University of Virginia Library

The Editing of Folio Romeo and Juliet
by
S. W. Reid

Few plays in the First Folio of Shakespeare appear to have received less editorial supervision from Heminge and Condell than Romeo and Juliet. Although the authoritative second quarto is sometimes regarded as one of the less satisfactory so-called "good" quartos, responsible scholarship has found no reason to believe that Heminge and Condell, or their agents, felt obliged to bring the text of its descendant, Q3, into conformity with a playhouse manuscript in order to have F1 represent the "True Originall"—as they frequently did with other quartos that provided Folio copy. Sir Walter Greg summed up the matter in 1955 as follows: "Q2 was reprinted in 1609 and from this third quarto F was set up without material alteration. There is no evidence that the copy was ever compared with a playhouse manuscript; indeed it certainly was not, for any such comparison must have left apparent traces. Nor can the quarto have been conceivably used as a prompt-book. . . . Italic for the Nurse's part was finally discarded. The stage-directions were slightly revised, sometimes intelligently, sometimes not; there is nothing to suggest that this was done anywhere but in the printing-house." He documents these conclusions by noting that "at IV.v.103 [tln 2680] 'Enter Will Kemp' is replaced by 'Enter Peter' in accordance with the prefixes. V.iii.71 [2924] is recognized as a speech and given to Peter, in accordance with the direction at l. 22 [2874], but there Peter is an error for Balthasar, and in fact the speaker is Paris's Page. Prefixes ('Nurse') are given to the marginal calls of 'Madam' at II.ii.149 and 151 [952, 954], and the muddled prefix 'Watch boy' at V.iii.171 [3036] is corrected to 'Boy'." Otherwise he finds that "the only differences appear to be" some eighteen changes in stage-directions, "apart from purely formal alterations." Greg's view of the matter has, predictably, been followed by a variety of subsequent scholars.[1]

Greg's verdict on the Folio text of Romeo and Juliet (despite his reference to "the printing-house") would seem to receive some support from other plays in the volume, which generally exhibit variants from their quarto copy that are more striking either individually or as a group. But his argument for Rom. itself rests on the evidence of F1's errors of omission—the failure to alter the erroneous entry at 2069 (III.v) —or of commission—the replacing of the erroneous 'Enter Romeo' "by the equally erroneous 'Enter Servant'" (I.v. tln 568), the deficient exit introduced at 1309 (II.iv), the wrong speech-prefix added at 2924 (V.iii), the unnecessary deletion of 'manet' at 2675 (IV.v).[2] Both ranges of evidence of course entail Greg's usual hypothesis that such inconsistencies must have been resolved in the prompt-book, and both depend directly for their worth on the validity of this hypothesis.[3] The evidential basis for his conclusions is obviously not as broad as might be hoped. Also causing some uncertainty about his assessment are two facts that have come to light since Greg wrote: (1) the "newly corrected, augmented, and amended" Q4 was printed before November of 1622, and (2) most of Folio Rom. was typeset by the least capable of Jaggard's workmen, Compositor E.[4] Neither of these discoveries is so revolutionary by itself as to overturn Greg's verdict, however unsure its foundations, that the Folio is not dependent on the prompt-book. Yet examination of their possible implications reveals that the changes introduced in the Folio

sometimes agree with certain readings of Q4 and Q1 and thus seem to suggest that F1 is somehow related to one or both of them, but that nevertheless Q3 alone provided printed copy for F1; that these changes could not have originated with the compositors; and that consequently we are left with several explanations of their origin which, though not necessarily irreconcilable with one another, are unanticipated in Greg's conclusions.

I

The dating of Q4 has led one recent editor to conclude that it was the source of "a number of passages" found in the Folio.[5] Indeed since Romeo and Juliet went through the press during the spring of 1623 (Hinman, I, 363-365; II, 513-529), the fact that Q4 had been printed by the end of the previous year raises the possibility that it, rather than Q3, actually provided copy for the Folio. The possibility is attractive because it would bring Rom. under Greg's general observation that "in most cases the particular edition used as copy was what may be assumed to have been the latest available at the time of preparation" (p. 159), and also would explain a number of agreements between Q4 and F1 that are not readily attributable to independent compositorial correction of Q3. In fact, however, another explanation for the concurrences of Q4 and F1 must be sought, for the evidence is sufficient to demonstrate that Q3 indeed served as Folio copy.

The evidence consists not only of errors shared by Q3 and F1 as against Q4, but also of those changes in F1 traceable directly to Q3's faults. A few of these faults have a typographical origin. The most interesting is in Juliet's speech at the end of IV.i, where the Folio has her take Friar Lawrence's potion with the line 'Give me, give me, O tell me not ofcare' (2416). Here Q2 and Q4-5 have the correct 'of feare', whereas Q3 has 'off eare'. This reading is obviously due to the absence of spacing material in the type that printed some copies, at least, of Q3, and either Compositor E or some other agent has taken it to be an example of foul case and variant spelling (of / off)[6] and has made a plausible though wrong

"correction." A somewhat similar instance is found in Juliet's last speech: Q2's 'This is thy sheath, there rust and let me dye' is corrupted by Q3 to 'Tis is', restored to 'This is' by Q4, but further altered to ''Tis in' by F1 (3035). Q3's omission of the -h- may have been caused by the compositor's fiddling around with the line to accommodate a turn-up from the long verse line that follows,[7] but at any rate this Q3 error has parallels in several other readings. In one case Q2's 'turnes' becomes Q3's grammatically undesirable 'turne', and F1 corrects to 'turn'd' (and, incidentally, reconciles this by making the 'becomes' > 'became' change at 1956), while Q4 simply restores Q2's form (1957). A few lines later (1961), the Q2-3 'puts up' (an error presumably for 'puts upon') appears in F1 as the awkward but metrically improved 'puttest up' and in Q4 as 'powts upon'. This pattern is repeated in the Q2-3 omission at 2380, where Q4 inserts 'shroud' and F1 the less imaginative and repetitive 'grave', and in the ambiguous abbreviation 'L.' at 951, which Q4 expands to 'Love' and F1 to 'Lord'. The same pattern of compound variation may be found in the speech-prefix at 2087, where Q2-3's incorrect 'Ro.' becomes 'Ju.' in Q4 but 'Juilet.' in F1; though substantively identical (even with Q3, which has 'Ju.' as a catchword), the different forms suggest F1's independence of Q4, since Compositor E is obviously reproducing (with a typical transposition) the form he found in his copy. (More on this point later.) Finally, there are the well-known cruces involving the assignment of speeches—the lines which close II.ii and open II.iii and the 'O Godigoden' speech in III.v (992-998, 999-1009, 2215-19). In all three cases the differences between Q4 and Q3 are unmistakable, and in all three F1 has accepted the Q2-3 readings whereas Q4 has altered them. The relatively numerous and persuasive examples of compound variation combine with these significant differences and with the more numerous instances of common error to show F1's derivation from Q3 independently of Q4.

For two reasons this conclusion must stand despite the fact that the Folio and Q4 agree against Q3 in having readings that are about as numerous as those just cited, though not so significant.[8] Aside from the usual agreements traceable to obvious correction or fortuitous identical variation, Q4 and F1 have in common some twenty identical readings that vary from their mutual copy, Q3. Several of these involve the Folio alterations that Greg cites—e.g., the substitution of 'Peter' for 'Will

Kempe' and of 'Boy' for 'Watch boy', and the righting of the redundant stage-directions in the last scene (2680, 3036, 3061, 3074+1). To these conspicuous changes may be added another speech-prefix and several perceptive corrections of dialogue that are common to F1 and Q4.[9] Most of these alterations represent intelligent attempts to remedy the defects of Q3 and would appear to be beyond the several compositors, particularly since those in F1 were typeset by the otherwise deficient E. Yet it is their very nature that makes them suspect as evidence of F1's dependence on Q4. As McKerrow some time ago pointed out, "it is a general rule that the less significant the readings varied are, from a literary point of view, the greater is their weight as evidence of the genetic relationships of the texts in which they occur."[10] As attempts to improve Q3, rather than common errors, the agreements of Q4 and F1 are much too significant (in McKerrow's terms) to establish the derivation of the 1623 Folio from the 1622 quarto, for the alternative explanation that they originated independently with the respective editors is equally plausible—or rather (as will emerge shortly) more plausible, in view of all the other changes introduced in F1 independently of Q4.

There is still another reason these twenty or so readings cannot counterbalance the weightier evidence already discussed. Since the Folio "corrections" traceable to Q3's typographical faults make it clear that an example of this quarto served as copy in Jaggard's shop, any argument for F1's dependence on Q4 would have to explain why he would have used this "newly corrected, augmented, and amended" quarto only sporadically in combination with Q3, when it would have been more convenient and more sensible to use Q4 alone as copy. Under ordinary circumstances, McKerrow's theory of dual copy—according to which "the master printer might have copies of the two preceding editions and it might be convenient to give one to each of the compositors to work from" (p. 105)—would admirably suit a book generally set up, at any given time, by two men working more or less simultaneously. But the "intercalary" formes of Rom. were typeset by only one workman at a time, almost always Compositor E. Moreover, such a practice as McKerrow envisages should have produced some sort of bibliographical pattern in the Folio's readings; but there is none discernible in its agreements with Q4, which are on the whole isolated ones scattered throughout the play.

The only alternative to the theory of dual copy is to postulate consultation

of Q4. By nature this theory is open-ended and difficult to test, since almost any instance of lack of agreement between F1 and Q4 is easily justified by an appeal to the concept of occasional use which is basic to the theory. However, conflicting evidence of positive disagreement does arise whenever—as in the instances first cited—the supposedly dependent F1 deliberately changes a Q3 reading but in doing so ignores the acceptable solution present in Q4. The only way to explain such conflicting evidence is to make a second postulate—viz., that the Folio editor sometimes supplied alterations independently of Q4 instead of consulting it. Yet each time such an independent change occurs, the second postulate increasingly gains strength, until it alone is sufficient to account for all the non-compositorial alterations in F1, whether or not they appear in Q4. On the premise that unity of reading necessarily implies unity of source, one could just as well argue that Jaggard's shop consulted the 1597 "bad quarto," behind which presumably lies a report of some sort of recollected performance.[11] As a matter of fact, the agreements between these two texts as against Q3 are more diverse and substantial than those between Q4 and F1, ranging as they do from several of the Folio stage-directions and speech-prefixes that Greg cites—e.g., the identical entrances for 'Tybalt', 'Mother', and 'Appothecarie' (1556, 2592, 2785)—through essential concurrence in directions like 'Exit. Mercutio, Benvolio' at 1242 (Q1: 'Exeunt Benvolio, Mercutio'), to numerous agreements in speech assignments and the wording of dialogue, all of which differ from Q3's readings. Indeed, F1's use of Q1 might explain some, though by no means all, of the agreements between F1 and Q4, since a few editors believe that Q4 "habitually consulted" Q1.[12] Still, the conflicting evidence of deliberate Folio alterations made independently of Q1 is equally substantial as that for F1's dependence on Q1, and even if one were to follow the consultation theory to its illogical end and postulate the Folio's use of both Q4 and Q1, such independent Folio alterations would remain to be explained. Here one may well paraphrase Greg (p. 160): we may claim the right to prefer probabilities to possibilities, simpler hypotheses to complex ones, and to hope that people performed their tasks in a more or less reasonable fashion. It goes without saying that had Heminge and Condell or any other playhouse editor consulted a document, it would have been a playhouse manuscript rather than another print. Moreover, the notion that Jaggard would have used another quarto to supplement Q3 raises a vexing question about his motivation for doing so, when generally he seems to have relied on Shakespeare's company to provide suitable copy. Such consultation of another quarto to supplement the Q3 copy approved by Heminge and Condell is quite a different matter from earlier printers' supposed use of bad quartos as an aid when setting from foul-papers. And the Folio's agreements with Q4 and Q1 in its attempts to improve Q3's text belong to an entirely different order of evidence from, for instance, the typography of the Nurse's speeches which presumably links Q1 and Q2 Rom. or the variable indention of speech-prefixes on the second page of Q2 Hamlet, which are apparently traceable to the inner and outer formes of Q1. We must have grounds more relative than such agreements in reading to establish the Folio's dependence on either Q4 or Q1. There can be no reasonable doubt that Q3 alone served as printed copy for Folio Romeo and Juliet. The conclusion must follow that the source of the alterations introduced in F1 is to be sought in the Folio's own transmission process, despite its interesting agreements with both Q4 and Q1.

II

The implications of this conclusion are naturally of considerable interest for the Folio, especially since all but one and one-half of its pages were typeset by Compositor E. If a copy of Q3, and not Q4, lies behind the Folio and if F1 nevertheless contains a number of variants that agree either with the reported text of Q1 (which, especially in its stage-directions, presumably represents a theatrical tradition of some kind), or with the apparently "edited" Q4 (which may or may not have consulted Q1 at a given point), or with both, then the question remains of the origin of the Folio's changes, which in fact are not unlike those that it introduces in other quarto plays. The two traditional responses would credit the prompt-book, against the considered judgment of generations of critics, or blame the compositors. The former is perhaps debatable at this point and cannot, in any case, be discussed until the compositorial alterations are factored out. As for the latter, invoking the Folio compositors simply will not work as an explanation in Romeo and Juliet. It might go to explain some of the more imaginative changes in the first page of the play, set up by B, if the traditional view of this workman

were followed.[13] However, it will not explain similar alterations throughout the remainder of the play, set by the less competent E, nor the curious evidence of the last part-page, which was reset by B at a later time. Altogether there are, by conservative estimate, some seventy such alterations, the majority of them in the stage-directions and speech-prefixes but many in the dialogue as well, and they are often found elsewhere than in the attempts to correct Q3's errors and in the agreements with Q1 and Q4 so far discussed. We have yet much to learn about Compositor E (and Compositor B), but whether or not we identify him as the apprentice John Leason who joined Jaggard in 1622, we can all agree, I suppose, that his work on Tit. and Rom. was subject to unusual scrutiny in the proofreading, that he was incapable of performing certain elementary compositorial tasks satisfactorily, and that many of the changes introduced in Folio Rom. would appear to be beyond either his powers or his experience.[14]

Furthermore, whatever his capacities, the mode of operation that he was forced to adopt effectively thwarted any ability or inclination he may have had to improve the play while setting it. Hinman has shown that E's role, at least early in his work on the Tragedies,[15] was that of a substitute for the regular compositors (B and generally A), who were engaged in setting other plays from manuscript. When either workman was called away, E was to step in at the vacant cases and set up a portion of Tit. or Rom. in order to keep the printing of the Folio going forward. His work on Rom. was, therefore, very intermittent, and discontinuity was even greater than that normally occasioned by the alternation back and forth between various parts of the plays caused by setting by formes. This discontinuity was augmented in the first third of Rom. by the fact that four of the pages in quire ee (sig. ee5-6v, tln 493-1016) shared formes with the last four pages of Tit. and that consequently E was here

alternating between two plays.[16] The same was true to some extent in the one-sheet quire gg. There current gg2v was originally a forme-mate of cancelled *gg5, which contained part of Tro., and the cancellation of this play caused a particularly long delay between E's composition of *gg2v ( gg2v) and of its eventual forme-mate gg1, though continuity appears to have held in the setting up of gg1v:2. Yet even in those formes not experiencing this kind of alternation between plays, the delay and discontinuity that plagued E's work was enough to disorient a more accomplished and experienced typesetter. In quire ff, for instance, which contains Acts II, III, and IV of Rom., E set the formes in "normal" order (i.e., from the inside, ff3v:4, out); yet he had to contend not only with the normal disorientation in the first half of the quire caused by setting these portions of the play in reverse order, but also with the discontinuities occasioned by delays: between ff6 and ff6v, for example, six formes of R3 and H8 set by B and others went through the press, by Hinman's calculations (II, 220-234), and it was thus about three days that E was away from the cases and from Rom. before resuming composition (with ff1 presumably).[17] These severe handicaps together with the weaknesses exhibited in his early work make it difficult to believe that many of the Folio's changes which aim at improving Q3's text originated with the main compositor of Rom., the tyro E.

These changes, which must therefore represent annotation of Q3 copy, include but go beyond those listed by Greg (p. 235). One added stage-direction, noted by Greg, appears in the ninth line of gg1, which E set up after the protracted delay occasioned by the setting and cancellation of part of Tro., about a week after he had completed ff6v and after he had composed what would have been the two last pages of Rom. and the first four of Tro.[18] This is the re-entry 'Enter Mother' (2592) omitted in Q2-3 and also in the intelligently edited Q4, which presumably failed to consult Q1 (which has it) or to notice independently that an entrance

was required for the subsequent speaker. Indeed, the complicated and quick action—which involves entrances and departures of Capulet, Lady Capulet, and the Nurse, as well as the business of music and servants with logs and baskets—is enough to elude an editor reading consecutively. E, however, began his setting of gg1 in the middle of Q3's sig. K2; for him to have comprehended the action, we must suppose that he not only read through the facing page K1v, where Lady Capulet appears in the stage-directions and speech-prefixes in her alternate function as 'Lady' ('La.'), but also leafed back to I4v and K1 to get the full sequence of action, where she appears as 'Mother' ('Mo.').[19] This entry is the only change introduced in this page of Q3, and to argue that E made it independently and thus alone amongst the reprints provided for this entrance is to attribute to him a comprehension of the action of the play matched only by the actor-reporters of Q1.

Another interesting re-entry cited by Greg is that for Tybalt, which E sets early in column a of ff3 (1556); it pairs with the 'Exit Tybalt' that F1 substitutes for Q2-3's unusual 'Away Tibalt' (1522). E would have begun setting ff3 with tln 1542 on Q3's F4, and unless he looked across to F3v and read through it in order to gain a sense of the continuity of the action, he would not have had the opportunity to understand the flow of the staging until at least a day later, when he returned to set up ff2v, working his way from Q3's sig. F2 up through its F4. To argue that in fact E did both these things is, as is often the case, to argue that he took more care with and made better sense of the play than the editor of Q4, who failed to mend Q3's ambiguously centered words (either to an exit or a speech[20]) and to provide a re-entry, in spite of the fact that Q1 does both.

A third example comes to hand in the two expanded exits at 2465 and 2477 (IV.ii), which show great pains in specifying the departures of 'Juliet and Nurse' and later 'Father and Mother', though failing to do so for the 'two or three' servingmen that also appear in the scene. Both directions occur early in column a of ff6v; E began setting this page with the last two lines of Q3's sig. I4, which also contains the centered entry for all the characters but Juliet (2423-24), whose entrance is provided in a rather hidden direction sandwiched between long lines of dialogue (2438). Between the time he set these entries in sig. ff6 and added the

specifics for the exits in sig. ff6v, some six formes went through the press in the course of about three days' work, before E returned to Rom. by setting up forme-mate ff1 and then ff6v. Hence, unless we suppose that he deliberately reread the rest of sig. I4 prior to setting its last two lines and digested them so thoroughly as to introduce the two sets of characters in the subsequent directions—where even the intelligently edited Q4 is satisfied with Q3's simple 'Exeunt' at 2465 (despite the 'Exeunt Nurse and Juliet' of Q1) and with changing its 'Exit' to 'Exeunt' at 2477 (more or less in accord with Q1)—we have to assume that these additions to Q3 represent annotations of a Folio editor.

Similar specification occurs halfway down column a of ff2 in the Folio's addition of 'Nurse and Peter' to Q3's 'Exit' (1309), which Q4 again found satisfactory even though Q1 offers a complete account of the departures—including Romeo's, which, as Greg points out, F1 fails to provide for. In this case, of course, E was to begin setting the preceding page (ff1v) four formes later and would have had to read not only the rest of Q3's E4—where he began with the short penultimate speech toward the bottom (1280)—but E3v as well and also would have then had to decide that Q3's 'Exeunt' at 1242 referred to both Mercutio and Benvolio, altering that stage-direction to read 'Exit. Mercutio, Benvolio' (substantially with Q1) when he returned to set it two days later. In this as well as the other cases, E faced considerable discontinuity owing to the mode in which he worked, and if we are to assign the Folio's alterations to him we must postulate his habitual reading of earlier pages of Q3 and attribute to him a comprehension of the play comparable to that of the Q1 reporters and exceeding that of the Q4 editor, who had the presumed advantages of working through Q3 consecutively and (perhaps) of consulting Q1.

This hypothesis, or set of hypotheses, is of course untenable, especially given E's observable failures in other, simpler tasks; and Greg's citations of these Folio alterations are virtual acknowledgements that none of the Folio compositors can be credited with such changes, even if seriatim setting were assumed.[21] Indeed just about all the stage-directions that he cites are open to one or more of the improbabilities so far discussed, if we attempt to trace them elsewhere than to a Folio editor. And this generalization applies to a few that he failed to notice. These must also represent annotation, although one or two that Greg cites are probably the result of compositorial justification.

As already suggested, the specification added to Q3's exit at 1309 inevitably

involves comprehension of the total scene and identification of the characters at 1242; similar elaborations of a Q3 exit and of a Q3 entry occur elsewhere (2758, 3061, the latter involving the deletion of 3074+1). The gratuitous change 'Romeo' > 'Servant' at 568 no doubt represents a misunderstanding of Shakespeare's staging and an attempt to clarify it and to provide for the subsequent speaker, but the introduction of 'their' in the preceding entry (567) shows some detailed attention to the wording that finds a parallel in F1's 'a Servingman' (446). Somewhat comparable is the substitution of 'Peter' for 'Will Kempe' (2680). But the extent to which the annotator is responsible for the alterations of the exits in this part of the play is problematic. The preceding speech-prefix (2679) was also annotated, and all that marking up may have somehow helped distract E from the 'Exeunt omnes' (2679+1) that appears in the direction line of Q3's K3. Certainly, the earlier omission of 'manet' (2675) in an exit that also affects this sequence of action is to be put down to E's justification. There is some doubt, too, about part of another stage-direction which Greg mentions, the deletion of Q3's 'or partysons' at 71: although it could be argued that F1's omission represents a book-keeper's decision on an "indefinite" stage-direction, it might well exhibit anticipatory justification on the part of Compositor B aimed at centering the entry, in accord with his custom.[22] On the other hand, the "imperative" instruction 'Fight' in the preceding line almost surely represents an annotator, and it finds parallels in the stage-directions at 938, 1881, and 1885 (which Greg cites), at 3035 (which he fails to cite), and at 952 and 954 (which he confuses in a curious way).[23] The Folio editor even exhibits concern with the details of exits. The fairly straightforward 'Exit Tybalt' < 'Away Tibalt' (1522) must be laid to his charge by virtue of its subtlety, its positioning, and its coordination with Tybalt's later re-entry (see above). There is, though, also some attention to the somewhat formal question of exit/exeunt, which is displayed not only in the otherwise annotated departure at 2477, but at 1643 and 793 as well. This suggests that the added exits at 959, 450, and 330 are also the work of an editor, either because other annotation of the Q3 pages containing them makes insertion of them more probable or because the context requires more attention than we have a right to expect from E.

If these stage-directions be acknowledged the work of a Folio editor, rather than Compositor E, then a number of the Folio's speech-prefixes must represent annotation also. Perhaps the clearest example is the series of regularizations in the divertimento between Peter and the musicians in IV.v (2679-2719), where F1 alters to 'Mu.' Q3's variable 'Min.' and 'Fid.' (and their formal variants). The last of these (2719) was set from Q3's K4 and alone appears on gg1v, which followed gg1 by some five formes (or two and one-half days). That E achieved this kind of uniformity on his own is incredible; the Folio's 'Mu.' is undoubtedly editorial, being derived from Q3's 'Musi.' on K3 (at 2676), which E would only have looked at a few days earlier, and the annotations were part of the same process that included the 'Will Kempe' > 'Peter' alteration (2680) if not the deletion of Q3's 'Exeunt omnes' (2679+1). To the editor must also be ascribed the long delayed romanization of the Nurse's speeches, which E first encountered at the top of B4 in Q3 and from then on uniformly altered from its italic, even though in setting the same column of ee4 he followed the italic of Capulet's "letter" as found on Q3's B3v.[24] No doubt the editor, rather than E, was also responsible for altering line 2924 to a speech, although as Greg remarks Shakespeare's own confusion at 2874 seems to have misled him into assigning it to Peter; this intelligent but probably wrong editing is comparable to the earlier mistreatment of similarly ambiguous words as a stage-direction rather than a speech (at 1522). Certainly the correct assignment to Juliet of the lines which conclude her soliloquy in II.v (1325-30) must be ascribed to an editor, and this finds a parallel not only in Q4's identical change but also at 2087. Here F1, like Q4 (and in accord with Q1), corrects Q3's 'Ro.' at the top of its sig. H3v, giving the speech to 'Juilet.' Since Q3's catchword is 'Ju.' and since the context makes the change relatively self-evident, the assignment itself is of less significance than the form it takes (even though crediting E with the change on these bases would seem unjustified). By and large E is rather faithful in reproducing the form of the speech-prefixes found in Q3,[25] and his only 'Juliet.' in a

speech-prefix occurs the second time he sets it (at 359)—under the influence of Q3's first 'Juliet.' two lines earlier and perhaps of the distractions which the annotation of the Nurse's initial speeches might have caused him. In any case it seems probable that the speech-prefix 'Juilet.' represents a reproduction of an editorial notation, the transposition being typical of E's response to marked-up copy at this point in his career (see below). More or less formal concerns are also exhibited at 2221—where F1 makes one attempt at regularizing the "functional" speech-prefixes by altering Q3's 'Wi.' to 'La.' in accord with the other speech-prefixes on the page, though ignoring the 'Mo.' on the following page (2249)—in Compositor B's '2. Wife.' < 'M. Wife 2.' (81), in the addition (also found in Q4) of the fairly obvious prefix for Juliet after her entrance alone (1645), and in the 'Boy.' < 'Watch boy.' normalization in the last scene (3036). Particularly interesting is the regularization to 'App.' of the various Q3 forms ('Appo.', 'Poti.', 'Po.'), since all appear to have been coordinate with the addition of the entry at 2785 and since the first (2786) is not really called for and depends on someone desiring to regularize those (2795, 2804, 2806) that occur on the following Q3 page. Q4 follows these various forms, and it seems unlikely that E altered them on his own. The Folio editor, therefore, appears to have been concerned with such matters. He may even be responsible for the mistaken '3. Cap.' < '1. Capu.' early in the play (612) and the two curious changes at the end in B's Gg1 (3155, 3146), though these two are rather complex (see discussion below). But apart from these three changes, there is abundant evidence from the regularization of 'Mu.' < 'Fid.' ('Min.') to that of 'App.' < 'Appo.' ('Po.', 'Poti.') of editorial attention to the speech assignments of the play.

Since this is so, it may come as no surprise that there is some evidence of an editor's attention to and annotation of the dialogue itself. Here we are generally on less secure ground, because the context of the dialogue is much sooner grasped by a compositor setting by formes than is that of the action and because the changes themselves are by nature less easily assigned to a particular agent. However, with E's early work we may make some headway in trying to distinguish what can be laid at his door from what must be attributed to annotation. Two of the most helpful examples (495, 763) occur early in his pages ee5 and ee6; each is separated by intervening formes from the preceding page (i.e., ee4v and ee5v) and each follows its intervening forme-mate, ee2v and ee1v, both of which contain the end of Tit. In both instances F1 succeeds in mending a subtle Q3 error derived from Q2 in a witty or sophisticated context. In the case of 'Abraham Cupid' < 'Abraham: Cupid' (763), the Q4 editor also succeeds in correcting Q3, but at 495 he fails to make F1's generally

accepted 'your' < 'you' alteration (Q1 provides neither correction). Moreover, F1's 'Lord' < 'L.' (951) seems to be superior to Q4's 'Love' (Q1: 'Lord'), and the Folio's attempt to deal with the Shakespearian duplication at 1843-45+2, derived from foul papers via Q2, shows attention to the dialogue superior to that exhibited by Q4, which follows Q2-3.[26] In each of these cases we would, I believe, be loathe to credit Compositor E with the alteration.

This view of E, which partakes no doubt of some prejudice, finds support in several other instances of change. One is particularly complex and somewhat inferential. It involves F1's curious lines

These famous lines appear correctly in no early text, in the view of most editors, Q1 having 'name' for 'word' but also being the only text to provide the query after the first 'name'. F1's obviously botched line is puzzling, given Q3's basically satisfactory 'What's in a name that . . .', until it is seen to represent an interchange of the '?' and the 's' (both types one-en thick). But, this acknowledged, the question remains how the text came about. Had E recognized what was wanted in the course of setting the lines, his text should have come out right. The explanation appears to lie in a characteristic failure of E's that is exhibited several times in his attempts to introduce proof-corrections, where he seems to have transposed the changes called for, sometimes simply by alternating types. Perhaps the most serious is at Rom., 348: here E's 'shall scant shell' was marked to be corrected to Q3's 'shall scant shew' but E confused the '-all' and '-ell' letters that follow the identical ligatures and produced the irredeemable nonsense 'shew scant shell,' that now stands in examples of "corrected" ee4.[27] Three other instances exhibit a similar proclivity to interchanging types in response to marked-up print:        

F1 (u)	F1 (c)
Rom. 229 wisewi:sely	wisewi : sely
Rom. 1218 [qua-|] th aGentleman:	[qua-|] t ha:Gentlemen,
Lr. 1787 skin. so'tis	skinso :'tis

Although other examples of E's botching annotation might be cited,[28] this particular disposition accounts, I suggest, for F1's curious line, which results from E's characteristic interchange in his attempt to carry out instructions given by the Folio editor in the form of an annotation in his Q3 copy.[29]

Other evidence of the editor's annotation of fine points in the dialogue are perhaps less inferential. Certainly F1's unique correction 'here, of' < 'hereof' shows sensitivity to subtleties comparable to that found at 763, as does 'counterfaits a' < 'counterfaits. A', though here Q4's 'counterfeits, a' is comparable (2010, 2170). Of interest, too, is the series of changes in lines 1956-61. Of the five, only 'or' is an outright spontaneous error, perhaps occasioned by the annotation that must have surrounded it. The last, 'puttest up', though wrong, was probably motivated by metrical considerations; Q3 needs correction and F1 does as well here as Q4 ('powts upon'). In the previous line both editors also take Q2-3's 'mishaved' for an error and alter it, F1 with perhaps a little more imagination (F1: 'mishaped'; Q4: 'misbehav'd). They discern that Q3's 'turne' is erroneous, but here Q4 restores Q2's 'turnes' while F1 alters to 'turn'd'; this presumably leads to the Folio change 'becomes' > 'became' in the previous line, which shows some concern with consistency of tenses and can hardly be attributed to E (who would have already completed the line before coming to Q3's 'turne', unless he memorized two lines at a time).

The attempt to correct Q3's 'puts up' at 1961 suggests a certain concern with meter, and elsewhere F1's 'Capulets' (also adopted by the Q4 editor) might yet stand had not Q1 provided the superior 'Capels are' (1433). Two other examples, a little less subtle, occur in ff2v, where Q3's grammatically satisfactory wording is deliberately altered ('Forbid' > 'Forbidden', 'both houses' > 'both the Houses') to make regular ten-syllable lines (1520, 1524). The even more gratuitous changes at 2153 nevertheless exhibit a similar, albeit misguided, concern with meter ('Shall happly make thee there a joyfull Bride' > 'Shall happily make thee a joyfull Bride'). Mild concern with censorship may be traced in the substitutions for 'zounds' and 'sounds', and this may also account in part for the alteration of the stage-direction 'Slud knocke' (1480, 1534,

1885).[30] Less easily classified alterations appear elsewhere. In the compound variant at 3035, F1's ''Tis in' makes tolerable sense of Q3's 'Tis is', though Q4 succeeds in restoring Q2's 'This is'; here an editor also added the stage-direction 'Kils herselfe' and simplified the next speech-prefix to 'Boy.' (3036). Another example depends on Q3's paging: the addition 'grave' (2380) is not terribly imaginative and could have been made with a glance at the previous line, if that line were not on the previous page of Q3; Q2-3's text makes tolerable sense ('tomb', or Q4's 'shroud', or some other grisly word associated with burial being implied), but its defective meter signals the need for a monosyllable, and the Folio editor, rather than E, appears to have supplied it. Less imaginative but correct alterations are the changes 'Sycamour' < 'Syramour', 'jaunt' < 'jaunce', and 'denote' < 'devote' (123, 1338, 1927). These are all readings adopted by later editors; it may be noticed that Q4 fails to get the first, Q1 and F1 alone providing it. More questionable, indeed gratuitous, changes are found in 'o'th Collar' < 'of choller', 'done' < 'dun', and 'Coarse' < 'course' (9, 492, 1782); though all are wrong (and the last, perhaps, still a mere spelling variation in 1623), they all exhibit some elementary concern with double meanings and a tendency to pin down one of them, and all three find parallels in either Q4 or Q1. Other examples of Folio changes of Q3 dialogue might be brought forward: for instance, the oxymoron created by 'fire' < 'life' is difficult to attribute to E because it produces a highly conceited reading and represents a conscious attempt to improve Q3's quite acceptable passage (2497). But from the evidence just cited it would seem plain that the Folio's text exhibits a rather surprising attention to its dialogue, particularly since all but one (tln 9) of these thirty or so instances occur in the work of Compositor E.

III

If such editorial annotation of Q3's stage-directions, speech-prefixes, and dialogue took place, then several questions about the treatment of the play as a whole arise. One is rather specific and concerns the relining of verse and the alterations from verse to prose or from prose to verse. Such changes are numerous in Folio Romeo and Juliet. In the lower half of gg1b, for instance, the dialogue between Peter and the musicians exhibits various alterations of this sort (IV.v, 2695-2708). Sometimes these changes can be seen as Compositor E's attempts to divide one line at the caesura to fit the verse to his measure, or as another mechanical

expedient. But often, as with the alterations in IV.v, the nature of the changes suggests the operation of some literary judgment while annotation of stage-directions, speech-prefixes, or dialogue in the area indicates the presence of an editorial hand. These alterations in Rom. are comparable to those in other Folio plays set from known quartos—though not nearly so numerous as those in 1H4, for example—and constitute a specific problem needing further investigation as to their origin and purpose.[31]

Of more general interest are the questions which the editorial annotations of Romeo and Juliet raise about the exact origin of the Folio's readings. Of course such shadowy and elusive matters cannot be explored in a vacuum nor decided in any final way without reference to other plays in F1, for the editing of this particular play was actually part of a larger process which remains somewhat obscure at best. Yet because this play has editorial changes that can be rather precisely differentiated from compositorial error and also exhibits probably the least complicated sorts of annotation found anywhere in the Folio, Rom. offers an unusually good opportunity for inquiring into the origins of the Folio's readings, however brief that inquiry and however provisional its conclusions must be.

The most obvious question is whether or not a playhouse manuscript, specifically the prompt-book, was the source of these annotations. The evidence is far from conclusive either way. Despite the verdict rendered by Greg and others who have followed him, it can be argued that the non-compositorial alterations in Rom. are not unlike those in other plays which have often been viewed as signs of the prompt-book. The attention to and correction of speech-prefixes are remarkable, particularly those changes which eliminate ambiguities that Greg (p. 114) believes would be "intolerable" in a prompt-book (e.g., 'Wi.' to 'La.' at 2221, or 'Appo.', 'Poti.', 'Po.' to 'App.' at 2786ff.). The substantial alterations of the stage-directions are also noteworthy: many of these are business-like, and some are of the imperative kind often cited as theatrical.[32] The

added 'Fight' (70) is an example, but of especial interest are the shortened stage-directions 'Knocke' in III.iii (1881, 1885); identical parallels to them occur in The Second Maiden's Tragedy, the earliest extant prompt-book of the King's Men, where the "prompter" has added the simple 'Knock' four times in one scene.[33] The provision of two re-entries, for Tybalt and Lady Capulet (1556, 2592), could certainly be traced to a bookkeeper concerned with such details (see Greg, p. 133). One could also assert that the removal of 'Will Kempe' (2680) is evidence of a revival of the play (see Greg, p. 114), and that the censorship found in F1 is likewise a sign of a post-1606 revival, though it is slight and generally not as thorough as in other plays (cf. 1H4, where there are many oaths and consequently many cuts). Finally, there is the rather substantial evidence of the correspondence between the Folio's readings and Q1's, which are often very similar and sometimes curiously so. These agreements range from being substantially to precisely the same, and the simplest way of explaining them would be to postulate a single source—the prompt-book itself.

With this argument there are naturally many difficulties, for the situation is not a simple one and is indeed ambivalent. One difficulty relates directly to Q1. As mentioned, F1 not only fails to incorporate alterations present in Q1, but often varies from it in the alterations that it makes. Some of this variation, presumably from the prompt-book, could as easily be traced to the reporters of Q1 as to the editor of F1, but in other cases—e.g., 'Serving.' > 'a Servingman', '1.Capu. > '3. Cap.', 'Capels' > 'Capulets' (446, 612, 1433)—Q1 supplies what appear to be the authentic readings (i.e., 'Clowne', 'Cap:', 'Capels are') and F1's look like editorial guesses.[34] Moreover, it is of some interest that where Q1 is totally defective F1 continues to exhibit alterations that are of the same general character as those it introduces where Q1 appears to represent a

theatrical tradition. Some of these Folio changes could be attributed to its sources's superiority over Q1's defects, but the failures of F1 just mentioned suggest that all its alterations might well have been made without reference to Q1's ultimate source—i.e., might not come from the prompt-book.

More general observations also make the theory of F1's dependence on the prompt-book tenuous. The Folio text of Romeo and Juliet simply lacks any certain sign of its derivation from prompt-copy; it has nothing comparable to 'They sleepe all the Act' or 'Tawyer with a Trumpet before them' found in Folio MND, nor to the music and sounds added to Folio MV, nor even to the new "fly scene" and speeches found in Folio Tit. (also set by E). Beyond this rather negative evidence, there are the errors of commission cited by Greg, some of which suggest not simply tangles that had not been completely resolved but outright ignorance of any source other than Q3. Furthermore, the many changes in the dialogue are not what we have come to regard as sure signs of the prompt-book, whereas Folio Rom. does, for instance, lack attention to props,[35] which the prompt-book presumably would have dealt with. The attempted correction of Romeo's fly speech (1843-45+2) suggests absence of a prompt-book, where the matter should have been resolved if it was given any attention at all (at least, so we have been led to believe). The same may be said of the Folio's failure to deal with the second thoughts and unclear speeches at the end of II.ii and the beginning of II.iii.

It might be possible to explain some of these inconsistencies in Folio Romeo and Juliet by postulating a playhouse manuscript that partook of some characteristics supposedly typical of the prompt-book and of some associated with foul-papers. For instance, a transcript used in mounting the play and containing random notes reflecting decisions made by the actors—including the author, presumably—would help explain some of the annotation of dialogue, the supplying and correction of speech-prefixes, perhaps the excision of 'Will Kempe' (assuming revival after 1602), the slight attention to oaths,[36] and particularly some of the stage-directions (like 'Kils herselfe') added by F1 but, because descriptive, supposedly not typical of the prompt-book or book-keeper.[37]

Yet even such a written source will not account for some of the more remarkable Folio alterations to Q3, as Greg's verdict against "a playhouse manuscript" suggests (despite his belief that "the printing-house" must have been responsible for the Folio's changes).[38] The errors of commission which he cites—the assignment of a speech to Peter rather than Paris's Page, the substitution of the erroneous 'Enter Servant' for the equally wrong 'Enter Romeo', the deficient exit in II.iv (2924, 568, 1309) —cannot be blamed on any sort of theatrical manuscript (nor even on messy foul-papers, which in this case of course had vanished with the printing of Q2). Moreover, even if the error of omission that Greg mentions (the neglect to alter the erroneous entry in III.v, 2069) is suspect as evidence, F1's failure to amend the glaring inconsistency in the 'grayey'd morne' passages at the end of II.ii and the beginning of II.iii (992ff.) certainly shows the Folio editor was not consulting a manuscript at this point. The fact is that many of the Folio's alterations in stage-directions, speech-prefixes, and dialogue are attributable to an intelligent editor attentively working through Q3 without the benefit of a manuscript. This has been suggested already in the discussion of Compositor E's role and the fact that his manner of setting the play (as well as his own incapacities perhaps) would have prevented him from making changes that require an understanding of the dramatic context. But a number of the changes could have been made by an editor with an eye to clarifying

the action and the speeches and with no aid but his own wits. For instance, several of the new entries—e.g., 'Servant', 'Enter Mother', 'Enter Appothecarie' (568, 2592, 2785)—appear to be motivated by a simple attempt to account for the presence of an immediately subsequent speaker, whether rightly (2592) or wrongly (568). A similar motivation probably lies behind some of the speech-prefixes added or corrected—e.g., 'Jul.', 'Pet.', 'Ro.' > 'Juilet.' (1645, 2924, 2087). The merely formal provisions of exits added to the ends of scenes can be put down to a scrupulous editor, as can many of the alterations of dialogue, which show either an inclination to regularize meter or a disposition to clarify speeches that is misguided in its smoothing-out of puns or else is simply mistaken (e.g., 9, 492, 1520-24, 1782, 2153). Then there are the curious alterations introduced for the first time in Compositor B's part-page Gg1, the last page of the play in the Folio as finally published. This part-page, set by B together with Timon of Athens some thirteen working days (and twenty-three formes) after E had finished typesetting the rest of Rom., exhibits two interesting variations in speech-prefixes where E's cancelled *gg3 follows Q3, and it thus affords a rather close view of the editorial process. Both speech-prefixes concern minor characters, Balthazer and Paris's Page, who are designated as 'Balth.' ('Balt.') and 'Boy.' in Q3 and E's *gg3 but as 'Boy.' and 'Page.', respectively, in B's Gg1 (3146, 3155). In an ultimate sense of course B's 'Page.' is correct, but elsewhere in this last scene F1 (set by E), generally following Q3, inconsistently identifies him as 'Pet.', 'Page.', and 'Boy.' (Q3: 'Watch boy.')—which makes the 'Boy.' of E and Q3 quite satisfactory at this point (and in the theater).[39] Perhaps even less called for is the alteration 'Balth.' > 'Boy.' at 3146: F1 (retaining the forms of Q3) elsewhere in Act V has either the speech-prefix 'Man.' or 'Pet.', and Balthazer is named only once in the dialogue (2735). These changes accomplish relatively little, if regularization in accord with the prompt-book is the presumed aim. But they do resemble editorial alterations elsewhere in the text. Moreover, both are curious and precisionistic alterations, and both suggest some general acquaintance with the play. It is now virtually impossible to attribute them to B (who was, incidentally, principally concerned at this time with the rather daunting task of setting Tim. from manuscript), because he seems to have followed his copy's speech-prefixes rather carefully when encountering material anew[40] and had not dealt with Rom. since composing its first page some seventy formes earlier. The alternatives are to assume that E overlooked both these annotations—which seems unlikely since other evidence suggests he was overly sensitive, rather than oblivious, to such notations—or that B's copy, whether Q3 or F1's *gg3, was subjected to further[41] annotation before he began typesetting. To make such niggling alterations, whatever their negligible virtues, it seems certain that the editor would not have gone to the trouble of re-consulting a playhouse manuscript. He must have introduced them while quickly looking over printer's copy once again before B began to set the last part-page of the play.

The cancellation of the last page of Folio Rom. has resulted in two new editorial alterations that seem on the whole consistent with the other changes found in the Folio text. But the circumstances themselves are unusual. If such an editor interceded elsewhere in the play, distinguishing his independent alterations from those which he drew from a playhouse source would often prove to be an almost impossible task (a fact that has some interesting implications for Folio texts other than Rom.). This would be true particularly if, as seems likely, the editor was either Heminge or Condell or another playhouse agent. Several features of the text can be taken as evidence of his connection with the playhouse. The Folio changes often exhibit a general familiarity with the play and an understanding of both the on-stage and off-stage action that is rather remarkable. This is true even where the first quarto is unsatisfactory, and where it is essentially sound, F1's frequent identity with or uncanny resemblance to Q1 suggests an editor well schooled in the play as performed. Moreover, ascribing the Folio's changes ultimately to the King's Men would help explain not only their general resemblance to Q1's version but their frequently theatrical nature. Some of the stage-directions are very business-like, and other changes—e.g., the additions of 'Within' (952, 954) that Greg, in one of his rare moments, misremembers and thus suppresses—are theatrically technical. Folio Rom. exhibits not one but two added re-entries, which, Greg remarks, are "exceptional in early texts,"[42] and it is of some interest that Edward Knight adds one to Massinger's carefully prepared fair-copy of Believe as you List.

Whether Knight or Heminge and Condell themselves would actually have annotated the Folio copy can hardly be determined. But even if one of the players was actually the editor, his general familiarity with the work, as distinguished from dim recollection of parts (which would

come into play only in certain speeches, if at all) would probably have been exhibited chiefly in the stage-directions and speech-prefixes—i.e., those features traditionally relied upon as evidence of a playhouse manuscript. Consequently, we can never be sure whether or not a playhouse manuscript was ever consulted. But it is relatively certain that an editor, rather than a compositor, is responsible for many of the changes introduced in the Folio text of Romeo and Juliet and that he made some of them at least without reference to a playhouse source (not to mention Q4 or Q1). This simple hypothesis will account for virtually all the non-compositorial alterations of Q3 found in F1, even if the question of the use of a theatrical source is in fact irresolvable. The general conclusion seems inevitable that in this Folio play we have an editor who worked through Q3's text with (for his day) considerable care, annotating the printer's copy where it struck him as deficient and relying mainly on the context to do so, though perhaps occasionally—and certainly not often enough—consulting a playhouse manuscript.

Notes