, in particular the replacement of directly quoted speech by indirect speech. Direct speech, I would point out, mimetically represents the act of speaking as well as offering a supposedly faithful transcript of the language of the quoted speaker; Nessa Wolfson
on which a speaker's statement is based. As currently understood, evidentiality covers a range of distinciions involved in the identification of the source of one's knowledge. Various languages have grammaticalized evidential markers indicating whether or not the speaker vouches personally for the information contained in a statement — the function Marchello-Nizia attributes to si though she does not use the term "evidential."
41 The term "shifters" was also introduced by Jakobson to designate a class of items which cannot be interpreted independently of the context in which they are uttered. These include personal pronouns, tenses (normally deictic to the moment of utterance), demonstratives, and certain adverbs of time and place ("now/then," "here/there") as well as terms like "mother" or "home."
12 Le signe et la mention: Adverbes embrayeurs "ci," "(a," "la," "iluec" en moyen francau (XlVe-XVe slides) (Geneva, 1988).
15 To avoid ambiguity, the term "cotext" refers to the discourse surrounding a particular utterance, "context" to its extradiscursive or situational setting. 44 Perret, Le signe et la mention, p. 264.
32 The New Philology
includes direct speech among the defining features of "performed stories" — a discursive rubric more appropriate to the verse romances than to their prose counterparts.45 Indirect speech reports only the content of a quoted speaker's utterance, transposed into the nonpersonal language of the narrative. In the indirect mode, which Cerquiglini links to the emergence of prose in the thirteenth century, all "interfering voices" (including that of a real author/composer) have been obliterated (p. 123).
These studies by Cerquiglini and Perret document the later stages of a linguistic trajectory through which the textual language of medieval France evolved from the discourse of a speaker to a discursive ecriture, a writing in which the voice of a subject no longer speaks. My own work on twelfth-century texts sheds light on earlier stages of this trajectory, when the text composer's presence as a speaking subject was more strongly imprinted on the language of the texts through various devices, including, but not limited to, direct speech. With the development of prose, we see the concomitant development of a discourse in which the authorial voice is silenced; the prose reworkings of the Lancelot-Graal material present themselves as a speakerless discourse — une enonciation sans enonciateur. Perret contends that this disappearance of the author as a deictic center for the text is what confers a performative function on the discourse articulator (i)ci: in the formulas "cy commence . . . ," "cy dit. . ." that punctuate Middle French prose; the locative adverb — which performs the act it describes — takes over the position formerly associated with a speaking subject.'16
Use of the locative adverb (i)ci as a discourse articulator is founded on a conceptualization of the text as an entity that unfolds through space: "Et cy se taist l'ystoire de lui . . ." (And here his story breaks off . . . , Melusine).'17 Middle French prose is alternatively "punctuated," at times copunctuated, by the adverb or(e) "now," which assumes a text that unfolds through time: "Ore retorne li contes a mon signeur Gauvain . . ." (Now the story goes back to Sir Gawain . . . , Lancelot).*6 To the extent that the spatial metaphor underlies a
15 Nessa Wolfson, "A Feature of Performed Narrative: The Conversational Historical Present," Language in Society 7 (1978), 215-37.
*16 Perret, Le signe et la mention, p. 265. The writings of such theoreticians as Roland Barthes, Maurice Blanchot, Michel Butor, and Michel Foucault link the concept of ecriture and its attendant grammatical markers (the passe' simple, nonpersonal pronominal reference — what Barthes calls "the third person of the novel," and indirect speech) to dominant themes of contemporary criticism: the impersonality of language and its artifacts, the absence of the author from the text (see Ann Banheld, "Ecriture, Narration and the Grammar of French," in Narrative: From Malory to Motion Pictures, ed. Jeremy Hawthorn [London, 1985], pp. 1-22). Through their respective discussions of indirect speech and desubjectivized deictics (notably "self-referential" ci), Cerquiglini and Perret locate the roots of these phenomena in the grammar of prose writing that established itself during the later Middle Ages.
" Jean d'Arras, Melusine, roman du XlVe slide, ed. Louis Stoff, Publications de l'Universite de Dijon, fasc. 5 (Dijon, 1932), p. 269.
"Lancelot, roman en prose du XHIe siecle, ed. Alexandre Micha (Pans and Geneva, 1980), p. 419. Cf. "Or se taist cy l'ystoire des deux chevaliers . . ." ("Now the story of the two knights breaks off here . . . ," MHusine, ed. Stouff, p. 82); "Or laisseray cy de parler de teste tres puissant noblesse des seigneurs . . ." ("Now 1 will cease here to speak about this very great nobility of the
The Discourse of the Medieval Text 33
notion of textual discourse based on writing (the text unfolds across the space of the page) and the temporal metaphor a notion of discourse based on speech (the text unfolds through the time of its recitation), it would be of interest to investigate a possible correlation between the passage from oral to written transmission of the cultural legacy and the use of temporal versus spatial adverbs as articulators of discourse.
It is not an arbitrary occurrence that a major focus of New Philological research is on the "little words" of vernacular grammar. This is in line with the redirection of attention from parts of speech to discourse, from an atomisticalty conceived morphology to the syntactic structure of texts. One fruitful area of syntactic inquiry that has now begun to be explored in Old French is that of discourse connectivity, that is, the language's resources for creating cohesive, connected "text."
utterances into discourse
A serious pragmatics should account for the functions of utterances with underlying textual structures.
— Teun A. van Dijk
Old French, like other medieval vernaculars, is a language in which the explicit connective tissue of discourse is limited. This is predictable from its status as a spoken language; speech in general requires less formal (i.e., grammatical) articulation than does writing. As Walter Ong explains, "written discourse develops more elaborate and fixed grammar because it is more dependent simply upon linguistic structure, since it lacks the normal full existential contexts which surround oral discourse and help determine meaning independent of grammar" (cf. note 10 above).49 Though no medieval vernacular displays the range of grammatical connectives (conjunctions, adverbs, particles of various sorts) that are developed in languages of high literacy to code temporal, causal, and argumentational relationships between utterances, a New Philology sensitive to the interplay between "grammar" and "discourse" will find it of value to study the "glue" that holds medieval discourse together.
As a first step in this direction, Harro Stammerjohann sets up a typology of connectives in the Oxford Roland, organized according to the level of discourse structure at which they operate: the clause, the sentence, the episode (i.e., the laisse), or more than one of these levels.50 Regrettably, he does not address the prior theoretical question of how one distinguishes a sentence from a clause in a text like Roland, whose manuscript syntax, like that of most orally composed verse texts, consists largely of independent units (clauses? sentences?), linked paratactically or asyndetic-ally and punctuated minimally. Is this decision left to the editors?
lords . . . ," Antoine de la Sale, jehan de Santre, ed. Jean Misrahi and Charles A. Knudson, Textes Litteraires Francais 117 [Geneva, 1965], p. 201).
19 Walter Ong, Oralily and Literacy: The Technologiiing of the Word (London, 1982).
50 Harro Stammerjohann, "Hierarchie des connecteurs dans la Chanson de Roland" in Operateurs synlaxiques et cohesion discursive, ed. Henning N0lke (Copenhagen, 1988), pp. 63-74.
34 The New Philology
Stammerjohann's brief paper represents a valuable first stage of research: identification and distributional classification of the data.3' From there, one might go on to investigate the pragmatic functions of the connectives that have been identified, as I attempt to do for the particle si.
Si occurs in virtually every French text (verse or prose) from the ninth-century Strasbourg Oaths through the fourteenth century. Its use declined in the fifteenth century, and by the sixteenth it was virtually extinct (in seventeenth-century texts it occurs only in passages of dialect or colloquial speech). Historical grammarians have in general regarded si as a piece of intersentential connective tissue (variously translated as "and," "thus," "and so," "and then," "therefore," or left untranslated), with meaning assigned contextually. As an alternative to the approach taken by Marchello-Nizia (above, p. 30), Noel Corbett suggests that si might profitably be studied from the perspective of the pragmatics of Old French oral literature. His proposal is founded on the belief that "systematic inquiry into the dynamics of discourse may help us to better understand the idiosyncrasies of medieval French."52
Corbett's suggestion, together with the Paris group's call for a "linguistics of the medieval text," poses a challenge to philologists who bring linguistic expertise to the study of medieval texts: a challenge to rethink certain of the received ideas about early vernacular grammar, to rewrite the grammars, in light of important insights afforded by a pragmatically based text linguistics. This is the approach taken in my own current inquiries into the operation of si, which draw on the discourse-based notions of "topic continuity" and "participant tracking" and which, accordingly, consider the rise and fall of .si in relation to the grammaticalization of subject pronouns.
telling it again and again
Repeating is a wonderful thing in being, everything, every one is repeating then always the whole of them and so sometime there will surely be an ordered history of every one.
— Gertrude Stein
The pragmatics of oral literature may also be invoked to shed light on a salient — and for literate readers disconcerting — feature of the compositional technique of Old French epic: the "similar blisses." In several chansons de geste, Roland in particular, the movement of narrative time is halted at strategic points in the story by sequences of two to three laisses that repeat the same basic information with variations of detail. Most frequently emphasized by literary analysts is the so-called lyric dimension of these moments of
51 Earlier studies along this line, which similarly emphasize distribution (syntax) over function (pragmatics), include Wolf-Dieter Stempel, Vnlersuchungen iwr Sativerknupfung im Altfranzosischen (Braunschweig, 1964), and two works by Jean Rychner, Formes et structures de la prose fraricaise medievale: L'articulation des phrases narratives dans "La Mort Arlu" (Neuchatel, 1970), and "Analyse d'une unite transphrastique: l.a sequence narrative du tnemc sujet dans La Mort Arlu" in Beilrdge vur Textlinguistik, ed. Wolf-Dieter Stempel (Munich, 1971), pp. 79-122.
M See Corbett's review of Marchello-Nizia, Dire It vrai, in Romance Philology 42 (1988), 92-96; the quotation is from page 96.
The Discourse of the Medieval Text 35
narrative stasis and their role in the elaboration of a poetic technique that has elevated the Oxford Roland to a position of preeminence within the epic corpus. With no intent to minimize the artistry of Roland, I have chosen to investigate this highly formalized repetition from the quite different perspective of its pragmatic function in oral storytelling.53
The similar laisses bear striking formal similarity to a narrative protocol observed in the "natural" storytelling practices of several nonliterate languages. This protocol has been labeled "overlay."54 In the languages in which it occurs, overlay serves two discourse-pragmatic functions crucial to effective story performance: monitoring information flow (i.e., controlling the pace of the discourse and the rate at which "new information" is introduced) and foregrounding (i.e., providing distinctive encoding for units of information that the storyteller wishes to highlight so as to focus listeners' attention on them). The formalized repetitions of the similar laisses can be shown to carry out these same discourse-pragmatic functions in the orally composed song poetry of medieval France.
An underlying agenda in this analysis, which draws on research into natural narrative protocols observed in non-European languages with strong oral storytelling traditions, is to make the case that certain strategies viewed by literary critics as hallmarks of poetic craftsmanship can and do occur in nonliterary discourse and in diverse language traditions. My intent is not to question the poetic effect of devices such as the similar laisses, but simply to elucidate their linguistic underpinning, their functional motivation in oral story performance.
il mist la main al'espee, si commence a ferir . . .
Un changement de temps suffit a recr^er le monde et nous memes.
— Marcel Proust
The title of this section is taken from a prose passage in Aucassin et Nicolette. What attracts our attention is the alternation of tenses, unpredictable from the rules of temporum conseculwr "he placed his hand on his sword, [he] begins to strike. . . ." La confusion des temps — as the switching phenomenon has been referred to in French — turns out to be extremely widespread in early vernacular narrative texts and over the years has spawned a considerable body of critical literature: stylistic, literary, philological. My own approach to the problem involves a radical shift in the premises and direction of inquiry, away from exclusive enclosure within the problematics of medieval textuality
53 See my two articles "Overlay Structures' in the Song of Roland: A Discourse-Pragmatic Strategy of Oral Narrative," in Proceedings of the Twelfth Annual Meeting of the Berkeley Linguistics Society, February 15-17, 1986, ed. Vassiliki Nikiforidou, Mary VanClay, Mary Niepokuj, and Deborah Feder (Berkeley, 1986), pp. 108-23, and "A Linguistic Perspective on the Laisses similaires: Orality and the Pragmatics of Narrative Discourse," Romance Philology 43 (1989), 70-89.
M Joseph E. Grimes, "Outlines and Overlays," Language 48 (1972), 513-24.
36
The New Philology
— or literary textuality in general — to an area of discourse linguistics that has of late drawn considerable attention: the analysis of "natural" narrative.55
Recently, tense switching has begun to be explored in a more rigorous and informed way by linguists concerned with understanding how we construct and organize the narrations that surface in everyday verbal interaction and with identifying linguistic strategies that make for effective storytelling. The tense-switching patterns that have been identified in natural narration are strikingly similar to those found in medieval narrative texts. As in the case of overlay structures, it was this surface similarity that prompted my "revisit" to the medieval texts, armed with the analytical tools used in studying natural narrative. Admittedly, the inner-city raconteurs interviewed by William Labov in the 1960s have little in common with singers of epic and reciters of romances — except for the pragmatic imperative common to all storytellers: to make their accounts of past events as lively and interesting as possible. Thus the cultural and time gaps narrow progressively the more we learn about the linguistic foundation of their shared expertise in oral story performance.
My work on performance narrative has led me to identify a number of pragmatic functions that tense and aspect are called on to perform in medieval narrative, functions that go beyond — indeed, may even cancel — their basic grammatical functions (establishing the time reference and situational profile
— bounded or unbounded — of reported events).
Among textual or discourse functions, tense-aspect contrasts are used for grounding (delineating the textual foreground and background),56 creating cohesion, marking boundaries of various types within the text (e.g., demarcating information "spans" corresponding to a participant,57 setting, or macro-event), and modulating the pace of the discourse.
Among expressive functions (communicating elements of subjectivity), we find tense aspect pressed into service for the encoding of evaluations and signaling of point of view.
Finally, tense-aspect categories have a metalinguistic function: to "announce a type of language," to situate a text or piece of text within a typology of discourse forms. This function is of particular relevance to primary epic, which universally chooses the present tense as the basic grammatical vehicle of its discourse. According to the theory of tense in narrative language developed in my book Tense and Narralivity, the metalinguistic function of the
M In addition to my book (cited above, n. 6), see my articles "Discourse Functions of Tense-Aspect Oppositions in Narrative: Toward a Theory of Grounding," Linguistics 23 (1985), 851— 82, and "Evaluation in Narrative: The Present Tense in Medieval 'Performed Stories/" Yale French Studies 70 (1986), 199-251.
See also Barbara Wehr, Diskurs-Strategien im Romanischen, Romanica Monacensia 22 (Tubingen, 1984), who documents the foregrounding function of the narrative present tense in several varieties of early Romance.
57 Rychner, "Analyse d'une unite transphrastique," p. 82, claims that the paragraph divisions of La Mart Artu serve this same discourse function of demarcating participant spans. He invokes the paragraphing as evidence that the primary concern of this text is the protagonists, not the action.
The Discourse of the Medieval Text
present tense is to announce language that is not narrative, albeit "according to the rules of narrative's own game." In several respects, which I cannot elaborate on here, the premodern textuality of the chansons de geste resembles the postmodern textuality of the nouveau roman.is This resemblance is not gratuitous; indeed it is predictable from the spectrum of properties, located at various levels of linguistic structure, that the present tense implicates. It is at the metalinguistic level that we observe most clearly how the choice of a tense in fiction is much more than just a grammatical agenda.
In the preceding pages I have sketched parts of an itinerary across a terrain of inquiry where modern linguistics and medieval philology converge. By surveying recent research that has in various ways brought the pragmatic concerns of a discourse-based linguistics to bear on the textual artifacts of medieval France, I have sought to identify several territories in which the New Philology might profitably stake a claim. If philology is to remain viable as the science entrusted with constituting medieval texts, then the New Philologist must proceed along different paths from those staked out by "Gaston Paris and the dinosaurs"59 and traversed by textual editors ever since. If it is to move beyond an atomistic approach to language and to grammar, aimed simply at "filling in little holes on the great map of knowledge," then the New Philologist must, insofar as possible, recontextualize the texts as acts of communication, thereby acknowledging the extent to which linguistic structure is shaped by the pressures of discourse. It is through these and similar gestures that we might ultimately reformulate philology's role in the field of medieval studies, adapting its praxis to the challenges of postmodernism.
58 The interested reader is referred to chapter 8 of my book Tense and Narralivity.
59 See n. 1, above.
Suzanne Flehckman is Professor of French and Romance Philology at the University of California, Berkeley, CA 94720.