9. Sourcing and Editing Shakespeare: The Bibliographical Fallacy

In the 1980s a debate was resuscitated around the two material instantiations of William Shakespeare’s King Lear, the text given in the Quarto printing (Q1) of 1608, as against the text contained in the posthumous Folio collection (F1) of Shakespeare’s dramatic works of 1623. Back in the eighteenth century, Samuel Johnson, as the acute intuitive critic he was, opined that the Q and F texts represented two distinct versions of this Shakespeare play. Throughout the nineteenth and far into the twentieth century, this remained an outsider view. Madeleine Doran in the early 1930s substantiated Samuel Johnson’s hunch in much detail, but her critical conclusions from precise analysis of the textual and bibliographic data were quelled by W. W. Greg. His text-critical conviction held sway that the two printed texts of Shakespeare’s work King Lear were derived by different corruption from one, and one only, Shakespearean original. Editorially, consequently, King Lear continued to be established as composite text conflating elements from the two material source texts in Q1 and F1.

But the text-critical orthodoxy on which such editions relied was eventually rocked. Bypassing preconceptions from traditions of textual criticism and consequent rules of editing, Shakespeare criticism turned directly to the givens of the material transmissions in the source documents to argue the differences between the divergent texts interpretatively. Initially, the renewed proposition caused great upheaval that behind the Q and F texts lay two distinct versions. Today, by and large, it is the default assumption with regard to King Lear. What editorial consequences might, or should, be drawn from it, is as yet still a relatively open question. To advance into that field may in fact first require a reconsideration of the might, as well as the historical rootedness, of analytical and textual bibliography in Shakespearean textual criticism. This essay wishes to gesture towards such reassessment. Arising as it does from my review of an amazingly untimely book: Sir Brian Vickers, The One King Lear of 2016,1 it is strictly an occasional essay. Vickers reargues strongly the tenet else ousted that Shakespeare’s ‘One King Lear’ descended via different routes of corruption to Q1 and F1. With a masterly command of the rules and techniques of the bibliographic game, he does so exclusively on bibliographical grounds. Conspicuously, the book refrains from developing editorial models by which to operationalise as text the text-critical proposition. This gives rise to the fundamental question whether analytical and textual bibliography, even though proven text-critical instruments, can be considered to provide sufficient grounding also for meeting the editorial challenge of the material transmissions they help to analyse.

There is no doubt that Shakespeare wrote one work King Lear—just as Milton wrote one work Paradise Lost; or Wordsworth, say, one work, The Prelude; or Henry James, Roderick Hudson, or Virginia Woolf, To the Lighthouse. From Paradise Lost onwards in this series, we have no difficulty in recognising that the works each materialised in two versions: Paradise Lost structurally distinct in at first ten (1667), then twelve books (1674); The Prelude (1805 and 1850), as well as Roderick Hudson (1875 and 1904–1907), respectively, attained deeply revised textual shapes. To the Lighthouse, even while by the author’s order published in London and in New York on the self-same day, 5 May 1927, came out, thus simultaneously, in two significantly distinct authorial versions. Preceding all these, we also have the case of Hamlet. It survives in three texts materialised in separate documents, Q1, Q2 and F, of which two—Q1 and Q2—are the extant witnesses to radically distinct versions; while F provides the material text for (at the least) a theatre arrangement of the Q2 version of the play.

It is a fundamental norm in cultural and literary history and tradition that works perceivable as one, and usually meant so to be perceived, nonetheless have their material existence in manifold texts studded with variants. These arise naturally in processes of composition, or indeed as variation constitutive of distinct versions. Seen from the compositional angle, therefore, variants and variation are concomitant with, are part and parcel of the creativity invested in writing. A writer, an author, engaged in writing lives and thinks in constant reciprocity vis-à-vis the text forming in the writing process. The variant is always a dialogic response to the text in the making, be it local and small or large-scale and constitutive of a distinct version. Nor, in truth, does the compositional, the creative urge cease with the attainment of text in states or versions committed to the processes of transmission, and agents and agencies other than the writer, the author, of origin. At any given moment, whether at pre-publication or post-publication stages, the author may respond dialogically afresh to the text as then encountered. The opportunities to do so are likely to be especially rich when the medium of transmission is the stage, its agents the performing players, and, as in the case of Shakespeare, its author of origin himself one of the company.

Engagement with the text resulting (potentially at least) in changes is also perceived as a duty of printers’ or publishers’ editors, book-keepers in Elizabethan playhouses—or whatever analogous functionaries may be recognised as intercalated between text materialisations of works and the reception of works through such texts. When the matter is seen from this angle, a noxious sub-class of variants and variation arises, it is true, from the transmission of texts and the investment of understanding and concentration on the technical accomplishment always involved in the labour of transmission. Scribes or typesetters, typists, as well as editors of every description (the scholarly editor not excluded), produce errors, be it inadvertently; or, since they, too, engage as readers with the texts transmitted, through varying them advertently and thereby creating their own sub-sets of variants through miscorrection, correction, emendation or conjecture.

Textual criticism focused on transmissions operates in the field of material documents, the material inscriptions on them and, by inference, the scribes and copyists who performed the inscriptions. Positing this framework has enabled structured reasoning for understanding the transmission and evaluating it in terms of the text(s) it represented. Traditionally, the argument led back from documents and text inscriptions materially extant to antecedent text instantiations which, though lost with their carrier documents, appeared, or were, reconstructable through application of logic to the comparative analysis of the instantiations in extant documents; and in the process, too, allowing for, defining, and eliminating disturbances in the transmission. Textual criticism by such rearward-directed methodology claimed to be capable of arriving at, and editorially realising, one stable source text, inferred as the text carried by one lost document. This text, by definition a materially non-existent construct, was posited as the node of origin, dubbed the archetype, of the multiplicity of the extant variant texts.

Within the game, the discipline, of textual criticism as initially so conceived and played out, authors as authors were not considered players. This was logical: for textual criticism so construed—which was textual criticism as applied dominantly to the transmissions from antiquity and the middle ages in manuscripts—text and texts existed, and could exist, only in material instantiations of inscription and their respective embeddings in documents of transmission that were (consequently) by definition not authorial documents. The mindset of the textual critic of written heritage from antiquity through to the middle ages was deeply ingrained, too, in the founders of Shakespearean, bibliography-based textual criticism in Britain in the early twentieth century. W. W. Greg, foremost and in the long run the most influential among them, was by original training a medievalist. The merger of bibliographical analysis with the inference in retrospection of the archetype text out of which he forged his methodology carried great, at times indeed overpowering, intellectual conviction. Its next-to-irresistible appeal arose from the logic of the text-critical argument leading to the inference—inference, admittedly, often close to, or even amounting to, positive proof—that it was now the author’s autograph writing that held the position of the archetype. In the case of Shakespeare, such autograph writing was posited predominantly to have been his drafts, or ‘foul papers’ (by Elizabethan nomenclature).

The evidence accumulated by Shakespearean scholarship is impressive that, taken as a whole, the surviving Shakespeare texts in print are at only a short remove from the author’s writing as it originally took shape under his hand. Yet conclusions as to the author’s text from empirical analysis of its extant variant derivatives in print are one thing; equating them with the concept of the archetype as a logical abstraction and, by such definition, an invariant stable text, is another. Whether it be legitimate to conflate the empirical with the abstraction would seem questionable. In Shakespearean textual criticism, however, the danger inherent in the equation has not been seen or heeded. Consequently, Shakespearean textual criticism as systematised by its founding fathers in the first half of the twentieth century left little or no room for conceptualising or text-critically and editorially handling variation, let alone versions. On the contrary: textual bibliography was, in the service of Shakespearean textual criticism, raised to the status of a procedural underpinning akin to science for the retrospective analysis capable of opening up vistas onto the existence and shape of texts in documents no longer extant. With sharpened powers of adjudication, it identified variants—yet it did so still under the default assumption that instances of variation indicated moments of textual error.

Thus inherited orthodoxies were shaped to fit patterns developed for text transmissions from pre-Gutenberg eras. Leaving, as said, little or no room for conceiving of variants other than errors, and even less for positing versions, they also saw no urgency at first to involve the authors of texts when assessing the nature of the material reality, evidenced or inferred, of transmissions. But the bibliography-based Shakespearean textual criticism that evolved from this seed-bed redirected, on the one hand, the strict text orientation (retrospective to the archetype) into a fervent author orientation (still retrospective, but to Shakespeare’s ‘foul papers’). On the other hand, it allowed little or no forward movement in terms of the growth and development of texts. It remained reticent towards authorial variants, and intensely sceptical towards the notion of versions originating from and shaped (in whole or in part) by authors, such as William Shakespeare himself.

Nonetheless, the notion that King Lear should be considered in terms of authorial versions began to attract insistent attention in the 1980s. That the impulse came from the realms of Shakespeare criticism is no doubt indicative of the division between (textual) scholarship and (literary) criticism that by the late twentieth century had become ingrained in the Anglo-American humanities. It was not the textual scholars who began to critique their own methodology and a priori assumptions; it was instead their ‘opposite numbers’, the Shakespeare critics who raised their voices by strength of their native concerns with the structures, aesthetics and meaning of literary works. Their voices were heard—though only under great storms of controversy in Shakespearean textual criticism, and but as a faint murmur in Anglo-American editorial scholarship at large that, for the establishment of edited texts, had proceeded to turn the retrospective author orientation into a prospective orientation towards the fulfilment of authorial intention. Elsewhere than in these precincts, which means in schools of textual criticism and editing outside the Anglo-American tradition, it was meanwhile the a priori assumption that text variation and versions were to be reckoned with as norm and reality and that they had, moreover, not uncommonly left material traces in text transmissions pertaining to given works. This amounted to a fundamental shift in the mindset informing textual criticism towards establishing edited texts: the normative assumption for texts in their material instantiation was no longer that they were error-ridden and that their readings had to prove legitimacy. On the contrary, texts were prima facie not faulty, and declaring individual readings to be textual faults or errors was admissible only after the strictest scrutiny; the default assumption for variants and variation was that they signalled revision and, in given cases, alternative versions. The focus on error built into the handling of the tools of analytic and textual bibliography were not amenable to such an understanding of texts and transmissions.

* * *

Against the understanding prevailing today that King Lear has come down to us in two texts reflecting two distinct versions of the work, The One King Lear positions itself still uncompromisingly within the twentieth-century mode of Shakespeare textual criticism based on textual bibliography. Positing a priori one singular and unique manuscript text penned by the author is tantamount to defining a conjectural archetype text for the play. This assumption rules out the possibility that variants evident between the Quarto and Folio texts, whether small or large-scale, originated with Shakespeare, the author. By prejudice of method, instead the text differences between the two extant printings are made to appear as nothing but transmissional corruptions. For these, the printer and type-setters (for Q), or the playhouse book-keepers and the King’s Men themselves (for F), are given the blame.

On such a premise, can case-by-case assessments of variants be grouped as either possible, probable, or demonstrable? The classification ‘demonstrable’, alas, is ruled out absolutely. The copy from which Q was printed, while in great likelihood a manuscript in Shakespeare’s hand, is lost and hence unavailable for comparison. Nor is direct demonstration feasible for textual specifics of F. The F text stands at more than one remove from whatever author-inscribed source spawned it. Intermediary document identities and text states may be assumed, but are indefinite. The Folio text can therefore not, by comparison with the Quarto text or otherwise, demonstrate the text state of any posited original manuscript.

Might the textual correlation between the Q and the F texts, if bibliographically not demonstrable, at least be argued as possible? A bibliographical investigation founded on strict logic, such as that exercised by Brian Vickers, makes its results seem on the whole possible. That they are, however, only seemingly possible is due to the fundamental flaw of textual criticism as analytical bibliography striving towards the fulfilment of its Gregian definition as ‘science’. This reduces ‘textual criticism’ to an analytical exercise and shears it of its critical dimension, specifically its potential as, dependence on, and integration into, literary criticism. The impasse becomes dramatically apparent in assessing specific correlations of the formal findings from bibliographical investigation to text and text meaning. This is where text-critical reasoning from bibliographical analysis must prove itself. Its probability must depend thoroughly on critical assessment and judgement.

* * *

For my own argument, I wish to cast into exemplary focus one detail from an essay on the two versions of King Lear that I contributed (collaboratively with my late friend and colleague Klaus Bartenschlager) three decades ago to the Deutsche Shakespeare-Gesellschaft West: Jahrbuch 1988.2 As subtitled, the essay explicitly assesses ‘[das neue] Verhältnis von Textkritik und Literaturkritik’ (‘the new relationship of textual criticism and literary criticism’). It gives weight straightaway—as has, true enough, been done repeatedly elsewhere—to Lear’s declaring, at the opening of the state scene, scene 1(Q)/Act I.1(F), his intention to divide in three his kingdom. I give the main part of the speech here in a fusion of the two document texts:

Meane time we will/shal expresse our darker purposes,/purpose.

Giue me the Map there. Know, that we haue diuided

In three our Kingdome: and ‘tis our first/fast intent,

To shake all Cares and Businesse of/from our Age,

Confirming/Conferring them on yonger yeares/strengths, while we

Vnburthen’d crawle toward death. Our son of Cornwal,

And you our no lesse louing Sonne of Albany,

We haue this houre a constant will to publish

Our daughters seuerall Dowers, that future strife

May be preuented now.

Rendered in bold type are pronouncements absent from the Quarto text. In close proximity, and in their vein, too, follow two further lines in the Folio text:

[…]

Since now we will diuest vs both of Rule,

interest of Territory, Cares of State

On the a priori assumption that there was only ever one text of Shakespeare’s original writing from which the Q and F texts differently descend, these text elements (in bold type) must be posited as present in that authorial manuscript and declared as abridgements, cuts (by whomever) or omissions by oversight in Q. This is a rationalisation entirely confined to, and in terms of, the bibliographical analysis, deployed not to test, but to suggest that the observed bibliographical data self-evidently prove, the premise. Genuine testing would need to resort to an evaluation at the level of criticism. A critical approach would, for example, take into consideration distinctions of meaning between the texts materially in evidence and, by interpreting the differences between them, venture beyond the materially and bibliographically manifest, so as to verify or falsify conclusions from bibliography at a complementary critical level. Critical approaches in our day would to this end be able to draw on text-critical experience gained from wide-ranging analyses of compositional and transmissional processes since the onset of the modern period in Western culture, which are more richly documented than in the particular instance of William Shakespeare. This would allow measuring the text divergences for King Lear as they have come down to us against what is empirically known elsewhere of authors writing in creative dialogue with their texts, as well as of the transmission of texts of acknowledged origin in authorial papers. Texts in the process and progress of composition and transmission are ever changeable. Whether at an author’s desk, in cooperation with printers and publishers or at moments of review after publication, while texts may always strive towards stability, they are at the same time also always open to modification and revision. In this manner, they implement throughout the potentiality of the stuff they are made of: language. Grounded in, and resorting to language to express itself, a text can always also be otherwise.

In the case of drama, in particular, a critical approach will be wise not to leave out of its account the fact that the texts that have come down to us are but the substratum of the play thereby represented, and are by nature akin to musical scores in ways not matched by literary texts of other genres. Play texts always remain textually open in particular to performance-oriented modification, exercised cooperatively by the author (in recent times one might think of Samuel Beckett, say, or Tom Stoppard) and ‘the theatre’: in Elizabethan and Jacobean terms, book-keepers and the fellowship of actors in the playhouse. To say so is of course to invoke the social dimension of textual criticism, which has gained ground in the discipline over the past decades. As a matter of fact, though, textual criticism has traditionally always acknowledged the ‘social dimension’ of transmission—only that it has over the centuries been seen exclusively as the source of error and text corruption. Before the advent of print, medieval scribes were—while in effect agents of social collaboration in transmissions—seen as perpetrators of error. Under the aegis of early book printing, scribes were in such detestable function joined by typesetters and both printing-house editors and correctors. Textual criticism by way of bibliographical print analysis did not revise its attitude to their social agency as corruptive. This remained a highly effective bar to appreciating variation and versions as being in the very nature of texts. By contrast, textual criticism under the umbrella of literary criticism, recognising that texts are ever changeable, will begin from assessing neutrally that texts, materially extant to represent works, by default manifest themselves as variant from each other; and will proceed from there to differentiate the nature and quality of non-identity in instantiations of text. To accept changeability as a basic given of texts defines variation and versions as fundamentally distinct from error. The age-old procedures of recognising and eliminating error from texts in transmission do not serve and cannot be adapted to realities of text variation. Texts are formed in language. Language is, and depends in essence on being, open and processual (or communication and understanding would be impossible). Consequently, variability that is of the nature of language is a natural condition, too, of texts. Their variability, moreover, is always also prone to leaving material traces in source and transmission documents whose texts become subject to the scrutiny of textual criticism. To do both text-critical and literary-critical justice to variation, such scrutiny demands an autonomous methodology, separate from carrying forward the traditional business of textual criticism and editing by methods and procedures of identifying and eliminating error.

To refocus on my example from the first scene of King Lear: a literary-critical argument joined to bibliographical analysis will begin its approach from a text-genetic vantage point. This provides us with the knowledge of patterns recurrently identifiable in authorial writing, widely observed to indicate that, with composition already far advanced, there will be a reworking of beginnings so as to prepare for, and bring into (as it were) prospective focus, later moments in the text’s development and resolution. Measuring the variation in Lear’s opening speech against such general empirical knowledge brings within ranges of the possible that the play’s first scene was so revisited: that, on rereading the play’s beginning, the protagonist’s sense of age became intensified, his will to abdicate and divide up the kingdom between his daughters’ husbands made more explicit, and all this together projected into, and heightened by, the image of divestiture—textually as well as bodily and theatrically performed as this had meanwhile been through unbuttonings and the taking off of clothes in the play realised since the first drafting of its opening scene.

But then again: the point here is not to put forward as inescapable a conclusion that between what was behind Q and what behind F the play’s text was revised—much as I personally am convinced that it was. The point is, rather, to emphasise that the one-text-only as well as the revision rationalisation of the text divergences as documented in Q and F for King Lear cannot but each be hypothetical. These material survivors from an originally (by all inference) wider and more diverse substance of transmission demonstrate neither hypothesis incontrovertibly. What is more: neither do they permit incontrovertibly defining and identifying the respective agents of text variation and versioning. Against the fact that Q and F furnish us with two textual representations of a play, and that materially, therefore, the Q and F texts as such simply are two versions: which they are, regardless of whether it be in any way conceded or not that they represent two authorial versions; against, furthermore, either the possibility or the probability (respectively) that either Q and F constitute two divergent corruptions of one, and one only, archetype text; or else that in the transmission’s lost past there existed already, and maybe at different points in time, two versions of the play to which Q and F derivatively bear separate witness—against any of these hypothetical assumptions, and so against all of them, the question or questions must take second place with whom, i.e., with which agent or agents, any given variant, any group of variants, or the variation in its entirety, originated. Given only Q and F as the material instantiations of text for King Lear that have come down to us, the differentiation of these versions remains hypothetical in the first degree. Positing agents for the text variation into versions is hypothetical at one further remove. What is more: here the forming of hypotheses encompasses, too, the author. While on all accounts originator of the play, that is of the work, it depends on the assessment of the variation between the two material instantiations of text for the work whether or not the author be proposed as one among undoubtedly several agents to have introduced the variation—and if so, what bibliographically, critically, or otherwise he should be declared responsible, or indeed should be given credit, for.

In consideration of all the complexities offered by the transmissions of Shakespeare’s plays in general, and King Lear in particular, scholarship that integrates text- and transmission-focused with literary-critical perspectives may profitably return to textual criticism in the discipline’s modes prevailing before the onset of its fully-fledged author-centricity of today.3 By its older traditions, textual criticism, and scholarly editing resulting from it, focused on texts in the given—meaning essentially the surviving—materiality of their transmission. If, for the old school of Shakespearean textual criticism and editing, the author’s foul papers as the one-and-only urtext usurped the systemic position of the archetype, then fresh systematics and methods of textual criticism and consequent editing would appear today to be called for in this specialised pocket of editorial scholarship, and a wider understanding of textual criticism than prevailed throughout the twentieth century.

* * *

Therefore, as said: William Shakespeare wrote one work King Lear. It has come down to us in two material texts. By objective definition, they are different text versions, one of them (Q), just on its own terms, apparently more deficient than the other. Opinions may still differ whether behind each stands a different version of the play, or whether both material texts derive by different routes from one-only original text. The Folio instantiation, more discernibly than the Quarto instantiation, betrays that distinct elements of theatrical origin were interlaced into text of authorial composition. One may assume that the opportunities to respond in creative dialogue to a text as at any given moment it existed are likely to be especially rich when the medium of transmission is the stage, its agents the performing players, and, as was the case for Shakespeare or Beckett, and is still for Tom Stoppard, its author of origin himself one of the company. This means no less than that in the surviving testimony to the work King Lear of William Shakespeare’s writing we already find distinct foot- and fingerprints bearing witness to its public life as drama. What the company of the King’s Men established over the years, until the documentation thereof reached stability in the 1623 printing, was their text of William Shakespeare’s King Lear. The Q text and the F texts, which together we possess, are records in print of a play as play in versions relating to texts for performance. The textual heritage of King Lear at its manifest sources—meaning: in its earliest materialisations before, downstream in transmission, they surface for posterity—on all accounts therefore offers itself already enriched with versioning potential for the theatre.

This, to conclude, begs trenchant questions of editing. The editorial enterprise will ever anew be obliged to sift the total materiality of a play’s transmission. Recurrent endeavour will need to be invested to differentiate the span of its variation, assumed to range from authorial input via collaborative modification to corruption in transmission. As a condition of its genre, a dramatic text as play will also ever prove dynamic and ever anew be caught playing in the theatre. Every realisation on stage in fact constitutes a performance edition of a score text underlying it. Against such score texts, text editions of works for the stage will ever, too, need to be adjusted, so as in scholarly editions to bring the dynamics of drama as caught in the dynamics of dramatic texts into the realms of experience of readers and users.

---