University of Virginia Library

A short while ago in these pages, G. Thomas Tanselle quite properly characterised the editors of mediaeval, biblical, and classical texts as having paid very much more attention to charting the transmission of these texts than to investigating the likely auctorial patterns of usage—particularly with regard to accidentals.[1] In this predilection, the editors of such "ancient" texts had behaved rather differently from their colleagues working in the post-Gutenberg periods, for with the frequent availability of later texts in potentially authoritative printed and manuscript forms, "modern" editors had quite appropriately argued over the relative value of these different stages of the text as representing auctorial intentions (both original

and final). From these concerns came, for example, Greg's famous distinction between substantives and accidentals. To the mediaevalist, biblicist, and classicist, however, such passions seemed either remote from or irrelevant to the conditions under which they laboured. "Ancient" texts were typically extant only in scribal copies often several hundred years removed from any auctorial intention, and the texts transmitted in these non-authoritative documents had often been copied and recopied so many times that the best one might hope for would be a resuscitation of the so-called "archetype" for the extant manuscripts. From these assumptions came, therefore, not accidentals and substantives, nor indeed an articulated copy-text theory analogous to Greg's, but rather a fascination with stemmatics and filiation, with genealogy, conflation and contamination. The "ancient" editors got Lachmann, Quentin, and Maas, not Greg, Bowers, and Tanselle. And beyond the archetype in Maas' charting[2] of the Lachmannian system, there was a terra incognita where the editor dared not stray.

These dispositions and theoretical interests, arising as they did out of differing documentary circumstances, seemed to offer little chance that editors of the older material would ever need to consider the problems of their authors' preferences (especially in accidentals), since the forms of the text had been so moulded by scribal accretion and interpretation. But in the editing of one ancient field at least—Middle English—there have been signs of a change. When, for example, George Kane and E. Talbot Donaldson invoked the concept of a "corrupt" B archetype in order to facilitate both their denial of authority to the B manuscripts of Piers Plowman, and their consequent smoothing of the B text to fit the pattern established by the A version of the poem,[3] they were making a leap beyond documentary evidence, beyond the infamous archetype, into this very terra incognita of auctorial intention that was supposed to be off-limits to the editor of mediaeval texts. The problem with this particular leap, and the source of the somewhat contentious debate that has surrounded the edition since its publication ten years ago, was that there seemed to be no readily-accepted "authority" (except for the editors'

proclaimed "knowledge" of their author—derived itself from the "corrupt" documents carrying his text) which could be appealed to as an arbiter of critical decisions. There were no auctorial drafts, no fair copies, no corrected proofs, no annotations on a published text—nor even any auctorial correspondence on such basic issues as whether each line was supposed to fulfil "perfect" alliterative requirements. In sum, one either had to accept Kane and Donaldson's word that they had created Langland for the first time (virtually in despite of the documented textual transmission), or seek to re-edit the work oneself.[4] Perhaps, after all, it might be safer to stay with a version of "best-text" theory, and not to wander into the uncharted territory beyond the archetype.

The ambiguous status of the archetype (and the question of editorial fidelity to its resuscitation) has recently been raised again in Middle English, this time with a very clear connection to the equally problematic status of the accidentals in mediaeval texts, and to the validity of copy-text theory as it relates to those accidentals. The editors of the Hoccleve edition (and specifically of his major work, the Regement of Princes) have been faced with an unusual question: is it possible, considering the peculiar conditions of Hoccleve's texts, to combine orthodox "classical" base-text theory (stemmatics) with orthodox "modern" copy-text theory to produce for the first time an edition of the text of the substantives and auctorial accidentals of a mediaeval work which survives only in scribal copies?

Clearly, the immediate and sensible answer to this question would be "No", for the forty-three scribal copies of the Regement, even though as a group displaying a remarkable degree of conservatism in copying, should not logically provide access to auctorial accidentals in a work which, like virtually all other mediaeval texts, does not exist in holograph, in rough draft, in annotated first edition, or in any other "authorised" form.[5] The major qualification to the question, however, is that nearly

all of Hoccleve's other works do survive in holograph, and that it might therefore be worth looking at the accidentals of these manuscripts to determine whether there were patterns of auctorial usage that could be employed in the editing of the Regement. This process, if successful, would certainly take us beyond the archetype of the Regement, but in accidentals not in substantives. The "classical" base-text, in collation with the forty-two other extant witnesses of the Regement, would be used in a perfectly conventional manner to establish the "meaning" or the "content" of the text; the paradigms derived from the holographs of the other works (if such paradigms were genuinely demonstrable) would not weaken the status of this orthodox base-text, but would provide it with an overlay of "surface features" that could be shown to represent auctorial intention with regard to accidentals. The entire enterprise would clearly be a hybrid affair and would no doubt provoke (and possibly antagonise) loyalists of both orthodoxies; but the possibility seemed worth pursuing, especially if it could be established that some degree of consistency in the Hoccleve holographs was at least likely. This supposition depended in part upon current research on the much wider question of the movement towards standardisation of orthography in the fifteenth century, and of course, Hoccleve's place in any such movement.

It has long been accepted that some degree of standardisation in spelling took place in England during the late Middle Ages. The main influences on this tendency towards a uniform orthography have been variously described as technological (the invention of the printing press, and the resultant fixing of "conservative" forms); religious (the wide dissemination of Wycliffite tracts, in which individual scribal idiosyncracies were levelled out by what was essentially a "factory" system of copying); literary (the dominance of Chaucer and his followers, with the resultant ubiquitousness of South-East Midlands forms in the corpus of literary manuscripts, and subsequently in their incunabula descendants); economic (the growth of London as the centre of trade and finance, coupled with the new sedentary—as opposed to peripatetic—authority of parliament and crown in Westminster); and, most recently, bureaucratic (the installation of a professional class of administrative civil servants to run the business of government). For the student of orthography and spelling history, this last area is clearly of great potential significance (if only because government agencies and their clerical staffs tend to generate so much paperwork). For the Hocclevean editor, textual or literary critic, the bureaucratic move towards standardisation is of paramount importance, for while Hoccleve was indeed an avowed and theoretically devoted Chaucerian (and could therefore have been greatly influenced by the supposed dominance of Chaucerian forms in the fifteenth-century copying of literature), he was also a bureaucrat, in fact, one of the clerks whose major responsibility it was to put into documentary practice the orthographic reforms which characterised the administration of Chancery and its related offices in the fifteenth century.[6]

Hoccleve worked as a clerk in the office of the Privy Seal for many years. Indeed, his labours as clerk, and his supposedly poor remuneration

as a government hack, become one of the pervasive topics of his poetry.[7] As in the case of Chaucer, who was also a government employee, we do not have any great body of auctorial documents of a professional nature (only one of the identified holographs is "bureaucratic" or "non-literary"), so we cannot directly examine Hoccleve's clerical contribution to the orthographic reforms. However, since Hoccleve was an author as well as scribe, he did leave a considerable corpus of literary materials in his own hand. While, therefore, the content of these manuscripts[8] is clearly very different from that to be expected in the documents typically copied in either Privy Seal or Chancery, their very existence is so unusual (perhaps unique), compared with the scant holograph materials otherwise encountered in Middle English verse,[9] that the Hoccleve manuscripts, whether or not they provided confirmation of the bureaucratic reforms in orthography, would offer a great opportunity to study at first hand the accidentals of a mediaeval poet. It was, therefore, with these two concerns in mind, (the place of the Hoccleve holographs in documenting the standardisation of spelling, and the evidence they might give of auctorially-preferred accidentals), that the Hoccleve editorial team began to chart the usage patterns in the identified holographs. It should be emphasised that the primary purpose of this investigation was always editorial, and not linguistic, palaeographic, or historical. As mentioned above, we were frankly concerned as editors to discover whether there was a sufficient degree of standardisation already present in Hoccleve's usage to allow us to employ the paradigms available in the holographs in the attempt to reconstruct the accidentals of the one major Hocclevean work (the Regement) which had not survived in an auctorial fair copy. The analysis of the accidentals of the Hoccleve manuscripts could obviously have significance to historians, palaeographers, and linguists, but our aims were more modest (or perhaps more specific): to overcome the editorial and theoretical limitation traditionally associated with Middle English studies (and outlined at the opening of this essay), that the most one could hope for was to resuscitate the substantives of the archetype.

The editorial philosophy and procedures of the Hoccleve edition are, therefore, a combination of "classical" base-text Lachmannian methodology and "new bibliographical" copy-text theory derived from the Greg-Bowers school.[10] However, while the "reach beyond the archetype" might seem to replicate the Kane-Donaldson intent and method mentioned above, in fact the Hoccleve edition possesses the one editorial and technical tool missing from Kane-Donaldson's critical equipment—a "control" derived directly from auctorial documents, which can then be used to substantiate the emendations in accidentals made in the copy-text (emendations which clearly extend beyond the archetype because they do not depend for their authority upon any document, copy-text or otherwise, carrying the text of the work being edited). The immediate problem for the Hoccleve editors was that the combination of two orthodoxies meant that within mediaeval studies, there was no developed or consistent body of textual theory relating either to the choice of copy-text on "modern" principles or to the normalisation of base-text by reference to documented auctorial preference.[11]

As practising editors, we are convinced that the paucity of textual theory in this area is directly the result of the typical character of the documentary materials with which most mediaevalists have had to work: scribal copies at several removes from auctorial accidence,[12] with no "control" group of auctorial holographs to consult. However, this does not mean that such general conditions should determine textual theory and practice where some external evidence is available. In principle, therefore, we would endorse the position of Martin L. West in his Textual Criticism and Editorial Technique Applicable to Greek and Latin Texts (1973):

This position has recently been most convincingly extended (via Greg) to all editing of non-auctorial documents, in the Tanselle SB article on "ancient" editing cited earlier. Tanselle, arguing that West and other biblical, classical, and mediaeval editors sharing this position are in fact merely fulfilling Greg's orthodox view in copy-text theory of printed texts, claims that Greg's theory may be (and ought to be) extended into other fields where the problems are, in fact, similar, even though the documents appear to be of a different order.[13] In other words,

that (as Housman had objected long ago, and as Greg was also to recognise) there would always be a temptation to yield to the "tyranny of the copy-text"[14] for all features of the text, and consequently to reduce the prerogative of editorial judgement. This temptation to succumb to such tyranny should be resisted, for, as Tanselle notes (p. 65), "any feature of the copy-text that one has good reason for emending can be emended without affecting the status of the copy-text as the text one falls back on at points where no such reason exists to dictate the choice among variants" (i.e., copy-text as "presumptive authority"). This means that even if only some of the "surface features" of the copy-text could be emended by reference to auctorial usage, these changes must be undertaken, admitting that the results will be imperfect, though less imperfect than if no such emendations were attempted. "Indeed, one must ask whether for a work of any period there is ever a justification, from a scholarly point of view, of any aim regarding accidentals other than the reconstruction of the author's own practice; however imperfectly that aim may be realized in many instances, it is the only aim consistent with the view that accidentals are integral to a text" (Tanselle, p. 43). A similar view (and one of particular value to the Hoccleve editors' articulation of a theory of accidentals in a field where such theory has not hitherto been confronted adequately) has been expressed by Fredson Bowers. "Unless Hoccleve were really random and indifferent in spelling, even a false consistency if based on your selection of his authorial mss preferences would conform to Greg's criterion of reproducing accidentals as close as possible to the author's time although admittedly artificial in consistency as his exact own." (Correspondence, 30 April, 1982).

If we were to aim for a greater success than this limited, but necessary, stage, Bowers believes that the Hoccleve editors would have to demonstrate two conditions: a) that Hoccleve is consistent enough that "normalizing would not distort his variability by imposing a false and quite artificial uniformity, and b) that on the evidence [we could justify] extending normalization to words not in authorial mss on morphological considerations" (Bowers, loc. cit.). While acknowledging Greg's warning that normalisation of an early text entails considerable philological difficulties (a position the Hoccleve editors would certainly endorse, both

in theory and in practice), Bowers nevertheless believes that Greg's copy-text theory, developed for conditions unforeseen by its originator, would tend to support "any attempt, whether or not synthetic, to reproduce accidentals that were at least in part Hoccleve's own and the rest consistent with the time, as against a late scribal copy" (Bowers, loc cit.).[15]

Because the conditions of Hoccleve's text are so very different from those encountered elsewhere in the editing of Middle English literature, it is clear that the proposal to normalise by reference to auctorially-derived paradigms of accidental usage will appear as an anomaly to other mediaevalists; but it may not just offend the practice of medievalists, it may also call into question some of the theoretical assumptions that editors of mediaeval texts have long cherished. I believe it would be fair to say that many mediaevalists still regard any "mixed" text with horror, preferring to trust to the authority of a single document, rather than risk exposing the fragility of editorial judgement. This belief explains in part why we are anxious to emphasise the dual orthodoxies of our editorial method, and to stress the continuity of a developed textual theory in perhaps unfamiliar circumstances. Despite the peculiarities of the Hoccleve text, we believe that the two lines of textual theory (one running from Lachmann to Quentin, Bédier, Dearing and beyond, the other from Greg to Bowers to Tanselle) may both find a logical relevance in the unusually "mixed" conditions of the Regement and its related documents. It is not so much that we are perversely iconoclastic in textual method, but that the Hoccleve text encourages a balanced fidelity to two principles, both of which are present in our procedures.

Having established that the normalisation of copy-text accidentals by reference to auctorial preference was a viable procedure at least theoretically, we then had to create a model for such normalisation, charting the potential relationships among copy-text, the holographs, and a final edited text. This model had to consider the various "levels" of consistency of usage that might be discoverable in the basic documents, and then determine the nature and degree of editorial prerogative that should be expected in transferring the findings of the textual analysis into an edited text that would be responsive to the author's observed intentions, the textual conditions of the documents themselves, and a reader's quite proper expectations of editorial as well as auctorial consistency. As will be explained in detail later in this essay, the holograph accidentals

were listed in a new computer program which we called HOCCLEX (the same name being used for the resulting lexicon and its statistics), and an additional dictionary of normalised forms was created on microcomputer under the name NORMLEX. These normalised forms, based upon well-attested patterns in HOCCLEX, would be used for the "morphological extension" which Bowers had rightly insisted would have to be a part of our editorial rationale. The resulting model (somewhat simplified in the chart below) attempts to codify the theoretically inevitable relationships that the data-systems embedded in HOCCLEX, NORMLEX, COPY-TEXT, and EDITED TEXT would create.