University of Virginia Library

UNCHARACTERIZABLE ENTITIES: THE POETICS OF MIDDLE
ENGLISH SCRIBAL CULTURE AND THE DEFINITIVE
PIERS PLOWMAN
by
Andrew Galloway

George Russell and George Kane's edition of the Athlone C text of Piers Plowman is available at last, and not only is it prefaced by nearly as long and provocative an introduction as Kane and E. T. Donaldson's controversial Athlone B edition, it has also been accompanied, in the same or recent years, by the completion of several other related projects on Piers Plowman and Middle English textual scholarship, including Kane and Janet Cowen's edition of Chaucer's Legend of Good Women, applying to Chaucer the principles of the Athlone Piers Plowman; Ralph Hanna's Pursuing History, whose longest (and only wholly new) essay investigates the question of how we are to conceive of the “versions” of Piers; and, just preceding Russell and Kane's C edition (and one presumes they did not hold off going to press until he had committed himself to his own C-text readings, although given his prior dependence on Kane and Donaldson's B text, the suspicion is hard to allay), A. V. C. Schmidt's parallel-text edition of the A, B, C and Z versions of Piers Plowman. For all of their intricate massiveness as studies and editions of specific texts, these works cumulatively present, contribute to, or complete projects presenting the most important demonstrations or provocations in the second half of the twentieth century for evaluating and characterizing the evidence of Middle English authorship and late-medieval scribal culture, in the center of which, not by coincidence, is the similarly long-term medieval textual endeavor that we now consider, almost without a pause, “Langland's poem.”

The appearance of these projects and studies thus allows some conspectus of complex materials and circumstances and some assessment of extraordinary labors in textual criticism, even in a field that is merely a “small farmyard attached to the great formal gardens of classical, biblical, and Shakespearian editing,” as Derek Pearsall has modestly described Middle English textual scholarship in a valuable survey of its trends.[1] Particularly in the Piers Plowman corner of the farmyard, the most significant gains in textual scholarship have come from exploring jointly or in some fashion as a continuum au-

thorial and scribal textual production, even though this poem is almost universally acknowledged to be, at some primary stage or stages, that of a remarkable individual writer. Because of the latter circumstance, such studies indeed cannot be annexed to exciting recent work in its more august neighbor, New Testament textual scholarship, where pursuit of an original text is so radically quixotic that it has lead to appreciation of the whole tradition of transmission as a “living text” bespeaking the “community of the Spirit,” but yielding no single, authoritative reconstruction of an original.[2] Yet Middle English and especially Piers Plowman textual studies have their own crises of origins and authorial inspiration, if less heralded and less momentous, and in the projects I have mentioned the arguments for and materials of these humbler crises now more than ever lie at hand.

Because the Athlone editors' overt commitment to tracing a portrait of the author and his distinctive poetic style is the guiding principle of their series, it is ironic that the greatest achievement of the instigator and now completer of the long project has been to make the distinctions between the poem and its copyists more difficult than ever—or at least to grant that difficulty the fullest attention of any editorial project on a major Middle English work. I admit that to show more fully what I take to be the larger value of the Athlone C text for Piers Plowman and Middle English textual studies is to offer a “defense” for its labors that its editors might well wish to do without. For one thing, I do not dispute that criticisms mounted against the Athlone editions over the years, most acutely by Charlotte Brewer and Robert Adams, are applicable in new forms to the Athlone C text: the chief objection is that decisions of past editions are supported by increasingly elaborate rationalizations, even when the original judgments no longer appear well-founded.[3] But I wish to emphasize the value of the results. For Kane's previous arguments and editions combine with the complex and sometimes contradictory schemes by which Athlone C characterizes the evidence and the author in such a way that the reader is continually forced to test the

categories of such characterization, a process yielding rewards far beyond the Athlone forays of reasoning, brilliant or implausible as they may be. For better and worse, and in contradiction to their stated achievement of distinguishing the author's ipsissima verba, the Athlone editions sustain constantly the difficulty of this poem as a textual product, from which we can productively disengage neither the entity of the author nor the multiple entities of scribal culture.

While Schmidt's commentary is still to come, his more consistent and far more readily usable text now available does not impose the same burden. For that and other reasons, it is a formidable competitor to Russell-Kane as a standard scholarly edition. Indeed, Schmidt's notion of a “standard edition” is what literary critics and advanced students have a right to expect. Schmidt's text is an expensive but, considering its scope, not prohibitively costly parallel text edition, critically edited including a claim to a full conspectus of the manuscripts; Schmidt's simple display omits a great deal but also increases the edition's usefulness. Schmidt, not Kane, has finally fulfilled R. W. Chambers' 1935 plan for what became the Athlone series: to produce a parallel, critical text of the poem to replace Skeat's. In contrast, the text of the Athlone volumes, access to which is guarded by massive introductions and apparatus and a blizzard of brackets and other signs of manipulation of the copy-text, have appeared with increasingly longer delays, in separate, non-cross-indexed —indeed, non-indexed—volumes. They are notable for subtle efforts to account for the full range of textual evidence in relation to a complex, developing theory of Langland's own textual production, an authorial style which the Athlone editions have been at pains to distinguish from scribal modes, often enough conjecturing authorial readings that exist nowhere in the documentary sources. In contrast, almost all the readings presented in Schmidt's text appear somewhere in the documents (if in partial forms that have never before provably existed in the splicings he presents), a posture of triumphant conservatism that Schmidt has recently adopted in repudiation of his earlier dependence on the most controversial Athlone volume, the edition of B by Kane and Donaldson. And Schmidt is willing to show clearly when he has drawn a reading from another version, while the Athlone editors have always buried such information in dense, small-point lists in introductions.

For comparative work with the versions of Piers Plowman, Schmidt's edition deserves to hold the field as long as Skeat's 1886 edition did—for nearly a century. And although the conveniences of Schmidt's texts of Piers Plowman may seem so many tacit criticisms of the cumbersome features of the Athlone texts, Schmidt's generous comments on the pioneering importance of those editions softens such implications. Indeed, the differences between the Athlone texts and Schmidt's are in part the result of the more than forty years that the Athlone project has spanned, as opposed to the concentrated labor of Schmidt's Parallel Text Edition; in greater part they are the result of Schmidt's ambitions for accessibility and a wide educational market, and, as his literary criticism of the poem shows, for an elevation of Langland into the canonicity of a great literary talent (see The Clerkly Maker). More

admirable still, Schmidt is able to free himself from past editorial assumptions, including his own.

Nonetheless, the subtle limitations of Schmidt's easily readable, accurately transcribed, and reassuringly empirically based texts of Piers Plowman ought not be ignored, since being accessible in modern, mass-educational terms may easily distort the fluid, inchoate nature of late-medieval vernacular literary authority, hiding the possibilities that on occasion no manuscript reading may present what the author wrote and that no one version exists as a finished, literary object. Both issues are made clear by the other recent Middle English studies and editions under consideration here, which show the sorts of difficulties that scholarship on Piers Plowman and other Middle English works should continue to attend to in order fully to appreciate Middle English poetics—even when the stated reasons in these works for such involvement sometimes contradict the results.

The reasons are, however, stated at unusual length in all these projects, and with obvious efforts at methodological definitiveness. Along with the first complete collation of the manuscript evidence of The Legend of Good Women, for example, Cowen and Kane present a textbook of the editing theories that Kane has elaborated through a life's work on Piers Plowman. Such theories can be summed up as means to isolate the “distinctive personality” of the author from the “banalities of scribal variation,” as Kane puts it in his terse manifesto at the opening of the new C text (iii), but also as oblique entries into late-medieval scribal culture itself. As with the Athlone editions, the basis for Cowen-Kane is the rejection of the view that the many extant manuscripts can ever be assessed in their totality as “inferior” or “superior,” or securely arranged in a genetic relationship. Instead, like the Athlone editors, Cowen and Kane focus on defining categories of scribal error (now numbering some forty) and the determinate principles of the author's meter (they list some fourteen governing Chaucer's final -e), in order to inform an otherwise free editorial judgment at every reading of all the manuscripts, venturing to classify the manuscripts themselves in a historical relationship (on the basis of persistent and exclusive shared variation or error) only after defining by these principles what should be taken as scribal variation, and thus creating a text that depends at every moment on editorial judgment rather than any overall assumptions about the particular value of whole manuscripts.

It is worth recalling that eclectic editions on this basis still thrive at all in Middle English thanks largely to Kane's 1960 edition of the A text of Piers Plowman, even though Kane there applied his own principles fairly conservatively. Kane adopted from Old French and classical textual scholars the strictures against using shared errors to reconstruct a lost archetype by preparing a branching stemma of existing manuscripts. Such a recension, as Kane early noted, requires first a conception of what an “error” is, a notion whose subjectivity vitiates any claims that constructing stemmata on these grounds can be an objective means to reconstructing the ancestor of all extant witnesses. It should be added that such attacks on recension were origi-

nally generated because of recension's earliest practioners' claims for scientific objectivity. A modern rather than a nineteenth-century recensionist might reply that since any method involves internal assessment of error, the use of a genealogy may still provide some use in establishing a text, provided it is not followed slavishly. But Kane, editing A, felt that the proportion of nongenetic shared variation was so high as to rule out recension altogether as a means to establish the text: some genetic dyads and triads could be identified with reasonable probability, but the larger genetic groupings that would reconstruct the archetype of all existing manuscripts were simply not ascertainable. In a brilliant sustained essay against Knott and Fowler's competing, stemmatic edition of A, Kane's careful and skeptical genetic conclusions in his 1960 introduction stopped with the presentation of seven groupings of the seventeen-plus A manuscripts: two large groups, two small, and three independent manuscripts. The circumstance was further complicated by such severe contradiction between these groups that only shifting archetypes and massive coincident variation could account for it. Stunningly, Kane found “not a single instance of these manuscripts agreeing in the same, clearly unoriginal variant, or in unoriginal variants demonstrably related” (104).

The B text showed another situation, the controversial solution to which in 1975 led the Athlone series down its path toward Cowen-Kane and, almost simultaneously, Russell-Kane. B's stemma could be traced back fairly clearly to two archetypes, but both were deemed pervasively corrupt. One whole branch of this stemma is represented by only two manuscripts, called R and F, which were full of such intriguing differences from the other branch that increasing numbers of scholars, including Hanna, consider the possibility that these represent a separate stage of the poem. But by applying with a new aggressiveness the notions of scribal error that Kane first began elaborating with A, Kane-Donaldson declared that the differences between the majority archetype and the readings of R and F were due to scribal changes in both groups: these groups simply represented two original witnesses to a single, also corrupt copy of the author's manuscript. Those “corrupt” readings found in this reconstructed B and in C but not in A could, of course, present either the author's revision of A, or a scribal error in the archetype of A not made in those of B and C. But the poet's further alteration in C of many of the unique archetype B readings, sometimes roughly returning to what is found in A, plus what Kane and Donaldson judged a scribal appearance of many of the other readings that were left unchanged in C, suggested to the B editors that the C poet inherited a text with many of these scribal variations, some of which he tried to correct back into something more closely resembling the original.

This circumstance of pervasive scribal corruption in B, and in the B text used by the poet for C, allowed Kane and Donaldson to emend the archetype for B very considerably in order to reconstruct the author's original B, even if its readings were found nowhere in the evidence—and, of course, simultaneously to define the “errors” of the vulgate B text whose further transformations in C, they thought, showed the C-text poet's superior aesthetic

tastes when confronted by scribal poetic efforts: the C-poet's changes show that he was “offended” by these passages (Kane-Donaldson 105), filled with “dissatisfaction” with them (108), finding them “objectionable” (108) and frequently “correcting” (109) or at least “firmly” responding to them (110). A lingering problem was that C still left many such archetypal B readings where Kane and Donaldson could “confidently identify scribal corruption” (111), and even including more, such as doubled lines, where new C passages were inserted into B. These last points, the aesthetic gaffes of the C text, are where the story of the Athlone C edition begins. In essays over the years, Russell has increasingly suggested that C was left incomplete and only finally assembled by a second individual, Langland's first editor, whom they here propose was “a sort of literary executor” (Russell-Kane 83n28). And almost simultaneously, the discriminations involved in such judgments are what Cowen-Kane involves, since with Chaucer's work too the question is distinguishing scribal variation and textual participation from authorial revision, a distinction that, according to the methods Kane pioneered, must be made reading-by-reading rather than manuscript by manuscript—even when, as with Chaucer's Prologue to the Legend, the authorial revision is witnessed by only one manuscript.

It is no coincidence that Walter W. Skeat was the first to edit Chaucer's Legend of Good Women (1889) as a two-version work in a parallel format, since Skeat had established such a format with the 1886 Parallel Text Edition of Piers Plowman, which he was the first editor to declare existed in three authorial versions. (The argument about The Legend of Good Women had been made in scattered footnotes in Arthur Gilman's edition of 1879, and perhaps this sparked Skeat's thinking about both Chaucer and Langland's works [see Skeat, Legend xliv].) The concept of presenting parallel editions for both Langland's and Chaucer's “revised” works started with Skeat's involvement in the Early English Text Society's published parallel-text manuscript transcriptions of the works of Chaucer and Langland; and it is clear that Skeat was far more tied to what the manuscripts of the works grossly presented than to hypotheses of what authorial text lay distantly behind such witnesses. In fact on the basis of gross manuscript evidence, Skeat called two additional states of Piers “versions” even though he accepted that these were not authorial: the first conjoins A's short text with C's long conclusion, the second combines B and C texts (actually only one manuscript that he knew of presented this “version,” Bodley 814 [=“Bo”], and it presents a hybrid AC Prologue followed by a B text [Skeat, Piers i, xxv-xxvi; cp. Kane- Donaldson 2]). Thus Skeat categorized manuscripts into “groups” or “classes” on the basis of specific readings (in the case of the Legend a single name, “Ouyde” or “Guydo,” in line 1396), and within these groups as “inferior” or “superior” on the basis of kind and pervasiveness of variation. By contrast, the Athlone editors developed a taxonomy of scribal variation as a way to hypothesize the author's always-lost original. Yet in spite of these differences in philosophy and method, Kane and the Athlone editors have never come to different conclusions from Skeat's claims about the authorial origins of these works'

versions or the number of those versions. At the time of the A text edition, there were polemical reasons to follow Skeat's profile of the author and his unified progress (the alternative seemed to be the crowd of authors that Manly conjured for the poem). As the Athlone series comes to completion, its editors' assumption seems to have hardened into defense of a portrait of the author against “any post-structuralist reduction” (Russell-Kane iii)— presumably into a mere Foucauldian convenience without any coherent or developing intention (even though in fact the attacks on three discrete versions have come from sources vigorously rehabilitating the author's calculation and intention).[4]

Of a piece with this continuity with Skeat's Middle English portraiture is Cowen-Kane's re-establishment of Skeat's view that the radically different form of the Prologue of the Legend found in Cambridge University Library MS Gg.4.27 is an authorial revision. But Cowen-Kane use far more rigorous principles to establish this hypothesis than Skeat and pursue its implications far more boldly. For the question in their endeavor is to decide just how much of the unique copy of “Prologue II” (G) is to be accepted as at least a blurred witness to authorial changes; how much to be taken as a better witness than any other manuscript to what originally existed in both “Prologue I” (F) and “Prologue II” (that is, lost original readings rather than authorial revisions); and how much totally rejected as the “banalities of scribal variation.” What may best be called Kane's “profile textual criticism” comes into play, a strategy of characterizing scribal consciousness that at times approaches a kind of antiheroic Middle English literary history. Thus Cowen-Kane go to elaborate lengths to profile the Gg.4.27 scribe (he is “relatively careful” but “dull,” as shown by his signs of struggling to puzzle out the meaning of the verse) while they strive to recreate a revised authorial text prior to that scribe's unique witness to this text, as well as readings lost by scribal error or simplification from all the manuscripts of Prologue I. Yet they emphasize that, as Greg's theory of copy text dictates, what they print at all points is not fully authorial but merely “represents authentically what run-of-the-mill fifteenth century scribes made of, did to, Chaucer's language” (Cowen-Kane 149). The text presented is still only the hypothetical trace, not the presence of the author's text. As always in Athlone editions, the constant bracketing of insertions and changes (but as in the Athlone texts never marking deletions from the copy text) further reminds the reader at every stumble of the eye how far such a text, any text, is from what the Middle English poet wrote, hence how extensive the need for the editors to emend it.

In practice, the heavily emended text Cowen-Kane print is notable for how often it takes small readings of G to present not evidence of authorial revision but, instead, as evidence of the correct reading behind the flawed archetype of the first version of the Prologue; thus both texts they print begin, “A thousand sithes I haue herd men tell...,” even though only

G preserves the reading “sithes” (instead of the “tymes” of all the other manuscripts, which Chaucerians will be more familiar with as the first line of the “F” version). In turn, Cowen and Kane often use the other manuscripts to correct Prologue II—whose distinctive features, of course, exist in only one manuscript—where their principles of scribal variation explain the differences better than authorial revision. Naturally enough, in at least one instance of such peeling of hypothetical textual layers the two editors cannot agree about the distinction between scribal and authorial variation. They present their disagreement publicly in one instance, Prol. I, 118, Prol. II, 225, conservatively allowing the peculiar G reading to be printed as the text of Prol. II. But surely Kane is right that a scribe, not Chaucer, changed F's remarkable reading “Vpon the smale softe swote gras” to G's blander “Vpon the softe sote grene gras.” Cowen balks, and prevails: a bit of reverence must always, perhaps, adhere to the solitary witness we have to Chaucer's revised version. But it is the consequence of working through Kane's bold editorial scholarship that such reverence for empirical evidence—even the plausible reading of a text that is proposed as the sole witness to authorial revision— begins to seem mere superstition. Why shouldn't Kane's logic of scribal habits be applied here too? Unfortunately, to change evidence on such bases may be a slippery slope, leading to the sorts of wholesale and often less supportable emendations of the Kane-Donaldson B text.[5]

For editors of Chaucer's Legend, the ultimate test of conviction that Gg.4.27 really does represent an authorial revision is to reject everything about that manuscript that appears more likely to fit the profile of its dull, careful scribe. This leads, of course, to the awkwardness and extravagance— or sign of relentless commitment to principles—of printing parts of the first Prologue twice, the second time in place of Gg.4.27's unique variation, dismissed to its own critical apparatus. Cowen and Kane pass the biggest test of nerve by printing at Prol. II, 127-138 precisely what the other manuscripts (Prol. I) read for the “fowls' song” at 139-152, rejecting on the basis of a complex explanation of hypotheses of mental errors and compensations any attempt (such as Skeat's and the Riverside Chaucer's) to claim for G's hodgepodge here some authorial revision, or to patch it into some grammatical and metrical form. The same level of intricate reasoning allows inserting in Prol. I what no text prints, such as the brilliant emendation at 815, when Thisbe is “ek forgladde that she was escaped,” where the unusual participial adjective (forgladden, i.e. `begladded') is preserved only as indirect traces (“ffor gladde,” “so glad”) which are not explicable on the grounds of variation from the majority reading printed by all other editors, “so glad.” Other notably ingenious emendations appear (e.g., 2508, 2522-23, and 2676).

Such salutary resistance to treating as sacred an individual manuscript and its scribe allows Cowen and Kane, and still more Kane and his Athlone co-editors, not just to edit texts but to depict sharply, if mostly quite pejoratively, the nature of scribal culture in late medieval England. Because Kane

is so unenamored of the “banalities of scribal variation,” his and his co-editors' portraits of scribal activity are all the more powerful and memorable. The consequent portrait gallery of scribal culture is what theorists of historical writing such as Hayden White would call `history in the ironic mode' —what is perceived so clearly because of what it is not. Thus the direct corollary of such transient, “ironic” profiles would be a full-blown exploration of scribal agency and circumstance.

Hanna's collection of essays directly undertakes this exploration, and makes a crucial complement to Kane's and his co-editors' editions and textual criticism even while Hanna critiques many of their assumptions. The bases for the new essay on Langland's modes of production lie in all the other essays, so their claims require some comment. Hanna's opening essays on booklet manuscripts, especially those constructed for Lollard readers, for example, instance his ability to elucidate with necromantic skill both the probable archetypes of existing manuscripts and the traces of original social meanings and social uses of manuscripts. Here he reconstructs both the booklet manuscripts' original physical forms and exemplars, and the social and religious uses implies by these forms—Lollard circulation of separate tracts, non-Lollard lay religious materials gathered in provincial centers of manuscript production, for example—located against broader sweeps of social history, such as the population movements that culminated in such works being collected in London, where the diverse dialects of the booklets can most readily be conceived to converge. Such studies are important because they read manuscripts as having shifting social uses, displaying a succession of temporally defined—and superseded—“agencies.”

Here Hanna broaches the very difficult issue of differentiating properties of agency in manuscripts and, in effect, of intention (although he rarely uses the word, it is usually what he means by the more fashionable “agency”). In the case of manuscript booklets, he specifies two kinds of agency—one embodied in manuscripts assembled by a book vendor or owner, to whom the book or booklet would be a found object, a given unit any further changes to which would exceed the intention of the original scribe; the other in manuscripts intended by the book producer to be sub-units, designed to be “joined with other booklets in the same or a similar format” (23). The potential of reading manuscripts as precisely differentiable layers of multiple intention has important significance for unfolding the social world embodied in manuscripts. Ultimately, it points toward the tricky issue of who should be credited with the layers of revision in a poem like Piers Plowman, where scribal variation clearly intervenes between the author's revisions, and where the question thus constantly intrudes of how much of such scribal participation should be granted the sanction of authorial “intention” when it persists in a revised version.

Hanna's differentiation of intentions (or “agencies”) in the case of booklet- manuscripts, however, is hard to sustain rigorously, and this shows the difficulties of any effort to carry out such a fine differentiation. For if he distinguishes manuscripts made of booklets into those that were “conceived

and produced as single volumes” versus those that “reflect the eclectic tastes of an owner” (23), he also collapses that distinction by positing that those that were “conceived and produced as single volumes” were in fact conceived as small units of texts to be assembled into large codices according to the tastes of the purchaser: “Booklets... provided a useful way of building up such large codices out of a series of small sections—a procedure that may have been useful in terms both of marketing and of flexibly planning production” (23). Thus considering booklet-manuscripts in terms of the intention of “producers” turns out to merge with considering them from the point of view of the “vendors or owner.” Nor could this problem be settled were Hanna's novel proposal to produce critical editions of booklet-manuscripts realized (47). The differentiation Hanna had suggested could never be actually made, since the final, “definitive” shape of the codex would be the vendor's or owner's, but the booklets which were designed for such ad hoc assembly would carry with them the general intention of such further assembly.

In fact, Hanna is more successful when arguing for a delineated uncertainty of individual intention in the relation between a text maker, his patron, and the text's subsequent owner. This is acutely hypothesized, for example, in his subsequent discussions of John Trevisa's massive prose translations and their subsequent medieval editions, and most effectively when Hanna considers not just Trevisa's text in relation to Trevisa's own patron, Thomas Berkeley IV, and his London network of distribution, but also when he turns to a second main textual tradition, where a later scribe's patron sought to present Trevisa's works in a specific Trevisa canon, one that frames the textual initiative by Trevisa's patron (as Trevisa does himself) in order to reveal the same theme, with later topical meaning, of a literary laborer's exemplary obedience to a “patronal order” (71). Such canny explication of a very careful medieval edition accomplishes a great deal: first, it shows how even a remarkably accurate text is faithful to the original for its own, socially implicated reasons; second, it suggests that such a social meaning behind an accurate witness could never be conveyed in a critical apparatus alone.

Indeed, Hanna's most epitomizing and suggestive dictum is “A manuscript work is at any given moment always socially single”—and what is “single” here for medieval manuscript agency is not an individual human being, the writer or scribe, but rather a single horizon of immediate interaction encompassing the work, the patron and the producer. The ideal object of editions is to recover a text's “public constituency” (74) that will fuse these individual agencies into a social Gestalt, an end accomplished only by “greater access to `the evidence' than standard formats based on collation forms derived from print books allows.” For only by this means can be appreciated how all manuscripts are “every bit as mediated in their own differing ways as the standard modern edition, each historically situated and incapable of being understood outside that situation” (75).

Hanna's long essay on the versions of Piers Plowman attempts to apply to the stages of Piers the notion of such “socially single” horizons which in-

clude not a patron but the scribal culture the author relied on, and the social pressures he responded to. Since A, as Kane's edition demonstrated, has no clear genetic stemma beyond seven separate groups, Hanna proposes that seven scribes had separate access to a version of the poem in its first, short draft; the regional dialects of the seven archetypes indicate a situation of London as the meeting point for provincial scribes who each took a copy back to their own areas. This represents a stage of the poem where the author did not control access to it—whether because the scribes made their copies after he had died, or because the poet no longer cared about his first draft of his poem. The two groups of B manuscripts Kane-Donaldson defined as witnesses to two archetypes suggest to Hanna a different situation: the large group witnesses (with some omissions and many other errors) Langland's only delibrately published version of the poem, while the two manuscripts R and F, representing the second tradition, contain evidence of sets of lines the poet inserted later, on the way to producing the C text.

Almost simultaneously with Hanna's work, the argument for a second B tradition in RF was closely argued by other scholars (see Taylor; Justice 5-9); Brewer had generally proposed it in 1991, in part following Donaldson himself in a view repudiated in the Kane-Donaldson edition.[6] The sudden dominance of this claim now has perhaps been overdetermined, by, on the one hand, doubts in many quarters about the Kane-Donaldson emphasis on a single, perfected, but profoundly scribally marred B text, and on the other hand, by a current interest in fuller historical literary inquiry further fragmenting such claims of authorial stasis. Once set forth, the hypothesis finds abundant confirmation and appears likely to constitute a permanent change to the paradigm that Langland scholarship for another period will generally accept. Working through Kane-Donaldson's list of the differences between the main B tradition and RF (66-67) has only convinced me further: in tandem with the other evidence of C's further attention to the topics (generally poverty and labor) of these passages, I consider at least 20 of these RF-only sections much more likely to be further poetic insertions than the kinds of eyeskip in the single common ancestor that Kane-Donaldson claim. (Some cases of omitting or including brief, syntactically self-sufficient passages, however, might be due to eyeskip from paraph to paraph in either of the archetypes—a principle Kane does not list.)

Of course, it might seem that if this claim is so easily shattered, then one must question and `reenact' (as Kane-Donaldson actually challenge us to do [220]) every instance the Athlone editors present of distinctively scribal errors rather than authorial changes; but in fact such questioning of their conclusions has not been so pervasive. The debate over the scribal or authorial “Z text” of Piers Plowman is the only significant ongoing instance. No one

has followed Charlotte Brewer's provocative question, for example, of how we can not prove that the A text appeared in even more minorly revised authorial forms.[7] Of course, the case for an authorial post-B ancestor to RF has a much simpler basis than that for various A texts slightly modified by the author, or even than that for an authorial ancestor to the short Z as a pre-A text. Whatever the Z text is, it can at least be printed, however its evidence is characterized.[8]

Hanna's characterizations of the poet's production allows some proliferation of versions, but not an indeterminate number. Z is an extravagantly scribal version drawing from all three versions, made after the poet's death. The poet's own development proceeded as follows: the B text's careful, line-by-line reworking of A, so different from the RF and the C stages of revisions that primarily involve large-scale insertions and relocations, indicates two different kinds of attention to the text—not “authorial” versus “non-authorial,” but different kinds of authorial intention. Hanna's distinctions here resemble what Hershel Parker has argued about revisions of American fiction, when preexisting chunks are not reworked in detail and are left as vestiges of previous textual intentions that collide with new, often less meticulously integrated layers of material. In the case of Piers, this helps explain how the B text was the poet's only meticulously thorough revision, while RF and C were much more topically specific productions responding to more specific, and in part externally generated pressures.

In Hanna's view, an initial scribal copy of B (possibly a fair copy) was made when Langland first released this version for publication by transmitting it to a scribe—and from this fair copy (with its scribal variations) descends the archetype (with further variations) of the largest group of extant B manuscripts. In RF and C, the poet clarified the text on topics that the rebels in 1381 had rallied around, separating his inward-turning conclusions from their revolutionary goals, while also agreeing with them on certain social points, especially “the need for a pious and [socially] corrective clergy, not a clergy of lord's servants” (243). Thus after Langland inserted the RF passages in his original manuscript, he sent this to have the new passages inserted or interleaved into the fair copy, in order to have the work published as it had been before—and from this scribally interleaved copy descends the archetype of the RF manuscripts. After thus “publishing” both B and the partly revised B (i.e., the RF tradition), the poet no longer wished or was able to use the original B manuscript, which he had already used once for RF additions; so Langland used the scribal fair copy with the scribal RF insertions as the basis for producing the further revisions of the C text—thus inheriting the scribal variations which that initial fair copy and its later

additions had generated, few of which the poet restored in their original form.

The possible goals of an edition of “versions” that Hanna defines are, then, themselves multiple but far from innumerable—although even these raise questions about whose intention was being preserved, and how the differences might be treated. The horizons of “complete” intention that an editor could present include A, B, RF, and C (although Hanna avoids the question of how to consider in an edition the scribal variations still in the B copy used for RF; would these be sanctioned once RF was added, requiring both an edition of conjecturally emended B and one of archetypal B plus RF?). Or, Hanna suggests, the entity to present as the only meticulously complete realization of poetic intention could be simply B (although this would lack the readings from RF, even if these improve or correct the text: here, Hanna avoids the issue of B's infelicities, the more casual alliteration than A, and the occasional clumsiness at insertion of new material, the last a point I discuss below). Or the entities to present as authorial horizons of intention could feature A-as-realized-in-B, and then RF-as-realized-in-C; or the entity characterizable as the fulfillment of the author's intentions could feature just C, as the last work, which the poet might have considered complete in some sense (even though it incorporated scribal errors from the B copy, as of course did the RF version of B). His narrative is thus far more attentive to specific social interaction than the Athlone conclusions are, but far from Brewer's more extreme hypothesis that any one of the “versions” may have received occasional alterations by the author.

In fact, a second line of Hanna's inquiry into the versions of Piers Plowman seeks to curtail the number of authorial versions of C. Hanna endeavors to characterize as scribal the two manuscripts that Wendy Scase first considered in tandem as evidence that a second tradition of some passages of C circulated in rough form as attachments to the B text and thus perhaps constituted some of Langland's preliminary forays into the full-scale rewriting that became the C text. These manuscripts, Huntington Library, San Marino, California, MS 114 (=“Ht”) and Ilchester (University of London S.L. MS V.88; =“I”), both present sections of C text that resemble one another more than they resemble any of the other C manuscripts, and both include the same portions of the C revision of the Pardon passus (C.9) and the curious C insertion of Conscience's only partly alliterative speech in the Prologue. In Ilchester this last passage is made to alliterate, but in a form that evidently is an attempt to regularize a text like that found in Ht rather than the rest of the C tradition. The two texts combine these materials from C.9 and the C Prologue in manifestly similar forms, although Ht (a dizzyingly composite manuscript of all three versions) locates them after B.4.153, and I (otherwise a good C manuscript) places them in the Prologue along with sections from the A Prologue—clearly, like all instances of Skeat's “fifth version” of the poem, a copy with a cobbled Prologue to complete a headless exemplar. Scase provocatively suggested that the focus in the new pardon passages against false hermits and “lollards” might have represented a poetic criticism

of such figures that was written (“made”) and circulated between the B and C versions, and that the narrator in the full C text looks back on when he states he was

(5.3-5)

Hanna disputes the claim to an authorial tradition behind the Ilchester / Huntington materials by demonstrating that their archetype must have been a mislineated form of the received edition. Three missing lines from HtI must, he argues, have been due to eyeskip occurring from a þer to þerof appearing at line end in the mislineation. Thus the HtI archetype, far from an authorial version, is a particularly corrupt scribal fragment, plugged into damaged copies of the poem to “complete” them.

Hanna's solution to the question of the HtI tradition is meant to clear the ground before he proposes his own authorial “version” midway between B and C: the RF revision. But his handling of Scase's argument leaves some nagging questions. For example, Hanna's claim that mislineation produced the eyeskip on þer-þerof in the HtI archetype (Hanna 210- 211) is convincing for only part of his evidence: Ht and I together track the mislineation he has shown from C.Prol.113b-116, but only Ht can be mapped onto such mislineated lines for C.Prol.109-110, while Ilchester appears either to summarize or present inchoately those lines and, it seems to me, in part the lines presumed dropped from the HtI archetype (C.Prol.111-113). For this section, Prol.109-113, Ilchester's evidence seems here to have some closer relation in lineation and content to the main C tradition than to Huntington's evidence, even while Ilchester has closer relations to Huntington for the rest of the shared materials.

Here are the readings for these lines (including more than Hanna presents to indicate the repetition of the lines “dropped” from HtI plus the point where Ht and I indeed together suggest a mislineated form of the main C tradition, as Hanna has astutely noticed). I have broken some lines with virgules and indentation to indicate the most plausible points of mislineation:

(f. 42v)[9]

(Russell-Kane Prol.105-115)

(Russell-Kane 191)

As my line breaks indicate, the argument of mislineation and eyeskip can be much more easily granted for Ht than I, which parallels Ht's mislineation only with the final lines. Ilchester's “and for þair sire soeffrede þe sennes þat þey wrogh::: / and chastised not his children of her euel chekkes” seems to point toward C rather than follow Ht's mislineation here; only Ilchester makes reference to “his children,” the point of the exemplum, while Ht lacks both phrases in C that refer to `chastising his children' (110, 115). Yet while C provides the most careful psychological explanation for the reaction by the father, Heli, all three versions make the exemplum's point clear enough. Eyeskip may of course explain the absence of C's three lines (though the repetition of Anon / anon at the beginning and end of the omitted lines seem a more likely way in which eyeskip could have occurred than þer, which actually appears only in Ht). But it is intriguing that the lack of C's 111-113 in both Ht and I separates HtI from the main C tradition precisely where that tradition is extraordinarily repetitive, as well as completely non-alliterative (“Thei were discomfited in batayle and losten Archa domini ... Were disconfit in batayle and Archa domini lorn”). Eyeskip seems to be only one possible explanation: others are active editing or expansion by some agent, either in the C archetype or the HtI archetype or in more than two stages.

The difference could, of course, be scribal editing out or eyeskip of repetition in a single archetype of HtI; but how is such gross repetition to be explained in the main C archetype itself? Is it not possible that the archetype of HtI appeared first, as Scase suggests, and that the repeated phrases in the main C tradition in this passage appeared in the course of inserting a version of the HtI materials? As I will show in a moment, such prolix repetitions, whatever their source, are most common where the poet inserted new materials in C and B. And what of Scase's observation that the HtI materials paralleling C.9, in HtI found in the Prologue, lack a long section of strictly B lines carried over into the main C tradition (9.163-187)? In this section (although not in the Prologue section) the HtI materials represent swatches of materials that were here wholly new to C, which, as Scase observed, a scribe is unlikely to have copied by omitting every line carried over in the main tradition from B to C. Hanna suggests that the lack of the B carry-over lines when HtI's archetype transmitted C.9 may have been “a deliberate suppression by the archetypal scribe, reflecting a perceived desire for hermit, not beggar, materials” (214)—that is, it's a matter of coincidence that this “desire” followed the fault line between C and B. But this cannot be the answer. For as Russell-Kane's new ultraviolet-assisted transcription shows, the Ilchester Prologue includes C.9.98 and 9.105ff., which directly concern beggars not hermits (Russell-Kane 186-92). Ht moreover shifts, at the end of its presentation of the C-only insertion on Hophni and Phineas, to other materials altogether (see the summary and transcription in Russell and Nathan 129-130). Scase's point about the absence of B lines in portions of Ht and I remains unanswered, and the difficulties of surmising the respective origins and sequence of, and overall relationship between, Ht, I and the main C tradition cannot be settled unquestionably as scribal mislineation and eyeskip. Once the process of revision—on topical, spiritual, literary and other bases— is itself defined as the primary characteristic of an author, it is difficult to claim any definitive entities.

For on some basis the very existence of such an apparently rough version or versions of Conscience's speech must be explained. What sort of poet or poetic process would leave a nonrhyming first draft tacked into the poem? If this passage in the main C tradition is a rough draft, to be later polished up, why is there no evidence anywhere else in C of similarly unalliterative drafts? Did the poet have such trouble with the alliteration here, but everywhere else manage to destroy his first drafts?

The non-alliterative and rather abruptly if not irrelevantly intrusive Hophni and Phineas passage in the Prologue, framed in a speech where Conscience condemns clerics for misleading the laity and thus allowing idolatrous worship but placed amidst the scene of Westminster, constitutes in essence one of the largest problems for the Athlone C editors. For they insist that Langland was (in Hanna's wry phrasing) a deft, Chaucerian poet, “always intent on getting things precisely just so, and taking his poem through three meticulous runs simply to achieve this end” (Hanna 242). To Russell-Kane, not only is the stretch of what would be “prose if so written out” a sign of

“the roughest of Langland's drafts” (87), the whole passage on Hophni and Phineas seems not to apply to its local context of satirizing administrative clerics. It constitutes the most “extreme” evidence of what Russell-Kane claim for the entire C text: the work of revising B was radically unfinished, and C was assembled—rather, misassembled—by a later individual. Thus the Athlone project, summarized in Kane's Foreword, of showing how “the distinctive personality of the poet... survives the banalities of scribal variation” and “affords an experience calculated to dismiss any post-structuralist reduction of him as `author'” (iii), is made unattainable from the outset of the C text.

Russell-Kane's profile of this later individual, the first editor of C, is rather different from Kane's portraits of other dull scribes producing endless banalities of variation. This figure is motivated by “love and piety or zeal for the message of the poem” (though not, be it noted, for the details of the work's poetic craft as the Athlone editors are), and the clumsiness of his insertions of new material show that he is “not necessarily knowledgeable, however well intentioned and close to the poet” (87). Thus while Kane-Donaldson found the B archetype to be pervasively corrupt at the level of local errors, Russell-Kane find the C archetype to be additionally corrupt at a deeper structural level. They bring to the bar lines in C that are redundant, and they argue these exist in part because cancellations and insertion marks in the dead poet's foul papers were, however piously viewed, grossly misread.

But even the poet's own efforts, they note, include readings like those in the Hophni and Phineas section that are marked by “signs of haste” (84), “unresolved indecisions about detail” (84), and a “postponement of attention to poetic form, whether through preoccupation with particular meaning, or a distracting sense of urgency about larger plans and changes” (83). In fact, what this exercise in “profile” textual criticism produces is not simply a distinction between author and scribe, but between an author who is less than his usual careful self and a scribe who is more than his usual dull self. The profiles of the zealous but hasty author and the zealous but clumsy literary executor acutely and non-idealistically define some of the real character of the C text; but by Occam's razor one may question why two figures are presented who seem hard to tell apart.

Is the profile of the “literary executor” not also plausibly a severe if partial portrait of Langland producing his final version—a more penetrating but not significantly different image from Hanna's view of the author of C? As clear evidence of the hand of the “literary executor,” Russell-Kane mention several instances of “incompleted revisions” (86) and “ensuing confusion on the part of the maker of the first revised copy” (84) which can be identified where “both revised and unrevised forms of a line within the passage” are present (85). They instance Piers' first appearance, C.7.185-193 (cp. B.5.540-549). But the authority of this passage may be preserved in several ways. First, the repetitions in C are almost entirely emphases on how Truth pays his hired helpers by means other than literal money, producing a drumming insistence that demands for fair pay, one of the central issues of fourteenth-

century labor laws and rebellions, are only truly met by the metaphorical “pay” that satisfaction of duty yields (just as Conscience's new C-text speech to Meed [C.3.344-406a] elaborates the nature of “right reward”). The repetitions in the revision of the B passage make the pun on “to paye” as `for pleasure' and `for profit' overwhelmingly clear, as do others of the differences from and repetitions of the B text that emphasize metaphorical rather than monetary recompense: And if C is to be criticized for repetitions in new passages, so can B. In B, for example, Meed promises to help refurbish her friar-confessor's priory, leading the narrator to break all literary unities and interrupt rather garrulously and repetitiously:

B.3.64-70)

The repetition occurs just at a point where B made an insertion into A; to be sure, it is cleaned up in C. But whatever the poet's ultimate or belated efforts to remove repetitions, their presence in B and C suggests a proneness to casualness about allowing repetition, perhaps in order to keep later expansions firmly attached to the original point, perhaps because of a taste for homiletic repetition, perhaps because of a specific technique of making or ordering insertions. At any rate, such repetitions are a tic of the poet whenever he made revisions in any version, not a habit only of a post-mortem devotee. At just the points where B adds new materials to A, one finds several other examples of the sorts of repetitions just cited at B.3.64-79; for example, B.1.114, 121 (expanding A.1.112-114), and B.4.30, 32, 42 (expanding A.4.14-23). So too, B's insertion of Wrath's confession into A's sequence of Sins is connected with an odd transition, announcing that the narrator himself is very quickly confessed (5.185-186); the transition is smoothed to a more logical connection in C (6.168-169). Even in A one finds the poet recollecting his point with repetitions (Prol.91, 95; 2.22-26)—who knows if these represent expansions of still earlier passages. A rather disjunctive C addition to B.18 (C.20.350-358) even returns to the B materials with a frank admission by the poet of his tendency to let second thoughts pull away from the original point: We always catch the poet in the process of carrying out his craft, but he never appears as ashamed of being caught in that state of ongoing activity as we are inclined to think he should be. Even Russell-Kane's very modern use of parentheses here registers a kind of editorial embarrassment at work-in-progress that the poet's text nowhere demands.

What basis, then, do Russell-Kane's many detailed conjectural emendations possess? On what profile of the writer do they depend? One of many ingenious examples of their work is 5.44, in the midst of the dreamer's moving self- justification: “And so y leue in london and [vp] lond[on] bothe,” where the manuscript readings divide between ”in london and upelond bothe” and “in london and upon/on/by/out of london both.” Russell-Kane's emendation seeks to ascertain direction of variation, to divine what scribes must have been responding to; and the sly, idiomatic reading they construct— in which the narrator admits to living parasitically on (“vp”) London as well as in it—is compelling indeed. At the same time, one might well ask how structurally perfect is the C text insertion of the whole 104-line self-defense at this point in the poem, a passage roughly three times the length of the Hophni and Phineas insertion in the C Prologue. Critical acclamation of the “autobiographical” insertion is widespread and perfectly justifiable. But the kinds of explanations for the placement of this insertion (for example, that the dreamer is “confessing” as the Sins in the following section do) look to thematic parallels and associations and ignore the literal inconsistencies. These principles and this licence could also be applied to explain the insertion of Conscience's speech in the Prologue, where Langland located, at a scene featuring officially employed clerks, Conscience's condemnation of how clerks are not attending to the laity, who are thus left to worship idols while the clergy itself worships the money gained in the king's civil service. The unalliterative form of Conscience's speech (which is not “prose”; it has the right metrical stresses) could be appreciated at a deliberate evocation of the homiletic style, a sermon, Langland inserting yet another genre into his poem in a bold formal experiment. Curiously, when critics treat B, Langland's un smoothed versions of passages find much critical support; one of the reasons B has traditionally been favored over C is its surprise and illogic, which C more often than not edits out. But when C presents new passages, they are not given the same approval as the “disjunctive” or “illogical” passages first appearing in B. In the twentieth century, the C text of Piers Plowman seems destined to bring out the contradictions of our post-Romantic aesthetic ideals, favoring spontaneity and supralogical texts on the one hand, but perfectly constructed transitions and literal continuities on the other. And as always, textual criticism, whether it intends it or not, is at the edge of confrontations between past and present aesthetic ideals and other principles of authority.

Even Hanna's distinctions between the attention of the poet carefully

revising A to produce B and the poet more architecturally revising for C appear too simplistic. At any point in his career Langland could be meticulous with composition or extraordinarily unconcerned about matters like repetitions; one might venture that this is why his work has both its centripetal progress, its inward absorption, and its constant ability to invite others to feel that they too can add to its idiom and participate in its impulsive spirit.

Since there is no officially nominated D text, Russell-Kane are forced to allow in C for something like this range of authorial attention, which Kane-Donaldson could ignore in favor of a single ideal intention, ascertained against C and A as required. Russell-Kane, however, must generate a spectrum of authorial agency in C, at one end of which is “prime Langland, crisp and to the point” (87), at the other, what the poet “sanctioned” of unauthorial readings from the corrupt B text that he used in making C (93-94), or left standing because of “haste” and impending death. Russell-Kane's portrait of the artist as an old man finally allows a complexity and variety in his poetic modes that Kane-Donaldson denied. Thus while Kane-Donaldson emend B's archetype freely to square with their unvarying standard of what the poet originally wrote as B (“prime Langland”), the C poet must be hypothesized at some points at least as potentially accepting the archetypal errors of B that he did not change. Russell-Kane weaken and finally abandon the category of such “sanctioned errors” in the case of the final three passūs, the last two of which they decide to treat as having only “unsanctioned” errors, because the passūs show no clear signs that the poet changed them at all. Since his eye was never cast over them, these passūs may be changed to accord exactly with the heavily emended B edition, where the text at least represents the editors' judgment of what Langland wrote earlier rather than what he simply held in his hand later.

But in the antepenultimate passus the argument of “sanctioning” is difficult to sustain rigorously. The passus (C.20, B.18) is slightly expanded, but most sections appear hardly touched. What to do here? In Russell-Kane, C.20 is sometimes treated as having sanctioned archetypal errors—or, more likely, readings just as plausible as those that Kane-Donaldson printed for B.18 and also found in the B tradition, which Kane-Donaldson emended away but which Russell-Kane now retain (e.g., 20.14 “Auntres” vs. B.18.16 “Auentrous”; 20.51 “Thei nayled hym” vs. B.18.51 “Nailed hym”). Sometimes, however, in this passus Russell-Kane print Kane-Donaldson's B readings over C, on the presumed grounds that the C text preserves archetypal B errors that are not sanctioned (e.g., lines 30 and 38). The implications of the two kinds of treatment at once in Passus 20 are blurred. Either the editors take the poet to have “looked over” the lines whose archetypal variants they decide not to emend to follow Kane-Donaldson, but not to have glanced at the nearby lines whose archetypal variants they do choose to amend according to Kane-Donaldson; or they assume that the poet made some tiny changes in Passus 20 that they approve of (“Auentrous” to “Auntres,” for example), but not others that they think scribal and thus emend back to Kane-Donaldson

(Kane-Donaldson's “leyeth” vs. archetypal C's “hath leide” [30]; Kane- Donaldson's “cryde `crucifige!' sharpe” vs. archetypal C's “cryede `crucifige!' loude” [38]).

Or, of course, Russell-Kane may be seen as confined in some ways by the choices of Kane-Donaldson and wishing to preserve, however awkwardly, the authority of previous decisions, just as Kane-Donaldson tended to keep A lines preserved in B in the form they had in Kane (in his review of Kane-Donaldson, David Fowler counted 684 cases of emendations to produce this result, the largest of any category of emendation in Kane-Donaldson that Fowler considered). In the same way, as the Athlone editions have preeminently shown, the poet was confined by his own, scribal copies of his antecedent texts. One measure of his greatness was his ability to accommodate or elaborate such constraints. A skeptic might claim that when Russell-Kane print a “sanctioned” form of scribally corrupt B which therefore differs from Kane-Donaldson, they are covertly admitting that Kane-Donaldson went too far, that, in effect, Schmidt was right to free himself from their bolder claims and return to the B archetype (see Russell-Kane 93-94). But the theories of the Athlone editions remain capable, if with difficulty, of containing the shifting texts they present, in large and small terms. More important, the strains visible in their presentation continue to provoke, not merely allow, full consideration of the alternatives. The efforts they exact from the reader, prompting that reader to contest and construct characterizations at any given moment and on the basis of a comprehensive array of scribal and authorial evidence, stands finally as the measure of the greatness of their editorial achievement.

As I indicated above, Schmidt's recent B and C texts are often more reasonable than the Athlone texts—Schmidt presents the archetypes of both B and C in a far less emended form, and the texts are much less prone to raise questions about their own authority. Yet Schmidt's efficient, reassuring presentation never forces the reader to confront the possibilities of a very complicated interaction between Langland and his scribes that both Hanna and Russell-Kane make clear.

To show this, it would be easy to align any of Russell-Kane's challenging and ingenious emendations with Schmidt's failure to emend. But the point can be made better by examining closely a point where their texts exactly agree. This point in the C text of both editions conceals a tiny editorial crux, so small it is not noted as a crux anywhere in the Athlone editions, though it has intriguing implications both for editing Piers Plowman and for appreciating the poetic capabilities of the poem's scribal context. Pursuing the passage through Schmidt's and the Athlone editions demonstrates the intractable difficulties of the manuscript tradition and the character of the editions—as well as the limitations of any edition of the poem especially at the level of minute readings.

At A.1.38-39, B.1.40-41, and C.1.38-39 Holy Church teaches Will something about the relation between spiritual dangers, his soul, and his heart; but the particular nature of her advice is varied in many of the manuscripts

and in the editions of the three versions. I start with Kane-Donaldson's treatment, because it is the most overtly scandalous in its emendation of the manuscript evidence. They present a reading found nowhere in the B manuscripts but drawn, rather, from Kane's edition of A where it is based on a majority of manuscripts. The world will betray Will, Holy Church declares, This is straight Pauline moral warning and injunction, modified by the topos that the three great dangers to the soul are the world, the flesh and the devil: these together follow the world, and “that besmirches your soul; set this in your heart,” as Donaldson presented the lines in his translation. But the B manuscripts Kane-Donaldson drew from present the second line differently, as a moment of nearly Spenserian faculty psychology in which the soul and heart dramatically interact: “that seeþ þi soule and seith it in þin herte.” The only exception in B manuscripts is that R and in part F (resembling here the Z text) read “that sueth thi soule and seith in thin herte.” (I present a full apparatus for the line below.)

Kane-Donaldson's boldly sustained view here is that the true B reading is preserved only in A because of the pervasive corruption of B. Indeed, none of the A manuscripts present a variation precisely like B's majority, so the principle of “direction of variation” seems in this perspective to point from B's majority reading back to A. Yet two A manuscripts seem to display some connection to the reading of the majority B tradition but construed in a different way, with a different portrayal of psychology: they state, “that schendith þi soule and seith it in þin herte”—that is, the dangers destroying the soul are what speak to the heart; here the lines define not the psychology of conscience but the mechanism of sin. C manuscripts, finally, generally parallel the majority B reading, as would be expected since Langland revised from a B text (even though Kane-Donaldson here emend the B text to what they take to be a more correct text than the one the poet himself used in revision). But elements of the A reading return to haunt three C manuscripts. One of these manuscripts (Add. 10574, “L”) is, apart from this passus and part of the next which include a C text, a B manuscript although it shows no consistent genetic relation with any other B manuscript (see Kane-Donaldson 10, 62n96). For the second line in this portion of its C section, it reads “For thei wolde shende thi soule be war of her wyles”—a paraphrase of the majority A tradition. The second C manuscript, Cotton Vespasian B XVI, “M,” reads “And þat sleþ þi soule and þow sette hem in þin herte” (that is, they will kill your soul if you place them in your heart—M's insertion of the subject pronoun insures that “and” will mean “if”). The third manuscript is “X,” Huntington 143, the copy-text of choice for all twentieth- century C editions: “And þat schenth þe soule and sayth hit the in herte.” In this instance the word “schenth” is tightly written, with suspended `n,' in a different but early hand over an erasure where “seeth,” for example, would

precisely fit (f. 4v). The scribe went to some trouble to change the reading in order to accord with what appears to us as the A majority reading.

L's reading could reflect memory of another version of the poem—although since L is apart from this C section a B manuscript, unless L's scribe were remembering a very unusual copy of B (close, in fact, to what Kane and Donaldson speculate was the original copy), that memory would here almost certainly involve memory of A as well, where the reading L has here is most abundantly found: L's scribe would be a true Piers scholar of the fifteenth century! X's correction to “schenth” probably represents another kind of collation found especially in C version manuscripts: careful scribal checking of other C copies, which are distinct from manuscripts of the other versions of the poem in having circulated in a limited, dialectically defined area, near Hereford and the narrator's stated home in Malvern Hills. X has been corrected in small ways like this throughout; indeed, it also presents in a back empty leaf the four opening lines of another, lost C copy of the poem, a version related to the P not the X family, so far as can be discerned from four lines (see Grindley). It is clear, thus, that some owner or corrector of X consulted another C manuscript from a family that preserved many variations from the X family, although C manuscripts related to P do not otherwise read “shendith.”

All editors face some embarrassing difficulties in treating this tiny crux. The language of the two main competing versions of the line differs so little in sounds and orthographic shape, yet so greatly in meaning, that calculated authorial revision is far less likely than scribal error, compounded by scribal ingenuity in decipherment and perhaps scribal zeal in seeking to recover the pure Langland, and perhaps in turn authorial interest in or “sanctioning” of what the scribes had generated.[10] But outlining a story for this is difficult; the manuscripts do not preserve unvarying loyalty to the main reading of their version—three A manuscripts here read as if influenced by C and B, and three of C read as if influenced by A. A tenet of the Athlone editions has always been the commonness of random and contradictory groupings of manuscript pairs in the Piers Plowman tradition, produced either by “coincident variation”—that is, the principle that errors that occurred once spontaneously might at the same point again spontaneously occur—or by memorial or other contamination, and paralyzing confidence in establishing genetic bases for editorial decisions. On the one hand, then, this appears to be a good case of the problem of coincident variation that Kane emphasized in order to reject establishing a certain stemma of manuscripts on the basis of shared error. On the other hand, the consistency of readings where the whole line jumps track from the sense conveyed by the A majority to the B or C majority, or vice versa, speaks against coincident variation, which might account for random parallels in one part of the line but seem much less likely

to account for parallels in the whole line. The other possibility is memorial or textual correction: X, for example, was corrected, as already mentioned. M or its ancestor was evidently prone to correction from a text like the archeypal B text, because in fifty places it uniquely against all other C manuscripts reads like archetypal B (Russell-Kane 102n23). L is already a “mixed” text, so contamination would seem likely, given the availability of other versions to L's scribe. But all of these would need access to the majority A reading, not the majority B, and since M and L already have evident access to B, this circumstance multiplies the complexity of how such a reading might have arrived in their texts. How the non-authorial variant (whichever it is) started and developed is easy to see and set in one's heart; how it circulated and destroyed the previous reading, nonetheless allowing each reading sometimes to reappear, is much harder to say.

Both main readings (i.e., shendeth/set and seeth/sayth) are sufficiently coherent and interesting to deserve printing as part of the “living text” of the poem, but the two main readings cannot be confidently assigned to different “versions” of the poem or even different productions by the author. In different ways each is “harder” than the other, each bespeaks subtlety. In the face of such difficulty at this small a scale, any edition falters, and a choice must hide far more complexity and doubt than any critical apparatus can immediately present.

In this perspective, how the Athlone editors have treated the passage, and the poem, across nearly forty years, may not seem so scandalous. At first glance, the editors appear flagrantly inconsistent and counterintuitive: editing A, Kane ignored the majority reading of B and C; editing B, Kane and Donaldson ignored the majority reading of B itself as well as C in favor of A. But editing C, Russell and Kane accept the majority reading of C, thus denying either its absolute validity or A's. But the Athlone editors at least make both readings available in the text, and in a sequence that has an intriguing and intricate basis in their theories of the different states of the poem and intentions and attentions of the poet. Whether or not they have accurately located the fine line between the C poet's intention and sanctioning of B archetypal errors they have “left in” C, and the lost B and C readings that their emendations have sought to recover, they have accomplished the task of forcing readers to confront the fluidity and almost instantly suprapersonal status of the poem. Although they nowhere comment on the line, their approach allows them to present both possible readings while implying an explanation in their subtle, provocative, if unprovable account of B archetypal scribal participation and C revision authorial sanctioning.

Schmidt makes the reader of the poem work much less hard. He chooses the simpler and more elegant correction that ensues by assuming that B and C should overbalance the reading in A: the corruption occurred in the archetype of A, never appeared in B, and of course was retained as the right reading in C. Schmidt's view on this reading is expressed with both consistency and economy in his Parallel Text edition.

Schmidt himself did not arrive easily at his solution, but less because of the difficulties of the text than those of finding his own editorial confidence. In his first edition of his B text in 1978, he accepted most of such readings that Kane-Donaldson had dramatically altered from the B archetype, but he seems not always to have known the reason for their emendations. In 1978, accepting their reading here as at most points, he even added a note by which in effect he tried to provide a rationale for what Kane-Donaldson had done, but which ends up in a circular logic that makes almost no sense at all and points to no recognizable editorial principle: “The complex variation in this line [B.1.41] easily explains corruption of an ostensibly easy word,” he states. As if this were not dizzying enough, Schmidt seeks support for Kane and Donaldson's reading from the solitary B manuscript, Cambridge University Library Gg.4.31 that had written “set” (it happens to be a very late manuscript: early sixteenth century); “as frequently,” he states, this manuscript seemed to be corrected from a B manuscript “superior” to the B archetype, “or from an A or C MS regarded as containing superior B-readings” (Schmidt, Complete Edition 267). How the “complex variation” in the witnesses to B.1.40 could “easily” explain “corruption of an ostensibly easy word”; how Gg.4.31 was so insightful as to know when to regard an A or C manuscript “as containing superior B-readings”—these are not questions Schmidt's textual commentary in 1978 could answer. So deeply was Schmidt in 1978 faithful to the text that Kane and Donaldson produced that he was willing to supply even absurd explanations where they from their Olympian heights had uttered none, except in all they had said about the B archetype's pervasive corruption.

But Schmidt has been able to renounce his own editorial past, to a degree of which the Athlone editors are incapable. In itself this is admirable; had Kane showed more of this trait, the Athlone series might invoke less suspicion of face-saving ingenuity. In his 1995 Everyman edition of B (identical to the B text in his Parallel Text), Schmidt boldly announced his pervasive revision of the poem's text: “all printings before the current one are to be regarded as superseded” (Schmidt, Critical Edition xi). In part this turns out to be his emancipation from Kane-Donaldson (see, e.g., lxxiv). In the lines in question he prints the almost unanimous B reading, with a more felicitous rationale for its variants: “satisfactory in sense and confirmed by C.” G now loses its evidence of scribal insight and becomes a case of contamination “from an A source of the r-family” (365). With this comment on the line, Schmidt thus rejects not just the Kane-Donaldson reading in B, but Kane's reading in A too, where “set” becomes simply the reading of “the r-family,” that is, the majority group of A manuscripts which provided Kane with his reading. The A majority reading, then, is not only to be avoided in B, it is wrong in A—a position that Schmidt carries out in his Parallel Text edition, where A, B, C and Z all read as the B and C majority reading, requiring only the emendation of one word in the A text (although Z's seuth, i.e. `sees' with -uth instead of -eth, looks suspiciously like a mistake for the sueth found elsewhere, or,

conversely, the source of sueth as an early misreading). In fact the only difficulty with this neat solution is that treating A's “shendith” as an error found in all the A manuscripts ignores everything Kane demonstrated and Hanna further elucidates about A's seven separate archetypes, agreeing in “not a single instance” of unoriginal variants (Kane 104).

Here, then, is a schematic view of the manuscript evidence and editorial choices for these lines. The chart lists the poem's “versions,” its manuscripts (as denoted by the normative Athlone sigla but not otherwise identified, except in the discussion above) and its editions (with Schmidt's 1978 B text designated as Schmidt1 and his 1995 Parallel Text as Schmidt2):

The most notable features are again, of course, how or whether the editions have manuscript support; but as I have already suggested, the manuscripts' further, partial or cross-versional readings are thought-provoking as well. As noted above, Schmidt2 presents the most consistent and economical strategy, and requires little struggle by the reader: now all four texts present the line identically, and in the form in which, it is true, most medieval readers, doubtless including Langland himself by the time he came to reread B, encountered the line. But consistency and economy are not necessarily guides to historical truth. To reduce clutter on a page, and avoid section after section of a dense introduction, rife with arguments for printing a known scribal error—that is, to eliminate all the obsessive attention that Russell and Kane so constantly demand—is an immediate relief. The commentary in Schmidt's Everyman B text has, even before his commentary on the Parallel Text is

published, already shown how he will spare us. But Piers is a tradition rather than an isolated “work,” the result of a complex, diachronic relation between Langland and his earliest readers that is nothing if not prolific clutter, which can be truly understood only by pondering the shifting authority of variants (whatever their origins) in Langland's own lifetime as well as the continuous process in which he was engaged.

Of course, in this small instance Kane and the associate Athlone editors may simply be wrong, Schmidt right. And how we describe Langland's kind of poetry will change as a result. Schmidt's choice of lection here is for a more cerebral reading, in keeping with Schmidt's view of the “clerkly maker” who seeks allegorical effects at least partly for their own sake. Kane and the Athlone editors, by choosing a more directly moral reading and more direct didactic posture, choose as it were a more old-fashioned poet concerned with moral principles, not the frisson of allegory. But Kane and his co-editors provide us more perspectives because they insist that we adopt a complex theory of social and authorial intentions behind the work even if we do not accept their particular version of that theory, and because they force upon the reader the evidence that complexity of intention necessarily obtains in the poem's living text. Just as Kane insisted more than once in his introduction to the A text that “Far from the text... here being secure, as a glance at the critical apparatus might suggest, it is actually seriously in doubt” (155), and just as Kane-Donaldson demonstrated by their aggressive and obtrustive emendations and introduction how thoroughly non-authorial the B text archetype was, so now Russell-Kane, in their theory of Langland's consistency of poetic ingenuity but inability to keep in his own control his final—and I would say every—version of his poem, potentially challenge the reader to test all the possibilities personally. In the Athlone C text as in Hanna's study and Cowen-Kane's edition of Chaucer, it is a challenge we are encouraged to keep rising to, even though few readers will be able to sustain this exhausting endeavor for long stretches. Not just Schmidt's choice of lection in how the soul confronts the world, the flesh and the devil, but Schmidt's presentation of this reading as a clear and identical text in every version removes the taint of a vigorous and intimate relation between Langland and the immediate world that his poem so often addresses, evidence of which is tightly compressed into a lean critical apparatus. It is an efficient and elegant editorial solution, and it is likely to move Langland's text toward the familiarity, perhaps even susceptibility for memorization, that throughout this century has been enjoyed by Chaucer's. In contrast, the shifting Athlone “profile” of the author is only awkwardly held together as their project arrives at its ponderous fruition. Yet, however contradictory and deceptive, that project potentially impels constant labor by the reader to consider the mutual invasions of scribal culture into Langland's poem, and of Langland into scribal culture, the very process that the Athlone theory expressly seeks to put an end to. Reading the Athlone edition, particularly reading it against its own theses, is a painful struggle. Both Langland and his world deserve such pains.

WORKS CITED

Adams, Robert. “Editing Piers Plowman B: The Imperative of an Intermittently Critical Edition.” Studies in Bibliography 45 (1992): 31-68.

Brewer, Charlotte. “The Textual Principles of Kane's A Text.” The Yearbook of Langland Studies 3 (1989): 67-90.

—. “Authorial vs. Scribal Writing in Piers Plowman.” Medieval Literature: Texts and Interpretation. Ed. Tim William Machan. Binghamton, NY: Medieval and Renaissance Texts and Studies, 1991. 59-89.

—. “George Kane's Process of Revision.” Minnis and Brewer 71-96.

—. Editing Piers Plowman: The Evolution of the Text. Cambridge. Cambridge UP, 1996.

Chambers, R. W. “The Manuscripts of Piers Plowman in the Huntington Library, and Their Value for Fixing the Text of the Poem.” The Huntington Library Bulletin 8 (1935): 1-27.

Cowen, Janet, and George Kane, eds. Geoffrey Chaucer: The Legend of Good Women. East Lansing, MI: Colleagues P, 1995.

Donaldson, E. Talbot. “MSS R and F in the B-Tradition of Piers Plowman.” Transactions of the Connecticut Academy of Arts and Sciences 39 (1955): 179-212.

—. Piers Plowman: An Alliterative Verse Translation. Ed. Elizabeth D. Kirk and Judith H. Anderson. New York and London: W. W. Norton, 1990.

Ehrman, Bart D. The Orthodox Corruption of Scripture: The Effect of Early Christological Controversies on the Text of the New Testament. New York and Oxford: Oxford UP, 1993.

Fowler, David. “A New Edition of the B Text of Piers Plowman.” Yearbook of English Studies 7 (1977): 23-42.

Green, Richard F. “The Lost Exemplar of the Z-Text of Piers Plowman and its 20-line Pages.” Medium Ævum 56 (1987): 301-310.

Grindley, Carl. “A New Fragment of the Piers Plowman C Text?” The Yearbook of Langland Studies 11 (1997): 135-140.

Hanna, Ralph III. Pursuing History: Middle English Manuscripts and their Texts. Stanford, CA: Stanford UP, 1996.

Justice, Steven. “Introduction. Authorial Work and Literary Ideology.” Written Work: Langland, Labor, and Authorship. Ed. Steven Justice and Kathryn Kerby-Fulton. Philadelphia: U of Pennsylvania P, 1997. 1-12.

Kane, George, ed. Piers Plowman: Volume I: The A Version. London: Athlone Press. 1960; rpt. with addenda, Berkeley and Los Angeles: U of California P, 1988.

—. “The `Z Version' of Piers Plowman.” Speculum 60 (1985): 910-930.

— and E. Talbot Donaldson, eds. Piers Plowman: Volume II: The B Version London: Athlone Press, 1975; rpt. Berkeley and Los Angeles: U of California P, 1988.

Middleton, Anne. “Introduction: The Critical Heritage.” A Companion to Piers Plowman. Ed. John A. Alford. Berkeley and Los Angeles: U of California P, 1988. 1-25.

Minnis, A. J., and Charlotte Brewer, eds. Crux and Controversy in Middle English Textual Criticism. Cambridge: D. S. Brewer, 1992.

Parker, D. C. The Living Text of the Gospels. Cambridge: Cambridge UP, 1997.

Parker, Hershel. Flawed Texts and Verbal Icons: Literary Authority in American Fiction. Evanston, Ill.: Northwestern UP, 1984.

Pearsall, Derek. “Authorial Revision in Some Late-Medieval English Texts.” Minnis and Brewer 39-49.

—. “Theory and Practice in Middle English Editing.” Text 7 (1995): 107-126.

Russell, George and Venetia Nathan. “A Piers Plowman Manuscript in the Huntington Library.” The Huntington Library Quarterly 26 (1963): 129- 130.

— and George Kane, eds. Piers Plowman: Volume III: The C Version. London: Athlone Press; Berkeley and Los Angeles: U of California P, 1997.

Scase, Wendy. “Two Piers Plowman C-Text Interpolations: Evidence for a Second Textual Tradition.” Notes and Queries ns 34 (1987): 456-463.

Schmidt, A. V. C., ed. The Vision of Piers Plowman: A Complete Edition of the B-Text. London: J. M. Dent, 1978.

—. The Clerkly Maker: Langland's Poetic Art. Suffolk: Boydell and Brewer, 1987.

—, ed. The Vision of Piers Plowman: A Critical Edition of the B- Text Based on Trinity College Cambridge MS B.15.17. 2d ed. London: J. M. Dent; Vermont: Charles E. Tuttle, 1995.

—, ed. Piers Plowman: A Parallel-Text Edition of the A, B, C and Z Versions. Vol. 1. London: J. M. Dent, 1995.

Skeat, Walter W., ed. The Vision of William concerning Piers the Plowman... (Text B). Early English Text Society, O.S. 38. London, 1689.

—, ed. Chaucer: The Legend of Good Women. Oxford: Clarendon P, 1889.

Taylor, Sean. “The Lost Revision of Piers Plowman B.” The Yearbook of Langland Studies 11 (1997): 97-134.

Notes