Source Text and Provenance

1The earliest surviving text of the play, The Tragedie of Cymbeline (listed as Cymbeline King of Britaine in the catalogue of plays in the First Folio of 1623), is also the only independent authority for a source text. The cleanness of the text, along with certain indications such as the number of parentheses, a predilection for hyphens, and reasonably consistent idiosyncratic spelling preferences suggest that Ralph Crane, a scribe who worked for the King's Men on a number of occasions, copied the text, possibly specifically for the publication of F1, from an unknown source.

2For a number of reasons, identifying the source from which Ralph Crane made his transcription presents a challenge, not the least of which is the fact that Paul Werstine and other scholars have demonstrated how such staple concepts in twentieth-century bibliographic studies as "foul papers," "fair copy," and "promptbook," not to mention the qualitative judgments attached to the various classes of documents, are not the absolute categories once believed. Nevertheless, some distinctions once applied to these groups allow us to discuss the characteristics of different texts.

3A drastic simplification of the categories as outlined by W. W. Greg aligns foul papers with authorial drafts, fair copy with a clean copy of the author's final revision, and the promptbook as the copy annotated with performance directions. For the last, none of the signs that bibliographers generally interpret as evidence of theatrical manuscript origins is present: Cymbeline is free from confusion between actors' and characters' names, early entrances suggestive of influence from a prompter's notes to warn an actor of an imminent entrance, and musical cues. Instead, the text features increasingly elaborate stage directions which seem descriptive rather than prescriptive: "Enter in State" (1374), for instance, renders an impression of royal pomp without providing the kinds of information believed necessary to staging a performance, such as the numbers and kinds of attendants or musical cues for a sennet; and

4Enter (as in an Apparation) Sicillius Leonatus, Father to Posthumus, an old man, attyred like a warriour, leading in his hand an ancient Matron (his wife, & Mother to Posthumus) with Musicke before them. Then after other Musicke, followes the two young Leonati (Brothers to Posthumus) with wounds as they died in the warrs. They circle Posthumus round as he lies sleeping" (3065-71)

5is full of details which would be interesting to a reader but whose wordiness would likely interfere with rather than promote the smooth function of staging the scene. Logically, then, if the provenance of F1 Cymbeline is not theatrical, it must be authorial. However, the nature of those authorial papers remains in doubt. Even Taylor and Jowett prefer the explanation that the few demonstrable orthographical inconsistencies between the two halves of Cymbeline result from its having been transcribed by two scribes (Ralph Crane and an "unknown" scribe) to the conclusion that Crane copied the entire play from a play written collaboratively or heavily revised by a second author (256). Thus, applying Howard-Hill's observations regarding Crane's practices to Cymbeline in order to evaluate the evidence that Crane prepared the whole script, in context of the preferences of the Jaggard compositors and the tendencies of Shakespeare and other possible co-authors, especially John Fletcher (known to have collaborated with Shakespeare on Henry VIII, The Two Noble Kinsmen, and the lost play Cardenio) presents a more fruitful avenue of exploration.

6Evidence on the work of Ralph Crane has been slow to accumulate. Since F. P. Wilson's article for The Library in 1926 on Ralph Crane's idiosyncratic spelling and punctuation, only Trevor Howard Howard-Hill's benchmark work in 1972 identifying five comedies transcribed by Crane, Ralph Crane and Some Shakespeare First Folio Comedies, and a few subsequent articles also by Howard-Hill have substantially elaborated on Crane's preferences and experiences as scribe. For instance, Howard-Hill identifies a number of features which are typical of (in some cases, invariant in) Crane transcripts, one of the most prominent of which is Crane's tendency to use hyphens in such unusual cases as to connect verbs and pronouns (e.g. "whip-you") or possessives with substantives (vertue's-sake) (40-41). The list of hyphenated words in Appendix A will give some indication as to their prevalence in Cymbeline; this list naturally excludes the use of hyphens indicating a word wrapping between lines. Because Howard-Hill limited his scope to the comedies, he does not consider the evidence for or against Cymbeline as a Crane transcript. In Shakespeare Reshaped (1993), Gary Taylor and John Jowett offer some arguments developing the suggestion from A Textual Companion that Cymbeline is a sixth Crane play, relying upon Howard-Hill's observations. Roger Warren's edition of Cymbeline (1998) extends the Oxford case to deliver the most complete analysis of the Crane elements present. Compositor studies such as Charlton Hinman's monumental analysis of the traits of the compositors in the Jaggard print shop who set text for the Shakespeare First Folio also provide an invaluable contribution to our ability to evaluate the possible kinds of alteration made to a text, as have those who have confirmed and in some cases refined Hinman's conclusions.

7Crane's idiosyncratic spellings also support the conclusion that he is the source of the scribal transcription from which Cymbeline was set, although it is necessary to examine the compositorial preferences overlying and in many cases obscuring Crane's habits, and doubt remains. His invariable spelling "guift" occurs five times in Cymbeline, with "gift" never. "Ceize," long noted as extremely rare outside of Crane manuscripts, occurs once in Cymbeline. His preference for a doubled medial "t" in "cittie" and "pittie" but not before other occurrences of "-ie" also is the rule in Cymbeline. Crane's tendency to include an "a" following a long vowel recurs in theame, groane, yoake, smoake/ie/s, choake/choak'd, but not generally in "stroke" or "strokes."

8Other characteristic spellings are obscured by compositorial preferences which could be representative of confirming or replacing the original. For instance, Crane prefers the spellings "I'ld" or "Il'd" over "Ide" or "I'de" (Ralph Crane 88); "I'ld" appears three times and "Il'd" once in Cymbeline, with no occurrences of the other two spellings. "Blood," either by itself or as "bloody/ie" is invariant, as is true of Crane's manuscripts but also generally of B's stints. "Traitor" and "vertue," other preferences, likewise occur exclusively. On the other hand, Crane's preferred forms "auncient" and "goe" may either have been covered up by B's preferences or not present in the first place. In either case, with very strong preferences by Crane and B, the evidence cannot be used as proof of Crane's presence or lack thereof.

9The most convincing orthographic evidence may occur when, despite B's usual practice, some of Crane's unusual favored spellings appear. "Extreamitie," a variation of Crane's "extreame," occurs at 1687, and "powre(s)" and "flowre(s)" both occur more frequently than their B-preferred counterparts. Single occurrences of "gon" and "goe," the first in a song and the second in a stage direction, may or may not be significant. Two instances of the contraction "it's" with an apostrophe, still a relatively new linguistic development at the time, appear. Crane's preferred spellings of answer, approach, behind, old, sun, war, and courtesie also appear with some frequency. The patterns seem to be similar to those Howard-Hill analyzes for the five comedies; nevertheless, B uses alternate spellings elsewhere in enough instances to make this orthographic data inconclusive.

10Crane's elaborate stage directions and use of parentheses have also been used in the past to identify Crane transcripts. Of the facts that can be taken to demonstrate that Cymbeline is as likely as any of the other five comedies in F1 to be considered set from Crane transcripts, the observation that "aside from one example in King John, only the five Crane texts and Cymbeline contain parentheses in stage directions" is one of the more persuasive (Textual Companion 604). John Jowett elsewhere observes that "One phrase in particular, '(as in an Apparation)', is striking in that exactly the same words occur, again within brackets, in at least two of Crane's transcripts of A Game at Chess." (113-14). He further considers the wording of stage directions (focusing on The Tempest, but with observations that pertain to Cymbeline as well), noting that the kinds and quantities of descriptions in stage directions, particularly in the use of specific vocabulary words, are similar in Cymbeline and in the stage directions in Crane plays (111). Furthermore, the word "solemn(e)," besides its two instances in Cymbeline, occurs only in The Tempest and the Fletcherian parts of Henry VIII. Crane did not restrict his fondness for parentheses to stage directions but is well known for his use of parentheses around names or titles in vocative address and in general dialogue. Howard-Hill ranks Cymbeline third among Crane comedies in frequency of parentheses and hyphenated-prefix words ("Shakespeare's Earliest Editor" 128, n. 51).

11Another valuable observation is that Crane often altered the spelling of his sources to ensure that rhymes also were eye rhymes, the multiple copies of A Game at Chess making it possible to identify variations and multiple stages in alteration (Howard-Hill, "Shakespeare's Earliest Editor" 119). Without multiple versions, knowing whether any change was made is difficult: if a rhyme is also an eye rhyme, there is no proof that the original author did not write it that way. It is only possible to note that, of the infrequent rhymes in Cymbeline, several do seem to be as close to eye-rhymed as possible, and that this is not something to which Shakespeare usually paid attention. For instance, in his speech to "Briton Lord" after the battle, Posthumus makes several rhymes. Some of them may simply be coincidentally preferred spellings; others seem to be eye-rhymed. Between 2986 and 3014, we find six couplets, ending as follows: lane/bane, end/Friend, doo/too, Britaine/againe, death/breath, and agen/Imogen. The rhyme agen/Imogen occurs once previously at 2020-21, and these are the only two instances of the spelling "agen"; all twenty-three other instances are spelled "againe." At 2610-11, it is possible that Crane's intervention in preference of eye-rhyme has created an editorial problem as it corrupts the grammar. The F1 text reads, "The ground that gaue them first, ha's them againe: / Their pleasures here are past, so are their paine." On the other hand, at 588-590, the triple rhyme do/untrue/you is evidently not adjusted -- but whether this should be taken as counter proof is less clear, since it is so problematical to identify a spelling which would accommodate all three of these words, all of which have accepted orthographical spellings that are incompatible. In the other couplets as well, determining whether the spelling is altered or not simply cannot be established. Nevertheless, even if it cannot prove whether or not Crane is the scribe behind the text from which Cymbeline was set, understanding Crane's practices can help explain the kinds of changes that might have been made, in at least one of which the editorial practices of a professional scribe may have created a wording that has largely been taken by later editors to be an error.

12Howard-Hill also notes that Crane occasionally wrote "Ex^t " for "Exeunt." In at least one place in Cymbeline, "Exit" and "Exeunt" are evidently confused. At TLN 587, a singular exit is given for a group exiting, "Exit Qu. and Ladies." An alternative explanation besides a misinterpretation of Crane's "Ex^t " is readily available, however. It is possible, although the line is not full as it stands, that the compositor altered "Exeunt" to "Exit" and possibly abbreviated Queen(e) to "Qu." in order to create enough room for the stage direction to stay on the same line as the text it follows. Again, at TLN 887 and 901, compositorial confusion over an abbreviation may have created some confusion. The scene contains Clotten, First Lord, and Second Lord; Clotten invites the lords to leave, saying, "Come: go." Second Lord responds that he will come soon, and an exit is recorded at TLN 887, followed by a soliloquy by Second Lord. What is unclear whether Clotten, First Lord, or both exit. If both leave, then the direction "Exeunt" at TLN 901 is incorrect. If "Exeunt" is correct, then either First Lord or Clotten must exit and the other remain, but there is no indication as to which would do which. (See note for TLN 901). The only other commonly emended exit occurs with "Exit Lucius, &c." (TLN 1912), where it is impossible to rule out the possibility that the unusually abbreviated direction led to the conclusion that "Exit" would be correct. (The ampersand in a stage direction is commonly associated with Ralph Crane as well, but this line is already full, so the ampersand, like "Exit," might well be compositorial, as might the unusual spelling, "happines," also consistent with Crane's other spellings.) Additionally, exits commonly suffer from irregularity. In a nearly 4000-line play, these few confusions should not be considered too unexpected.

13Beyond the textual evidence characteristic of Crane, his autobiographical comments in The Works of Mercy and The Pilgrim's New Year's Gift (1621 and 1625?) support the possibility that he could have transcribed Cymbeline. Howard-Hill cites these notes as evidence that Crane's association with the King's Men was broken off during the printing of F1, apparently because Crane, at around seventy years of age, was slowing down too much to be of use to a dramatic company where speed was often of the essence ("Shakespeare's Earliest Editor" 125-29; see also Howard-Hill's discussion of the creation and copying of Fletcher and Massinger's Sir John van Olden Barnavelt in "Crane's 1619 'Promptbook' of Barnavelt and Theatrical Processes" for his assessment of a time-sensitive composition.) After the printing of the First Folio, no further evidence of Crane's working for the King's Men exists, although he took commissions from individual authors, including Webster, Middleton, and Fletcher (Howard-Hill, "Shakespeare's Earliest Editor" 127). If we merely extend Howard-Hill's timeline slightly to allow Cymbeline's inclusion as a Crane transcription, it would provide a more tidy timeline for the end of Crane's association with the King's Men. (Of course, it usually pays to be suspicious of tidiness when reconstructing events.) Even without altering the timeline, it is possible that Cymbeline was set from a Crane transcription, given that a transcription could have been made at any point from the beginning of its composition until the printing, though the occasion of the publication is often taken as the occasion of transcription as well. Crane's own testimony that his writing speed was diminishing may help explain the delinquency of The Winter's Tale text in getting to the printers, and possibly why Cymbeline, like The Winter's Tale, is printed last in its section of the folio.

14Because only five other plays from F1 (the first four comedies and the final comedy, The Tempest, Two Gentlemen of Verona, The Merry Wives of Windsor, Measure for Measure, and The Winter's Tale) show convincing evidence of having been copied by Crane, it is tempting to establish a link between this group of plays and Cymbeline that would indicate why only these six required copying. As the tragedies (except for Troilus and Cressida, inserted as the first tragedy but in fact the final play set for the Folio) were included in the Folio in the order in which they were set, with Cymbeline being the last play, genre and printing order can be dismissed. Chronological order of composition can be eliminated: although Cymbeline, The Winter's Tale, and The Tempest are among the last of Shakespeare's independently written plays (assuming Cymbeline was, in fact, independently written), Two Gentlemen of Verona is early, and The Merry Wives of Windsor and Measure for Measure are from the middle of Shakespeare's career. Nor does it appear that these plays shared a remarkable textual history such as having been heavily censored or manifestly rewritten. It is possible that some unknown common factor such as being favorites of a patron is at work, but it is also equally possible that they were transcribed for different reasons. The Tempest could have been a presentation copy, Merry Wives could have been so heavily censored that it was difficult even for early modern compositors to set, and so on. Alternatively, the five comedies might share some common feature, and Cymbeline might have been transcribed for a different reason. Perhaps all thirty-five plays were to have been transcribed before printing, but the cost and time became prohibitive, given Howard-Hill's estimate that it took Crane around three-quarters of an hour to transcribe one page of about fifty-five manuscript lines ("Crane's 1619 'Promptbook'" 155). Or, perhaps the first four were printed first precisely because clean copies already existed, for various reasons, and some difficulty with The Winter's Tale and Cymbeline was discovered which required transcription later, during the printing. A problem with the text, whether copious and confusing revisions, unintelligible handwriting, or some other difficulty, might account for the printing difficulties which could be solved by a professional copy. Given the paucity of evidence regarding theatrical documents and the controversy over the questionable utility of traditional bibliographic terminology, the likelihood of ever discovering or recovering knowledge that would provide a causal link is small.

15One of the primary unexplained questions about Cymbeline is the break of some kind between 2.4 and 2.5 (TLN 1336 and 1337) as evidenced by the sudden shift from the spelling "O" to "Oh," which both Compositor B and E follow in their respective stints. This evidence is available because Crane appears to have had no preference between "O" and "Oh" but to have preserved whatever orthography he found. Taylor and Wells conclude that "the Folio copy was a Crane transcript, itself copied from a manuscript in which a second hand took over at 2.5.0.1/1154.1" (Textual Companion 604). Because the shift is not affected by the alternations between compositors, compositorial preference can be eliminated as the source; and because Crane had no preference and because signs of his presence are present in both halves, it appears not to be a scribal change. Another option -- the one which Taylor and Jowett apparently favor -- that two scribes copied the text of a single author, with Crane perhaps copying the first half from authorial papers and the second half from a copy made previously by a different scribe, is suspect on the basis that, spellings aside, a number of vocabulary choices mark the dichotomy (not to mention the illogic of paying a scribe to copy a copy). For instance, forms of "choose" ("chuse," "choose," and "chose") occur only in the first half (four times), while forms of "never" ("ne're" and "nere") occur only in the second half (seven times). More strikingly, the author of the first half exclusively prefers "betweene" and "betwixt" (seven times) while the author of the second half only uses "among" and "amongst" (five times). Other preferences support this split: "either" occurs in the first half only twice, while "either," "eyther," "neither," and "neyther" appear a total of thirteen times in the second half. The second half also demonstrates a marked preference for forms of "taken," with "t'ane" appearing only once in the first half and "'tane" and "tane" appearing nine times. We see "ha's," "has," and "hast" in the first half only in E's stints, and "hast" fourteen times in the second half to one "ha's."

16With most words, the sections do not divide so clearly by orthography or usage, and slight or even moderate preferences do not preclude the use of the word in question. Intriguingly, a few discrepancies suggest the possibility that the two hands are not entirely separate after TLN 1336. The speech prefix for Pisanio, for instance, is "Pisa." through 654 (13 times), heavily favors "Pis." from 1129 to 2750 (thirty-two times, compared to four of "Pisa."), and then returns to exclusive use of "Pisa" (six times between 2779 and the end). Counting by act, Act 1 and 5 favor "Pisa.," with Act 3 favoring "Pis."; Pisanio has only three lines combined in Acts 2 and 4.

17Unfortunately, Taylor and Jowett do not speculate on the nature of the two hands they detect in the text; the problem they are concerned with is determining whether this observation throws doubt on the conclusion that Shakespeare's preference was for "O" (255). If their surmises are correct regarding Crane's participation and the "two hands," it would be logical to extend those assumptions to say that Shakespeare's is the first hand, since the single piece of orthographical evidence they bring forth, the preferred spelling of "o(h)," associates Shakespeare with the first section, through TLN 1336. This agrees with Howard-Hill's provocative assertion that the copies Crane made were frequently of collaborative plays ("Shakespeare's First Editor" 126).

18While not entirely decisive, the weight of evidence makes reasonable the conclusion that the Folio text of Cymbeline was printed from a transcription made by Ralph Crane, probably during the printing of F1 but possibly earlier, from authorial papers of unknown kind and mixed provenance. The shift between the two parts not only in orthography but in some matters of usage as well suggests that the differences represent a collaboration between Shakespeare and another author, conceivably John Fletcher, with whom he later collaborated on three plays.

Printing-Shop Work: What Do We Know about the Printing of the First Folio?

Identifying the Compositors

19Charlton Hinman's The Printing and Proof-reading of the First Folio of Shakespeare (1963) uses bibliographic data to determine, to the extent possible, the date and conditions of the publication of the First Folio. His identification of five compositors who set portions of F1 has been modified slightly in the intervening years but not overturned. Of the five, only B and E are of primary importance to Cymbeline: Compositor A, who also set type in the tragedies, was apparently working on another project since he stopped setting type for the tragedies about halfway through, while Compositor D worked only in the comedies, and C worked little in the tragedies. Hinman argues that Cymbeline was set entirely by B, supported by his conclusion that Compositor E, generally supposed to have been an apprentice, was only allowed to work on texts for which print copy was available after inept attempts at setting copy from manuscripts. Trevor Howard Howard-Hill corrects Hinman's assessment to suggest that Compositor E set pages zz4v, zz5v, zz6v, aaa3, and aaa1v, or 633 lines ("New Light" 159). As we have no record in the Stationer's Register of Cymbeline having been licensed for printing before this time and no previous printed text is extant, it appears Compositor E was probably allowed to set Cymbeline from a scribal copy prepared by Ralph Crane (see "Source Text" for more on Crane's work).

20A new test, not proposed by Howard-Hill, confirms his ascertainment of E's presence. E invariably sets the semicolon immediately after the preceding word, whereas B prefers to set an intermediate space. On the pages accorded to E, nearly forty occasions confirm his preference (with the exception of a semicolon in a prose line at 966 where the space probably marks his attempt to justify the line). Unfortunately, this test may not be as useful in other texts as in Cymbeline; the frequency of semicolons in E's stint here is markedly higher than elsewhere in the text and in other plays set by E. The ratio for B is not quite as dramatic, but it is significant: B sets around 110 semicolons with a preceding space compared to only around 23 without, some of which are in tightly justified lines. Perhaps E is reproducing semicolons in a frequency found in the text, as one occurs nearly every sixteen lines, on average, in his stint, in contrast to one every twenty-four lines in B's stint.

The Ralph Crane Comedies and Cymbeline

21Of the six plays thought to have been copied by Crane, four are the first four comedies printed, one is The Winter's Tale, which was not printed with the first four and also was out of order for the comedies, being set after work on the histories had begun, arguing that the text was not ready to be set when expected. The sixth is Cymbeline, which, like The Winter's Tale, is printed last in its section among the tragedies; there is no indication, however, that work was delayed in order to wait for Cymbeline. The superficial correspondence might suggest that Cymbeline was reserved for the final play because the text once again arrived late, but there is no foundation for this argument.

The Order of Printing the Folio Plays

22Cymbeline is the last play in the Folio, but it was not the last of the text to be set: the preliminaries -- the title page, note to the reader, epistle dedicatory, tributory poems, list of the principal actors, catalogue of plays, and so on -- came after, as, Hinman argues, did post-cancellation Troilus and Cressida. In addition to the bibliographic information Hinman adduces, including type recurrence studies, some evidence supports the conclusion that Cymbeline was set prior to Nov. 4, 1623: on the last page of Cymbeline, a colophon lists William Jaggard as a publisher, while on the title page's list of printers, he is replaced by Isaac Jaggard. William Jaggard died sometime before Nov. 4, setting the final date before which Cymbeline must have been printed. On the earlier side, Hinman points to August, when type recurrence studies indicate that the Jaggard shop printed the "Heralds' Visitation Summons" partway through quire vv, after which most of Othello, Antony and Cleopatra, and Cymbeline must have been set (quires xx-bbb) (Hinman vol. 2, 25). As Compositor B was mostly working alone or with the aid of the slow Compositor E in setting the text at this time, a slower rate of production must be allowed for, but even so, it would have been possible to print those plays in the time allotted (vol. 1, 360-62). Thus, the majority of Cymbeline was probably set in October 1623.

The Order of Printing Cymbeline's Pages

23The order of composition follows norms for the Folio: 3v and 4, 3 and 4v, 2v and 5, 2 and 5v, 1v and 6, and 1 and 6v for quires zz (shared with Antony and Cleopatra) and aaa. In quire bbb, however, the last in the Folio, Compositor B switches the final two pairings (following the order 1 and 6v, then 1v and 6). Work then began on the preliminaries and Troilus & Cressida (Hinman vol. 2, 517).

Proofreading

24In terms of the proofreading, Hinman cites Cymbeline as one of a few plays which received a relatively great amount of attention from the proofreader, judging from the number of press variants. That is, given Compositor E's lack of skill in setting type, it is not surprising that much of the proofreader's attention would go to correcting E's errors, occasionally glancing at the work of the skilled compositors but primarily allowing their work to go unchecked. For instance, in Timon of Athens, set entirely by Compositor B, the proofreader corrected two pages, demanding only six changes. In Macbeth, set nearly equally by A and B, three pages were corrected, again producing six variants. In the pages set by E, a predictably greater number of variants occurs, and a greater number of pages are checked. Over a third of the pages of Titus Andronicus, set almost entirely by E, were corrected, with 63 variants; twelve of twenty-five pages of Romeo and Juliet, again set predominantly by E, were checked, producing 47 variants. The anomaly appears in the final work by B, in Othello, Antony and Cleopatra, and Cymbeline, where the pages which were proofread produced more corrections and more variants. Of the six proof-corrected pages in Cymbeline (zz4v, zz5v, zz6, zz6v, aaa3, and aaa6v), four were set by E. With Howard-Hill's correction to Hinman's work, the high number of variants in Cymbeline no longer seems irregular.

25The kinds of corrections are, with a few exceptions, representative of common Folio errors: turned letters, printed spaces, spelling alterations (some changed due to preference, others having been mis-spelled or mis-set), and a number of punctuation changes. Of the changes to spelling, a few are notable, especially as possible evidence of transcription by Ralph Crane: "Continwes" was corrected to "Continues," "Tallents" to "Talents," "dampn'd" to "damn'd," and "Incke" to "Inke" -- all four of which are typical of Crane's well-known idiosyncratic style. This is where using a "best page" facsimile can be problematic, as these clues can easily be missed if not correlated with Hinman's careful work in collating those variants here.

Textual Provenance

26The fact that two compositors worked on setting the text might be taken to complicate the matter of determining the provenance of the text. As it stands, we have at least one author, Shakespeare, and the hand of Ralph Crane seems evident, and then Compositors B and E. In fact, having a second compositor's contribution has strengthened the argument that the text from which Cymbeline was set contains a break between TLN 1336 and 1337 (see "Source Text" for further discussion). Also, E is known for his tendency to reproduce his copy more exactly; several typical Crane spellings are found in E's stints, and the proof variations reveal that the accepted policy would be to alter Crane's most unusual spellings. The four spelling changes listed in the previous paragraph are all on pages set by E; presumably, the experienced B conformed voluntarily to the proof-reader's preference for less uncommon spellings.

27Since Compositor E did in fact set 633 lines in Cymbeline, or somewhat less than five full pages, (Taylor and Jowett 254, citing Howard-Hill "New Light"), some in both halves of the text, this raises an interesting point. As demonstrated by Hinman, Compositor E had been barred from setting any copy not based upon earlier printed text. It is unknown why E set part of Cymbeline's text when it had not been printed before, but it is likely that the professionally transcribed copy was deemed clear enough to allow even an unskilled compositor to set with fair success.

Editorial Principles and Practices

28Even the relatively objective process of identifying and reproducing what appear to be invariable data is, ultimately, fuzzy. Press variations and corrections account in part for this difficulty, but it is certainly enough to challenge the desirability of claiming objectivity or factual basis as a goal. Unfortunately, the impossibility of producing anything but a single text, even in hypertext, currently renders the theoretical objection nearly moot. With the best faith in the world on the editor's part, what the reader must perceive at first is a single stable text, no matter how variations with alternate texts undermine that fixity after the fact, or how the author prefaces the edition with cautions regarding the fluidity of texts. Yet, even as I mock these efforts as too little, I must acknowledge that a little is better than none at all. My edition will rely on technological advances and explicit declaration to make its statement (and my own) that neither the Folio nor the modern text should be considered authoritative.

29Conservative editing practices such as those advanced by W. W. Greg, Fredson Bowers, and G. Thomas Tanselle offer a schema by which to produce an edition with at least internal consistency in emendations; at the same time, Jerome McGann's concerns regarding the suppression of the collaborative origins of any published text are well taken. Whereas the former theory can be reformed, however, I see the inability of the otherwise attractive work of social theorists to construct a rationale of editorial practice as an insuperable obstacle to espousing such a belief as the guiding principle to producing an edition. Fortunately, these seemingly opposed views can easily and productively be reconciled by editing conservatively while considering the multiple influences upon the text in the introduction and commentary.

30Many of the questions surrounding the text's provenance and authorship touch upon theoretical concerns central to editing. Even the most pragmatic of studies is fraught with resonance. For instance, as Trevor Howard Howard-Hill summarizes his argument in "Shakespeare's Earliest Editor, Ralph Crane" that

31No part of his [Crane's] activity is so important for us as his involvement with Mr William Shakespeare's Comedies Histories, and Tragedies, one of the greatest editions of dramatic works ever assembled. When we regret how much evidence of the stage was lost from Crane's transcripts as he prepared them for printing, we must also recognize his assiduity in supplying, according to good models, the best texts for the best readers. Fully to identify the effects of his ministrations, many of which are unalterable, requires close and patient study, (129)

32we have to wonder what the multiple functions of such knowledge would be. As we more fully identify Crane's practices and preferences, what is our obligation? If the editor attempts to strip the symptoms of Crane's influence from a modern edition, she risks errors on both sides: inaccuracy, in removing authorial material as interpolated material; or falsity, in the failure to remove unrecognized non-authorial work while claiming to do so. The same arguments are true of, for instance, identifying and studying the patterns of the different type compositors.

33In either case, the works of cultural historians, theorists, and editors have clarified the intellectual pitfalls attendant upon over-valuing the single and singular author as a Romantic relic. Especially because the practices of emendation, annotation, and collation tend to be conservative, emphasizing the traditional aspects of textual editing, it is incumbent upon the editor to help to alert the reader to the multiple social forces present in an early modern text. Although it has been commonplace for decades that collaboration of different kinds was the norm rather than the exception in authoring dramatic texts, the dominant values systems still demonstrate fascination with Shakespeare as a solitary author, ignoring his collaborative works or relegating them to the bottom tier according to supposedly aesthetic tastes. Scions of the New Bibliography implicitly support this view as the best work is considered to be the author's final intention -- a formulation which falls into confusion as soon as a second author is introduced. Authorial shares are notoriously difficult to determine even when the authors are known; often, collaboration is suspected but the hands cannot be identified with certainty, making it more difficult to separate the shares. Plays were subject to revision or addition at multiple stages, including for revival a decade or more after the original performance, but relatively few plays were published in multiple versions recording the different iterations. When such versions do exist, their provenance is often entirely unknown.

34Even when two authors worked together more or less simultaneously on a project (which also was not always the case), they cannot be considered to have originated the text alone. McGann and Randall McLeod have written eloquently and at length regarding the interlocking factors at play in the creation of a text in a virtually limitless web such as the presence of literary influence; direct and indirect sources; multiple authors; commercial pressures; possible revision by a censor or for performance for a revival, for a specific occasion, or for publication; scribal transcription; the editorial process involved in selecting works for publication or choosing a source text from among potentially similar sources; compositors and proof-readers; the vagaries of printing and publishing; and, finally, the opinions, and interventions of the actors who, according to some legends, had the right to modify a proposed play after reading the first two or three acts but who in any case would have a great influence on the writing of a script if the author had in mind a given set of actors, whose skills, flaws, and preferences he knew intimately. While at some point these complexities become non-issues for an editor, at others they carry considerable influence.

35John Jowett's examination of The Tempest reveals the temptations of editing an author rather than a product. Certain of the stage directions in The Tempest appear to fall fairly clearly within parameters we can establish as non-Shakespearean, through vocabulary tests, theatrical problems, and so on. Without an alternate text of any authority, however, changing the wording to reflect our understanding of Shakespeare's practices seems counter-productive. Even with alternate texts of competing authority, editors sometimes base their decisions on capturing the reader's interest instead of on strict adherence to editorial principles. With Crane, since we know from the evidence in the transcripts of A Game at Chess, as discussed by Howard-Hill, that Crane objected to oaths which neither most playwrights nor the censors had a problem with, an editor of a text based on a Crane transcript should be aware that oaths may have been altered, potentially changing the meter or imagery, although Crane seems to have been alert to the desirability of retaining the meter and, in the A Game at Chess witnesses, would often splice together short lines and elide words with apostrophes in order to preserve (or create) regular lines of iambic pentameter. One result of such alteration would be that stylometric analyses might deviate more than anticipated from Shakespearean norms. If Shakespeare's rate of line fragments was typically a certain percent, and a Crane-transcribed play contained a lower percent, a change in Shakespeare's composition practices should not be the foremost explanation. Thus, knowing a text's provenance, while it might not affect the text that is generated, would affect at least the annotations, where the editor can highlight the collaborative nature of the text.

36In terms of practice, I follow the I.S.E. "Guidelines for Editors," which can be reviewed at http://web.uvic.ca/shakespeare/Foyer/Guidelines/index.html. The folio text, as a diplomatic transcription, is as close to the original as the pragmatic constraints regarding computer typographics allow. Catchwords, running-titles, and signatures are preserved, as are, again, to the extent possible, the size of the text and the text's location on the line. Exits in F1 typically share lines with speech and are set toward the right margin but are not necessarily flush right; nor is the amount set in from the right margin consistent throughout. I have attempted to replicate the position on the line, but such positions should be taken as approximate, not precise. All aspects of the text including those which are not at present visible online have been recorded and will be available through full search capabilities, including archaic letter forms such as the long "s," ligatures, diacritical marks, and abbreviations such as w^t and y^e.

37The reader should be aware that individual computer settings will affect the display of the text. The browser settings regarding the font type and size may affect the way the lines wrap within the columns. The best results will be obtained with at a medium setting, but in some cases the display will vary from browser to browser at similar settings. Thus, lines may appear to break at different places despite the most careful transcription. On the other hand, despite the problems it presents for faithful reproduction, the ability for individual users to be able to change the size of the font is not a capability we would want to withhold, as this would unhelpfully restrict the site's usefulness to vision-impaired users.

38Other features of the Folio text which it is impossible to replicate -- and which will probably remain impossible to replicate for the foreseeable future -- include the "fuzzy" state of some of the printing variants: faintly printed letters, letters out of proper alignment (printed high or low), printed space quads, damaged or broken type, or variations in font cannot be reproduced except in scanned pages. A particular gap in a word in which it is impossible for me to tell whether an "e" is printed but inked so lightly as to be imperceptible to the human eye or whether, for some reason, no "e" was set demands a choice: with shame at using such a hackneyed pun topic, I recapitulate this struggle as "to 'e' or not to 'e.'" Non-editing friends would not believe how much time I spent attempting to determine the "true" state of the text, hunched over multiple copies of F1 with my magnifying glass, especially given that the word is indisputable; no editor I can imagine has ever or would ever challenge that "pen trate" should be anything but "penetrate." But the point of a diplomatic transcription is to reproduce what actually was printed, not what was intended. Such decisions do not occur on every line, but when they do occur, some meaning is always lost because one reading is always excluded or marginalized in the most literal sense.

39These ephemera might not be missed at all by the vast majority of readers of a diplomatic transcription, but to one who is accustomed to the textured variation in page color and ink darkness, and who befriends damaged pieces of type, sadly watching the top of the "f" gradually separate until it floats separately like an island, or who cannot help but notice the "o" with a distinctive crack recur over and over, any loss is to be mourned. The three kinds of question marks in the First Folio of Cymbeline are precious; calling one the roman type and one the italic type and losing the third represents a painful loss.

40The other major classification of irreproducible text is the corrected variant. An editor must choose whether to reproduce the original form in the diplomatic transcription or to offer the corrected version. Again, either form represents an emphasis on a different aspect of textual reproduction. If the corrected press variants on zz5v are reproduced, some of the evidence of Ralph Crane's transcription is suppressed; if the uncorrected variants are replicated, the kinds of errors typical in early modern type composition in general and the errors common Compositor E's stints in specific are lost. The presence of a scanned copy of the Folio text linked to the diplomatic transcription offers some consolation for the inevitable losses in appearance that the conveniences of searchable text necessitate.

41Modernizing a text carries its own challenges. The ISE Guidelines establish procedures for a number of issues to provide continuity throughout the editions. In regard to spelling, for instance, the decision has been made to modernize unilaterally rather than to retain certain archaic spellings, no matter how entertaining. The few exceptions include when the modernization would prevent a reader from perceiving word play which the old spelling would convey, or when the spelling is significant in itself, as with a dialect. Neither of these issues comes into play in Cymbeline particularly, with the exception of the ubiquitously problematic travel/travail combination, which I render "travel" at TLN 2018 and "travail" at 1589, relying primarily on the reader's familiarity with the knowledge that the word could mean either or both and on my annotating the matter. The other exception to the rule, which occurs one time in Cymbeline, is the use of the anachronistic, fabricated word "an" for "if," when the original spelling is "and." The rationale is that "the familiarity of the usage recommends 'an' not 'and'" (12), but I would argue that this is one instance in which the Renaissance clearly ought to be unedited: if readers can become habituated to one mode, they can certainly learn another. The ambiguously used "neer(e)" and "farr(e)" (either "near" or "nearer" and "far" or "farther") do not appear ambiguously in Cymbeline. Emphasis capitals are likewise eliminated in the modernization, except when referring to a character by title instead of by name.

42The Guidelines leave it to the discretion of the editor to choose, consistently, to regularize verbs ending in "-est" either to "-est" or "-st"; I have chosen to use "-st" throughout, which matches the meter in F1. Elisions are used as infrequently as possible, with indications of "-èd" when it is pronounced as a separate syllable, which only occurs a handful of times in Cymbeline. The most common cause of elision is not within a single word but when two words are elided (such as "of the," "in the," or "do it" as "o'th'," "i'th'," and "do't") and the elision in the modern text can preserve the meter.

43In terms of punctuation, modern practices generally employ substantially less. The primary difficulty in punctuation is that, to modern readers, many of the sentences appear quite long, with multiple clauses. It is not always possible or desirable to break single sentences into as many individual complete sentences as possible, even if this would make the experience of reading the text more comfortable to the reader. The sentences -- and not only the longer ones -- might also contain grammar or usage errors according to modern standards. I do not believe that correcting grammar is a necessary part of the editorial process, preferring to preserve the language of the F1 text whenever possible. In one of the primary areas of editorial change historically, I have decided to leave grammatically incorrect passages alone whenever the text is intelligible without alteration. The slippery slope of grammatical emendation begins with, for instance, making verb and subject number agree, but extends to exchanging farther for further because of usage preferences; when alterations for usage are allowed, the edition becomes a translation.

44The temptations of such tinkering are evident in the emendation patterns of the eighteenth-century editors -- although not all succumbed to the same degree, as Ann Thompson reports in "Making Him Speak True English"; and while most editors mock the tactics of such grammatical extremists as Alexander Pope, the temptation remains to continue the process of making the text more readable to a modern user by eliminating the confusion imposed by the error. This strategy is familiar to translators, who, in translating from one language to another, are spared the necessity of replicating the errors. Indeed, there is something to be said for the viewpoint that all modern editions are of necessity already translations inasmuch as meaning is not only lost but unavailable when colons for pauses are silently eliminated, when all emphasis caps go by the boards, when the exclamatory question mark is changed to a question mark. Fortunately, the I.S.E. texts, with their links between modern and F1 texts, help to foreground at all times the knowledge of the originary text behind the modern text.

45Another area where alteration of the original under the guise of modernization or correction occasionally appeals is in the characters' names. The Guideline provides a brief and clear injunction regarding the spelling of characters' names: "Retain the traditional modernized forms of characters' names. Avoid the pedantry of the New Oxford in its adoption of such spellings as Petruccio for Petruchio" (12). Despite this, I have given a great deal of thought to the matter for my edition, especially since the New Oxford spellings have been adopted more frequently in critical publications. Imogen, Clotten, Posthumus, Iachimo, Philario, and even Polydore could arguably be spelled differently, which the New Oxford editors do in most cases. I sympathize with the desire of the New Oxford editors to correct tradition; however, usually the case is that eighteenth-century editors have adopted an emendation that has persisted but which should be eradicated, not that they have retained the consistent, even invariant, folio spelling despite arguments that the folio is "inaccurate." While it is a relief to have the Guidelines as a defense, I would retain Imogen and Iachimo in any case. Because they present such interesting test cases, it is worth discussing them here in addition to in the collation and annotation.

46The New Oxford editors alter "Iachimo" to "Giacomo" on the principle that this more accurately reflects a modern Italian spelling of the original. The first question is whether the "I" in "Iachimo" represents an "I" or a "J," since the early modern "I" represented both sounds. Often, the case is that "I" before a vowel is modernized as "J," while "I" before a consonant is rendered "I," as is true in Iohn/John, Iuliet/Juliet and Isabel/Isabel. Exceptions obviously exist, as is the case with Iago, where the metrical demands for a tri-syllabic pronunciation preclude the spelling "Jago." As Iachimo is pronounced in three syllables throughout, except once where it is elided to two, meter cannot be the deciding factor. Roger Warren considers but ultimately dismisses the alliteration of "yellow Iachimo"; I think it is a weighty factor as the only internal evidence as to its pronunciation. It is countered, however, by the contemporary use of "Iachomo" or "Iacomo" as a character name in dramas (in Marston's What You Will and Fletcher's The Captain, respectively) -- which names descend to us as "Jachomo" and "Jacomo" in editions which differentiate between capital I and J. Despite the fact that "I" and "J" were becoming commonly distinguished in print around the time the second Folio was published in 1632, the spelling "Iulius" reveals that the F2 compositors were not yet making that distinction in names, although they did use "j." F3 (1664) is the first that distinguishes in proper names, and preserves the spelling "Iachimo"; F4 (1685) is the first to suggest "Jachimo." In the final analysis, although Jacomo seems a reasonable conclusion, I believe that preserving "Iachimo" better accounts for the greatest number of factors.

47The factors involved in the decision between "Imogen" and "Innogen" are even more complicated. The heroine of Cymbeline, called "Imogen" from 1623 to 1986, is now frequently referred to as "Innogen." There are four primary arguments in favor of Innogen. First, Simon Forman's diary clearly refers to her as "Innogen." Roger Warren scoffs at the idea that Forman could have misheard, that any actor could enunciate so badly. So it is possible to conclude, based on a single eye-witness report, that "Innogen" was the name used in performance. Ralph Crane, when transcribing Shakespeare's bad handwriting, might have mistaken the "nn" in Innogen for an "m" and assumed that it was the same at every subsequent iteration. Likewise, in Holinshed's Chronicles, which Shakespeare consulted, the legendary eponymous founder of Britain, Brutus, had a wife named Innogen. Shakespeare also apparently planned a character named Innogen in Much Ado About Nothing, as the name is included twice in entrances, although she never has any lines. The thematic significance and similarity to the names of other Shakespearean heroines from late plays appeal to many critics. As has frequently been pointed out, Innogen, sounding like "innocence," is similar in kind to Perdita, Marina, and Miranda.

48Upon closer examination, these arguments are not as convincing as they may seem at first. Because we so rarely have eyewitness information regarding a production, it is tempting to elevate this evidence over other evidence, but I feel that doing that too much might be a problem; eyewitness evidence and first-hand reports are notoriously unreliable, as the play itself acknowledges. For instance, the New Oxford editors produce this evidence, wherein the princess's name is clearly spelled "Innogen," not "Imogen," as proof that it should be spelled "Innogen." However, Forman's names do not always match. He uses "Clotan" once and "Cloten" once; this inconsistency does not change the pronunciation much, as the unstressed final vowel would likely be pronounced as a [schwa] in any case, at least according to modern rules. However, he also calls Caesar "Octavius," which name is never used in the play; "Augustus," the preferred name in Cymbeline, occurs five times. Forman refers to Cymbeline as "king of England," even though "England" is never used in the play either, "Britain" and "British" being used over forty times, counting the stage directions (of which Forman obviously would not be aware). It is possible, I would think, that the substitution of one remembered name for another which "Octavius" represents leaves reasonable doubt that "Innogen," perhaps familiar to him as the name of Brutus' wife, is the result of a similar substitution for "Imogen." Also, all of the names Forman mentions are historical names: Lucius, Cymbeline, Octavius Caesar, Clotan, and Innogen. Other characters are tagged descriptively, such as "her love," "the Italian," and "an old man whom Cymbeline had banished."

49Another apparent error includes his statement that Clotten was "banished for loving his daughter." This could be a complicated referent error to Posthumus in the next sentence, but that would be difficult to understand, since the sentence as is reads, "And how [one] of them slew Clotan that was the Queen's son, going to Milford Haven to seek the love of the king's daughter, whom he had banished also for loving his daughter," and the following sentence begins with a reference to Iachimo, not to Posthumus at all, who is mentioned after that. (This confusion between Clotten and Posthumus might partially be explained if the part had been doubled by the same actor; Forman obviously does not tell us specifically.) Obviously, Posthumus should be the one who was banished for loving Cymbeline's daughter. A final mistake is that Forman says that Iachimo hid in a trunk containing plate which was a present for the king. In the first place, Iachimo is lying about having plate, but even the lie states that it is for the emperor, i.e. Augustus Caesar, not Cymbeline, the king. Forman makes similar errors in his account of Macbeth, where his identifications of such characters as "Mackbeth, king of Condon," "Bancko," and "Mackdoues wife" again challenge the reliability of his diary as a phonetically accurate record of Shakespeare's names.

50While some of the differences between Forman's account and ours are natural and should not be overread, such as calling Cymbeline the king of England instead of Britain, in combination they add up to suggest that, like any single eyewitness account, it should not be taken as of greater significance than other evidence; and in fact, the lack of date and confusions of details might even suggest that he recorded his impressions some time after seeing the play rather than immediately, lending further weight to the suggestion that his memories are suspect.

51In Holinshed, the conjunction between Innogen, Brutus' wife, and the other kings of early Britain is undeniable. However, various historians before Holinshed record Brutus' wife's name as "Innogen," "Innoges," and "Ymogen," possibly recording the process of Anglicization, dialectical variants, or simple transcription errors. The two versions may have been linguistically equivalent at some point, such that the historical "Innogen" had evolved in at least one dialect to "Imogen" as "Penbroke" shifted to "Pembroke," so that even if Shakespeare's reading Holinshed suggested the name to him, he might have preferred an alternate form.

52As for the argument that Shakespeare was more likely to have used "Innogen" because he had already used it before in Much Ado About Nothing, logic demands the consideration that it is equally if not more likely that the name "Innogen" occurring only twice is the result of a minim error than that every time Shakespeare wrote "Innogen" in stage directions, speech prefixes, and dialogue throughout Cymbeline, around 150 times, Ralph Crane, a professional scribe, misread the name.

53"Innogen" may sound a little bit like "innocence," but etymologically it is unrelated. Breaking the name into component parts, "innocence" comes from the Latin "in-" meaning "not," and "nocens" or "harmful." Changing the "c" to a "g" not only changes the sound but renders the meaning nonsense. "In" would remain the same, but "nogen" has no resemblance to any relevant roots. Imogen, on the other hand, can be construed as conveying thematic meaning as well. If considered aurally, with "o" pronounced [schwa], it sounds as close to "image" as "Innogen" does to "innocence." Equal arguments aside, it is not necessarily a great advantage to have a heroine whose name is a typically feminine virtue. Arguably, the character "Imogen" has achieved more recognition of complex characterization than might have been accorded to a character with a name like "Chastity." And without being burdened by comparisons to the historical Innogen (whose personality, if any, has at any rate been erased by the process of historicizing), Imogen has been free to develop in ways uncircumscribed by external forces.

54Although often used as a defense for retaining tradition, the argument that a particular word should be preferred simply because it is familiar lacks persuasiveness. Just because Imogen is more familiar to audiences and the general population than Innogen, just as Gertrude is more familiar than Gertrad, Gertrard, or Gertred from Hamlet, does not mean that we should retain it if it is demonstrably incorrect. Yet, even if it were true that Imogen was not the form Shakespeare intended or that was performed in early productions -- which, as I hope I have demonstrated above, is not necessarily the case, preserving the spelling in the most authoritative published version available seems only responsible. "Imogen" is important as a record of the collaborative and material realities of textual production in a way that the purportedly "correct" Innogen is not.

55In other areas, Cymbeline presents fewer problems. Probably thanks in large part to the attentions of Ralph Crane, act and scene divisions and stage directions are fairly regular, with a few exceptions. Following the scene in which Posthumus surrenders the wager to Iachimo, the stage is cleared and Posthumus returns for the misogynistic soliloquy. In F1, no scene break is marked, but because the stage is cleared, I have opted, following the majority of modern editors, to insert a new scene break at TLN 1336. Although I do so with some hesitation, I reverse this practice at 2109, where I remove a scene break. When Imogen discovers the cave of Belarius and the princes, the stage direction for her to exit occurs before Belarius, Guiderius, and Arviragus enter. Technically, the stage is clear, thus calling for a scene division. However, because there is some indication that the cave is considered a special entrance or exit (see the stage direction at 2244 where they enter "from the cave," I consider the action sustained.

56Act 5, scene 2 also contains some controversial scene divisions. As is typically the case in Shakespeare's battle scenes, the stage is cleared several times between separate incidents. To provide a scene break every time the stage cleared between action would be overkill (so to speak). (I once saw a staging of one of the history plays where the lights were dimmed between each scene, and no provision was made not to dim the lights between battle scenes, so the battle ended up being difficult to watch or concentrate on, what with the distracting strobe-light effect of the stage lights blinking on and off every five seconds.) Not to provide scene breaks gives the appearance that all scenes are necessarily set in the same locale, which may or may not be true. Alternatives such as creating new scene breaks between character speeches prove unsatisfying; this attempt at compromise would create even more confusion, since the character speaking would not necessarily be associated with the action before and after.

57At TLN 3032, I remove the scene break creating the original scene 4. The situation is not completely unambiguous, but there is obviously some confusion. After the battle and Posthumus' "arrest," the stage direction reads, "The Captains present Posthumus to Cymbeline, who delivers him over to a Jailer," (3030-31) followed by the scene break. No exit is provided before the scene break; however, one is evidently necessary for Cymbeline and the Captains. Immediately following the scene break is a direction for Posthumus and Jailer to enter, even though two jailers have lines. A less intrusive change would be to have the two groups, Cymbeline's court and Posthumus and the Jailer, exit separately, with Posthumus and the jailers re-entering to represent a shift in location to a different part of the British camp, but such re-entries are somewhat rare. I have opted instead to follow the majority of modern editors by removing the scene break and Posthumus' and 1 Jailer's entrance while adding an exit for Cymbeline and the others.

58Concerning stage directions, I have attempted to skirt the fine line between directing and describing rather than inserting every stage direction implicit in the text. I have added a number of directions for asides, more for characters who are speaking apart, and some for directing a speech to a particular character or characters; I am not positive that all of these are necessary, but I do have a high degree of confidence that they are accurate. Few of the entrances or exits require moving. At 574-75, Pisanio exits before the Queen has an opportunity to tell him, "Think on my words" and begins insulting him behind his back and planning his death; as "Exit Pisanio" would not have fit on the following line, the reason for the misplaced exit is obvious. Character reactions that seem incontrovertible and obvious likewise receive no direction; props, however, are noted, as are directions for actions that are not instantly obvious to the reader.

59I prefer to be inclusive rather than exclusive in my stage directions, as with my annotations, suggesting complex, multiple, or alternative readings. At TLN 887, as I do elsewhere, I have chosen to give a choice in the stage direction rather than make the decision for the reader. I feel that the benefits of being able to choose outweigh the potential risk that readers might think that when I say "Exit Clotten or 1 Lord," "or" means that it doesn't matter. On the contrary, I feel that, while either is defensible both editorially and in performance, the kind of society depicted will differ radically. Depending on which character stays onstage to await the previously humorous Second Lord's serious soliloquy about Imogen, the balance of the play shifts from one in which people who have power frequently abuse it, as Clotten and his mother (and, arguably, several other characters) do, to one in which the abuse of power by authorities is rendered even colder by the complicity of sycophants who are willing to enforce the dominant party's will.

60At TLN 913, for similar reasons, I add a "reader's choice" stage direction for Imogen's lady. After Imogen asks Helena to wake her, Helena patently is no longer a physical presence on the stage, but no stage direction is given for her exit. While it is possible that an exit was simply omitted, it is equally possible, given both the social practices of early modern England and precedent in at least one source, that Helena sleeps in Imogen's chamber. Once again, the difference is palpable on stage. Without another woman's presence, the intimacy is perhaps greater, but this effect is offset in the alternative by how Iachimo's boldness and audacity are emphasized by his sneaking around in Imogen's chamber without waking either her or her lady. Also, if Helena were onstage throughout, this might increase the feeling of voyeurism which I feel is undeniably part of the author's goal.

61Although it is often pleasant to be surprised when reading or watching a play, I have opted to follow the folio text in referring to the disguised Belarius, Guiderius, and Arviragus by their true names, as do most editors, with parenthetical references to their assumed names. This is the more defensible here as Belarius so quickly reveals their true identities. Because they not only have false names but false identities, however, it is preferable to remind the reader of the fictitious nature of their identities by repeating that they are in disguise as Morgan, Polydore, and Cadwal. This also provides the reader plenty of opportunity to contrast the Roman nature of their true names with the British flavor of their false names.

62In only a few places have I felt obliged to alter a speech prefix. At TLN 1187, I follow tradition in giving a speech originally assigned to Posthumus to Philario. Then, at TLN 2308-09 and 2312, Arviragus is assigned two speeches in a row. I have opted to add the first of these onto the end of Guiderius' preceding speech rather than giving it to Belarius. It is taking Belarius and his charges a long time to get out the door in order to go hunting since Imogen is sick and they are taking leave of her. Guiderius has already tried to offer to stay home with her and take care of her, and both brothers have insulted Belarius by saying that they would rather that Belarius die than Fidele. Belarius' first line urges them to go hunting; after he is insulted, he prods them again, reminding them of the hour; he repeats "To th' field, to th' field" (2297), in his next speech; and his following speech reminds them, "It is great morning. Come away" (2324). Between his comments to Guiderius and Arviragus trying to get them to leave, he is polite to and about Fidele, wishing her well and complimenting her upbringing, but he does not seem to me to be in the mood to hang about rhapsodizing with the others. The two speeches I have put together have the advantage of both being about Imogen's cooking, whereas if I joined the two speeches prefixed to Arviragus, Guiderius would have one brief hemistich and Arviragus would discuss her cooking for a few lines and then, without transition, her sadness for several more.

63A choice occurs at TLN 981, when the music Clotten wants for Imogen is performed. After Clotten says, "First, a very excellent good conceyted thing; after a wonderful sweet aire, with admirable rich words to it, and then let her consider" (977-80), the direction merely reads, "Song." Presumably, the musicians would not have been brought onstage if they were not going to play, but it is not clear whether Clotten joins in. Michael Shapiro points out that serenades became watchwords for humorous attempts at seduction, virtually never successful. At the same time, noble characters rarely sing, and only after much persuasion. No new speech prefix is given between Clotten and the song, but this cannot be read as conclusive in one direction or the other. If the convention holds that nobles would not sing onstage, Clotten would merely hire the musicians and listen to them play. However, Clotten is certainly a clownish enough character to illustrate that point by singing in contravention of proper behavior. (An accompanying question is, if Clotten sings, whether he sings humorously badly or whether he sing well. Clotten is, after all, not entirely a fool.) Thus, I have chosen to indicate the uncertainty by inserting the speech prefix, "Musicians and possibly Clotten."

64I have relineated verse in several areas, but only a few require special commentary. In particular, I would draw the reader's attention to at 3.1, predominantly in verse, where Clotten speaks in prose. This is not by itself unusual; Clotten, like Iachimo, often speaks in prose rather than verse. What is worth noting, but has not, to my knowledge, been commented upon before, is that if Clotten's speeches were removed, the lines preceding and following (TLN 1388 and 1393, and 1412 and 1424 ) would combine to make whole lines of iambic pentameter (apart from Cymbeline's order, ignored by Clotten, to let the Queen finish speaking, which is essentially injected into one of Clotten's speeches). This suggests that Clotten's speeches are interpolations, though whether they simply record interjections that the actor playing Clotten originated, whether they were inserted in a revision, or whether they are in fact part of the original composition and Clotten's speeches are meant to be felt as the interruptions they are by making the speeches seamless around him, we cannot know. Clotten's third and final speech of the scene follows a whole line of iambic pentameter, and Lucius' response, "So, sir" (1464), seems added only as a brief response. Reading Lucius' speech, skipping Clotten's speech with Lucius' terse response, then reading Cymbeline's next speech creates a kind of continuity not accessible with Clotten's speech. This helps to explain why Clotten's first speech is set as verse even though it utterly fails to scan. If I am correct and Clotten's lines are interpolations, this alters the international politics of the text. Many scholars have cited Clotten's speeches in this passage as explaining why Cymbeline decides to pay tribute again even though Britain defeats Rome. Clotten's speeches, which seem to echo the nationalistic patriotism of, say, Elizabethan England, in context of the presumed date and James's pacifistic tendencies, are taken as being too overbearing. Patriotism may be well and good, but his aggression and defensiveness are unnecessary. Of course, since we do not know the nature of my hypothesized interpolations, it is possible that they were present as far back as the first performance of Cymbeline; and even if I suspect them of being added after the first phase of composition, revising our entire understanding of the relations between the two countries might be premature.

65By far the largest category of lineation differences between the folio and the modern text is in the arrangement of hemistichs. In the folio, of course, no indication of half-lines that can be combined to form a whole line of iambic pentameter is present. It is modern practice to attempt to preserve the verse by giving a greater physical appearance of verse: the second half of a line is indented in order to show that it completes the iambic pentameter of the first; three thirds are spaced across the line. Unfortunately, in a few series of part-lines particularly, all of the parts do not add up to a given number of whole lines but have one or more part-lines left over. Such occurrences happen at TLN 218-24, 650-57, and 680-84.

66A different lineation difficulty arises with the ghosts' scene. They speak in fourteeners, but each fourteener is usually set as an 8- and a 6-syllable line. Not only do they switch to fourteeners on the next page when it appears that the compositor is not constrained by the need to space out the text, but the 6-syllable line always begins with a small letter, not a capital, following the convention for carrying over a verse line. The argument for setting the lines as broken in two is that they do have regular ballad-like 8-syllable breaks, although they do follow a balladic rhyme scheme. The caesura, when present, is usually after the 8-syallable, but it is not always present and varies somewhat. I prefer to follow the typographical evidence of F1 that the lines were intended to be fourteeners than to adhere to tradition by breaking the fourteeners into ballad meter.

67I have aimed for consistency in my emendations and annotations, but at the same time, the individual circumstances of different cases in context make it difficult to be unswerving. In many instances, even the editor most desirous of consistency must acknowledge that individual situations require judgment calls. Most questions are of degree rather than kind; and the weight of tradition is surprisingly difficult to ignore. Nevertheless, I hope that my attempts to use stage directions and the annotations to reveal the possibilities of the text rather than close them off will help open the editorial practice to allowing -- or forcing -- the reader to make some of the kinds of decisions that editors have previously made.

68Not incidentally, I am torn about the question of re-humanizing the editorial process. Arguably, promoting the identity of the editor -- not in the sense of advertising, but in the sense of highlighting the fact that there is an editor -- would play an important part in encouraging the reader to participate in the collaborative process of making meaning. Just as it has become mandatory to realize that histories, for instance, are not neutral portrayals of fact but are always inherently biased in content and presentation, so the editing process is not neutral. Insofar as an editor is a narrator -- in the introduction, in the annotations, in the primarily silent textual emendations -- it is helpful to highlight the presence of the figure whose judgment determines what the reader reads. A fine line exists between making obvious the fact that a human editor is present and promoting the identity of the editor, but when what is said and how it is said are changed, the opportunity not simply for more knowledge but for new kinds of knowledge increases, which is one of the most valuable goals of literary and textual studies.