Medieval Autographies: The "I" of the Text

The Conway Lectures in Medieval Studies

2008

The Medieval Institute gratefully acknowledges the generosity of Robert M. Conway and his support for the lecture series and publications resulting from it.

PREVIOUS TITLES IN THIS SERIES:

Paul Strohm

Politique: Languages of Statecraft between Chaucer and Shakespeare (2005)

Ulrich Horst, O. P.

The Dominicans and the Pope: Papal Teaching Authority in the Medieval and Early Modern Thomist Tradition (2006)

Rosamond McKitterick

Perceptions of the Past in the Early Middle Ages (2006)

Jonathan Riley-Smith

Templars and Hospitallers as Professed Religious in the Holy Land (2009)

A. C. SPEARING

MEDIEVAL AUTOGRAPHIES

The I of the Text

UNIVERSITY OF NOTRE DAME PRESS

NOTRE DAME, INDIANA

Copyright © 2012 by the University of Notre Dame Press

Notre Dame, Indiana 46556

www.undpress.nd.edu

All Rights Reserved

E ISBN 978-0-268-09280-1

This eBook was converted from the original source file by a third-party vendor. Readers who notice any formatting, textual, or readability issues are encouraged to contact the publisher at ebooks@nd.edu

CONTENTS

Preface

ONE: The Textual First Person

TWO: Autography: Prologues and Dits

THREE: Chaucerian Prologues and the Wife of Bath

FOUR: Why Autography?

FIVE: Hoccleve and the Prologue

SIX: Hoccleve’s Series

SEVEN: Bokenham’s Autographies

Afterword

Notes

Bibliography

About the Author

PREFACE

This book originated as the Robert M. Conway Lectures in Medieval Studies given at the University of Notre Dame in October 2007. I was honored to be invited to give these lectures, and I am most grateful to Tom Noble, then director of the Medieval Institute at Notre Dame, to his wife, and to his colleagues for their generous hospitality during my stay there. I am also grateful to those who heard the lectures for their searching questions and valuable suggestions, which have helped to make the book less inadequate than it would otherwise be. I owe special debts of thanks to Roberta Baranowski, associate director of the Medieval Institute, for much good-natured practical help and many entertaining e-mail messages, and to Barbara Hanrahan, then director of the University of Notre Dame Press, for her warm encouragement and shrewd guidance when I was struggling to plan the book.

The tortuous process of converting and enlarging three lectures into a book that often bears little resemblance to its original form has been eased, and the book itself much improved, by the kind colleagues and friends who have read drafts and discussed problems with me. My obligations are too many to be recorded in detail, but I should like to thank Peter Baker, Cristina Cervone, Deborah McGrady, Gary Saul Morson, and especially Elizabeth Fowler. I am most grateful to Elizabeth Spearing, who stepped forward at a crucial moment, read the whole, and made invaluable suggestions for improvement. Some more specific debts are recorded in notes to the text. It goes without saying that the book’s faults are my responsibility alone.

Some parts of the book’s argument and occasional ideas and sentences have previously appeared in the following: The Poetic Subject from Chaucer to Spenser in Subjects on the World’s Stage: Essays on British Literature of the Middle Ages and the Renaissance, ed. David G. Allen and Robert A. White (Newark: University of Delaware Press, 1995), 13–37; "Textual Performance: Chaucerian Prologues and the French Dit," in Text and Voice: The Rhetoric of Authority in the Middle Ages, ed. Marianne Børch (Odense: University Press of Southern Denmark, 2004), 21–45; and Was Chaucer a Poet? Poetica 73 (2010): 41–54. I am grateful respectively to Associated University Presses, to Professor Marianne Børch, and to Professor Toshiyuki Takamiya for permission to reuse this material here. Some material also derives from A. C. Spearing, Dream Poems, in Chaucer: Contemporary Approaches, ed. Susanna Fein and David Raybin, 159–78 (University Park: Pennsylvania State University Press, 2010); copyright © 2010 by the Pennsylvania State University Press; reprinted by permission of the Pennsylvania State University Press.

In quoting from medieval texts I have silently modified editorial spelling and punctuation where I thought that would aid understanding; this goes against the grain of medieval scholarship, and with good reason, but I persist in hoping that some nonspecialists might be willing to learn more about premodern literature. Further, in the hope that the book might find a few nonmedievalist readers who are interested in the theoretical issues I discuss—issues that I believe ought to be the concern of others besides medievalists—I have added modern translations (my own unless otherwise specified) of all medieval texts quoted in the original, except those, generally very brief, whose meaning seemed obvious.

CHAPTER 1

THE TEXTUAL FIRST PERSON

In this book I attempt to bring into focus a category of medieval English writing that has not previously been recognized as such. I call it autography, and, put simply, it consists of extended, nonlyrical, fictional writings in and of the first person. A more precise sense of what this involves and why it matters will, I hope, emerge from studies of specific texts in the following chapters, but it may be helpful to begin by indicating how my recognition of this category—perhaps better called a supergenre than simply a genre—relates to work I did in an earlier book entitled Textual Subjectivity. Its subtitle was The Encoding of Subjectivity in Medieval Narratives and Lyrics, and in it I investigated some of the linguistic and formal features by means of which subjectivity is built into texts in the two supergenres of narrative and lyric, and I tried to show how attention to these features might affect literary interpretation. Much of the argument of Textual Subjectivity was negative, illustrating how, as it seemed to me, failures of attention to the way language works in specific medieval texts had led to widespread misinterpretations. Dissatisfaction with accepted readings of major works such as Troilus and Criseyde, The Man of Law’s Tale, and Pearl, and a growing conviction that they were indeed bad readings—and often perhaps worse than bad, because they seemed contemptuous of the actual achievements of great poets—led me to question the assumptions on which I came to see they were based. Underlying my argument was a distinction between the representation of subjectivity and its encoding in the written form in which all the medieval literature we know has come down to us.

To explain this as briefly as possible: An assumption of long standing is that writing is a representation of speech, with the consequence that, as one medievalist sweepingly puts it, no tale can be interpreted except as the product of a human speaker.¹ This axiom goes back to Plato but has been increasingly taken for granted over the last century and more, and, because unquestioned, has often remained unmentioned. Thus the assumption has been that in principle every word of a written fictional narrative is to be interpreted as representing the utterance of a fictional speaker distinguishable from the author—and this is no less true of lyrics and indeed of imaginative writing of almost every kind. As a typical statement of this assumption puts it, The writer creates a fiction when he attributes what he writes to another speaker; . . . he attributes the performance of his speech acts to a speaker he creates.²

Reading aloud to listeners was a common practice in the Middle Ages, as in the famous scene in Chaucer’s Troilus and Criseyde in which Pandarus visits Criseyde in her palace,

And fond two othere ladys sete and she,

Withinne a paved parlour, and they thre

Herden a mayden reden hem the geste

Of the siege of Thebes, while hem leste. (II 81–84)³

——

[And found her and two other ladies seated in a paved parlor, and these three were listening to a maiden reading the story of the siege of Thebes to them for as long as they pleased.]

But what is being read aloud is a book, and it is important to grasp that in medieval linguistic theory it was not assumed that writing must be a representation of speech; on the contrary, as Martin Irvine and David Thomson put it,

In medieval grammatical theory, language is unthinkable outside writing, and even the theory of speech was modelled on the properties of writing.

The model for articulate speech is writing, not spoken utterances. The vocal utterance considered the materia of the grammatical art was written, and, conversely, articulate speech was understood to bear the marks of writing. . . .

Speech bears the imprint of writing; indeed, speech is considered meaningful only as it manifests the distinctive features of writing. . . . In grammatical discourse, the Platonic sense of the secondariness of writing has been erased; speech and writing become dual manifestations of a single activity—signifying or the production of meaning.⁴

It is understandable that writing should have been central to linguistic thought in a culture in which the universal medium of intellectual discourse was Latin, a language that, to be sure, was often spoken but was spoken exclusively by those who could read it in its written form. And long after the rise of the vernaculars as media for abstract thought and for literature (in English from the fourteenth century onward), the attitudes belonging to Latinitas continued to hold sway.

A further assumption in modern interpretation of the literature of the past has been that an author’s purpose would be to produce a text coherent in perspective and ideology and that he or she could normally be expected to be perfectly in control of the text in fulfillment of this aim. In medieval literary studies, an article or chapter concerned with a single text typically argues that the text is more coherent than has previously been supposed—almost never that it is less coherent. Our standard academic tools for literary analysis are not well designed to make arguments against coherence, and the professional consensus is so strong that it would be hard for such arguments, if made, to achieve publication. Once you come to think about it, this seems strange. Reviewers of contemporary novels have no difficulty in noticing flaws in them, understood to be failures on the part of their authors. In any particular case the reviewer may be mistaken, but I suppose nobody would deny that even the most gifted and admired contemporary authors do sometimes fail; yet somehow it seems that almost every text surviving from the past, if analyzed with sufficient subtlety, can be shown to be a success in the sense of being more coherent than previously thought. The fictional speaker is a crucial part of the machinery of analysis used to transform past writings in this way. The author’s true meaning and his success in expressing it can be understood, it is supposed, by identifying the fictional person who is the text’s speaker and recognizing the gap between what he or she says and the coherent meaning really intended by the author. (If there were no such gap, it would of course be hard to see why the fictional person had been invented.)

Modern interpretations of medieval texts along these lines have usually concluded that the speaker of the text is designed to be unreliable, or at least limited in knowledge and understanding by his or her location and personal characteristics (as of course people are in everyday life), and that the medieval author who controlled the whole work intended a meaning different from that expressed by the speaker. That intended meaning then almost invariably turns out to be one that appeals to the taste of the twentieth- or twenty-first-century academic interpreter. Thus texts that apparently celebrate a warrior ethic turn out to be really pacifistic, those that apparently express misogyny turn out to convey feminist values, those that apparently satirize unnatural sexual behavior turn out to sympathize with it, those that apparently admire powerful rulers turn out to condemn them as tyrants, and so on. Such transformations of meaning are our modern equivalents to those brought about by medieval commentators in their interpretations of classical texts. The kind of truth that medieval commentators believed themselves to reveal about the intentions of classical authors is neatly illustrated as follows by a modern scholar: "Homer had intended to dissuade people from unlawful union, which—as in the case of Paris and Helen—incurs the wrath of the gods; Ovid had intended to reprehend inchastity [sic] and to commend legal and just love; Lucan had intended to discourage his readers from engaging in civil wars."⁵

It goes without saying that we cannot help reading the cultural products of earlier ages through the preconceptions and desires of our own culture, and also that the preconception of the fictional speaker is by no means the only factor in the way this has been done to medieval literature over the last century; but that particular preconception makes it especially easy to follow the natural inclination to see ourselves reflected in the mirror of the past. My cautionary argument in Textual Subjectivity was that, whatever might be the case with writings of the last hundred years and more (many of which do have unreliable speakers), medieval writings rarely represented the distinct subjectivity of a text’s fictional speaker, and their habit was to encode subjectivity in textual form by means such as deixis—subjectivity not usually that of a specific, self-consistent person but broadly and variously diffused throughout the text. It is hard to determine any specific point at which it became normal or common for writings to have fictional speakers. One interesting recent theory is that the change occurred in England in the late seventeenth and early eighteenth centuries, and was associated with the emergence of a public of critical size for whom silent reading was the norm and who therefore read not as speakers themselves but as hearers of a fictive persona addressing them in the physical absence of the writer.⁶ Whatever the cause, I believe the change to fictional speakers as the norm occurred long after the Middle Ages. It is true that in England, by the late fourteenth century, Chaucer’s Canterbury Tales shows the beginning of a move toward representing the subjectivity of fictional speakers, but it is only a beginning, and the meaning of many of the Tales has been distorted by attempts to see them in more modern terms as consistent expressions of the personhood of their pilgrim-tellers.

The argument I have just been sketching also underlies the present book, but here, as I have said, I am concerned with a third supergenre besides narrative and lyric, one that began to emerge in French in the thirteenth century but in English not till the fourteenth. As the following chapter will show, it has been identified by French scholars as the dit, but in English it has not previously been recognized and has therefore not been given a name. The term autography is not one I invented, though I may be among the first to apply it to medieval literature;⁷ in due course I shall give a more detailed account of its meaning and previous uses. There is of course no single right way to divide writing into different categories, and I shall need to approach this concept of autography rather circuitously, pausing from time to time for definition and explanation. Ultimately I can do no more than to invite readers to share in the experiment of seeing texts of a certain kind, some well known and others less well known, as constituting a significant cluster or family, and thus of reading them in an unfamiliar literary context and interpreting them in a different way. I am not purporting to offer a comprehensive theory of subjectivity in discourse, nor am I even proposing a program for detailed interpretation. I am only offering an invitation to try out a different kind of reading, one less hobbled by presuppositions about matters such as narrators or speakers, the comprehensive planning of literary works, and their achievement of aesthetic and ideological unity.

I have described autography as fictional, and to apply that term to a kind of medieval writing is already to beg some important questions. Medieval writers sometimes distinguish sharply between the categories of fiction and history, or of romance and chronicle, as the fourteenth-century Scottish poet John Barbour does at the beginning of his Bruce, claiming the merit of truth and the pleasure that truth gives to readers above and beyond that of fictional narrative:

Storys to rede ar delatibill

Suppos that thai be nocht bot fabill;

Than suld storys that suthfast wer,

And thai war said on gud maner,

Hawe doubill plesance in heryng.

. . . . . . . . . . . . . . . . . . . . . . .

Tharfor I wald fayne set my will,

Giff my wyt mycht suffice thartill,

To put in wryt a suthfast story⁸

——

[Stories are pleasurable to read even if they are only fictional; so stories that were true, if they were well told, ought to give double pleasure to listeners. . . . Therefore I’d be glad to apply myself, if my ability should be adequate to the purpose, to put a true story in writing]

Here the contrast between fabill and suthfast story, corresponding respectively to the Latin terms fabula and historia, seems clear enough; yet it is also true, as Christopher Cannon puts it, that "the Middle Ages prized historia for its invention, classing it with poetry among the arts of grammar."⁹ It is often hard for modern readers to grasp the principles governing medieval assignments of material to the two categories,¹⁰ and so I am going to use the term fictional here in a broad sense to refer to any writing that is not simply informative or didactic. My definition of fiction approximately corresponds to the definition of literature by Julia Boffey and A. S. G. Edwards, understood in the crudest possible terms as textual matter conceived or read for purposes not primarily connected with information, or with religious precept or devotion.¹¹ Autography in this sense differs, as I shall explain, from what we now call autobiography, in not being based on a claim of any systematic relation to documentable truth; it is first-person writing in which there is no implied assertion that the first person either does or does not correspond to a real-life individual. In freestanding form autography includes dream poems and prologues, along with the group of poems known in French as dits (of which, I suggest, dream poems and prologues may be regarded as species). It can also take the dependent or supplementary form of first-person narratorial commentary on heterodiegetic narratives—that is, narratives in whose events the teller of the story does not participate—but that will not be my main concern in the present book. (I use the term heterodiegetic, borrowed from Gérard Genette,¹² rather than the more familiar third-person narrative, because, as one theorist has put it, "Every narrator is a first person narrator.¹³) This kind of medieval writing, in both its forms, is nearly always in verse, so in discussing it I shall generally refer to poems and poets," using those words descriptively rather than evaluatively.

It is a striking fact that there is almost no writing of this kind in English until three centuries after the Norman Conquest: nearly all English fictional writing in those three centuries consists of heterodiegetic narratives and first-person lyrics. Few probably would deny that this is so, yet the fact has evidently not been regarded as significant; certainly it has not been a topic of discussion. One narratologist, Monika Fludernik, has noted in passing that in fourteenth-century English first-person narratives only occurred in dream poems,¹⁴ but this perceptive observation does not seem to have been taken up by literary historians—and indeed recent histories of medieval English literature have been conspicuously uninterested in the formal characteristics of their subject matter, preferring to focus on its ideological content.¹⁵ Before the Norman Conquest a kind of autography did exist in English, in the so-called elegies—poems such as The Wanderer and The Seafarer—but there is little continuity between these and later writings, and certainly no evidence that poets of the fourteenth and fifteenth centuries read them or indeed would have had the linguistic capacity to do so. Pre-Conquest autography is a separate field of study that I leave to those better qualified to discuss it; when I use the term medieval in this book, it will refer to the period between the Conquest and the Reformation. In that sense, then, there is almost no medieval English autography until the third quarter of the fourteenth century. At that point, rather suddenly, comes the emergence of a great variety of autographic writing in English: Chaucer’s dream poems, Piers Plowman, Pearl, prologues by Chaucer and his followers, along with extensive first-person commentaries on heterodiegetic narratives, again by Chaucer and his followers. The chief origins of this new kind of writing lie in France, and in many cases it seems to enter literature in English through the Anglo-French literary culture that developed at the court of King Edward III. That much seems fairly certain, though I cannot give any watertight explanation as to why autography developed in French and gained popularity in English. But that tends to be the case with important cultural shifts: explanations of them are speculative and often tendentious, and probably the best we can do, at least initially, is to describe them. That is what I intend, though in chapter 4 I shall allow myself to speculate about the nature of the appeal that autography may have had for medieval writers and readers.

Medieval autography is a kind of writing, and the word writing deserves emphasis, for, as we shall see, autography also has a strong tendency to be about writing. Much medieval literature was orally delivered—spoken, chanted, sung—and some may have been orally composed or at least composed in forms that bear strong traces of speech. One example of the latter is the occurrence in alliterative verse of alliterations that disregard word boundaries marked in writing but not in the spoken language, so that, in Sir Gawain and the Green Knight, "another can alliterate with newe and neghed."¹⁶ Nevertheless, all the medieval literature we actually have comes to us in written form and, along with many traces of the features of speech or song, has some of the distinctive features that belong to writing.

A telling example is The Book of Margery Kempe, which, according to the account it gives of its own composition, originated as a woman’s speech, but in the form in which it has survived bears marks of (male, clerical) textuality from beginning to end.¹⁷ One of the distinctive features of textuality is the way it uses the first person, if it uses it at all (The Book of Margery Kempe almost never does, except in quoted speech). The first-person pronoun works differently in writing from the way it works in speech. In this it is like other deictics, those lexemes often called shifters, which have no fixed referential meaning but are used in the spoken language to indicate persons, objects, or events in their relation to the speaker—their spatiotemporal relation and thus by extension their epistemic and emotional relations. Where and when does a person, object, or event stand in relation to me as originator of the utterance in which it is referred to? What kind of knowledge do I have of it? How do I feel about it? The answers to those and many similar questions are indicated or implied by deictics in speech. Deixis is thus one of the most important means by which subjectivity is encoded in language. It includes many pairs of terms such as I and you, here and there, this and that, now and then, to refer to what is nearer to the speaker and what is further from her, or, in technical terms, what is proximal and what is distal.

Let me quote a few sentences of what I said at this point when I delivered the lecture on which the present writing is based:

So when I say I, as I’m doing at this very moment, now, you rightly assume that the pronoun refers to the person from whose mouth the spoken word emerges, a being to whom you’re probably willing to attribute consciousness. As Emile Benveniste famously put it, Ego is he who says ‘ego.’¹⁸ And similarly, when I say "you rightly assume," you rightly assume that I’m addressing you, the people actually in my presence in this room.

I might have added that users of ASL can indicate I by simply pointing to themselves¹⁹—a striking confirmation of the way deixis works in interactive, communicative contexts—but when I write I, the word does not emerge from anyone’s mouth, and its deictic energy—the energy of pointing, looking, feeling, imagining—is freed for a wider variety of expressive purposes.

People often think of writing as restricting the freedom that belongs to speech, but writing also liberates discourse by distancing it from the communicative, I/you context in which the spoken word originates. As Paul Ricoeur puts it,

Writing is not simply a matter of the material fixation of discourse; for fixation is the condition of a much more fundamental phenomenon, that of the autonomy of the text. A threefold autonomy: with respect to the intention of the author; with respect to the cultural situation and all the sociological conditions of the production of the text; and, finally, with respect to the original addressee. . . . The peculiarity of the written work . . . is . . . to transcend its own psycho-sociological conditions of production and thereby to open itself to an unlimited series of readings, themselves situated in socio-cultural contexts which are always different. In short, the work decontextualizes itself . . . It follows that the mediation of the text cannot be treated as an extension of the dialogical situation.²⁰

The same point is made by Suzanne Fleischman: Only with the advent of the book—with writing, that is—can discourse become detached from a speaking subject and a context of origin.²¹ Some medieval poets, perhaps more conscious than we have reason to be of the break between the spoken and the written, dramatized the detachment of the text from the speaking subject by devices such as the Go, little book topos, as Chaucer does near the end of Troilus and Criseyde. The text is sent out into space and time, a world and a future in which its meaning may no longer be governed by the occasion and purpose for which it was composed, and in which its originator will no longer be able to control how it is understood by tone of voice, facial expression, gesture, and other forms of interaction with listeners, only through the words he has written:

Go, litel bok, go, litel myn tragedye,

Ther God thi makere yet, er that he dye,

So sende myght to make in som comedye!

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

And red wherso thow be, or elles songe,

That thow be understonde, God I biseche! (Troilus V 1786–88, 1797–98)²²

——

[Go, little book, go, my little tragedy, and may God yet grant your author, before he dies, the power to compose some comedy! . . . And wherever you may be read or else recited, I beseech God that you may be understood!]

As Fleischman notes, this idea of the detachment of discourse from its context of origin, made possible by writing, is seen by some as violating one of the basic tenets of the ‘communication model’ of language (i.e., that every utterance presupposes a speaker and an addressee . . .) and therefore has been a topic of intense debate among discourse analysts.²³ Most scholars concerned with the interpretation of medieval literature in English have been uninterested in the debate and perhaps unaware of it—unconscious, so it would appear, even of the possibility of conceptualizing written texts as anything but representations of spoken utterances.

Many therefore would argue, and more would assume, that the distinction I am drawing between the ways deixis operates in speech and in writing is of little importance, because writing can only be thought of as a representation of speech. In that case, a written I would always be the label of a represented speaker, a fictional being potentially in possession of all the characteristics, including consciousness or subjectivity, that we normally attribute to human persons. As I have stated, I believe this to be a mistaken assumption. Certainly writing can be designed as a representation of speech, but the supposed necessity of interpreting all writing in that way was, I believe, convincingly dismantled by Jacques Derrida in his essay Linguistics and Grammatology and then, more playfully, in his engagement with John R. Searle’s speech-act theory in Limited Inc.²⁴ Derrida’s argument in its full scope applies not just to writing in its conventional sense but to speech as well (archiécriture, as he calls it), and Searle’s apparent failure to grasp this is one of the themes of Limited Inc. I do not dispute this larger and more difficult claim, but in a book about medieval literature, which is never directly available to us as speech, it would seem unnecessary for me to discuss it further, even if I possessed the philosophical competence to do so. My concern will only be with the more limited question of the functions of the first person in writing, and what I believe to be the case is this. The first-person pronoun in written form in the narratorial discourse of a poem or novel may refer to a fictional individual, a speaker or narrator distinct from the author, an individual whose consciousness the writing purports to represent, as in a dramatic monologue such as Robert Browning’s Bishop Blougram’s Apology or a novel such as Kazuo Ishiguro’s The Remains of the Day. It does not necessarily do so, however, and in the Middle Ages it rarely does so in any clear-cut or systematic way. This practice corresponds to the theoretical understanding of personal pronouns in the work of late-classical and medieval grammarians, explained as follows by Daniel Heller-Roazen:

Pronouns, for the grammarians . . . articulate not distinct subjectivities presupposed by language but, rather, purely discursive functions . . . which are determined by the grammatical structure of language alone. According to such a conception, to say I is in no way to express oneself as an existing and substantial individual: it is neither to incarnate oneself in literature, to invoke Zink’s phrase, nor to elaborate a first-person (subject) position, to return to Kay’s formulation of her subject.²⁵

In practice, the textual first person differs from the third persons on whom a narrative or lyric confers a virtual existence. The third persons have to be represented in language and thus characterized, if only in some minimal way through the assignment of narrative or relational functions. In medieval texts the characterization is often very minimal: a third person need only be female or male, young or old, friend or enemy, ruler or subject, proz or sage, and so on. The first person, on the other hand, may be represented (as, for example, the I of a love lyric is likely to be characterized at least as male or female), but it need not be represented at all. In narrative especially it need be assigned no gender, no age, no social position, or it may be characterized to some extent in some passages (and perhaps differently characterized in different passages within a single text), while remaining elsewhere no more than an anonymous and unobtrusive channel of narration or discourse. I would argue that this difference in the world of writing should not surprise us, because it corresponds to a phenomenological difference: I experience that which I refer to as I, my self, quite differently from the way I experience the being of others. To myself, I am not a character in the way I may be to others who perceive me as a third person.²⁶ Characterization of the first person in writing is an optional literary artifice, by no means a necessary reflection of human experience. Or, to put it differently, the first-person singular pronoun need not be referential (referring consistently to an individual who uses the word I); it may only be deictic, its function being to convey proximality and experientiality without specific reference to a pragmatic center or origo.²⁷

Consequently, although the textual I, consisting of ink on parchment or paper, not of flesh, blood, and consciousness, cannot be literally identical with the author, that does not mean that it must represent a self-consistent imaginary person distinct from the author. Yet even medievalists are likely to have had their ideas about literature shaped by their reading of that dominant mode of modern prose fiction in which understanding of a story’s meaning depends on the interpretation and assessment of an unreliable fictional narrator. (And the very term narrator, unknown in the Middle Ages but indispensable to modern literary interpretation, substitutes a third-person expression for the first-person subject of narration, treating that subject as a character who just happens to be referred to as I rather than he or she.²⁸) Many of