The Lost History of Piers Plowman: The Earliest Transmission of Langland’s Work

In a famous forty-line invective printed as B 15.533–69, Anima inveighs against the ill effects of the Donation of Constantine, the apocryphal act that was believed to have ceded the Lateran in Rome to the papacy.¹ This passage has long figured prominently in textual and editorial work on the poem, and its culminating lines (564–67), an explosive call for disendowment, have served as an important touchstone for studies of Piers Plowman and topics as diverse as Wycliffism, anticlericalism, Franciscanism, vernacularism, English historiography, and the politics of the C revision.² Here is the passage as it appears in National Library of Wales MS 733B, which although always deemed a witness to C here (sigil N²), in fact agrees with “B” at every point of divergence between the two versions, including the absence of four C-only lines (“B” agreements against C in bold):³

“It’s the date, stupid”: Anne Hudson’s “first line of attack” with regard to the topic of “Langland and Lollardy” applies with particular force to this passage.⁴ Whereas the most common assumption identifies the call to disendow the church in lines 564–67 with Wyclif’s De civili dominio of c. 1377, Hudson instead finds “a clear echo here of the demands voiced by Ball and by Wat Tyler during the progress of the rebellion” of 1381.⁵ Yet others have pushed the dating of these lines (and thus of B) in either direction, Aubrey Gwynn suggesting that they reflect “London opinion” of the winter of 1370–71, when “popular feeling against the clerical magnates and the possessionati was acute,”⁶ and Pamela Gradon saying that Anima “would have been a credible participant in the famous scene depicted by Walsingham as taking place in 1385, when Parliament responded to Courtenay’s refusal to accept lay taxation of the clergy cum summa furia, and demanded that the clergy should be deprived of their temporalities as a remedy for their overweening pride.”⁷

These associations of Anima’s lines with major historical events of the fourteenth century show that, whichever dating might be right, Langland criticism urgently needs to come to terms with the passage’s extraordinarily complicated textual status. Even to call it “B 15.533–69,” as we will see below, is to beg the question. The treatment of this passage has already been the fulcrum of any edition of Piers Plowman, in the absence of any awareness that N² had anything to do with the problem. The dissolution of that ignorance brings home the point that the very identity of “the B version” as an integral poem available by c. 1380 is now at risk. My fundamental claim was laid out in Chapter 2: this forty-line passage originated in the Nx line of transmission of C, not in B. Hudson’s proposal that this passage responds to the Rising of 1381, and even Gradon’s quip about Anima as a would-be parliamentarian c. 1385, are entirely reasonable. But a comprehensive study of the textual problems of B 15 does more than just re-affirm my general point. It also provides new evidence concerning the two modes by which the C version left its mark on B: in the first stage loose sheets went to Bx as copied by the scribe of the W~M ancestral copy; in the second, a final set of revisions of individual lections were recorded in Bx as copied by the RF scribe. The poem’s transmission history was not one of three successive versions, each of which resisted combination with other versions so as to make later editors’ jobs easier, succumbing only later to conflation and cross-versional contamination. The early scribes of Piers Plowman, like their successors, saw themselves as transcribing that vital poem, not a self-contained “version.” This is an unexceptionable idea, but it means we need to start over in our engagement with the production and meaning of Piers Plowman.

In his discussion of these lines Aubrey Gwynn noted that in Skeat’s text, the twenty-five line passage concluding with received 567, “And purgen hem of poysen or more perel falle,” “is in fact a digression, interrupting the main argument about the Church’s mission to the heathen which is resumed” in the following line. Gwynn allowed that “Langland was always liable to digress from his theme under the stress of his emotions,”⁸ but Kane and Donaldson, perhaps prompted by his comment, saw here the “inconsequence, taking the form of discourse interrupted for no apparent homiletic or dramatic purpose or effect, and with unsmoothed transitions” as a sign of a major pre-archetypal error in the transmission of the text.⁹ Their rearrangement of the text from what they see as archetypal (501–3α, 533–69, 504–32, 570–613) to authorial (501–613) is the most dramatic and large-scale example of the editorial reconstruction of a corrupt archetype. And the error, they believe, already marred the pre-archetypal C reviser’s B manuscript, since “the differences between the two versions have the appearance of attempts to improve a defective sequence, as if the C reviser had experienced the same sense of inconsequence which first led us to suspect textual corruption.”¹⁰

If this diagnosis of “inconsequence” in the received text is accepted, as it has been by most if not all critics,¹¹ other questions arise: how did this inconsequence come about? Is the Athlone solution accurate? And how is the problem related to other problems in B passus 15? Kane and Donaldson’s answer to the first of these is remarkable for being so imaginative and so particular: a shuffling of papers, with certain lines in certain places, during a scribe’s coffee break, whose activities upon returning to his task are recoverable almost down to the level of the letter:

The dislocation will have occurred through misfolding of the inner bifolium of a quire of the unbound exemplar after some of its content had been copied. This bifolium had 20 lines to a side, as follows:

496–513 / 514–32 / / 533–51 / 552–69.

The scribe copied from it up to the Latin line 503α and then broke off. During the pause this bifolium, uppermost in his open, unbound copy, was disturbed. When replaced it was incorrectly refolded so that it now ran:

533–51 / 552–69 / / 496–513 / 514–32.

Resuming work the scribe remembered that his place was on the first side of the middle fold of the uppermost quire. As this now lay he knew that he had not yet copied the first line and went ahead (possibly even encouraged by the presence in 533ff. of several themes which, from memory of 504–32, he was expecting). He copied two sides, to 569. Glancing at the third (496–513) he noticed the Latin Querite et inuenietis (503α), which he knew he had copied. He therefore did not copy 496–503α, but did copy 504–32; line 570, to which he then came, was on a new bifolium. If he signified his error by indicating the correct order whoever used his copy as an exemplar ignored or missed his signals.¹²

This is a brilliant reconstruction of the material circumstances of manuscript production, and Middle English studies is only the richer for it. It is also wild and far-fetched, not just because it relies on such an extensive series of particular events, but also, even more so, because it comes up against the uncomfortable fact that the RF family of B does not witness to Bx for precisely the lines that Kane and Donaldson deem erroneously located in that copy: “The fact that RF’s omission is of the matter presumed dislocated is not significant,” they claim: “there is no connexion between the events. RF’s omission is typical of a scribe of their exclusive common ancestral tradition” and so forth.¹³

A problem even more pressing than this arose in Chapter 2: the fact that N², a supposed witness to C, happens to agree with W~M for precisely these forty lines. The immediate impulse among many critics might well be to maintain the current paradigm by saying that N² must have been contaminated by a manuscript of the W~M group for these lines. Those who would do so might welcome what appears to be support from Ralph Hanna and A. V. C. Schmidt, who, while remaining silent about this crucial passage, argue on other grounds that N² bears obvious signs of contamination by a manuscript of the W~M family.¹⁴ The problem, as we saw in Chapter 2, is that the supposed source of the contamination is just too obvious: how is it that Nx or N² took on this manuscript’s readings so overwhelmingly where an entirely unrelated manuscript was absent or spurious—for instance, this forty-line passage and not a line before or after? Are we to imagine that three separate scribes by coincidence faltered with regard to precisely the same forty lines, and did so only here? The eighteen lines of new C matter (17.233–49) somehow were available to the Nx scribe, who knew where they belonged and yet for some reason returned to his source of contamination for two, barely revised lines before then going back again to his normal exemplar, and in doing all this exhibited no signs of cobbling or confusion: how can this be?¹⁵ The omission of the four new lines in C’s version of the “poison of possession” passage is a real sticking point, too. As Donaldson said in 1955, with regard to RF’s assimilation to A in passus 8 and 9, the middle passages of the program outlined in Chapter 2: “Once again we are faced with contamination, though in this case the contamination is not only downward, from the greater to the less, but entirely negative, from something to nothing. Whereas I might reluctantly accept the possibility of the former, I find it utterly impossible to accept the possibility of the latter”—which, though, is precisely what this proposal must do with regard to C 17.202, 212, 228, and 230.¹⁶

Donaldson’s alternative proposal “is that none of these passages, including the B-version of the one in Passus 3 [lines 51–63], was in the original” exemplar used by the RF scribe—that is, in Bx—“but that they were introduced into the B-tradition later than the vast bulk of the text that we now have.”¹⁷ The status of our forty-line passage cements the case for him: it “is the most suspicious” of the longer passages not found in RF.

It is a long passage to have been omitted accidentally, although it could represent the material contained on a single side. What is suspicious about it is particularly its relation to the C-text. In B the passage comprises lines 495–531 [KD 533–69 after their reconstruction]; in C, 17.194–251: up to C233 C and B are very similar, though C contains four additional single lines; C233–34 are parallel to B530–31 [568–69], as are C250–51, the intervening material having been added by C. The significant fact is that in C, as in RF, the line parallel to B494 [503] (C254) is immediately followed by the line parallel to B532 [504] (C255), as though no intervening passage ever existed. In C the passage equivalent to B495–531 [533–69] appears between the C-lines parallel to B489 [497] and 492 [501] (B490–91 [498–99] are omitted in C). This suggests either that in the B-tradition the passage was introduced late and inserted erroneously five lines later than the poet intended, or, perhaps more probably, that the poet wrote the passage after he had completed the context it now appears in in B, inserted it after B494 [503], and then later, in his C-manifestation, plucked it out and shifted it to the earlier position. It should be observed that in the latter case there is evidence that the poet himself thought of the passage as a movable insertion.¹⁸

In my judgment this analysis holds up much better than does either the misfolded-bifolium narrative that would replace it twenty years later or the proposal, first put forth by Skeat and achieving widespread attention finally in 1996–97, that it is, rather, RF whose unique passages are late additions, which simply ignores the force of this argument.¹⁹ And in a major recent essay Robert Adams has repeated the idea that the lines “probably filled one side of an inserted, attached leaf” in Langland’s B papers, and that the scribe of Bx either acquired this sheet after he had copied the surrounding passages or was confused by where to place them, “so he left the inserts as mere attachments, just as they had been in his exemplar.”²⁰ In mounting this case Adams points out that many “heavily corrected” fair copies of other works (as he takes Bx to have been) “do, in fact, incorporate revisions by means of attached slips.”²¹

This might seem to raise the possibility that this loose sheet of B revision matter somehow made it into the hands of the Nx scribe. Before adjudicating that proposal it is worth noting just how difficult is this idea that Bx was the repository of loose sheets of authorial material—difficult enough to prompt Donaldson to abandon this straightforward account of 15.533–69 in favor of one that renders these lines’ absence from RF purely coincidental. There is no reason to think that Kane bullied him, or that the editors “could [not] have edited the B-version at all—other than as parallel [RF] and [W~M] texts—had they come to any other conclusion.”²² The Athlone editors’ conviction that the C reviser’s B manuscript was already corrupt reveals the problem. The 1955 essay describes Bx (Donaldson’s “Copy 1”) as “a single MS which was not the poet’s autograph, but a rather hasty and inaccurate copy of it”;²³ the 1975 edition, though, puts forth the compelling case (part but by no means all of which relies on what they see as C’s response to the dislocated 533–69) that Bx was not a copy of the holograph, but one further stage, at the least, removed from that document. If Langland was already working on his C revisions before the production of Bx, he could not have added these “B” versions of these lines to a manuscript descended from the copy he was using for those very revisions.

Adams for his part does not face this particular problem, since he does not assent to the Athlone postulate of the corrupt C reviser’s B manuscript. That alone will be a worry to those convinced by Kane and Donaldson’s account, but even those sympathetic to his cause might well have difficulty agreeing with Adams that an explanation of B 15’s production “simpler than Kane-Donaldson’s” is “ready to hand,” requiring only as an enabling assumption his hypothesis “that Bx was not entirely a fair copy and that the second half of Bx was extensively emended and corrected from an authorially revised scribal copy.”²⁴ If I understand this right, this explanation is that (1) Langland wrote a B holograph, (2) a scribal copy was made, to which Langland added some passages, (3) the Bx scribe used that scribal copy as an exemplar, but he either decided not to copy the text on the loose sheets that Langland had added, not bothering even to sew them in his copy, or he finished before Langland added still further passages, which were now added to (but, again, not sewn into) the copy he made, (4) the W~M scribe made a final copy—the first in which Langland’s lines were finally recorded. Meantime, the RF scribe’s omission of the sequence of passages in passus 8 and 9 was a series of independent instances of eyeskip that only coincidentally happened to bring that text into line with the A version at each point, while passus 3’s assimilation to A is also coincidental, being the result of self-censorship.²⁵

This does not strike me as very simple, even aside from its appeal to large-scale coincidence. It would help if there were any precedent for the idea that authorial revision matter was added not to the holograph itself, to enable the production of a fresh transcription, but to the copy that had used the holograph as exemplar. Or for the concept that a scribe ever simply left loose authorial matter in that state rather than inscribing it into his copy. None of the examples Adams cites of manuscripts enhanced by revision matter comes close to either of these postulates, and I, for one, have trouble understanding the motivations of Langland in the first instance or the Bx scribe in the second. Added to the difficulty is that this narrative does not account for the indications, manifest in C-character rubric of MSS LR at the end of passus 20, that Bx was the product of the C era, in which case any proposal that places Bx in the earliest stages of B’s authorial production falls apart.

And we still have not exhausted the list of problems posed by this approach. How plausible is the notion that such a loose sheet with lines 15.533–69 escaped RF’s notice but ended up in the hands of a random C scribe—presumably a decade or so later—who replaced his forty-four C lines with these forty, but still retained his exemplar’s new eighteen lines? How could it be that in doing so the Nx scribe created a perfectly smooth text where the only other scribe who copied them, W~M, marred his own text with inconsequence and unsmoothed transitions? As Adams says, any simpler explanation is obviously preferable. Fortunately, we do not need to create Bx in the image of Langland’s B > C revision manuscript to find one. That revision copy, say Russell and Kane (who receive Adams’s endorsement on this point), contained “well more than a dozen separate single leaves or bifolia of new material, interleaved or loose.”²⁶

This large collection of problems vanishes if the loose sheet was among the matter postulated by Russell and Kane, making its way to Bx. Again, there is powerful precedent even for the idea that such loose C matter circulated independent of its origins in Langland’s revision copy. Derek Pearsall has proposed the possibility that the lines from C 9 in the Ilchester Prologue “were ‘floating fragments,’ attached by scribes where they thought fit,” an idea Wendy Scase would later confirm by noting that Huntington 114 (Ht) attests this same variant material in its passus 6.²⁷ In my account, the sheet was part of the authorial exemplar available to the ancestral scribe Nx (the first post-holograph copy, which possibly turned into Cx upon further revision), confused the W~M scribe since there was no precise equivalent in his copy to the text’s location in C, and had been removed or lost by the time RF copied Bx. An added benefit is that, because of the late origins of this postulated document, it does not demand an attempt to explain away the evidence of LR’s C-character rubric as discussed in the Preface.

There is still additional support for my proposal, intriguing if indirect, which inheres in the indications that Langland originally envisioned the passage’s home neither where W~M attests it (after 503α) nor at the 532/570 juncture (Kane-Donaldson, Schmidt), but at the 343α/344 juncture. The reason for the passage’s absence from that location, I will argue, cannot be that a loose sheet of matter belonging to B was misplaced during the production of that version; and what is more, the C version’s character in the equivalent locations strongly supports my claim that these forty lines were a part of its production, not B’s.

First, a basic point: the arguments for the accepted location between 532 and 570 are thin on the ground, and all of them raise more problems than they solve. The only positive evidence that Kane and Donaldson offer for their reconstruction is that their 504–32 “would more naturally follow 501–3α” since both passages criticize prelates for not evangelizing—as both RF and C order the text, the “significant fact” adduced by Donaldson in 1955.²⁸ Yet it is equally true that in all forms in which this passage survives—C, RF, and W~M—the line parallel to 15.532 (C 17.282) is immediately followed by the line parallel to 15.570 (C 17.283), also as though no intervening passage ever existed, and making for a perfectly smooth—in my judgment, necessary—juxtaposition. The topic is St. Thomas Becket’s role as model for bishops:

The episcopal imperative to “walk through one’s province” is integral to the ideal reflected in St. Thomas of Canterbury, and the repetition of the terms “bishop(s)” and “cross” on either side underscores the pertinence of lines 570–71 to those culminating in 532.³⁰ The Kane-Donaldson and Schmidt reconstruction does not remove the inconsequence, then: just moves it.

Given how important is this reconstruction for such foundational ideas as the corruption of the C reviser’s B manuscript and the relationship of the two B-tradition manuscript families, it is startling to realize that the editors put forward no other supporting evidence. The rest of their discussion focuses on the diagnosis of inconsequence, not the reconstruction itself, as in the claim that in C “three revisions of detail relate to precisely the shortcomings we sense in the passage.”³¹ In a footnote, Kane and Donaldson observe that one C-version manuscript, Cotton Vespasian B XVI (M), transposes C 17.259–85β and 188–258, “that is to something like our rearrangement,” because of sophistication or good correction,³² but they misread the situation. It is clear that the scribe responsible for the error simply copied the passage with these lines as the second and third,

And perel for þe prelates þat the pope makeþ

Þat beren name of neptalym of nynyve and of damasc

(C 17.260–61, the revision of B 15.509–10)

in place of one that had these nearly identical words at the equivalent location,

So many prestes to preche as þe pope makeþ

Of nazareth of nynyve of neptalym and of damasc (C 17.188–89=B 15.493–94)

and eventually returned to correct his error, leading to the transposition of 259–85β and 188–258.³³ The W~M scribe of B made an almost identical error concerning these titular bishops: having copied “That bere bisshopes names of bedlem & babiloyne” (B 15.510), he resumed at the line just after “sovereynliche to suche þat of surrie bereth þe name” (528), an error that Donaldson cited as the best example of that scribe’s proneness to eyeskip.³⁴

None of these problems attends the proposal that the forty lines were originally intended for the 15.343α/344 juncture. Indeed, in this scenario both sites of the transition would be perfectly smooth. On the one side, the climax of a long consideration of the “Religiouse þat riche ben” would easily lead into lines 533–37:

The most felicitous effect of the juxtaposition would have been its dense focus upon the responsibilities of “the rich,” similar to that upon the responsibilities of the episcopate at 523–32, 570–71.³⁵ Other motifs, too, resonate across this divide: before, the rich “robeth … hem þat haen as ȝe haen” (340–41); after, the righteous, by contrast, walk about “evele yclothed” (535). On this side, the rich religious should feed beggars “as þe boek techeth” (343); on that, the righteous have “no boek but conscience” (536). This whole contrast between the rich and poor began with the speaker’s consideration of the examples found “in legenda sanctorum, þe lif of holy seyntes” (269); here Anima continues to expound that collection, remarking that it is piteous “to rede how riȝtwise men lyvede” (533). Likewise the other side of my reconstruction, especially if we accept Kane and Donaldson’s emendation at 566 (Charite), based on C 17.231 (it seems clearly preferable to N²/W~M’s non-alliterative “Good”), is marked by repetition of four charged terms across the juncture, with a perfect transitional term:

This reconstruction of Langland’s intentions with regard to “533–69” is further supported by C’s revisions at this site, where Langland newly invokes St. Lawrence, who “gaf goddes men goddes goodes and nat grete lordes” (C 17.67). Teresa Tavormina’s claim that St. Lawrence “may well have been the last ecclesiastical administrator free of the ‘poison of possession’” indicates his manifestation of the themes of our lines.³⁶ Langland continues by attempting to fix up the disjointedness of the received B passage. In B, immediately after counselling Christians to conform to Charity at my proposed juncture appears what Mary Carruthers calls “an odd non sequitur” in the shape of Langland’s comparison of the false soul to the false coin from Luxemburg, the lushburg:³⁷

Ac þer is a defaute in þe folk þat þe fayth kepeth,

Wherfore folk is þe febler and nat ferme of bileve.

As in lossheborwes is a luyþer alay, and ȝut loketh he lyk a sterlyng;

The marke of þat moneye is goed ac þe metal is feble. (B 15.347–50)

Langland seems to have anticipated Carruthers. Perhaps as “a draft of a new introduction to the figure to replace its abrupt appearance in B 349” he turned from Lawrence back to this figure of the lushburg, so that Anima’s original reference to bad money is no longer so abrupt:³⁸

Me may now likene lettred men to a loscheborw oþer worse

And to a badde peny with a gode printe:

Thus ar ȝe luyþer ylikned to lossheborwe sterlynges. (C 17.73–74, 83)

If the original plan had been activated, the image of the crux denarii, the cross on the coin, would have precipitated Anima’s lushburg simile, which, like the crux denarii, is associated with “werre and wikkede werkes” (356): “And now is werre and wo, and hoso why asketh: / For coveytise after cros; þe croune stant in golde” (542–43). Here the simile would have served as a natural development of a figure that itself had appeared in an appropriate context only a few dozen lines earlier, rather than the 200-odd lines earlier as in Kane and Donaldson’s text. The C revision, in other words, accomplishes what the “poison of possession” passage would have done had it remained in the location that first sparked it.

Might such indications be interpreted as evidence that the C reviser was here reconstructing an original no longer available to him—that is, that B 15.533–69 was part of the B version, but for some reason, such as the loose sheet’s absence from his copy, unavailable to him as he revised? Not very easily, which I say not just because this proposal replicates the problems canvassed above regarding B scribes’ inability or unwillingness to copy passages into their texts, but also, and more so, because it is clear that the lines were not absent at all. Else they could not turn up some 150 lines later in C 17. It makes much more sense to interpret the appearance here of the new Lawrence material instead as serving as a replacement for the “poison of possession” lines upon his decision to place them in that later location. Langland saw the 15.343α/344 juncture as the site of insertion of the Constantine lines when he initially sat down to revise his poem—but his plans changed, bringing about C 17 as we now have it, and enabling the production of B 15 as now printed in Kane and Donaldson. In sum, “B 15.533–69” was no such thing, with regard both to the line numbers and the version itself.

The manner of W~M’s assimilation to Nx, then, is relatively straightforward. But what explains RF’s lack of access to these lines? I will here argue that RF was produced after W~M, by which time the new Nx lines available to that scribe were no longer in Bx. In itself, this conclusion is not particularly noteworthy, at least if my interpretation of the LR rubric is accurate, for we already knew that Bx itself was the product of the C era, so its descendants must have been as well. But the evidence that leads to that conclusion is remarkable for uncovering yet another form of C > B contamination—a form that will play a central role in the argument of Chapter 4 below that the final two passus in toto came into B from C.

Some recent discussions of RF’s lack of 15.533–69 have assumed its connection with the absence of 15.511–28 from W~M, but these strike me as forced and arbitrary.³⁹ For one, as noted above, these lines’ absence from W~M is so easily explained as eyeskip from “That bere bisshopes names of bedlem & babiloyne” (B 15.510) to “sovereynliche to suche þat of surrie bereth þe name” (528). While appeals to local scribal error run the risk of obscuring larger-scale patterns, there are no indications of any such pattern here. Indeed, if this instance is to be connected to any other manuscript, the obvious candidate is not RF but MS M of C, which as we have seen features a major dislocation brought about by eyeskip from prelates who bear names of Nephthali, Nineveh, and Damascus to priests who bear names of Nephthali, Nineveh, and Damascus. The two separate absences of 15.511–28 and 533–69 from the respective B traditions are no more obviously related than are those of, say, B 12.139–47α from RF and B 12.116–25α from W~M,⁴⁰ or of, over the space of 300-plus lines of C 11 alone, some forty-four omissions totalling over sixty-five lines by seventeen manuscripts or groups.⁴¹ The proximity of any two errors does not indicate anything about the nature of either error.

Matters get more interesting at line 529, where W~M return, thus reading “That bere bisshopes names of bedlem & babiloyne / That huppe aboute in Ingelond to halewe men auters” (B 15.510, 529). As part of their separate cases that this W~M juncture is authorial, with the RF-only 511–28 representing a separate stage of composition, both Steven Justice and Sean Taylor have cited the syntax of RF’s 528/29 juncture (“sovereynliche to suche þat of surrie bereth þe name, / And nat to huppe”) as evidence that Langland revised this passage during the era of B-production. “What the Athlone edition (following the RF tradition) presents as its lines 526–30—which contrast today’s shirking bishops with Thomas, the apostle and martyr—is a syntactical mess,” claims Justice: “For the awkward ‘And nauȝt to’—which lacks a first term for the implied grammatical parallelism—occurs just at the splice where RF includes lines that are not present in W,” whereas the W~M version, by contrast, “supplements a grammatical parallelism with continuity of sense”; after noting the 533–69 phenomenon, he concludes simply that Langland wrote the passage twice, once with and once without St. Thomas.⁴² How this supposed hash supports such a conclusion remains murky: the passage would have made perfect sense if Langland had retained “That” in both supposed versions. In any case, this syntax appears at least twice elsewhere in B: “That is þe professioun apertly þat apendeth to knyhtes, / And nat to fasten o friday in fyve score wynter” (B 1.100–101); “For when a werkeman haþ wrouht, þan may men se þe sothe, / What he were worthy for his werk and what he haþ deserved, / And nat to fonge before for drede of dessallouwynge” (B 14.137–39).

There is, however, an intriguing explanation for RF’s syntax here. Over these four lines (15.529–32), RF (or (R ?= RF)) agrees with Cx against Nx/W~M six times, beginning with this phrase and nat to.⁴³ This is the most concentrated instance of a larger pattern. Of the 131 readings where RF alone agrees with Cx from B 8/C 10 through B 18/C 20, N² agrees with W~M ninety-three times (71 percent), by contrast supporting the expected RF/Cx reading only twenty-eight times (21 percent; for the remaining ten N² is idiosyncratic).⁴⁴ This pattern extends from variants that in other contexts might appear scribal, as do hundreds of Langland’s revisions, to occurrences for which coincident substitution is a very unsatisfactory explanation, such as RF’s spurious derivation at 11.437, “no man loveth his felachipp,” of C 13.243b, “uch man shoneth his companye,” where N² and W~M have “for evry man him shon(y)eþ,”⁴⁵ and RF/Cx’s “dobest” and “dronke aftur” for N²/W~M’s “dowel” and “toek þe coppe & dronke” (B 13.103–4/C 15.110–11).

The N²/W~M agreements here, it seems clear, are peers to those sixty-four inherited from Bx, which simply had not been revised to their final C form by the time Nx was produced, and away from which RF would move at a subsequent stage of the poem’s transmission. If these readings were instead the result of contamination in either direction, that is, W~M > Nx or Nx > W~M, it would be impossible to explain how it occurred so often where RF and Cx happen to agree. Likewise this can hardly be the result of RF > Cx contamination, unless the source of Bx’s C-contamination, as discussed in the Preface, was a pre-archetypal text, since C was already circulating by the time RF’s exemplar was produced. In any case, such an idea would not explain how these sites of contamination happened to correspond so overwhelmingly with those where Nx diverged from Cx. If, say, Nx did not support Cx for 5 or 10 percent of these readings, no one would be very surprised: but 80 percent?

My proposal, then, is that it is C that contaminated RF, after the earlier stage of C (i.e., Nx) had sent over sheets with entire passages to Bx as copied by W~M. Perhaps the RF scribe was being recompensed for his lack of the new materials to which his peer, the W~M scribe, had access. We know that between W~M’s copying and the production of Cx these forty lines were supplemented by four new ones, together with nearly thirty other small revisions. It would be a reasonable, if not inevitable, supposition that the loose sheets that went from Nx to W~M were, if not Langland’s only copies, very closely affiliated with him, so that they were recalled to enable the production of the final, C version of these passages.

Two other interesting and previously unrecognized phenomena bolster my proposal that C has contaminated RF on the level of individual line and lection, and are very difficult to account for otherwise. The first comprises RF’s substitution of spurious lines for N²/W~M agreements, discussed in Chapter 2. At these junctures Bx seems to have been marked as superseded by passages not available to RF, prompting that scribe to exercise his imagination a bit. Yet such marking could hardly have occurred before the new material existed. So the RF scribe must have used Bx as exemplar after, not before, W~M, who was able to follow up on those marks’ significance. Second is the pattern by which either of RF’s constituent surviving manuscripts, R or F, agrees with N²/W~M where the other agrees with C, which seems just too coincidental in the received account to be allowed to stand unchallenged now that its existence has come to light. Among the dozen or so instances between B 8/C 10 and B 18/C 20 are the attestation at B 8.6, by R alone, of C 10.6’s “in þis world” for N²/W~M/F “as I went,” and in the passage discussed above, R’s “amonges” at B 15.531 for N²/W~M/F “in.”⁴⁶ Kane and Donaldson explain W~M/F agreements in error as coincident variation, W~M/R ones either as correction in the immediate tradition of F or as coincident variation.⁴⁷ (R’s reading at 8.6, strictly speaking, falls into neither of the editors’ categories, since they judge it to be the sole erring B manuscript.) A simpler explanation is that the RF scribe recorded some or all of these new C readings as corrections to his manuscript, sometimes leading one of this exemplar’s subsequent scribes (e.g., R) to attest the correction where the other (e.g., F) preferred the original.

The scars of C revision upon RF are suggested elsewhere as well, as in the RF-tradition’s loss of some lines in a few instances where N² does not provide a check. Consider this Bx passage:

Thenne y courbed on my knees and criede here of grace;

And preyede here pitously to preye for my synnes;

And also kenne me kyndly on crist to bileve,

That y myhte worchen his wille þat wrouhte me to man.

“Teche me to no tresor, but telle me this ilke,

How y may save my soule, þat saynt art yholde” (B 1.79–84)

Kane and Donaldson quite reasonably explain RF’s loss of line 82 as “omission of intervening syntactical unit through attraction of kenne me … to 81 to Teche me to 83.”⁴⁸ But this is the only line of B 1.51–97 to be excised from C 1.47–93. It would now seem at least as likely that the line disappeared because it had been marked for cancellation in C, and thus in Bx as copied by RF as well. Likewise, of this passage,

Or who worse dede þan David þat Uryes deth conspired,

Or Poul þe apostel þat no pite hadde

Muche Cristene kynde to culle to dethe? (B 10.429–31)

the Athlone editors say that RF omitted the final line “through subconscious suggestion of the parallelism Or … Dauid þat, Or Poul … þat, that 430, like 429, was a completed statement.”⁴⁹ Yet whatever physical aspect of Cx prompted all C manuscripts save one to run that line (11.267) together with its predecessor (i.e., “Poul þe apostel þat no pite hadde cristene peple to culle to dethe”), this physical phenomenon, if signaled in Bx as copied by RF, could easily have caused his trouble. Finally, this proposal accounts well for the large-scale loss of lines that afflicted both subarchetypal B scribes. A number of W~M’s lost passages, in particular, can be mapped precisely upon the pattern of C’s structural revision.⁵⁰

The surviving evidence shows not only that Bx was a post-C production, but also that it bore heavy marks of the C text, far beyond the bizarre closing rubric about Dobest, in each of the two states in which it survives, W~M and RF. The “poison of possession” lines are the most spectacular instance we have encountered, yet it is the more mundane RF/Cx program that directs attentive readers to the most dramatic instance of C > B contamination, site of the erroneous “Dobest” rubric and of the heaviest concentration of RF/Cx agreements: that of the final two passus, which are at the core of the very identity of an integral “B version” c. 1378.

533/194	It is rewþe to rede how ryȝtwyse men lyved, How þei defowled her flesche, forsoke here owne wille Fer fro kiþþe & fro kynne yvel cloþed ȝeden, Badly ybedded, no boke but conscience, Ne no rychesse but þe Rode to rejoysse hem Inne: Absit nobis gloriari nisi in cruce domini nostri & c And þo was plente & pes amonge pore & ryche, And now is rewþe to rede how þe rede noble
540/201	Is reverenced ar þe rode, receyved for þe worþier
541/203	Than cristes cros þat overcome deþ & dedly synne. And now is werre & wo, & who-so why axeþ: For coveitise aftre crosse; þe corone stant in golde. Boþe ryche & religiouse, þat rode þei honoure Þat in grotis is ygrave & in golde nobles. For coveitise of þat crosse men of holy kyrke Shal torne as templeres did; þe tyme approcheþ faste. Wyte [ȝe] nouȝt, gode men, how þo men honoured
549/211	More tresore þan trwþe? I dar not telle þe soþe;
550/213	Reson & riȝtfuldome þo religious demed. Riȝt so þise clerkys for [ȝoure] coveitise ar longe <Shal deme [d]os ecclesie & ȝoure pride depose: Deposuit potentes de sede.> Ȝi[f] kniȝthode & kynde wit & comune conscience To-gidre love lelly, leveþ it wel, ȝe bisshopes, þe lordechip of londes for ever schal ȝe lese, And lyven as levitici as oure lorde ȝow techeþ: Per primicias & decimas. Whan Constantyne of curteisye holy kyrke dowed Wiþ landes & leodes, lordeshipes & rentis, An angel men herde an hiegh at Rome crye, “[D]os ecclesie þis day haþ ydronke venyme And [þo] þat have Petres power arn apoysoned alle.” A medycyne mote þerto þat may amende prelatis. That sholde prey for þe pes, possessioun hem letteþ;
564/227	Take here londes, ȝe lordes, & lat hem lyve by dymes.
565/229	ȝif possessioun be poysen & inparfiȝt hem make
566/231	Gode were to deschargen hem for holychirche sake, And purgen hem of poysen or more perel falle.
568/250	ȝif prestehod were parfyt þe peple scholde amende Þat contrarye Cristes lawe & cristendome despise.