University of Virginia Library

II

Some years ago Professor J. Dover Wilson edited the New Cambridge MND on the theory that an original of 1592 was revised in 1594 and again in 1598, perhaps then for the marriage of the Earl of Southampton to Elizabeth Vernon.[12] The copy for Q1, he believed, was the prompt book, which was written in Shakespeare's autograph and which contained manuscript pages from all three periods of composition.

Wilson's theory has not won acceptance in its entirety; Sir Walter Greg, for example, found difficulty with Wilson's notion of multi-level revisions in which there are perceptible (and dateable) stylistic differences and suggested that the Q1 copy was probably foul papers.[13] If so, however, there could easily have been revised passages present, even though they may not have been composed in accordance with Wilson's scheme. As Greg said, Greg does not comment on Wilson's contention that other parts of the text were also revised or added to, but if the copy was foul papers with occasional marginal additions, we could explain the compositor's difficulties as resulting from his being slowed down by patches of difficult material. If the quarto is in part mislined, there is reason to think that the copy may have been similarly mislined, and an important part of Wilson's argument for revision is based on inference from quarto back to MS mislineation. Yet, as the quarto was almost completely set by formes, the compositor may sometimes have tampered with the MS lineation, deliberately compressing or expanding the text in order to make it conform with the space limits established by casting-off. Since we have some idea of how each sheet was set, however, we should be able to tell to what extent the compositor was likely to alter MS lineation as he dealt with mechanical problems created by his method of composing.

In B(o), the first forme composed, there are two minor instances of mislineation. The first occurs in the first eight lines of the fairy's speech (B3,14-17;II.i.2-9), which appear in the quarto as follows (correct verse line endings are indicated by a stroke):

The second is found on the last page of the forme (B4v,20-21; II.i.115-116):

About the first of these one cannot be positive, but it looks as though the lines probably stood in the MS as they do in the quarto, the quatrain being lined as a couplet and the couplets as single lines of verse. If this is true, the couplets of which the rest of the speech is composed, so lined in the quarto, probably were written in single lines in the MS, but were divided by the compositor who, having had to turn over the last word of B3,17, decided that the MS lines were going to be too long for his measure. The compositor could, of course, do about as he chose with this material since he was working on the first forme of the sheet. In the second case, the mislineation may have resulted from the compositor's carelessness; but later in the play (at F3,26; G1,22; G3v,10; and G3v,18) we find other examples of very much the same thing, which just possibly suggests that the MS rather than the workman was at fault.

When B(o) was completed, the limits of B1v,2 and B3v,4 had been established; thus we have more reason to think that the compositor might juggle the text of the inner forme to make it get in these limits. Almost certainly he was doing just this when he set a short speech of Bottom's and one of Peter Quince's in a single line of type at the foot of B2 (B2,33; I.ii.52-53). But it appears that Bottom's Ercles speech (B2,15-18; I.ii.27-34) which is printed as prose was probably written as prose in the copy, since no provision was made in the casting-off for it to be set as verse. One further mislineation occurs in B(i): in the middle of B3v (l. 16; II.i.42-43), we find the beginning of one of Puck's couplet speeches in this arrangement:

As the speech goes on for fifteen correct lines, the compositor must have planned to set it in couplets and allowed space for them when he established the place in the text at which to begin B4v even though, by analogy with the fairy's speech on B3, the couplets were written in single lines in the MS. In this instance, however, he apparently thought that he could squeeze the first complete verse line into the same line of type with the half-line of verse which begins the speech, a calculation which, as the turn-over shows, was none too accurate.

Since, as we have seen, the compositor seems to have had to adjust the MS lineation in sheet B, we can better understand his decision to shift to seriatim setting in sheet C even though in doing so he penalized himself in his time relationship with the press. C4v was set out of seriatim order, but no serious problems seem to have arisen. The only instance of mislineation is found at C1v,12 (II.i.175-176) where the compositor chose to set a line and a half of verse in one line of type. It is in this part of the text, however, that Wilson thinks heavy revision to have been made—II.i.1-147 belonging to one level of composition, ll. 148-187 to a second, ll. 188-246 to a third, and ll. 247-268 (end of scene) and II.ii.1-42 to the second again. As C1 begins at II.i.130 and C3 ends at II.ii.38, the seriatim setting of the first five pages of the sheet may indeed suggest that the workman was confronted here with a particularly nightmarish piece of composite copy.

In D(o) we find two more minor instances of mislining—a prose speech of Peter Quince's at D2v,8-9 (III.i.93-94) set in part as verse, doubtless under the influence of Thisbe's immediately preceding verse speech, and at D4v,17 (III.ii.48-49) a line and a half of verse set in one line of type. The textual material included in this sheet runs as follows:

       

D1	II.ii.141-III.i.11	D1v	III.i.11-III.i.47
D2v	III.i.85-III.i.120	D2	III.i.48-III.i.84
D3	III.i.121-III.i.153	D3v	III.i.154-III.i.190
D4v	III.ii.33-III.ii.67	D4	III.i.191-III.ii.32

It is interesting that all the material in D(i), the later forme, is perfectly lined, although it is set solid and includes on D1v and D2 much more prose than verse. According to Wilson, all of III.i and the first forty lines of III.ii belong to the same stratum of composition and underwent small, if any, revision.[15] The exact fitting of the material in the quarto suggests that the manuscript at this point was regular enough to permit quite accurate casting-off. Wilson also believes the half-line at D4v,17 to mark an abridgement, but this is a matter upon which our analysis provides no additional information.

However, the piecemeal distribution of E(i), beginning with E1v after the setting of F2 or part of F3v, suggests that work on F(o) and the first two pages of F(i) went slowly; and this is another part of the

text (III.ii, ending near the foot of F2v) which Wilson thinks to have been considerably worked over. He sees the irregular lining at E1,13-14 (III.ii.80-81), a shift from couplets to blank verse at E1v,27 (III.ii.127), and a short line at E3v,19 (III.ii.256) as indications of abridgement, addition, and cancellation. The first is the only mislined passage in the sheet, and, since it occurs on the first page of the forme and does not affect the total number of lines in the page, it seems virtually impossible for the passage to have been deliberately mislined for mechanical reasons. Nor does there seem to be any likelihood that the two subsequent aberrations in the text, even though they are found in the inner forme, arose from the compositor's tampering. At F1v,31-32 (III.ii.396-399) we find Puck's four-line speech set in two lines of type, an indication, according to Wilson (pp. 125,130) that they were squeezed in at the foot of a page of revised MS. As two lines only must have been allowed for this speech in the casting-off, the rest of F1v and F2 being set solid, they were doubtless written in two lines in the manuscript, but it seems much less certain that they were written so as to crowd them on an MS page when we remember that the fairy's speech on B3, which is in the same meter, seems to have been similarly lined in the MS.

IV.i begins near the bottom of F2v and ends seven lines down on G2. About this scene Wilson says in part:

The first of these speeches is at F3, 18-19; it appears as follows: The second is at F3, 26-27: Because he did not turn "loue" over, his usual practice when a verse line was too long for the measure, the compositor probably thought the first of these speeches to be prose. We have seen him at D2v,8-9 line prose as verse, and it does not seem unlikely that here, under the influence of Bottom's immediately preceding prose speech, he would have mistaken verse for prose, especially since the line was probably written in one line in the MS. In the second case, however, the turning-over of "hoord" and the capitalization of "And" make it clear that the compositor knew he was dealing with verse, although he handled it incorrectly. We have here a case very like that previously observed at B4v,20, only in this instance I think it would have been very difficult for the compositor to have lined in this manner as the result of carelessness. The lines probably stood in the MS as they do in the quarto, but neither this mislineation nor that at F3,18 provides much evidence for the recopying of an earlier draft. However, the material in F3v and F4 (IV.i.44-IV.i.110), all verse, is perfectly lined and set solid, an indication that the MS, recopied or not, was regular enough for an accurate line count.

Sheet G includes the following textual material:

       

G1	IV.i.144-IV.i.177	G1v	IV.i.178-IV.i.212
G2v	IV.ii.22-V.i.15	G2	IV.i.212-IV.ii.22
G3	V.i.15-V.i.48	G3v	V.i.49-V.i.83
G4v	V.i.118-V.i.150	G4	V.i.84-V.i.117

In this section of the text the most serious mislining occurs, and, as we have noted earlier, on this evidence Wilson argues most strongly for revision in the copy which underlay the quarto. From the point of view of printing mechanics, the sheet has another feature (in addition to the mislining) which is distinctly odd: the number of lines of type per page, earlier a consistent thirty-five, was reduced to thirty-four on G1, G1v, and G2v and to thirty-two on G2. Because the mislining is present on G3v (which contains thirty-five lines of type), it is clear that there is no absolute relationship between the incorrect lineation and the reduction of the number of lines of type per page in the first four pages of the sheet—that is, the lines of type per page were not reduced simply because the compositor compressed material that would have occupied more space if it had been lined correctly.

Let us look more closely at the passages in question. At G1,21-23 (IV.i.164-166) we find:

Here, as at B4v,20 and F3,26, the compositor had nothing to gain by altering the lineation since his version occupies just the same number of lines as the correct version, nor does it seem very likely that his error resulted from a failure of memory after he had seen correct lineation in the MS. Wilson suggests (p. 135) that the lines were written in the margin of the manuscript "to take the place of a longer cancelled passage." Certainly the indications are that the quarto lining reflects the lining of the copy.

On G2v and G3, where there are two lengthier mislined passages, the effect of the mislining is different. The first is found at G2v,26-G3,3 (V.i.5-18); asterisks indicate lines which Wilson (pp. 80-81) thinks to have belonged to the earlier layer of composition:

The second is at G3,16-24 (V.i.29-38): In both instances the text is compressed, a total of twenty-five lines of verse (according to the quarto's standard of lineation) being set in twenty-one lines of type. Neverthless, as was true on G1, there seems to be no mechanical reason here for the mislineation, since the compositor was setting the first forme. Once more it seems likely that the lineation of the quarto reflects the lineation of the copy: there is no mechanical reason why the lines should not have been, as Wilson believes (p. 85), "written on the margin of the MS, in such space as could be found."

In G(i) there are further instances of mislineation. At G1v,7 (IV.i. 184-185) a short verse line is set in the same line of type with a complete verse line, in just the way we have seen earlier at B3v,16 and C1v,12. There is no evidence to tell us positively whether the copy stood this way—the compositor may either have been setting what he saw or hedging against faulty casting off—but the occurrence of the same kind of mislineation on C1v, which was set seriatim, creates a presumption that the copy was so lined. Similarly, we find at G1v,22-23 (IV.i.197-198) that there are two lines of verse set as prose (cf. F3,18-19), probably under the influence of Bottom's immediately following prose speech; no space is saved by the mislineation. G3v contains four mislined passages, the first at G3v,10-12 (V.i.58-60):

the second at G3v,18-22 (V.i.66-70): the third at G3v,29-30 (V.i.76-78): and the fourth at G3v,34-35 (V.i.81-83): Since we have seen this sort of mislining in the outer forme, we are probably safe in thinking that here too the mislineation arose from the copy and not from the compositor's efforts to compress his material into a fixed space. Had the workman been making adjustments for purely mechanical reasons, he probably would have done so on G4 as well as G3v, but the text on G4 is correctly lined.

Although detailed reconstruction of a compositor's procedures must be in part speculative, I believe we can see about what happened if we try to follow our man step-by-step through his work on the sheet. We know that he set G(o) first and that he was probably having to work as quickly as he could to keep up with the press. Before beginning G1 he

must have cast off the whole sheet, counting out thirty-five lines for each page and marking page-endings in the copy. Had he been under less pressure, he probably could have worked out the correct lineation of the mislined passages (we saw him making correct adjustments in the lineation of his copy in sheet B) which, as Wilson has noticed, were probably marginal additions; but as it was he counted them as they were lined in the MS. If this was his procedure, however, he evidently counted in some material that he later did not set, and it does not seem unlikely, if he was working hastily, that he may have failed to notice that some lines here and there were supposed to be cancelled. Wilson detects abridgements in the text at G1, 22-23 (IV.i.165-166) in the outer forme and at G1v,22-23 (IV.i.197-198) in the inner. It seems quite possible, then, that the compositor discovered during the course of setting G1 that he had in his casting-off included a cancelled line. He thus reduced the number of lines of type on that page by one, preferring to set the page short than to readjust the subsequent pageendings already marked. I take it that the lack of one line of type on G2v arises from similar causes, although there the miscalculation probably sprang from a failure to count out the marginal material correctly; and that on G1v a cancelled passage of about four lines was provided for in the estimate of space but was not set up, thus requiring the omission of one line of type on G1v and the opening up of the material on G2. An incomplete prose line at the foot of G2 (l. 31) indicates that the workman had to strain a little to marry that page with G2v, which had already been composed.

The instances of mislineation in sheets H and A (with one exception) are so like those we have already examined that we may summarize them:

• H1,33-H1v,1 (V.i.182-185): Prose mistaken for verse.
• H1v,7 (V.i.193): Two lines of verse in one line of type; probably MS lineation.
• H2,35 (V.i.262): Two lines of verse in one line of type; probably compositor's lineation (cf. B2,33).
• H4,7 (V.i.394): Couplet set in one line of type; probably MS lineation.
• A2v,33) (I.i.53): Two short lines in one line of type; responsibility doubtful.

At A2v,2 and 5 (I.i.24 and 26) Stand forth Demetrius and Stand forth Lisander are centered rather than printed as part of the text as meter requires. Since he set the proper names in italics rather than in roman, the compositor evidently did not take these phrases as stage directions, but lined them about as they were lined in the copy. Wilson (p. 105) notes that "they were probably written as separate half-lines in the original MS, to denote deliberate utterance . . ." or, we may add, to cue obvious stage business.

We have seen that the quarto was set largely by formes under adverse conditions probably caused by sections of difficult copy. In only two places (B2,33 and H2,35) does there seem to be any likelihood that the compositor juggled the lineation of the text in order to fit copy to a predetermined space.[17] The bibliographical evidence seems to me to point toward the kind of heavily revised manuscript described by Professor Wilson as copy for the quarto, although it does not, of course, in itself lend any support to his distinction of different levels of style in the revisions. There is, as Sir Walter Greg has noted, no indication that the manuscript ever served as a prompt book, and the presence of revision may lead one to think of late-stage foul papers as copy.