University of Virginia Library

Medieval scribes are most appreciated when utterly anonymous. Since it is generally editors who are concerned with scribal behavior, the ideal scribe is a pellucid medium for his text.[1] Assiduously (and dully) such a person copies what he sees before him, neither adding nor (excepting inadvertence) subtracting. His errors are those of inattention and mechanical fatigue, and the text he produces is a relatively faithful representation of the exemplar from which he copies.

In the context of textual criticism, where slavish fidelity is a virtue, woe be unto the scribe who thinks. For editorially, the gravest sin a copyist can perform is to wilfully refuse to transmit the text of his exemplar. Worst of all are those scribes who not only change the exemplar but who rewrite it or who conflate it by comparison with a second source. Such unfair practices, in which the scribe arrogates to himself the powers of a modern eclectic editor are, in the modern view of scribal conduct, substantial breaches and earn for the enterprising medieval copyist the greatest opprobrium. He cheated—broke faith with his authors—by not attempting to pass on what he should have believed they wrote.

That individual who has become known as "the scribe of Huntington HM 114" has a certain notoriety as such a guilty fellow. Skeat examined his Piers Plowman in preparing his great edition; he rejected the text as "Not of much value," refused to collate it, and recognized that it conflated the B and C traditions of the poem. But the actual extent of the scribe's transgressions became clear only when R. W. Chambers visited the Huntington Library in the 1930s. Chambers dismissed this copy of the poem as not simply conflated (as it is) but riddled with scribal substitutions and indifference to Langland's sense and metre (as it also is).[2] Chambers's initial chagrin has since hardened into a critical dogma: all appeals denied, this scribe is an inhabitant of that peculiar purgatory reserved for those untrue to their exemplars.[3]

Yet I hope, through analysis of his work, to argue that this scribe's copying procedures show far more extensive and complex features than his simple piecing together of the three texts of Piers. In fact, study of this copyist may prove instructive, not because he is a devious and conflating scribe, but because of the variety and combination of responses to Middle English texts which he exemplifies. For his stints of copying, of which three separate samples survive, range from the bizarrities of the Piers through some copies of moderate fidelity and on to others of the kind thought almost ideal—a slavish attention to the exemplar.

The scribe of HM 114 is in fact responsible for two full manuscripts and a fragment of a third extended by later hands. He copied:

• 1. San Marino, Huntington HM 114
• 2. London, British Library, Harley 3943, ff. 2-7v, 9-56v, 63-67v
• 3. London, Lambeth Palace 491, part 1 (ff. 1-290v)[4]

This surviving output includes a certain amount of overlap—two copies of Chaucer's Troilus, a copy of a version of The Three Kings of Cologne along with a brief excerpt from that piece. Particularly outstanding, the scribe is one of the most prolific known copyists of alliterative poetry—in every case of poems which one would guess on the basis of manuscript survival to have been among the most popular such works: Piers, The Siege of Jerusalem, Susannah, The Awntyrs off Arthure.[5]

The surviving codices offer substantial evidence for placing this scribe. He can be dated in two ways—by his script and, more narrowly, by the paper stocks on which he worked. The single hand which wrote all three manuscripts, a rather splayed anglicana formata, cannot be dated very precisely;

it belongs to s. xv med., i.e. c. 1430-60, perhaps to s. xv2/4. Although none of the papers used in the Huntington and Lambeth manuscripts can be associated absolutely with precise watermarks cited by Briquet, the marks do frequently resemble types which apparently had limited periods of popularity. Thus the Lambeth manuscript appears on three papers which resemble those produced in the teens of the fifteenth century; the Huntington manuscript includes two papers probably from the early 1420s. Given that individual paper stocks had to be transported from their continental places of manufacture and that they would require a substantial period to become exhausted, each manuscript might have been produced as much as fifteen years after the date assigned the paper.[6] Thus the scribe's activity should probably be placed in the later 1420s or early 1430s.

The relative dating of Lambeth before Huntington is confirmed by disposition of contents. Huntington includes a brief excerpt from The Three Kings, probably inserted to supplement a copy of Mandeville's Travels in that typically defective English form which lacks much of the description of Egypt.[7] For the excerpt to be supplied, the scribe or his director had, first, to know Mandeville and to recognize that parts of the French text were not reflected in the English version at hand. Second, they had to know an English source for some of this omitted material; the logical inference is that the scribe knew Three Kings because he had earlier copied it in Lambeth.

Similarly, Harley must have preceded Huntington. The latter text presents a version of Troilus so meticulously similar to Harley that it is hard to conclude anything except that they were copied from the same archetype. But Huntington, after the copy was complete, was apparently collated against a different exemplar: in addition to a very few minor corrections, some substantial omissions of the original archetype, omissions traditionally seen as typifying the alpha version of Chaucer's poem, were corrected on inserted leaves. Knowing the defects of the archetype, it is difficult to imagine the scribe copying the text a second time without adding corrections: on this basis, Harley probably preceded Huntington but cannot be dated relatively vit à vis Lambeth.[8]

The scribe can also be placed geographically. Three kinds of evidence prove useful for this purpose: dialect information inherent in his spellings, evidence of early manuscript provenance, inferences from the form of the manuscripts themselves. These in combination suggest a person from southeastern Essex probably employed in the London booktrade.

The scribe's language appears fairly consistent and allows one to place at least his original training. The spellings point to a small area along the Thames estuary in southeastern Essex. However, that the scribe shows such an origin does not necessarily imply that his manuscripts were produced in that area.[9]

In fact, both the appearance and contents of the codices themselves, especially Lambeth and Huntington, suggest that the manuscripts were produced after the scribe had migrated to the metropolis. Both, although large in their current numbers of folios, are composites, put together from booklets

of various sizes. These range from four quires (the "oriental" materials which comprise Huntington, ff. 131-92)[10] to the thirteen quires necessary to hold the prose Brut in Lambeth (ff. 1-205). Moreover, the booklets look cheap and as if produced in the expectation that they might remain unbound for a protracted period: the paper leaves which form most of the texts are protected from outside wear and from ripping out of any interior stitching by single vellum sheets folded in quarto to provide the outermost and central leaves of each quire. These small packets of quires, minimally decorated (the scribe did nearly all his own rubrication, including running titles), look as if they might form a small in-house bookseller's stock—cheap copies of popular items in heavy demand. One has difficulty conceiving of any locale with so consistent a demand for Middle English texts as to merit even so modest productions except for the capital.

Moreover, contents suggest that either the scribe or his director was intensely aware of textual contents and capable of assembling a number of archetypes with ease. The conflated Piers Plowman presupposes a scribe who knew that there were three versions of the poem, who knew how to recognize them with some facility, and who knew where to find copies of them. Similarly, the added materials in the Huntington Troilus indicate someone in a position to acquire two different archetypes of Chaucer's poem. And the effort at filling the "Egypt gap" in Mandeville implies that the scribe may not have been limited to a knowledge of English literature (although he might have known the "Cotton Mandeville") but also perhaps knew French. The scribe's work bespeaks an extraordinary literary sophistication that one would expect within a cosmopolitan environment.

This urban localization is confirmed by the Lambeth manuscript, for which there is extensive evidence of fifteenth-century ownership. Many leaves bear the repeated fifteenth-century scrawls of a very few people, among them Thomas Patsall, John Patsall, Thomas Sharp, and John Pysant. Particularly noteworthy is Thomas Patsall's inscription on f. 22v: "in the tone of barakyng," a probable reference to Barking, Essex.[11] At some point in the first sixty or seventy years of its existence the manuscript was still owned by persons in the metropolitan area.

This metropolitan localization challenges some venerable commonplaces of Middle English literary history. The scribe's access to a substantial range of alliterative texts should lead one to query conventional assumptions about the "provinciality" or the "regionalism" of this poetic mode. In fact, the scribe provides an outstanding early example of what may be termed "the fifteenth-century East Midland trickledown" of alliterative verse. Since Angus McIntosh's demonstration that Robert Thornton received his exemplar for the alliterative Morte from a south Lincolnshire source,[12] the theory that alliterative poetry was an exclusively western movement has been untenable. But whereas McIntosh hypothesized a northerly movement of alliterative materials in the east—Thornton in north Yorks. received an archetype which had migrated from a north-central Midlands locale—the abundant evidence for north to south movement of alliterative verse in the east has remained undiscussed.

Indeed the manuscript popularity of the texts this scribe copied rests on the fact that they seem to have escaped "regional" circulation to become part of the literary canon generally.

Two of the poems which the scribe copied are stanzaic, a type of alliterative verse which seems to have circulated with greater persistence than poems in long lines. So far as one can tell, this thirteen-line stanza first appeared no later than 1375-85 and probably in south Yorks. or north Lincs.; such at least is the best provenance one can assign to the earliest of these poems, Susannah.[13] Such an identification allows one to generalize about the dissemination of all the alliterative verse copied by the scribe, excluding Piers Plowman: all of it relies, so far as one can trace it back, on Yorkshire archetypes which gradually penetrated southward through the east Midlands.

The dissemination of The Siege of Jerusalem is fairly typical, both of this history and of the scribe's handling of alliterative poetry—which is considerably more careful than the published accounts, which concentrate on the Piers, would have one believe. The scribe of HM 114 received The Siege in a form which had been current since the turn of the fifteenth century, a quite debilitated scribal version which extends from the opening at least up to line 900 of the text. Readings which generate this version first appear in the Yorkshire Petre MS., now in the Robert Taylor collection at Princeton. In the Petre/Taylor MS., the scribe either stopped his textual distortions around line 900 or shifted to a better exemplar. This same split form of the text with readings more advanced in error than Petre, and resembling those of yet other manuscripts, was also available to the most dishonest and latest copyist of The Siege, the central East Midland scribe who prepared Cotton Caligula A.ii.[14]

This version of lines 1-900 with extensive errors derived from readings like those of Petre provided a koine used by other copyists. Beyond the scribes of Caligula and Lambeth 491, Cambridge University Library Mm.v.14, copied by the known London scribe Richard Frampton, shows this form of the text.[15] And so does Huntington HM 128, which introduces an entertaining fillip in this narrative: this rather complicated manuscript was apparently copied (this portion perhaps as late as the 1440s) in a Warwickshire religious community. Such a provenance suggests that later western circulation of The Siege (the earliest copy, Laud Misc. 656 is appropriately western, from north Oxfordshire) went on only after the text had made a circuitous journey from the north, down the eastern side of the country, through London, and thence to what has usually been considered the alliterative homeland.

Similar transmission histories can be constructed for the other poems. Although Susannah shows signs of Yorkshire origin, no northern copy survives. The earliest texts are those in the Worcester Vernon and Simeon manuscripts, but these, as A. I. Doyle has suggested for other works included in these codices, may derive from exemplars acquired from the metropolis.[16] Otherwise, Susannah shows only an East Midland circulation—not just HM 114, very closely related to the Vernon-Simeon archetype and again a relatively

faithful copy, but Morgan M 818 from south Lincs. and again the ubiquitous Caligula A.ii.[17] And the Lambeth 491 copy of The Awntyrs off Arthure clearly came to the scribe of HM 114 etc. from Yorkshire; it is palpably (and again relatively faithfully) related to that used by Robert Thornton to produce his copy in Lincoln Cathedral 91.

Although it carries one some distance away from considering the actual behavior of the HM 114 scribe, it is worth pausing a moment to consider some further evidence for the East Midland alliterative connection. First, there is simply the vast amount of surviving alliterative verse from clearly eastern locales and composed in stanzas. As any reader of The Castle of Perseverance or the "Ludus Coventriae" will recognize, verse in the thirteen-line stanza with greater or lesser alliterative decoration is a common métier of fifteenth-century East Anglian drama.[18] In that context it persisted throughout the century and as far south as Essex, the locale where someone composed a few odd stanzas in the unpublished moralities of Winchester College 33.

Moreover, eastern transmission of alliterative verse and metropolitan interest in this transmission extend past the end of the middle ages. Wynkyn de Worde published versions of two alliterative poems; for one of these, William of Palerne, probably a Gloucestershire poem and one which we know otherwise only from a mid-fourteenth-century western copy, he apparently commissioned a prose redaction.[19] But the second alliterative publication shows a by-now familiar pattern: it is a full text of The Quatrefoil of Love, another stanzaic poem.[20] Moreover, it is again a stanzaic poem earliest traceable in Yorkshire (Thornton's Additional 31042) and again a text with at least some later Northeast Midland distribution through the second known manuscript, Bodleian Additional A.106, probably from more southerly parts of Yorkshire.

Such an East Midland dissemination has substantially affected the quality of the alliterative texts presented by the scribe of HM 114. But textual quality and fidelity of transmission are not altogether the same things, for scribes may be precluded from transmitting good texts by the nature of those archetypes available to them and yet they may reproduce those archetypes with some care and accuracy. The Lambeth 491 text of The Awntyrs, which is conceivably older than Thornton's copy, must remain an open question: there the denorthernizing of the text may be the activity of this scribe. But in the case of the other, more widely attested texts, collation indicates that the scribe transmitted the archetypes with which he was working with some care, and perhaps more care than colleagues with access to similar copies.[21]

Quite simply, the East Midland pre-circulation of the archetypes available to the scribe had a disastrous effect on textual quality. This tradition of copying was of a predictable sort, primarily concerned with normalization of lexicon, occasionally with recasting misunderstood lines. In the case of The Siege of Jerusalem, the agreement of all four East Midland manuscripts in readings (and their frequent agreement with the Petre/Taylor copy) can establish fairly accurately the readings of the debilitated archetype available to the scribe. Thus the correct version of Siege 50, "Ouer þe Grekys grounde myd þe grym yþes" was subjected to three substitutions of synonyms and

emerged in all these copies as "Ouer þe Grekys see wiþ þe grym wawes," a nonmetrical form. (Petre reads "grounde," "on," and "waghes.")

But if the scribe's archetype allowed him only limited access to a genuine text, he remained generally fastidious about the text he did receive. Especially when compared to the copyists of HM 128 and Caligula, he is frequently careful and retains readings probably archetypal which other scribes had no qualms about suppressing. Thus at Siege 248, he preserves apparently authorial (and archetypal) diol, while HM 128 and Caligula substitute sorwe and Cambridge drede; in the next line, Cambridge and Lambeth retain correct authorial vis/visage, while Caligula reads face and HM 128 nose; and at Siege 258 Cambridge and Lambeth preserve the northernism garte, but Caligula and HM 128 substitute made it.

Moreover, on some widely spaced occasions, the Lambeth scribe shows himself capable of decent innovation. Unlike many East Midland copyists, he did understand the metrical and stanzaic conventions of the works he copied; and he did make some efforts at correcting archetypal deficiencies to preserve these conventions.[22] That these are frequently wrong guesses (and further examples of that tinkering which has merited opprobrium in the past) should not lead one to minimize their potential intelligence. Thus at Siege 118, "To take careyns kynde of a clene mayde," the East Midland archetype had rendered the sense absolutely irrecoverable through two stages of error. First, as Petre and Thornton's Additional 31042 indicate, "careynes kynde" was converted to the compound "carlemannes kynde." The four East Midland manuscripts reduce this reading still further and provide only "To take mannes kynde . . . ." The Lambeth scribe alone worried over the failure of alliteration in the line and was aware enough of alliterative style to hypothesize "To cacche mannes kynde." Again, at Siege 680, six of the seven surviving copies present a nonmetrical line "As sone as þe rede day ros on þe schye." Given the universality of the wrong reading, the Lambeth scribe is almost certainly emending; but he emends to the otherwise hypothetical authorial reading, "As rathe as þe rede day . . . ."

Further testimony to the scribe's frequent interest in fidelity comes from his most extensive copying stint, his two reproductions of Chaucer's Troilus. Once again, one here needs to distinguish textual quality from textual fidelity. The two texts, HM 114 and Harley 3943, deviate from all other copies of Chaucer's poem; R. K. Root believed their version of the text so distinctive that he identified it as a Chaucerian early draft, which he called alpha.[23] But these distinctive scribal errors (for there seems to me little likelihood that Chaucer is responsible for much, if any, of this variation) are almost certainly not the product of the scribe himself. First, they can be partially parallelled in the slightly earlier copy in Cambridge University Library Gg.iv.27.[24] Second, neither HM nor Harley was copied from the other since each has unique, although usually very minor, readings not in the other.[25] But in the overwhelming majority of instances, in excess of ninety percent, where the texts vary from Robinson's or Windeatt's textus acceptus, they

vary together. These two facts suggest that the scribe again had access to a relatively poor archetype; the persistent agreements of the two copies demonstrate how scrupulously he tried to reproduce it.

Some further proof of this accurate rendition, and of a general interest in accuracy on the part of the scribe, comes from corrections in the HM 114 Troilus. The text was read over, perhaps twice—if so, once from the archetype also used to produce Harley, as well as (that deviousness taken as typical of this writer, once again) from a second source not textually related to the first. A first layer of corrections generally makes Harley and HM 114 more alike and presumably attends to passages the scribe had botched in his original copying. For example, words accidentally omitted are supplied (so "me" I:714, "nay" I:770, and "more" I:819). Given the triviality of the errors corrected, however, one cannot be sure of their source, whether free guesses, from the original archetype, or from an independent source.[26]

The remaining corrections are more interesting. This sporadic group of readings is probably the result of spot consultation and corrects minor errors (although very few of the distinctive variants Root called alpha) which HM 114 shares with Harley. These are converted to the correct readings of other copies. For example, both manuscripts omit the word "loue" in I:436; HM 114 interlines it. (Similarly the shared omissions II:529 "syche" and II:1074 "were" are corrected.) On both copyings the scribe seems to have found no line in the archetype corresponding to II:509 but recognized the loss on the basis of disruption of the stanza form; in HM 114 he supplies the line later in the blank he left for it. (In contrast, at II:250 the archetype lacked a line, a fact the scribe failed to recognize in either copying; he added it marginally in HM 114, and another scribe similarly corrected Harley.) The act of substantive correction itself indicates an unusual fastidiousness about the text; what is striking is that this care and interest here takes the form modern editors regard as their own prerogative (and thus as a scribal sin), the consultation of extra copies.

This is especially true of the major correction effort undertaken by the scribe. That is, he (and not R. K. Root) discovered that the alpha archetype did not include Chaucer's entire poem. Root distinguished alpha, at least initially, by its failure to provide three extensive passages, Troilus's hymn to love (III:1744-71), his temple speech on predestination (IV:953-1085), and the description of his apotheosis (V:1807-27).[27] These were not originally included in HM 114; either the scribe learned informally of parts of the poem he hadn't copied or in his use of a different archetype for corrections, he discovered these.[28] He carefully copied them out on leaves from the same paper-stocks used elsewhere in the volume and inserted them into his already completed quires.[29] At this point, I should think, textual scrupulousness and infidelity to the received text have perplexingly merged. Rather than being a figure guilty and immoral, as traditionally perceived, the scribe of HM 114 proves to be immoral (if that is the right term) precisely because he is so fastidiously committed to the accurate presentation of his texts.

Yet in every instance, such fastidiousness is qualified: the scribe is not a modern editor. Thus, although he corrects many features of his received alpha version of Troilus, it never occurred to him to correct them all, to do a full collation and enter it in his copy. Similarly, however conscious of alliterative mannerism and alliterative requirements, he left many palpably unsatisfactory lines in his copies of those poems. The main charge which might be leveled against the scribe, then, would not be unscrupulousness, but inconsistency. Although he often attends to details of the texts, that attention never achieves the status of a principled and persistent scrutiny.

The scribe's prose copying offers support to this finding. There the copyist displays the same variable responses to the received texts already discussed, although efforts at fidelity to the larger textual tradition are lacking. The Brut of Lambeth 491, for example, appears a generally faithful copy; so far as I have collated it, its readings correspond with those printed by Brie as the textus acceptus, and variation is minimal and minor. The less extensive Lambeth Three Kings text seems to be unique.[30] When the scribe copied an excerpt from the archetype into what became HM 114, he began with much the same scrupulousness as he exercised on the Brut, but, as the copying proceeded, he grew more and more prone to paraphrase and summary; the concluding pages come to resemble his rendition of Mandeville's Travels, as described by Seymour. Possibly, the scribe felt that prose did not demand the same attention to ipsissima verba as poetry did; in any event, none of the works shows anything comparable to the various efforts at textual recovery and recuperation involved in the verse texts.

The scribe of HM 114's various approaches to his texts suggest that conceptions like fidelity and independent error, the tools of the editor, do not provide absoute ways of judging scribal behavior. Certainly, insofar as scribes are merely the vehicles for transmitting literary texts, the frustration which modern editors experience may extenuate, if not justify such attention as the HM 114 scribe has received. But clarity of transmission is not, after all, necessarily the raison d'être of scribal behavior. Scribes are also directly implicated in literary history. The forms in which they choose (or are directed) to transmit texts reflect conscious choices and thus inherently attitudes toward or readings of the works they receive.

Certainly completeness of reproduction is one favored textual response of the scribe under discussion. He apparently found versions of texts he knew incomplete and went to some lengths to fill them out into composite versions. This is not such an unusual response; it can be widely parallelled in A-Text manuscripts of Piers, where B- and C-Texts are often subjoined. The HM 114 scribe perhaps most resembles the copyist of Bodley 851: both share an awareness of deviant textual detail which needs to be concorded to produce a full and thus acceptable text.[31]

But elsewhere, the scribe shows other sorts of judgements, often ones highly inconsistent and perhaps text-specific in their operation. Thus the scribe retains, so far as he is able, certain contours of alliterative poetry which one might label "provincial." At the same time, he modernizes the text

of Piers with some rigor and there seems to desire a standardized and familiarized readable text.

Attention to all these features (and to their inconsistency) is lost by concentration upon the scribe as simply a textual vehicle. He becomes uncollatable and untrustworthy. Moreover, since editors typically deal with only single texts, they lose any overall sense of the scribe as a transmitter. He becomes unrecoverable, since his procedures are fragmented into the dispersed collations of separate critical editions. Attention to scribes as the creators of literary history is a desideratum, but it can be accomplished only by studies which look over a broad cross-section of a total oeuvre.