University of Virginia Library

Foul Papers, Compositor B, and the Speech-Prefixes Of All's Well that Ends Well
by
Fredson Bowers

About All's Well that Ends Well Sir Walter Greg wrote, "There can, of course, be no doubt that behind F lie the author's foul papers."[1] This conclusion was based in some part on the preservation of what Greg took to be authorial stage-directions and even of a memorandum like 'Parolles and Lafew stay behind, commenting of this wedding' (I.iii [TLN 1089-90]), various inadequate entrance directions and numerous omitted exits, but chiefly the wide variety of designations in the speech-prefixes for the Countess, Bertram, the two French Lords, and in part for Lafeu. Greg considered but rejected the possibility that some intermediary agent might have come between Jaggard's compositors and Shakespeare's holograph papers. He remarked, it is true, that "Some long dashes in the printed text and some broken lines suggest cuts or alterations, and in occasional inconsequences and contradictions and in imperfectly assimilated chunks of prose it is easy to find evidence of botching. No Shakespearian spellings have been noted and the text is in worse form than we should expect from Jaggard's compositors working on an autograph manuscript."[2]

He concluded: "It has been conjectured that the foul papers had been transcribed by a rather incompetent literary scribe—anyone connected with the theatre must have done more to tidy them up—and if so it might be possible to blame him for some sententious fustian. But what such a transcript could have been doing among the playhouse manuscripts is unexplained."[3]

Obviously Greg is right that Shakespeare's foul papers lie behind this play. Whether they were holograph or had been literally transcribed is basically undemonstrable but the available evidence suggests strongly that the play was set direct from a manuscript in Shakespeare's autograph. J. Dover Wilson's evidence in the New Cambridge edition (pp. 103 ff.) for transcription is as faulty as his evidence for a late revision of an early play by another hand, and it can be safely disregarded. In every respect the textual difficulties and anomalies are consistent with those to be expected from compositors tackling a manuscript that could be called 'working papers.' The manuscript contained at least one revised passage, several additions, and very likely a quantity of verbal alteration whether during or after inscription. All evidence that can be reconstructed for non-prompt theatrical transcripts such as lie behind the Folio texts of Julius Caesar and Twelfth Night, for example, indicates that they were designed to normalize just such irregularities as the highly variable forms of speech-prefixes that are preserved in All's Well.[4] That anything approaching 'Shakespearian spellings' are hidden by the overlaid texture of the three Folio compositors is not surprising as late as 1623. It is easier to believe in the concealment of Shakespeare's orthographical

characteristics (whatever they were) by compositorial styling of an old-fashioned manuscript than to posit a highly unusual form of literal non-theatrical transcript made for an unknown purpose and utilized by the theatrical editors as printer's copy for the Folio. Later opinion as illustrated by G. K. Hunter, for example (New Arden edition [1959], pp. xi-xiv), takes it—correctly in the opinion of the present writer—that Shakespeare's own foul papers served Jaggard as the printer's copy.

In addition to the now discarded hypothesis of a scribal transcript, the shadow of a book-keeper's markings interfering with the text, possibly in a serious manner, has hung over this play. Sir Edmund Chambers thought that "the book-keeper has possibly added the letters G. and E. to the 1. and 2. by which the author discriminated the brothers Dumain, who are indifferently described . . . as Lords, Captains, or Frenchmen, and are apparently also the Gentlemen of iii.2. The Letters may indicate the names of actors; Gough and Ecclestone have been guessed at."[5] Again, G. K. Hunter analyzing the evidence points out the weaknesses of the hypothesis for intervention and argues plausibly that the initials are Shakespeare's own means of identification.[6] The old-fashioned idea that a book-keeper would annotate Shakespeare's foul papers and cut them in preparation for direct transcription into a prompt-book is inherently improbable for plays of this, or perhaps any date, instead of working over a clean intermediate transcript made for the purpose; moreover, the evidence for cutting and other such preparation is weak in the extreme. However, it seems likely that the manuscript received a minimal 'editing' for publication. If Jaggard's copy was indeed the working papers, as seems to be assured, it is not certain that the act division would have been written in by Shakespeare. Since a compositorial division of the acts, given the order of printing the pages, is quite impossible to argue for, the implication follows that whoever prepared the

copy for Jaggard probably inserted the notation, at least for Acts II-V. If so, it is theoretically possible that the theatrical directions for flourishes at the King's entries and exits were also additions, and just possibly the directions for a tucket in II.v and alarums in IV.i as well, although there is nothing inherently improbable in attributing them all to Shakespeare. To sum up, one can dismiss with considerable confidence suggestions that the text itself has been interfered with in any significant manner by an outside agent, barring always the normal compositorial sophistication and corruption that accompanies the transmission of an early seventeenth-century text from manuscript to print, exacerbated in this play to some extent by more than usual difficulties with the handwriting but relieved of the secret and undetectable departures from copy of a scribal intermediary.

As indicative of working papers, the stage-directions betray the author's hand not only in their occasional descriptive cast, despite their generally businesslike nature, but also in their failure always to agree with the text. Description is present in the much-discussed 'memorandum' in II.iii Parolles and Lafew stay behind, commenting of this wedding (1089-90)[7] and in such directions as II.i (594-596) Enter the King with divers yong Lords, taking leave for the Florentine warre . . ., the odd III.vi (1730-31) Enter Count Rossillion and the Frenchmen, as at first,[8] or IV.ii (2016-17) Enter Bertram and the Maide called Diana.[9] Some directions are precise about minor persons, such as V.i (2592-93) Enter Hellen, Widdow, and Diana, with two Attendants, but some are vague about characters who would need to be specified in the prompt-book, as

the divers Attendants in I.ii and the divers young Lords in II.i, both of which include the two speaking parts of Lords G and E. The number of attendants may be permissively given: IV.i (1911-12) Enter one of the Frenchmen, with five or sixe other souldiers in ambush, a direction that also fails to specify Lord E as the Frenchman. Only two of the five or sixe souldiers have speaking parts. A similar direction occurs at IV.iii (2106-7) Enter the two French Captaines, and some two or three Souldiours.

Various directions do not agree with the text and would need to be set right in any preparation of the prompt-book. In II.iii (944) the direction reads Enter 3 or 4 Lords, but the following dialogue requires four lords for Helena to pass over before she chooses Bertram. According to a possible interpretation of Helena's 'Where are my other men?' in II.iv (1366), the Clown or some other attendant should have been designated as entering with her at 1325 although she alone is mentioned.[10] The familiar difficulty of the direction in III.i (1446-47) Enter Hellen and two Gentlemen when these are in facts Lords G and E (assigned the speech-prefixes French G. and French E.) needs no remark here. In IV.iii (2181) the text reads Enter a Messenger, but this is Bertram's servant who had accompanied him from Paris to Florence (1491-92, and the prefix Ser. at 2183) and he has no function as a messenger. In V.iii (2876) Parolles is listed in the entrance direction with the Widow and Diana but he does not in fact appear until his properly noted entrance at 2960. At least sixteen exits are not marked, only four of these at the ends of scenes where the omission is of no great concern. No directions are present for

the staging of II.i in which the King seems to withdraw apart at 621 and come forward again at 661. These are all matters that indicate working papers, not a text in any way prepared for acting.

Various internal inconsistencies exist that might or might not have been adjusted on review. The time scheme by no means permits Helena to have arrived in Florence on her way to Saint Jacques, to have performed the bed trick that very night, but on the same night to be reported dead at Saint Jacques as testified by the Rector's report. Lords G and E arrived in Florence in III.i and in III.ii appear in Rousillon on their way back to Florence from a visit to Paris. In Rousillon they receive a letter from the Countess to Bertram. Inexplicably the letter is not delivered when they meet Bertram in Florence and plan the exposure of Parolles; and this letter is in fact not mentioned as delivered until IV.iii (2108-11) and then on some vague occasion difficult to reconstruct. It is at least arguable that the gentle Astringer in V.i (2601) to whom Helena delivers a letter to the King disappears from the plot thereafter, for it is reasonably clear that the Gentleman in V.iii who delivers Diana's petition to the King at 2842 is not he, and this petition seems to have taken the place of Helena's letter which so far as we know is never received by the King. In V.iii (2924-25) the King speaks to Diana, 'Me thought you saide | You saw one heere in Court could witnesse it' (her intimacy with Bertram), to which Diana responds, 'I did my Lord'; but she has said nothing of the sort.[11] Whether or not these anomalies would have been worked over before production—and they are no more than the normal Shakespearean carelessnesses also observable in plays in copy more highly developed for the theater—the directions would have needed sharpening, correction, and addition for acting, and the faulty speech-prefixes at 615, 2190, 2233, and between 2227 and 2384 (perhaps even to 2417 in the manuscript) would have required correction. And Helen's broken speech at 169 might have been observed and a transition introduced.

The variable speech-prefixes constitute the chief evidence, perhaps, for the belief that All's Well that Ends Well was set from Shakespeare's working papers. Bertram's prefixes may be forms of Rossillion or of Bertram, and once he is Count. The Countess may be Mother, Countess, Old Countess, Lady, or Old Lady. Lafeu is generally Lafew but in II.i he starts as Lord Lafew though then changing to Lafew; and in III.iii he is Old Lafew, once Old Lord (perhaps a misreading of an abbreviation for Old Lafew), but before the end of the scene he has reverted to Lafew.

The two French Lords lead a varied existence: they may be Lord G and Lord E, 1. Lord G and 2. Lord E, French G and French E, as well as Captain G and Captain E. At 1466 1. G. is more likely an artifact prefix referring back to the stage-direction identifying them as two Gentlemen despite the fact that before and after 1466 they are prefixed as French G and French E and so 1. G. could just possibly have been a misprint for Fr. G. At 2233 Cap. is intended for Cap. G in terms of the erroneous reversed prefixes that had started shortly before.

McKerrow suggested that in Romeo and Juliet a case could be made for some variation in prefixes to derive from the function of a character in a given scene: for example, Lady Capulet could be Mother in relation to Juliet, Wife in relation to her husband, and Lady Capulet as mistress of her household.[12] But in All's Well Shakespeare's variety of nomenclature both within stage-directions and speech-prefixes seems to have resulted more from writing the scenes at different times than from any immediate association of the descriptive prefix with the function of the character. It is true, however, that the First Soldier's prefix changes within IV.1 to Interpreter as he assumes his new function and that in III.vi, IV.i, IV.iii (but not in III.i) the two French Lords in their military capacity become Captain G and Captain E though reverting to Lords in V.ii back at court. It is also true that in I.i the Countess for the only time is Mother in direction and in prefixes, reflecting her chief function in the scene in relation to Bertram. But later there is nothing in the scenes in which she appears as Lady or as Old Lady to warrant any functional difference from the usual Countesse; nor indeed do any of Lafeu's functions account for his variations between Lafew, Lord Lafew, and Old Lafew. The best case, then, is for the Countess as Mother in the opening scene. What might seem to be an even stronger case for the shift of the Frenchmen to Captain is made slightly uncertain, perhaps, by some question whether originally Shakespeare had intended these military characters to be the two Lords.

Any study of the compositorial treatment of speech-prefixes with a view to determining what specific forms, or variations, of prefixes were present in the manuscript copy, and what were compositorial and without authority, must take account of two somewhat associated guidelines. The first is that one may assume (other evidence wanting) that a dramatist wrote his working papers in a generally seriatim order of scenes, or if the manuscript were not holograph that a scribe would certainly copy the sheets in order. Thus authorial forms and variations in prefixes should be roughly chronological in their positions. However, when the

Folio compositors came to work on a play, owing to their habit of setting a quire from the inside out, the order of composing the Folio pages might differ widely from the seriatim order of the manuscript sheets. It follows that the compositorial treatment of the prefixes must be traced by following the order of typesetting, not that of authorial progression. The second caveat for an inquiry is that some attention must be paid (in connection with the order of setting) to the possibility that (a) the form of a character's name in a preceding stage-direction might influence a compositor to desert the form of his manuscript copy, and (b) eventual familiarity with a character could lead a compositor to adopt a standard form of prefix without regard for what he found in his copy.

That compositors (including Compositor B) could be influenced by copy even in minor matters of the abbreviation of prefixes was demonstrated some years ago by Professor Brents Stirling, who most ingeniously identified not only the revised section of the duplicate announcement of Portia's death in Julius Caesar IV.iii but also a revised section, or addition, in the conspirators' first conference in II.i with Brutus.[13] His evidence was exact. Throughout the play both Compositors A and B invariably held to the expanded prefix Cassi. for Cassius. But in these two sections the form changes to Cass. for Compositor A and Cas. for Compositor B. The inevitable inference is that the standard form of prefix in the underlying scribal manuscript had been carefully chosen as Cassi. in order to distinguish him from Cœs. (readily confused with Cas.) and from Cask. for Casca, but that the revised parts of the manuscript in a different hand more carelessly read Cas. or Cass.[14] and that this shorter form was faithfully followed by the two compositors.

The bibliographical evidence based on the order of typesetting is the same as that which proves useful in All's Well. The first page of Julius Caesar that Compositor A set was sig. kk3v in which at TLN 711 Cassius by name heads a group entrance and, speaking immediately at

713, is given the prefix Cass., this proving to be the revised section of the manuscript. Compositor A then went back one page in the cast-off copy and set kk3r, the end of Act I, in which he encountered a character not named in any direction and hence unidentifiable but with a copy-prefix that he set as Cassi. from the regular manuscript. That Compositor A followed copy in this distinction is clear; that he did so because under the circumstances he could not know whether this character in I.iii was the same as that on the page of II.i he had set just before, prefixed Cass., is certainly probable.

Compositor B worked on sig. kk4r while A was setting kk3v, this also being B's first introduction to the play. No entrance for Cassius is on this page; hence B must have followed copy Cas. or Cass. when he set Cass. as the prefix in text that continued II.i after A's catchword on kk3v. Hence in copy taken to be in the same revised form as A's, B set the same shorter form of the prefix. When in his order of setting B next encountered Cassius it was in III.i on kk5v where Cassius is part of a group entrance opening the scene and speaks first a dozen lines later, at 1215, where he is Cassi. and so continues until the second revision in the play in IV.iii. In this latter scene the explanation of strangeness, as in II.i, will not work, for B set sig. ll3v (within IV.iii) with the prefix Cassi., continued with the same tag on ll4r (which also began V.i), and on ll4v. He next set ll3r. In the first column the regular prefix is Cassi, but after the entrance of the Poet (2108) a few lines down in column b the first prefix in the column shifts to Cas. (Cass. once at 2159), this text containing the revised version of the announcement of Portia's death. When on ll3v (set earlier) the regular form of the manuscript returns, containing the undeleted second announcement, the prefix is the usual Cassi. always set from this manuscript. Between the two columns on ll3r no question could have arisen in B's mind that he was dealing with a different character from the familiar Cassius. Hence at this point B must have followed copy in the shorter form of the prefix, influenced by the variant form in a part of the manuscript written in a different hand even though this meant his abandoning the established Cassi. prefix. On kk3v-4r the fidelity in both compositors is readily explained by the strangeness of the character and the copy-prefix at the start of typesetting, but not on ll3r apparently.

This example illustrates that Compositor B—at least on the evidence of Julius Caesar—may be studied in All's Well with some general confidence that he would not be likely to go against copy except perhaps in mechanical matters of abbreviation to justify a line or, possibly, in the establishment of a favorite short form of the same prefix instead of a longer form.

Evidence for compositorial fidelity to copy may be found, in fact, on

the various occasions when the speech-prefixes of All's Well exhibit independence from the forms found in the preceding stage-directions. In II.i the entrance direction reads Enter Lafew but the prefix in the very next line (662) is L. Laf. and again two lines later (664). When Compositor C at the head of the next column then shifts to the formula Laf., which he maintains for the rest of the scene, the normal conjecture is not that his copy changed (although this is quite possible) but that he preferred the short form of the prefix when no ambiguity was present.[15] However, the essential point here is that his initially set prefixes did not agree with the immediately preceding direction. Similarly, at II.ii the direction is Enter Countesse and Clowne but in the first speech Compositor C set the prefix Lady which he followed six more times plus the catchword on V4v. When after setting three other Folio pages Compositor B came to set V5, he also chose the prefix La., presumably under the influence of the copy catchword marked as such in the manuscript.[16] The direction heading II.iii on V5 reads Enter Count, Lafew, and Parolles but B set the opening speech with the tag Ol. Laf., a form demonstrably from copy because of the error Ol. Lord at 997 heading V5v that follows although on a page that had been set earlier.[17] In III.iii the opening direction reads Enter . . . Rossillion but in the fourth line Bertram's prefixes begin with Ber. (1544). So also at III.iv the direction reads Countesse but the prefixes are La., and in III.vi after the direction Count Rossillion the prefix is Ber. In IV.iii a direction reads, in error, Enter a Messenger (2181) but the single speech of this character is correctly prefixed Ser. for Servant (2183).

On at least one observable occasion a compositor (in this case B) was influenced, seemingly, by the preliminary stage-direction to go against what was probably the copy-prefix. In III.ii (sig. X1) the opening direction reads Enter Countesse and Clowne and the initial prefix (1402) is Count., followed by a second Count. (1406); but beginning with 1422 she is Lad. and so continues in the shortened form La. for the rest of the scene. Because Lady as prefix had already appeared (following Countesse

in the direction) in II.ii set by C and continued by B, and was to re-occur in III.iv and IV.v, it may seem possible to take it that Compositor B started by drawing on the direction form as his prefix but shortly reverted to what he saw in his copy.[18] However, Shakespeare's own variation cannot, of course, be completely discarded although as a lesser possibility.

The possibility must be examined, also, whether in V.iii the opening direction could have influenced the following speech-prefixes. In this direction on sig. X6 the Countess is old Lady (2695), the same identification that had appeared in a direction in IV.v at 2481 on sig. X5, followed by the speech-prefixes La. On X6 in V.iii after the direction at 2695 she speaks soon after at 2701 and is given the prefix Old La. On X6v at 2799 (no further direction intervening) and again at 2872 the same prefix appears. At first sight one could take it that the distance is short enough so that the prefix at 2701 could have influenced that at 2799 and that, in turn, have affected 2872. However, between X6r and X6v Compositor B had withdrawn from the Folio to set some amount of text in another book, the extent of which is unknown. It is only natural to believe that B would probably not have kept in mind during this interval what in fact was a unique prefix for the Countess. Obviously, it is simpler to conjecture that the reason why B set Old La. when he returned to X6v after an interval between it and X6r was that he was following copy. No exception could be taken to this conclusion were it not that the Countess' only remaining speech, which occurs at 2919 in sig. Y1, has the prefix Coun., which must almost certainly be a copy-form.[19] Whether this variant is an authorial aberration or whether Coun. had been the copy-prefix earlier in the scene on X6 and X6v, changed by Compositor B, is a question not to be tackled in isolation, for another anomaly in speech-prefixes

is present on Y1 in connection with Bertram, and the two may be related.

Whether by accident or design, despite the widely varying identifications of Bertram in the stage-directions as Count Rossillion, Rossillion, Count, and Bertram, Compositor B throughout quire X (1375-2913) with but one exception invariably assigns him the prefix Bertram, usually in the form Ber. This decision, if decision it were,[20] might even have been made somewhat earlier, since the common prefix begins, in fact, in II.v (1269-1370) on B's sig. V6v, where the direction is also Bertram. The history of this prefix is interesting. Printing of the play started with the two simultaneously set pages V3v (by Compositor D) and V4 (by Compositor C). Bertram does not appear on sig. V3v except in the stage-direction for II.i where he is confused by D as two characters, 'Count, Rosse', an error that indicates clearly the form of the abbreviation in the manuscript. Compositor C set twenty-three lines of type on V4 before he came to a speech by Bertram, which he prefixed as Rossill. twice (626, 630) before reducing it to Ross. (649) with one short Ros. in a tight line preceding it (637). Given the form of the manuscript abbreviation Rosse in the direction as interpreted by D when he subsequently came to complete his page V3v, one may readily conjecture that C's initial two settings of Rossill. could not have derived from the direction (if he had consulted it on the same sheet of manuscript) and hence that they reproduce copy.

Bertram does not appear in the two pages of the outer forme set by C and D. When after a delay Compositor B took over the exclusive setting of the play with V2:5v, he started with V2 but encountered Bertram for the first time on V5v where there was no stage-direction to give him a clue. Here the prefixes in the dialogue with the King are Ber. It would seem most probable that in these prefixes B was following copy, for he had no other hint as to Bertram's identity except within the text, and that he would change the Ross. prefix in the manuscript (if it had been so written) when he started Bertram's prefixes at 1005 because in 1003 the King had called him Bertram is scarcely credible. Only after completing V5v did B turn to V2v where for the first time in B's experience Bertram is so identified in a stage-direction (as Bertram at 262) and is given the prefixes Ber. again. Since Ber. had already occurred without identification on V5v, there is no reason to take it that B was doing any-thing but following copy on V2v. In the opening direction for II.iii on the next page V5 the form is Enter Count but the single prefix for Bertram on this page is Ros. (901), again presumably from copy since at this

time B had set no text identifying Bertram as Rossillion. When B then carries on with V1v:6 (1-92, 1124-1249), he sets Ros. throughout I.i on V1v; and so on V6 concluding II.iii when Bertram makes his re-entrance as Count Rossillion in the direction (1171) he is Ros. in the succeeding prefixes (1174-1202). When B then begins Ber. as an invariable prefix in II.v on V6v (1272) after the direction Enter Lafew and Bertram, we have no means of knowing whether the manuscript prefixes continued mixed between Ross. and Ber. thereafter or whether it was Shakespeare, not B, who switched over to Ber.

Given this history it is difficult to explain the conflicting evidence in sigs. Y1 and Y1v (2914-3078) set in seriatim order in formes with pages of Twelfth Night after a considerable delay during which B had set various quires in the Histories. When he came to these pages B would long since have forgotten whatever forms of prefixes he had decided on earlier in the play. The text of Y1 starting with 2914 continues V.iii in mid-speech. On X6v, as throughout quire X, the sole prefix form had been Ber. with the exception of an aberrant Count. at 2412 on X5 in IV.iii.[21] However, in this continuous text abruptly on Y1 at 2930, followed at 3046 on Y1v, the prefixes change to Ros. Since Ros. had been one of the two prefixes for Bertram in quire V, the normal inference would be that starting in quire V and continuing in quire X Compositor B had got in the habit of setting Ber. for Bertram regardless of the copy-prefix, which was perhaps a mixture of Ber. and Ross. according to scene. General evidence suggests that when a compositor started to set a new play on $3v:4, usually well within the text, he was likely to follow the copy-prefixes at least until the characters were established in his mind. If B had behaved in the same manner, on returning as something of a stranger to complete the play after the lapse of several weeks, he also would have been likely to follow copy. This appears to be the only reasonable explanation for the change in prefix from Ber. to Ros. in continuous text but separated into two distinct bibliographical units.[22]

The case would be comparatively simple, therefore, were it not for another change in prefix. In this scene V.iii on X6 and X6v the few speeches by the Countess had been prefixed by Old La., a descriptive

phrase first found in the stage-direction in IV.v on sig. X5 (2481) but followed in the scene by the prefixes La., and then again in the direction in V.iii (2695) on X6 followed by the prefixes Old La. on X6-6v. However, her only speech in the next quire comes at 2919 on Y1, where she is Coun., almost certainly a copy form not only because of its abrupt change but also because B had usually set this prefix as Cou. Reasons have been given above for the conjecture that Old La. was the form of the manuscript prefixes in V.iii on X6 and (after a delay in setting) on X6v. The difficulty, then, is whether Coun. on Y1 is an authorial variation without significance or whether it aligns itself bibliographically with the shift from Ber. to Ros. [23] in the same pages, even though the corollary would follow that Old La. was in fact not the copy-prefix but that the stage-direction at 2695 had affected B's treatment of copy Coun. [24]

In such a dilemma no certainty is possible since all evidence is wanting and only speculation can take its place. The one hard fact is that the prefixes on Y1 must be copy-forms since no consultation seems to have taken place with previously printed sheets. This being so, there are difficulties either in assuming that the shift to Ros. is significant but that to Coun. is not, or that both are significant. One possibility suggests itself which as a speculation would resolve the problem. We know from other evidence that when Shakespeare revised the text, as in I.iii between the entrance of Helena at 450 and the end of the discussion with the Countess about her status as daughter at 512, he could adopt different prefixes: the Countess is Cou. in the original scene but Old Cou. in the revision, for example. Correspondingly, when he left off writing a scene and returned to it at a later time, he might also adopt different prefixes, as conjectured to be the explanation for the shift in prefix from Old Laf. in II.iii to Laf. after the curious memorandum to himself that Shakespeare seems to have written, Parolles and Lafew stay behind, commenting of this wedding (1089-90), marking a break in the writing of the scene.[25] If such a break, or revision, appeared in V.iii, we should have a

reason for the shift in speech-prefixes between X6v and Y1, the difference in the manuscript being preserved by the bibliographical fact of delayed setting of Y1 so that Compositor B was forced to follow copy instead of adopting whatever standard forms he might previously have utilized, as perhaps with Ber.. for a mixture in his copy of Ber. and Ross. On X6v Ber. as a prefix exists down to the next to last speech on the page (2910), to be replaced at 2930 on Y1 by Ros. However, if we conjecture that the Ber. form is not significant in quire X since B had become used to normalizing any Ross. prefixes in the copy, it is not to Ber. but to the unique Old La. prefix that we must look for possible evidence. At 2869 Bertram is brought back to face the charges in Diana's letter, with the direction Enter Bertram and after the remaining two lines of the King's speech the Countess, with the prefix Old La. calls for justice on those who have harmed Helena (2871). The King in 2873-75 addresses Bertram, breaking off with the demand, 'What woman's that?', followed at 2876 by the direction Enter Widdow, Diana, and Parrolles. This entrance introducing Diana's plea to the King marks the last visible point where Shakespeare could have stopped writing the scene at some natural break in the action and dialogue before the new set of speech-prefixes begins on Y1 in the midst of the give and take of the following speeches. Whether or not it is evidence of a break, this entrance direction at 2876 is in error, for Parolles does not come on the stage with Diana and her mother but instead must be sent for and makes his real entrance only at 2960 on Y1. If in the interval of stopping work before resuming at some later time Shakespeare had jotted the direction, intending Parolles to enter with the two women, but changed his mind when he resumed writing,[26] the error could mark some disruption that would support the speculation that the change to Ros. as a prefix for Bertram marked a manuscript shift at the direction for the continued scene that has been obscured by Compositor B continuing at 2886, 2889, 2899, and 2910 on X6v after Diana's entrance at 2876 to set the pre-entrance prefix Ber. to the end of the page and his stint, even though in the copy it had shifted to Ross. Conjecturally, therefore, Shakespeare after the entrance at 2876 changed to the prefixes Ross. and Coun. as found on Y1 but the evidence does not reveal itself until the end of the X6v page and the return of B on Y1 after a period of working on the Histories. If this conjecture is at all plausible, we should have an answer to the anomalous prefixes in continuous dialogue between X6v and Y1 both for the Countess and for Bertram. We should also have evidence of what has earlier seemed suspicious, the uniformity in quire X of B's use of the Ber. prefix when in quire V good evidence exists that Shakespeare varied between Ross. and Ber.

In the order of setting, Lafeu is first encountered on V4, set by Compositor C, where he is L. Laf. twice at the foot of the first column (662, 664) after an entrance as Lafew, but he shifts to Laf. in the second column, more likely a compositorial than an authorial shortening, perhaps. He next appears, in the order of setting, on B's V5v where at 997 he is Ol. Lord. He does not speak again until 1091 after he has been named Lafew in the memorandum stage-direction at 1089-90, and then he is Laf. for the rest of the page and in all text thereafter. However, as set, after the switch on V5v from Ol. Laf. (Ol. Lord is apparently a misreading by B on his first setting of a Lafeu prefix) to Laf., B went on to V2v where at 262 Lafeu enters as Lafew but is mute. Sig. V5 follows, in which after the Lafew of the entrance at 892 the first speech at 893 is prefixed Ol. Laf., this form continuing on the page seventeen more times, the last at 991. Thus when, in the order of the manuscript although not of the setting, Ol. Laf. at 991 is followed by Ol. Lord at 997 on the next page, there can be no doubt of the copy-form in this scene between 893 and 997, especially since Ol. Laf. differs from the immediately preceding direction at 892.[27] However, when he shifts to Laf. later in the scene, signs exist that the difference was present in the copy and that the change at that point is not a case of arbitrary compositorial shortening. Some evidence is present that Shakespeare himself often used the shorter Laf. form as well as Ol. Laf. For instance, on V1v where B is cautiously following the copy-prefixes on the first page of the play, Lafeu is Laf. (11 ff.) although Lord Lafew in the opening direction. Moreover, the fact that, as in quire X, he remains Laf. in the bibliographically discrete pages Y1 and Y1v (set after a considerable delay from X) also suggests the copy-form there, especially

in connection with the evidence from the prefixes for Bertram and the Countess. Finally, on V6v the change from Ol. Lord (995) as an indication that copy read Ol. Laf., to simple Laf. occurs in the continuation of II.iii at 1091 where it is a reasonable conjecture that Shakespeare resumed work on the play only after an interval indicated by the memorandum direction. Hence although B's tendency to shorten prefixes could have operated when Old Laf. alters to invariable Laf. at 1091 for the rest of the play, the general trend of the evidence suggests that it was Shakespeare more than B who was responsible for this change.

This is not the place to go into the intricacies of the prefixes for the two French Lords.[28] That the major distinctions in their forms such as 1. Lord G, Lord G, French G, and Captain G are authorial seems practically demonstrable since the stage-directions could have exercised small influence, usually identifying them merely as Frenchmen. It is at least arguable that Shakespeare's first intention was to distinguish them merely as 1. Lord and 2. Lord, for on their introduction in I.ii they are not identified among the divers Attendants (239) although speaking in the scene, nor are they for the second time identified among the divers yong Lords who are taking leave of the King in II.i (594), although again they are speakers. In I.ii on V2v set by B they are 1. Lo. G and 2. Lo. E except for the slip L. 2. E at 315. In II.i the first and only prefix set by Compositor D, on sig. V3v, is Lord G for G's first speech. Simultaneously, but setting from a sheet of the manuscript that very likely contained D's first prefix, Compositor C was continuing the scene with V4; in fact, his prefixes would have been set before D came to the start of II.i and G's first speech. Compositor C first set L. G but then 1 Lo. G and 2. Lo. E until in G's last speech (646) he is Lo. G. Both compositors, then, omitted the numerical designation in the initial prefix of their setting under circumstances that should indicate they were following copy both with and without the number, just as B must be assumed to have followed the manuscript in I.ii, set later. The numeral form with initial appears only once more, in the prefix I. Lord E in the first speech of IV.i (1913), uninfluenced by the immediately preceding stage-direction Enter one of the Frenchmen, although in the remainder of the scene he is Lo. E. Whatever the initials G and E signify (names, like George and Edward?), it is at least arguable that they were substitutes invented during the writing of the play for characters who had originally been denominated simply as 1. Lord and 2. Lord, the initials being added irregularly in the early scenes with or without displacement of the numerals. The 1. Lord example

in III.i (1378) may be a prefix unaltered by oversight, for afterward in the scene he is Fren. G and E is Fren. E. In III.ii (1466) the anomalous prefix 1. G is less likely to be a misreading of Fr. G, coming as it does after the setting of three prefixes for French E and French G, than a correlation with the opening stage-direction identifying the characters as two Gentlemen (1446) and thus it is perhaps another artifact.

There are three occurrences when the compositors copy what appears to be a manuscript repetition of a prefix within the same speech. The first of these is particularly significant because a change in the form of the repeated prefix appears to mark the beginning of a Shakespearean addition which accounts for the repetition. This occurs in I.iii at TLN 451 where after Helena's entrance at 450 the Countess' speech preceding the entrance continues without interruption but with a second prefix, changed in its form to Old Cou. from Cou. at 443, the start of the speech. This marks an addition from 451 (or 450) to 512 before the original text resumes with the prefixes Cou. at 513.[29]

However, the second and third examples appear to be related and they are troublesome. In II.ii the Clown has boasted that he has an answer to fit any question, and he begs the Countess to test him: 'Aske mee if I am a Courtier, it shall doe you no harme to learne' (859-860). To this the Countess returns, in a speech prefixed Lady., 'To be young againe if we could: I will bee a foole in question, hoping to bee the wiser by your answer.' With this speech Compositor C concludes sig. V4v and his stint in the play. The last word of the speech is divided 'an-|swer.' so that only the second syllable forms the last text-line of the column; then in the line below—the direction-line—C set the speech-prefix Lady. as the catchword. When after a delay in work on the Folio Compositor B took over the typesetting, he set sigs. V2, V5v, V2v, and only then came to V5 and the continuation of C's text. The first line on V5 is 'La. I pray you sir, are you a Courtier?' The temptation exists to suggest some bibliographical

explanation for this irregularity in what is continuous speech, but one is hard to find.[30] The coincidence of the repeated speech-prefix with the end of a Folio page and a change in compositors seems fortuitous, and there would appear to be every reason to take it that C was faithfully following what he thought to be the intent of his copy, which had two consecutive speeches by the Countess written in order, each prefixed by her tag. This also seems to have been B's understanding when he picked up the copy some days later on sig. V5. Unless one were to speculate that C had marked his manuscript wrongly after instead of before a now lost rejoinder by the Clown, and that B perforce followed the mistaken marking, the catchword ought to indicate that no undeleted rejoinder was present in the manuscript and that the two speeches by the Countess did indeed run consecutively in the final version. Whether revision and the deletion of intervening material caused the repetition of the prefix is an open question.

The conjecture that nothing is missing here owing to faulty compositorial marking of copy at the page's end may be strengthened by the parallel case that occurs in II.iv on V6. Here Parolles has been worsted by the Clown in an exchange, and with affected tolerance remarks, 'Go too, thou art a wittie foole, I have found thee.' To this the Clown has two consecutive replies (1243-47) each prefixed by Clo. In the first he enquires, 'Did you finde me in your selfe sir, or were you taught to finde me?' In the second, immediately succeeding and with no intervening reply from Parolles, he adds (with repeated prefix), 'The search sir was profitable, and much Foole may you find in you, even to the worlds pleasure, and the encrease of laughter.' Editorial opinion has tended to conjecture a lost line or so by Parolles between the Clown's two speeches,[31]

but the example of the Countess' two speeches does not encourage the hypothesis. The better proposition would seem to be to conjecture that both for the Countess and the Clown we have, instead, a marginal addition with prefix. Whether, instead, the Clown's second speech was intended to replace the first is barely possible but not probable. However, there would appear to be no question of any intended substitution in the Countess' two speeches and this fact may weigh in considering the Clown's two consecutive speeches, especially when their context is considered.

The conclusions that may be drawn from this study would have more general interest if further investigation were to show that the characteristics of copy as reflected in the typesetting of this play could be paralleled elsewhere in the Folio, including other work by Compositor B. It is true that the copy for All's Well—if it were indeed Shakespeare's own foul papers as seems almost certain—is not of a kind commonly found in the Folio, for evidence is accumulating that Sir Walter Greg's frequent assignment of foul papers as Jaggard's copy has been too liberal. Indeed, insofar as the evidence of this text bears on other Folio plays, it is a natural inference that uniformity of speech-prefixes, especially, is more likely to indicate the smoothing effects of a scribe transcribing Shakespeare's papers than it does any substantial interference of the compositors with variable authorial copy, if that was what they were setting. The marked differences between the varied prefixes of All's Well and the uniform ones of such a play as Julius Caesar can represent only the distinction between authorial working papers and a scribal transcript.[32]

If in All's Well the evidence has been correctly interpreted for what the compositors did with their copy, the strongest indication is present that the compositors were completely conservative in the treatment of names and titles in the stage-directions and that copy was followed in these respects with fidelity. Similarly, in the speech-prefixes there is every

indication that the variation in the forms of names and titles is Shakespeare's own and present in the manuscript, including even some variant spelling of the names. It is important to establish that these variables in the manuscript copy were not suppressed and the names or titles normalized on their earlier appearances by Compositors B, C, and D.

It is true that two levelling influences may occasionally be detected during the course of typesetting. The first is for the compositor to feel a slight tug toward setting prefixes in the form of the name or title in the immediately preceding stage-direction. In III.ii where the direction reads Countesse Compositor B, not Shakespeare, may have been responsible for the two immediately following prefixes of Count. before reverting to the copy's La. There are indications that in some part under the influence of copy, in some part affected by stage-directions, B came to set the prefix Ber. uniformly although the copy might have been a mixture of Ross. and Ber. Certainly on sig. X6v in V.iii this seems to be the best explanation for the appearance of Ber. as prefix at the foot of the second column although on the evidence of the continuation of the scene on sig. Y1 Shakespeare was likely to have written the prefixes as Ross. or Ros. after the entrance on X6v of the Widow and Diana. Nevertheless, Compositor B appears to have followed with fidelity the divagations of the prefixes for Lords G and E, and his interference with variants in more readily apprehensible characters is much less than might have been expected from a workman who on his record was likely in other respects to deal with his copy somewhat too firmly. Statistics about the incidence of his wanderings from copy in the text cannot be applied to directions and prefixes.

The second tendency associated with this minor drift toward uniformity is a general compositorial preference for short forms of prefixes over longer ones when some detail in the long one is not necessary for identification. For instance, in II.i after Enter Lafew Compositor C twice set L. Laf. before going over to the invariable short Laf. The shorter the prefix the fewer ems a compositor needs to set; but perhaps as important in setting dramatic verse, the fewer problems he may make for himself in justifying long lines. Yet the history of the number of times that long forms like 1. Lord G were set, or Ol. Laf., in All's Well indicates that author as well as compositor may be responsible for many shifts from long to short forms. Although in the order of typesetting B had composed II.iii with its Ol. Laf. prefixes after a Lafew direction, it seems probable that when later he set I.i in which the prefix is Laf. after the direction Lord Lafew, he could as readily have been following copy as not, given the care he seems to have shown elsewhere in the forms of this first scene.

Moreover, when in I.iii Compositor D in continuous text followed copy from Cou. to Old Cou. and back again to Cou. after the revision, he resisted any impulse to forsake his copy in favor of uniformity. The evidence of All's Well, thus, appears to indicate that however much compositors may have styled text to suit their own spelling and punctuation habits,[33] the movement toward normalizing the dramatic appurtenances such as stage-directions and speech-prefixes was slight. As a result, evidence about the nature of the underlying Folio copy based on the characteristics of directions and prefixes can be evaluated with some confidence as representative of the copy, not of the compositors' predilections.

Notes