Theory of Texts

Jerome McGann

  • Bibliography and the Sociology of Texts: The Panizzi Lectures 1985 by D.F. McKenzie
    British Library, 80 pp, £10.00, December 1986, ISBN 0 7123 0085 6

A quiet yet profound change has been taking place in literary studies during the past ten years or so. Initially it was obscured by the successes and celebrities of Deconstruction, where idealist hermeneutics – this century’s dominant textual and interpretative program – led its own tradition into a theoretical impasse. The more recently touted ‘return to history’ has also obscured the nature of this change – obscured it because the so-called new historicism (I give the name un-capitalised) comprises a variety of historicisms, some new and some not, some idealist and some otherwise.

The return to history is one signal, or perhaps one aspect, of something more fundamental, something which D.F. McKenzie’s three lectures on bibliography and the sociology of texts call more clearly to our attention. In a series of initial, critical remarks on theory of bibliography, McKenzie suggests that ‘historical bibliography’ should now probably be placed at the centre of the study of texts. This idea, which lies at the core of his lectures, seems innocent enough. In fact, its implications are far-reaching.

The most immediate object of McKenzie’s critique is that line of positivism which has underwritten most work in bibliography and textual criticism (properly so-called) during this century. McKenzie has Fredson Bowers particularly in mind here, whose position – that ‘historical bibliography is not, properly speaking, bibliography at all’ – McKenzie stands opposed to. In the course of McKenzie’s critique, however, he glances at the larger issues involved in this apparently lil-liputian struggle of the professors: ‘I am not bold enough to speak of paradigm shifts, but I think I am safe in saying that the vital interests of most of those known to me as bibliographers are no longer fully served by description, or even by editing, but by the historical study of the making and the use of books and other documents.’ McKenzie spends a fair amount of time in his lectures on the problems which have arisen in our age as a result of the media explosion. Film, television, magnetic tape and computers are only the most prominent forms of the many new types of information storage and transferral that have been invented or developed in this century. One part of his argument is that the traditional tools and frames of reference developed in bibliography will play an important role in the definition as well as the management of the archives created through these media.

More philosophical considerations arise in such a context, however. Keenly aware of the interpretative power which is embodied in these media themselves (‘the medium is the message’), McKenzie is brought to reflect upon how blind traditional bibliography has been to the analogous character of the book. In a series of trenchant illustrations, he unfolds a profound truth about ‘the book’ itself – and thence about every kind of possible text: that it is meaning-constitutive not simply in its ‘contained’ or delivered message, but in every dimension of its material existence.

This point of view entails the collapse of the pragmatistic rapprochement which has traditionally governed the theory of texts. According to that theory, the study of texts is to be structured along those two separate (but not quite equal) lines we call ‘interpretation’, on the one hand (the upper house, as it were), and ‘bibliography’, on the other (the lower house). McKenzie correctly sees this split to have been epitomised in the mid-20th century by the division of the literary kingdom between the New Criticism and the New Bibliography, those first-generation Anglo-American versions of hermeneutical idealism and textual positivism. Each appears here in a representative figure: the bibliographer Fredson Bowers and the interpreter William K. Wimsatt. McKenzie singles them out for tactical purposes – that is to say, in order to place each of them at the centre of his double-focused critique of the traditional theory of texts.

‘There is nothing outside of the text.’ That well-known Derridean aphorism has licensed every type of text-centred hermeneutics from the New Criticism to Deconstruction. The statement might have been made by McKenzie as well, for he too observes a human world which is thoroughly textualised. But because Derrida’s thought comes out of the modern hermeneutical tradition, and particularly out of Heidegger, his use of these words does not correspond at all to what McKenzie would have meant. For McKenzie, all texts are objective and materialised.

So if we were to imagine McKenzie saying today that ‘there is nothing ouside of the text,’ the statement – coming, as it would be, in a Derridean context – would not merely involve a different ‘meaning’ as regards the status of texts: it would carry as well a critique of Derrida.

What is it that generates differences of meaning when two statements are linguistically identical? Borges played with that question most famously in his witty story ‘Pierre Menard, Author of the Quixote.’ Common answers would be ‘context’, or ‘the use to which the words are being put’. Derrida says that these differences of meaning are built into the text’s inherent structure of deferrals.

The importance of McKenzie’s work is that it is operating out of a theory of texts which has supervened – which has synthesised – the traditional and the Derridean answers. This new theory of the text involves the deployment of a Derridean process of reading in a field which has exploded beyond the bounds of Derrida’s founding metaphoric frame: that is to say, beyond the bounds of the metaphor of language as that metaphor was delivered over to, and received by, later theorists from its Saussurean fons.

Of course McKenzie still uses linguistic metaphors, but his sense of the physique, the materiality, the socio-historical action of language and texts completely transforms the significance of his metaphors. Nowhere is this transformed view clearer than in his second lecture when he discusses ‘the sense in which the land – not even a representation of it on a map, but the land itself – might be a text’. McKenzie draws our attention to the traditional lands of the Australian aboriginal tribe, the Arunta: ‘every prominent feature of the landscape in the Arunta country is associated in tradition with some totemic group. It is not simply a matter of their being sacred objects ... but of their having a textual function. These visual, physical features form the ingredients of what is in fact a verbal text, for each one is embedded in story.’ I think myself that McKenzie’s language here is still too committed to traditional views about the character of language and texts. To speak of ‘the ingredients’ of this Arunta text, or of its ‘story’, is to slight the text’s agenting integrity, what makes it inseparable from the people who belong to the text that is the land (and one would not want to say here ‘to whom the text that is the land belongs’). In this case, where land is text, the text is also magic (the Western equivalent of the latter being what we call ‘technology’).

For students of literature, the most practical and immediate interpretative consequence of such a theory of texts is a procedural imperative: that one has to take the entirety of the language event as the object of interpretation. Indeed, one has to proceed from the position that language is more properly conceived as an event than a medium: not a container or even an avenue of meaning, but an extended field of communicative action.

McKenzie illustrates this by a series of readings in which he takes the entirety of ‘the book’ as the field in which his interpretation must operate. His discussion of the Prologue to Congreve’s The Way of the World is a paradigm of what gets involved when the field of the text is opened in this way. The interpretation, in fact, is hardly located in the Congreve ‘text’ at all. It starts, for example, from another ‘text’ altogether: Wimsatt and Beardsley’s influential 1946 paper from the Sewanee Review, ‘The Intentional Fallacy’, where the lines from Congreve are quoted (or misquoted) as an epigraph.

As the commentary develops, it becomes clear that to pursue the meaning of a text entails the pursuit of the text’s entire socio-historical field, that the range of such a field will stretch across large reaches of space and time, and that the field cannot be properly approached – cannot even be seen – if one’s vision is hemmed within the linguistic confines that have so dominated 20th-century hermeneutics.

As a bibliographer, McKenzie is particularly alive to the ways in which meaning gets constituted and reconstituted in the books and physical media through which (or rather ‘as which’) texts are transmitted. He is also a good reader of texts within a narrower focus: that is, he moves easily in stylistic and intertextual modes of criticism. But these are always subordinated to the more comprehensive frame of reference: so words are seen through the books where they are printed, and intertexts are deployed as one dimension of the larger socio-history of the literary work.

To see and work with texts in this way is to have left behind for good the hermeneutical categories which have governed literary studies, in the academy at any rate, for nearly a hundred years. The development of modern philology from late 18th-century Biblical and Classical studies initially moved within, and toward, the idea(l) of a unified investigative field, a marriage of what used to be called the lower and the higher criticism. That idea(l) eventually collapsed with the development of textual and interpretative specialisation in the later 19th century.

The most consequential division of this literary kingdom came when textual criticism, bibliography and editing were separated-off from ‘literary criticism’ and interpretation. The split was codified in René Wellek and Austin Warren’s normative handbook. Theory of Literature, published in 1942, where textual studies and editing were defined as ‘Preliminary Operations’ – investigative lines which seemed too materialistic and positive to be included in what was thought of as ‘literary’ criticism. And even those (properly so-called) ‘literary’ studies were splintered into the now well-known categories of ‘extrinsic’ and ‘intrinsic’ hermeneutics. Thence developed the many ‘fallacies’ of the 20th century’s literary fancy: the affective fallacy, the fallacy of imitative form, the intentional fallacy, the biographical fallacy, and so forth.

As literary studies now begin to understand better the socio-historical frame of reference within which those ideas developed, students are moving beyond the limitations which they once imposed. These developments may be observed in a variety of areas. Most apparent, I suppose, is the line which Shakespearean textual criticism has been vigorously pursuing for about ten years, and which has resulted in a new approach to the problem of Shakespeare’s texts.[*] What is most important to see here is not simply the bibliographical and editorial significance of such work, but its interpretative salience as well: and beyond the interpretative issue lies the profound conceptual change, that editing and interpreting, that bibliography and hermeneutics, are only heuristically separable.

Let us make that heuristic separation for a moment and observe some of the hermeneutical consequences entailed in this new theory of the text. For one thing, we have to reconstitute the very notion of ‘the text’ as an object, or field, of reading; and in carrying out that procedure we will perforce be driven to rethink the structure of the reading process itself. ‘The text’ will not be located as the words on the page immediately before one’s eyes. ‘The text’ is far less transparent than that, as one may see very quickly by performing a few simple tests. The ‘text’ of Don Juan Cantos One and Two in the first (quarto) edition of 1819 is virtually identical to the ‘text’ of the early (duodecimo) piracies of that work. But the differences between these two texts are profound, as one realises from a glance at certain early (conservative) reviewers; who are able to praise the wit and verve of the poem as it appers in the quarto text, but who feel compelled to denounce the duodecimo texts. The difference lies in the audience they perceive as present to and in the two texts, and in the general socio-historical context which those perceptions call attention to. The reviewer reads the piratical Don Juan (but not the quarto text) as immoral, blasphemous and seditious because it is a cheap work that is destined to fall into the hands of working-class readers, and to hold in ridicule the ideology of the dominant culture. The quarto text is, on the other hand, an expensive and handsome book whose audience is what Byron sometimes called ‘the twice two thousand’. In fact, that first edition was printed in 1500 copies, whereas the piracies were issued in their tens of thousands.

This dramatic example, which I have given elsewhere, might be multiplied with ease – in the sense that the works we read are all produced in forms that execute their meanings through systems that are far more complex than those we engage with when we simply study the linguistic text. Indeed, after Bakhtin we have become acutely aware of how multiply-voiced the linguistic text must be, and how ridden with the ideological conflicts that attend such multiple voicing.

If there is a ‘text’, then, we shall have to distinguish the different faces, as it were, which it always turns to us. I do not say the different ‘levels’ of the text because that word is freighted with the concept of the text as a container or transmitter of meaning rather than as a meaningful agent in itself. In this sense one might usefully distinguish ‘the text’ (or the poem as a purely linguistic event) from the ‘version’ (or the immediate and integral physical object ‘through which’ the ‘text’ is being executed), and make yet a further distinction of ‘text’ and ‘version’ from the ‘work’ (the term to stand for some more global constitution of the poem). There is a ‘work’ called Paradise Lost which supervenes its many texts and its many versions; to William Blake that work was one thing, whereas to William Empson it was something else; and of course to any one of us the work we call Paradise Lost can be, will be, reconstituted once again.

These texts, versions and works entail, it seems to me, corresponding structures of reading. The most common form of reading is the reading of the linguistic text – that is to say, a linear form of reading in which one moves from word to phrase to sentence and thence on through the many larger rhetorical and generic forms by which the linguistic text is constituted. Linear modes of reading have dominated hermeneutics from Heidegger and the New Criticism through de Man and Derrida. (Derrida, however, is the one linear reader in this tradition who has deployed, in order to carry out his readings, a number of non-linear texts.)

Though most people imagine that linear reading is the only one we practice, at least two other reading modes can and should be distinguished. To read Blake’s illuminated poems, for example, is to be forced to follow a spatial mode of reading as well as a linear mode. The eye scans the page, or is driven to move around it in odd angular ways – stopping at various points (perhaps to deal with interlinear visual materials), veering off at other points (into the margins, the headers or the footers, for example), backing up, jumping ahead, or just pulling back entirely to register what Vince De Luca has aptly called Blake’s ‘wall of words’. Nor is spatial reading in Blake simply a function of those plates with large or dramatic illustrations. Those extraordinary plates in Milton and Jerusalem which are filled from top to bottom with packed lines of seriatim verse are every bit as spatially-organised as the plates which carry large illustrations.

The play of linearity and spatiality is by no means peculiar to Blake. A poem like Herbert’s ‘The Altar’ may remind us that every verse form is a spatialised form once it is committed to a written mode. ‘Free verse’ re-emphasises this fact by using the space of the page as a framework for organising the linguistic text. We are all familiar with Apollinaire’s Calligrammes, Pound’s Cantos, the spatial innovations of Marianne Moore, William Carlos Williams, John Cage.

But spatiality functions in other less obvious, if no less important ways. The formating of a printed text (book or otherwise), the conventions of book production which dictate what sorts of framing materials will be included (for instance, the material and information supplied on the title page or in other parts of the front matter): both of these are aspects of a text which play significant roles in the constitution of the text’s systems of meaning. The title pages of the early editions of Lyrical Ballads, the fly-titles in the first and later editions, and the crucial matter of the arrangement of the poems: all these will be ‘read’ by the reader because all are meaning-constitutive. Not every reader will read these forms of the text self-consciously, however.

Or take a more particular example. Shelley’s ‘Hymn to Intellectual Beauty’ was first published by Leigh Hunt in the Examiner in January 1817. The poem is placed in the immediate context of a long article dealing with political power and its abuse by the government. As a consequence, the (initial) Examiner text of Shelley’s poem stands in a graphic context where the word ‘power’, which might otherwise have been confined to a Neo-Platonic range of signficance, is pushed to include as well more quotidian meanings.[†] (The event also shows the often powerful authority which may be exerted on a text by persons other than the author.)

Examples come easily to hand. Byron’s attack on the Prince Regent, ‘Lines to a Lady Weeping’, was first published in the Morning Chronicle in March 1812, anonymously. The poem caused a small uproar at the time, but when he re-published it in February 1814 under his own name (among a group of short poems he printed with The Corsair) the effect was explosive. The ‘meaning’ of the poem had changed with the change of format and its attendant change of circumstances. Today one registers those originary changes of meaning in the same way that the different meanings were executed at the time: through a reading of the poems that is both linearly and spatially-organised.

The Byron example calls attention to yet another form of reading which operates along with the linear and the spatial modes. The 1814 version of ‘Lines to a Lady Weeping’ has the effect that it has only when it is read in the context of the (absent) 1812 version. That is to say, the explosive 1814 version is also being read as it were radially, with the reader’s eye moving between two texts which are not in fact present to each other.

This sort of thing happens all the time. A famous modern instance is Marianne Moore’s ‘Poetry’, which Moore sent out in three successive and different versions. The last, in three excellent and witty lines, represents a severe process of reduction from the previous two – much longer – versions. The force of the three-line version lies exactly in one’s ability to read it linearly while one’s mind is adverting to the two absent texts.

This example illustrates only one elementary type of radial reading, however. The Shelley example cited above was produced by an act of radial reading – that is to say, by a reconstitution of an originary textual situation from a much later point of view. What we call scholarship is the territory where radial types of reading are constantly put into practise. Critical editions – for instance, any of the editions (good or bad) in the Oxford English Texts series – are typically structured so as to enforce spatial and radial reading processes along with the linear process. One does not simply move through such an edition, starting at the beginning and then proceeding on page by sequential page. Rather one moves around the edition, jumping from the reading text to the apparatus, perhaps from one of these to the notes or to an appendix, perhaps then back to some part of the front matter which may be relevant, and so forth. The edition also typically drives one to other books and acts of reading, ancillary or related materials which have to be drawn into the reading process in order to expand and enrich the textual and the reading field.

This is a process by which the entire socio-history of the work – from its originary moments of production through all its subsequent reproductive adventures-is postulated as the ultimate goal of critical self-consciousness. That the goal is in fact an unreachable one is clear. A practical move toward its attainment is essential to the critical enterprise, however. Such a move appears as some particular version of a work – say, Hyder Rollins’s Variorum edition of Shakespeare’s Sonnets, or Stephen Booth’s more recent critical edition. Each is a particular attempt to define a comprehensive socio-historical field for the sonnets. Their respective virtues are as much a function of their limits as of their particular strengths, both of which they execute as corresponding sets of presences and absences.

In this respect, the critical edition is a kind of analogue computer designed to reconstitute past texts and versions in forms which makes them usable in the present and for the future. The special hermeneutic advantage of such an edition over, say, a more simple ‘reading edition’ lies in its theoretical comprehensiveness: the complexity of the critical edition allows one to imagine many possible states of the text, including various types of ‘reading’ and ‘student’ and ‘modernised’ editions. Non-critical editions are much more limited in this respect.

Critical editions take many forms, of course, and some are better – are more accurate or more complex – than others. Nor can the computer that is the critical edition be programmed to regenerate anything but analogues of the texts it is interested in. This is one of its key structural limitations; the other is the selection processes which will be built into its programs. The excellent recent edition of Ulysses edited by Hans Gabler is a good case in point, and a case brought to our attention by McKenzie: ‘Given the evidence which it chooses to present, what the new edition could not do was to represent the physical form of Ulysses as it was first published.’ This limitation turns out an important one, because Joyce appears to have used the page sequences and lay-out of the 1922 edition as part of the work’s semiotic system. McKenzie cites the work of Dr John Kidd in order to show ‘Joyce working to make textual meaning from book forms, re-writing in proof in a creative interplay with the fall of the text on the page, and nudging it into patterns of page-to-text, which offer markers, boundaries and divisions directly related to its final “book” form.’ Once again we are reminded that there is no such thing as a definitive edition, only editions which serve different purposes, different audiences. Thus the limitations which are defined in the texts we reproduce come to reflect, and reflect upon, the systems of presences and absences which play themselves out in the texts we receive for reproduction, including the originary texts.

The example of the critical edition is useful for clarifying the theory of texts because it is, up to this point in time, the most socio-historically self-conscious of all possible texts. Because it also emphasises, in itself, the constructed and agented character of a text, it has the additional advantage of opening itself to critical reading, and thence of breaking down that spell of self-transparency which hovers over all the texts we read.

What McKenzie calls ‘the sociology of texts’ was first clearly, and critically, grasped in Germany at the end of the 18th-century, by those now almost legendary Early Modern philologists such as Wolf, Eichhorn and Gabler. At that point sharp divisions had not been drawn between textual criticism, anthropology, history and ethnology, so that a comprehensive socio-historical hermeneutic was able to be set in motion. The tendency toward specialisation which would later splinter that program has been giving way, in more recent years, to various kinds of interdisciplinary studies in the field of the human and the social sciences. These are the developments which made a book like McKenzie’s possible – indeed, which made it inevitable.

[*] See Michael Warren and Gary Taylor’s collection of essays by various hands, The Divisions of the Kingdom (Oxford, 1983), for a convenient gathering of relevant essays on the seminal problem of the ‘two texts’ of King Lear.

[†] I have this example from a conversation with Professor Terence Hoagwood.