Year's Work in English Studies 2004: Shakespeare: Editions and Textual Studies

    Three substantial editions of two Shakespeare plays appeared this year. Suzanne Gossett edited Pericles for The Arden Shakespeare and Roger Warren edited the same play for The Oxford Shakespeare, and John Jowett edited Timon of Athens for The Oxford Shakespeare. The coincidence of two Pericles editions is fortuitous since it allows for direct comparison of the latest work from the two most important series, although in the event the dating of Warren's edition is not certain: the review copy claims that the book was first published in 2003 while the publisher's website gives January 2004 as the occasion of both hardback and Oxford Worlds Classic paperback versions. Precise dating of books will soon matter greatly to academics in the United Kingdom because of the state's audit called the Research Assessment Exercize. We may expect a rush of books officially published in December 2007 (to fall within the current census) that are not available until early 2008, but this cannot explain why The Arden Shakespeare should choose, in October 2005, to sell copies of its third series Much Ado About Nothing with an imprint claiming it was "First published 2006". Its American editor Claire McEachern has nothing to fear from the Research Assessor General.

    In her preface Gossett announces that she wants to maintain a "typically postmodern diffidence" about proposed solutions to the problems of the text of Pericles, for which the only authoritative early edition is the manifestly corrupt quarto of 1609. Because Pericles is the only one of the seven plays that were added to the third Folio in 1664 to be accepted as Shakespeare's, and because it is the only 'bad' quarto for which there is no corresponding 'good' quarto, Gossett deals firstly with the text in her introduction. The 1609 quarto was Henry Gosson's first attempt at publishing a play, although in this case the text had already been entered into the Stationers' Register on 20 May 1608 by Edward Blount, who did not publish it. Perhaps, Gossett wonders, the reason was that once Nathaniel Butter published George Wilkins's prose novelization of the play The Painful Adventures of Pericles Prince of Tyre in 1608, Butter could claim precedence on the story. Philip Edwards showed that the sheets of the quarto printed for Gosson fall into two groups: ACDE and BFGHI, distinguished by different running heads, lines-per-page (37 versus 35) and founts. Variants in speech prefixes, capitalization, spelling, and punctuation indicate that the first set of sheets was set by one compositor (X) and the second set of sheets was set by two (Y and Z). From the ornaments on the title-page and the first page W. W. Greg identified the printer of the first set of sheets as William White. From the distinctive pica roman type, Peter W. M. Blayney identified the printer of the second set as Thomas Creede, and subsequent work has suggested that there might have been a third compositor at work on F4v in Creede's shop. By successive damage that is worst on sheet B, S. Musgrove showed that this was printed last by Creede, probably meaning that a whole initial run of B by White had to be discarded and redone by Creede, although this raises the problem of how Creede fitted 8 x 37 lines (White's norm for a gathering) into 8 x 35 (Creede's norm for a gathering), a difference of 16 lines. Gossett wonders if the start of 1.2 is garbled because some lines were cut to fit the material on B; against this is the fact that space is wasted on B3r, which is the same forme as B1r (where 1.2 starts). Blayney pointed out that Creede's type was smaller, so fitting the extra 6% (2 lines per 35-line page) would not have been especially difficult. That two printshops and three compositors did the work on the quarto is good reason to suppose that the problems with the book (which sprawl across these divisions of labour) come from the underlying copy not the printing work (pp. 18-20).

    Turning to specific sources of error, Gossett shows some examples of how mishearing errors could look like misreading errors, and vice versa, and notes that what is heard in the theatre can be intentionally ambiguous. Thomas Heywood's complaint about his plays being "copied onely by the eare" (Epistle 'To the reader' in his Rape of Lucrece 1608) is exactly contemporary with Pericles. If Pericles was so copied to make the manuscript underlying the quarto, we cannot tell where or why (illicit or authorized?) nor much about how it was done (using shorthand? by one person or a group?), so we cannot tell whether a supposed feature of the text (say, its being short) is due to failure in transmission or is accurate transmission of text shortened for specific performance conditions (pp. 21-22). Gossett summarizes the New Bibliographical theory of memorial reconstruction for 'abridged rural prompt-book' creation, and this theory's recent dismissal--not least via Paul Werstine's rejection, on the evidence of the Records of Early English Drama, of the inferred link between touring and 'bad' quartos--brings her to Laurie Maguire's conclusion that the quarto may be based on a memorial reconstruction, to judge from its mangled verse. If it is a reported text, Maguire maintained, then it is a good one and although it is a bit short it is longer than any other 'bad' quarto. Gary Taylor's view was that the text underlying the quarto was reported by the boy who played Marina and Lychorida and the hired man who played a number of small parts including Pander and a fisherman, helped out by the boy borrowing or stealing the cue-script of his master playing Gower. For Taylor, Wilkins's prose novelization is also a report of the performances, and where it differs substantially from the play as we have it (most obviously in the brothel scene dialogues of Marina and Lysimachus) the reason is censorship of the stage version. On balance, Gossett does not think the text of Q supports the claims of censorship or of particular actors doing memorial reconstruction (pp. 23-25).

    The claim of stenography (shorthand copying) of the play has come up a number of times, and Gossett does not reject it altogether, but she rejects the idea that publishers would have pirated at all since Blayney has shown how relatively unprofitable play publishing was. (This 'demonstration' by Blayney is refuted in an article in Shakespeare Quarterly 2005 by Alan B. Farmer and Zachary Lesser, which will be reviewed here next year.) Gossett is gently mocking of Doreen DelVecchio and Antony Hammond's claim (in the New Cambridge Shakespeare) that Q was printed from authorial foul papers and that its cruces are generally explicable (no, they are gibberish) and of their claim that the whole of the play is by Shakespeare. Gossett asks how come there are a couple of speech prefixes of "Omnes" in Q, since Shakespeare's authorial papers did not use that Latinate prefix? Gossett thinks we will not solve the mysteries of Q's transmission, and that in all likelihood a number of different corrupting forces are simultaneously at work: the copying of difficult handwriting, reporting by actors, dictation, shorthand, additions/revisions, and relining by compositors necessitated by faulty casting off. Indeed, if we accept the evidence for widespread manuscript circulation, there is little hope of pinning down the copy for particular printings (pp. 26-28). This is indeed typical postmodern diffidence in relation to editing. Gossett goes on to survey the early reprints of Q (pp. 30-38), noting that Q2 (a direct reprint of Q1) was again divided in labour between White and Creede, with the latter using quite a bit of standing type in places. Q2 was made late in 1609, to judge from how Creede's distinctive pica roman type becomes contaminated with other types he owned. Q3 was printed by Simon Stafford and another unknown printer for an unknown publisher and is unremarkable. Q4 is one of the Pavier quartos and shows considerable effort to improve the text, although the printer was working from just Q3 and his own wits; Q5 and Q6 are unimportant.

    Gossett considers three reasons that have been offered to explain Pericles's absence from the 1623 Folio: i) Heminges and Condell could not get copy, ii) they did not like the state of the text, and iii) they knew it was only a collaboration. The first is improbable: the men involved (Blount, Jaggard) clearly had access to the rights and to copy: the former entered the play in the Stationers' Register and the latter printed Q4 for Pavier. Certainly as actors Heminges and Condell would have spotted how bad Q is, but Gossett observes that this consideration does not seem generally to have stopped publishers from printing plays. True, but when one considers how bad a Folio printing of Pericles based on Q would have looked alongside other texts in F, perhaps the explanation has some merit. After all, of the bad Folio texts only Timon of Athens is about as bad as the 1609 quarto of Pericles, and as John Jowett discusses in his edition (reviewed below) Timon of Athens got into F by unusual circumstances. Having excluded the first two possibilities, Gossett decides that the collaboration issue is most likely what kept Pericles out of the 1623 Folio. Henry 8 and The Two Noble Kinsmen are collaborations too, of course, and they were included in F. But perhaps these were collaborations of a different sort: Henry 8 being a Shakespeare play that Fletcher completed and The Two Noble Kinsmen being a Fletcher play that Shakespeare completed, rather than being (as Pericles clearly is) a collaboration from the ground up. The conditionals rather pile up in this speculation, and Gossett's characterization of the natures of Henry 8 and The Two Noble Kinsmen could usefully have been elaborated.

    Given the mess of the play's only authoritative source, the best one can do in such a case is make a "credible, bibliographically defensible, reading and performance script" that steers between the freely emended reconstruction that appeared in the Oxford Complete Works of 1986 (which versified passages of Wilkins's prose novelization) and the absurdly staunch adherence to copy shown by the New Cambridge Shakespeare, which follows Q in almost all its demonstrable errors (p. 39). Complaining of the New Cambridge Shakespeare, Gossett repeats the views given in her essay within the collection In Arden: Editing Shakespeare reviewed here last year, and complaining of the Oxford Complete Works's liberties she minds especially the invention of scene 8a in which Pericles asks for a musical instrument upon retiring and then plays it. This the Oxford editors justified on the basis that Wilkins's novelization and the sources have Pericles playing and singing and that Wilkins probably would not have made up such a scene if it was not in the play as performed. Gossett, on the other hand, thinks that Simonides's morning-after reference to Pericles's music might have been "inserted specially to cover the absence of a song". Over against such speculation, Gossett's approach will be what she calls "moderate" (pp. 40-42). Importantly, Gossett thinks that new interpretative readings can make old emendations redundant, for example Edmond Malone's alteration of Marina's wondering "Why would she have me killed now? As I can remember . . " to "Why would she have me killed? Now, as I remember . . ." that removes a "precise adverb of time". Generally, the examples are the ones discussed in Gossett's In Arden: Editing Shakespeare essay. A particular problem that Gossett identifies is the first meeting of Lysimachus and Marina, which in the Oxford Complete Works reconstruction is patched from Wilkins. Surely, she argues, this is self-contradictory: if the boy playing Marina was one of those doing the memorial reconstruction (as the Oxford editors claim), how come he left out bits of what he himself said in this scene? Gossett also thinks that padding this part of the play contradicts the principle that one should use Wilkins more confidently to patch the parts of the play that he wrote (Acts 1 and 2) than to patch the parts he did not. We can use Wilkins to make a stronger and more feminist Marina, but then we are in danger of imposing our values, not the authors', Gossett writes (p. 44).

    Concerning the circumstances of the play's creation, Gossett gives a compressed account of the reasons for pinning the date of first performance to April-June 1608 in which some slight weaknesses are glossed over (p. 55). For example, the dating depends upon determining when the wife of the French ambassador to London was present in London and her not being recorded there until April 1607 does not mean that she was not around, so we cannot with certainty exclude the possible early dates for first performance identified by Leeds Barroll of May-June 1606 or one week in April 1607. On the other hand, as Gossett quotes Barroll noting, the title-pages of Wilkins's 1608 prose novella and the 1609 play quarto make the performances sound recent, which would tip the balance in favour of the last of the three slots identified by Barroll: from April to mid-July 1608. This third slot Gossett silently changes to "between April and June 1608" as though the first two weeks of July were not possible dates. Presumably she means that we must allow time for the play to become a hit before the closing of the theatres by plague (to explain the contemporary allusions to its success on the stage) in which case the latest possible date for first performance was late June, ending its run just over two weeks later in mid-July. Gossett recounts Shakespeare's knowledge of Wilkins and the fact that his dramatic career took off around 1605-8, so he must have been intensely frustrated by the stalling of that career by the plague closure that began in 1608. Gossett thinks that Wilkins's play The Miseries of Enforced Marriage might be a play-long expansion upon the potential of the first scene of A Yorkshire Tragedy, in which case the extra dialogue for the Lysimachus/Marina brothel meeting in Wilkins's prose novelization of Pericles might also be due not to the incompleteness of Q but "the collaborator's attempt to fill out the logic of the conversion". That is to say, pace Taylor, the novelization includes extra material Wilkins wrote to fill out the scene (pp. 55-57). Likewise, when Wilkins "elaborated on some elements abbreviated in the play" he might also have added "Pericles' singing", which Taylor took to be an omission from Q caused by faulty reporting. Gossett also wonders whether "a hungry actor or group of actors", impoverished by the long closure extending from mid-July 1608 to January 1610, "cobble[d] together the playtext as they remember it from performance" to get some money from the publisher Gosson in 1609. Wilkins may also have had a hand in the piracy of Pericles for Gosson; they certainly were acquainted. This would give the King's men a second reason (the first being the novelization) for disliking Wilkins (pp. 61-62).

    Gossett's consideration of "Evidence for collaboration" (pp. 62-70) adds nothing new to what has already been shown in Defining Shakespeare by MacDonald P. Jackson reviewed here last year. Gossett's look at the "Sources of Pericles" (pp. 70-76), however, refreshingly follows up the previous section's premise by considering Wilkins's sources, including his connections to those with first-hand experience of the famine at Barbary that underlies the famine at Tarsus in the play. Regarding "The theatrical context" (pp. 76-81) Gossett contextualizes the play's style (travel/travail play, Romance, brothel drama) within what other dramatists were doing; apparently there was a vogue for choruses and brothel scenes around 1604-10. Gossett thinks that the goddess Diana probably descended to make her entrance in the play and quotes C. W. Hodges (From Enter the Whole Army) who thought the effect would have been like the descent of Hecate in Macbeth. It is worth recalling, however (and Hodges admitted it, p. 124) that Hecate's scenes in Macbeth are probably Middletonian interpolations for a revival in the 1610s. No play text that can be reliably associated with the Globe before 1609 has a character descend from above, so presumably there was no descent machine there until one was fitted as part of the alterations that allowed for regularization of indoor-hall playhouse and open-air amphitheatre practices (having musicians sit in the stage balcony rather than inside the tiring house, having four intervals instead of continuous performance) when the King's men got use of the Blackfriars as well as the Globe.

    The issue of intervals matters because in her analysis of the play's structure (pp. 81-86) Gossett shows that it is essentially "constructed in seven sections" punctuated by Gower's choruses. Q only erratically marks Gower's entrances and exits, and this inconsistency the Oxford editors used to justify their decision in a number of cases to merge a chorus into the succeeding scene and in others to make the chorus a whole scene on its own (for example, their scene 10), based on the continuity markers of i) properties on the stage (heads, tomb, altar), and ii) the entrance of characters while Gower is talking about them. As Gossett points out, the properties and the entrances are editorial anyway, so this argument is circular. True, but it is not quite a failure of logic (the circumstance where circularity becomes a sin) since once one has accepted the conjectured properties and entrances the merging of the choruses with the scenes follows as a consequence. Gossett thinks that the erratic marking of choruses in Q is not necessarily dramatically intentional at all; after all, the horizontal rule that sometimes accompanies Gower is the work of compositor X in White's printing house; Creede's men Y and Z do not do it. Gossett explains her edition's "retaining the conventional five-act division" as a "convenience of readers", although to indicate "the narrative structure" she numbers the choruses from 1 (the opening lines of the play) through 8 (the closing lines of the play). Also, Gossett regularizes Gower's entrances and exits so that he consistently comes on just before he speaks a chorus and leaves as soon as he has finished one. Gossett's summary of the "Production history" (pp. 86-106) need not detain us, nor her examination of the "Interpretation and critical history" (pp. 106-61). Both are exemplary.

    Turning to the text of the play itself, it is worth noticing that Gossett's decision to have the collation and the notes on the same page give more opportunity to explain her emendations than Roger Warren has with his collation tucked away (as in the Penguin edition) within an appendix at the back of the book. Indeed, by comparison with Gossett's scrupulous documentation, Warren seems hardly to bother explaining his particular choices, and where he departs from Q in the same way that the 1986 Oxford Complete Works does he does not even record the fact in his collation. On the textual minutiae, Gossett's is much the better edition. Here I propose to consider only the most interesting of her emendations of the dialogue--not the stage directions, since practitioners ignore those anyway--with special attention to the ones where Gossett offers a new solution not tried before. Or rather, I am concerned with just one stage direction: at 1.1.0.2 when Antiochus, Pericles, and followers enter, Gossett adds "[, including Musicians]", apparently thinking that because Antiochus calls for music there have to be musicians on the stage. This is mistaken, since Paulina calls for music as she pretends to wake the supposed statue in The Winter's Tale but there are no musicians present. ('Present', of course, is a slippery notion here: once the musicians are lodged in the stage balcony their position in relation to the stage and the supposed location is productively liminal.) Gossett thinks that the musicians entering here is "simpler" than having them enter when they are needed, and yet she does not call for musicians in scene 3.2 (Cerimon raising Thaisa) even though music is called for and Cerimon's commands indicate that he is not making it.

    I found just two errors in Gossett's collation. The first is that she gives "[PERICLES] Will 'schew no course" (1.1.137) where Q has "Will shew no course" and her collation reads "137 'schew] this edn (Theobald); shew Q; shun Malone". In fact, the 1986 Oxford Complete Works also has this reading, and since the collation is supposed to show "the earliest edition to adopt the accepted reading" Gossett's collation should reflect that. If this error is to be corrected in reprinting of Gossett's edition, the opportunity might also be taken to fix the line-broken ellipses (that is, the three dots appearing on different lines) in the note explaining this choice and in the notes to 1.2.30. Making up for this omission is the correct collation for "[ANTIOCHUS] for the which" (1.1.144) where Q has "for which". The Oxford Textual Companion, oddly enough, attributes the emendation to "This edition (G[ary] T[aylor])" whereas in fact (as Gossett rightly records) it was introduced by Malone (1778-80), and it has been widely used ever since. The second error is that she gives "[THALIARD] He scaped the land to perish at the seas" (1.3.28) whereas Q has "at the sea", and claims this emendation for her edition, from a suggestion by Richard Proudfoot. In fact, the emendation is in Peter Alexander's 1951 text and is so mentioned by C. J. Sisson (New Readings 2:289). Gossett prints "[PERICLES] The rest--hark in thine ear--as black as incest. | Which by my knowledge found" (1.2.74-75) where Q has only a comma between these lines and justifies this by claiming that the stop is needed to "allow time for Helicanus' reaction". This seems a bit over-prescriptive of the acting and moreover produces an unwanted sentence fragment: "Which by my knowledge found, the sinful father | Seemed not to strike but smooth". Equally informal and at the risk of confusion is Gossett's use of the colloquialism "quotes" to mean quotation marks (1.2.90n). Where Q has Helicanus say "or till the Destinies doe cut his threed of life" Gossett has "Or the destinies do cut his thread of life", which solution was proposed for this edition by Richard Proudfoot and requires some gobbling of syllables if the metre is to be retained (as though it were "Orth' destinies . . .").

    Gossett has Cleon seek "To know from whence he comes and what he craves" (1.4.79) where Q has "to know for what he comes, and whence he comes, and what he craues". Q is repetitious ("for what he comes" means the same as "what he craves") but it makes perfect sense--for that reason Warren adopts it in favour of the Oxford Complete Works's wild emendation--and it is hard to see why Gossett emends at all. Gossett, following Theobald's emendation, has Gower end his second chorus with "this 'longs the text" (2.0.40) where Q has "this long's the text". Gossett's discussion of the possible meanings ('this belongs' versus 'this lengthens') leaves out Jeffrey Masten's suggestion (Textual Intercourse p. 89) of 'this long is the text', meaning 'I've no more to say'. Richard Proudfoot supplied the emendation, new to this edition, that makes the 1st Fisherman say "I'll fetch'ee with a wanion" (2.1.17) where Q has the allegedly unpronounceable "fetch'th'". In the same scene, the 2nd Fisherman says "If it be a day fits you, search't out of the calendar and nobody'll look after it!" (2.1.52-4) where Q has "if it be a day fits you | Search out of the Kalender, and no body looke after it?". This is essentially Gossett's own invention (with a hint from Theobald) based on the uncertain guess that the line means something like 'if this day suits you, you can take it from the calendar and nobody will object'. In the crucial tournament scene, Gossett sticks to Q in having Simonides not expound the meanings of the imprese of the second knight (of Macedon), and the third (of Antioch), and the fifth (no place named) (2.2.16-45). The collation becomes frantically busy here, but the main thing is that the Oxford Complete Works (followed by Warren) fills in the missing explanations using Wilkins's prose novelization, and gets from there the placenames (not given in Q) for the fourth knight (Athens) and the fifth knight (Corinth), although they are not in that order in Wilkins book. Warren, a man of the theatre, rejects the "unbelievably clumsy" (6.16.3n) Oxford Complete Works's stage direction that makes Thaisa hand each shield to Simonides. Gossett and Warren agree on the 'improvements' of the foreign-language mottoes.

    At 2.4.30-32 the 1st Lord says "If in his grave he rest, we'll find him there. We'll be resolved he lives to govern us, | Or dead" where Q has the nonsensical "If in his Graue he rest, wee'le find him there, | And be resolued he liues to goueverne vs: | Or dead". The latter seems to say that if they find him dead they will resolve that he lives to govern them, which is nonsense; the emendation is Gossett's own. In the same scene, Gossett has Helicanus say "A twelvemonth longer let me entreat you | To further bear the absence of your king" (2.4.45-6) where Q has "A twelue-month longer, let me intreat you | To forbeare the absence of your King". Gossett's is the first edition to put into practice Samuel Bailey's suggested emendation. At 2.5.24-27 Gossett explains why, although Simonides here thanks Pericles for his music, she chooses to have him not play any. In the same scene (2.5.72 SD, 76 SD) there are stage directions indicating that Simonides speaks aside, and these comes directly from Q. Gossett's note relies on Alan Dessen's work: ". . . in the Shakespeare canon, only here, in 'a suspect part of Pericles', and in the bad quarto of Merry Wives does aside mean 'speaks aside' Dessen, Recovering, 51)". Unusually for him, Dessen is wrong about this: there is an "aside" direction accompanying Tamora's "Why thus it shall become . . . anchor in the port" (4.4.34-8) in the Folio text of Titus Andronicus and editors universally retain it as a correct marker of how she speaks these lines. To clarify the opening direction of the scene of Thaisa's recovery, Gossett prints "Enter Lord CERIMON with a [Visiting] Servant [and a Poor Man]" (3.2.0.1-2). This makes clear that the servant is not Cerimon's own (the ensuing dialogue is about this man's dying master) and she adds another man that the dialogue makes clear must be present. While this clarifies the situation, it is slightly odd to provide an adjective that cannot aid an actor: how can one convey being on a visit? To explain Cerimon's description of the apparent corpse being "Shrouded in cloth of state" (3.2.63), Gossett glosses this as "material reserved for royalty. Sumptuary laws dictated what classes and categories of people could wear certain clothes or fabrics . . .". To my ears at least, the simple past tense here implies that the laws were still in force when the play was written and first performed; in fact they were repealed in 1604.

    In the essay reviewed here last year, Gossett argued for Q's reading in which Pericles swears that "vnsisterd shall this heyre of mine remayne", and yet in her edition she prints "Unscissored shall this hair of mine remain" (3.3.30). Gossett's introduction (pp. 46-48) discusses this crux and fails to settle the matter, so one would expect that sticking to Q were preferable to using George Steevens's emendation as she has done. Gossett has Thaisa say "I well remember, even on my groaning time" (3.4.5) where Q has "my learning time"; the emendation is Adele Davidson's and Gossett is the first to use it. At 4.1.13-15 Gossett has Marina say "To strew thy grave with flowers; the yellows, blues, | The purple violets and marigolds | Shall as a carpet hang upon the green" where Q has "to strowe thy greene with Flowers, the yellowes, blewes, the purple Violets, and Marigolds, shall as a Carpet hang vpon thy graue". Although the first change (greene -> grave) happened with F3, the corresponding and matching second reversal (grave -> green) originates with Gossett here. In the next scene, Q has the Bawd repeat himself: "and they can doe no more then they can doe, and they with continuall action, are euen as good as rotten". Removing the third 'they' (4.2.8) is Gossett's innovation, but I cannot see why it is necessary at all: the Bawd's speech may remain awkwardly repetitious without problem. In this scene Gossett makes Marina say "Untried I still my virgin knot will keep" (4.2.139) where Q has "Vntide", which is the opposite and impossible meaning (she wants to keep her knot tied and unbroken, surely). The emendation is Proudfoot's suggestion and Gossett is the first to adopt it. In a note Gossett rightly complains that editors have overlooked this illogicality and she suggests that the reason may be "the complex of contradictions, real and semantic, surrounding virginity".

    In having Cleon say "Thou . . . dost use thine angel's face | To seize with thine eagle's talons" (4.3.45-47) instead of Q's "Thou . . . doest with thine Angells face ceaze with thine Eagle's talents" Gossett is originating a new emendation. For Gower's "So with his sternage shall your thoughts go on" (4.4.19) where Q has "So with his sterage, shall your thoughts grone", the change of grone to go on comes from Malone, but sterage to sternage is Gossett's own emendation and for support she cites the parallel moment in the chorus to that exhorts the audience to "Grapple your minds to sternage of this navy" (Henry 5 3.0.18). In both cases, the audience are to hook their minds to imagined ships. Finally among the emendations are Lysimachus's "For me, be you bethoughten that I came" (4.5.113) where Q has "For me be you thoughten, that I came"; this is Proudfoot's suggestion and Gossett is the first to adopt it. The justification is that thoughten is the OED's only example of this word and nowhere else does Shakespeare use it, while he frequently uses bethink and bethought and from those bethoughten is not much of a leap. In David Hoeniger's Arden2 edition of the play, Harold F. Brooks suggested that Pericles should say "Night-oblations to thee. Dear Thaisa" (5.3.71) where Q has "night oblations to thee Thaisa", but Gossett is the first to action the change. Among the longer notes that Arden3 puts at the back of the book (pp. 407-12) there is nothing significant to record except that Gossett acknowledges F. Elizabeth Hart's article from Shakespeare Quarterly 2000 (reviewed in YWES 81) as the reason she retains Q's call for Cerimon to use "rough and woeful music" (3.2.87) rather than making the music "still and woeful" as has been the common emendation. In an appendix on casting and doubling (pp. 413-19) Gossett reckons that the play needs 18 men and 4 boys. Speculatively, and with appropriate discussion of the point, she has an adult man playing the Bawd; a forthcoming essay in Shakespeare Survey 2005 by David Kathman will show that adults males (that is, those over 21 years) did not play women.

    The title-page of Roger Warren's edition of Pericles for The Oxford Shakespeare makes the textual provenance abundantly clear: this is "A RECONSTRUCTED TEXT" edited "on the basis of a text prepared by GARY TAYLOR and MACD. P. JACKSON". As with the title-page of Stanley Wells's Oxford Shakespeare edition of the quarto History of King Lear (as opposed to the Folio Tragedy), Warren acknowledges that it builds upon an existing edition (in both cases, the 1986 Oxford Complete Works) rather than working from the ground up. As such this edition can be reviewed more briefly than Gossett's thorough reexamination of Pericles's textual condition, despite Warren's claim to have "reconsidered every detail of the Quarto text and of the Oxford reconstruction" (p. v). Warren admits that what he brings to the project is not editing knowledge but theatre knowledge: "Where my text differs from Oxford's, that usually reflects the practical experience of using their version in rehearsal, first in Stratford, Ontario in 1986, and then at Stratford-upon-Avon in 1989". As is usual with The Oxford Shakespeare, Warren's introduction is less than half the length of the competing Arden3 edition (80 pages to Gossett's 163), and the first section ("Theatre, Text, Authors", pp. 1-4) is a whistle-stop tour of the dating, early performance history, and popularity attested by allusions and reprintings. Warren asserts that the manuscript underlying Q was made by memorial reconstruction but rather than arguing the case in detail he refers the reader to his 2002 edition of Shakespeare's 2 Henry 6 (reviewed in YWES 83) for a general defence of this explanation of textual provenance. As part of his defence of the practice of inserting into the play material from Wilkins's prose novelization, Warren asserts that theatre has led the way: directors have "regularly drawn" on Wilkins's book to "provide themselves with a more performable script" (p. 3) than Q affords. Not to reconstruct the play behind Q is, he thinks, "an evasion of editorial responsibility".

    Warren holds that Wilkins wrote scenes 1-9 and Shakespeare wrote scenes 11-22, and in his brief consideration of the collaborative pattern (pp. 4-8) Warren notes that if the play was written from a skeleton outline (as Henslowe's Diary suggests plays were) it might well have been created by Wilkins, to whom the credit for the play's fine structure must go. Warren identifies the bits of Wilkins's translation of The History of Justine [ie Justinian] (1606) about Pericles, Lysimachus, and Antiochus the Great, some of which Wilkins plagiarized from Arthur Golding's translation. Wherever in his edition Warren draws on word-usage in The History of Justine (say, to defend a particular emendation as the kind of thing Wilkins would write), Warren assures the reader that he has checked that it was not simply stolen from Golding. Warren surveys the play's sources (pp. 13-20), and why the names (especially Pericles and Lysimachus for Apollonius and Athanagoras) were chosen from what certain of those sources had: the new names are easier to fit into verse, especially Gower's 8-syllable-line choruses that have to say 'Pericles' many times. Also, of course, Pericles is very like the Pyrocles (and indeed the Musidorus) in Sidney's Arcadia. Warren's "selective sketch" of modern revivals (pp. 20-30) need not detain us other than to observe that it is worth knowing that the music of the Royal Shakespeare Companies 2002 production was meant to recall "Solveig's song from Grieg's incidental music to Peer Gynt". For this reviewer it recalled the nursery song 'The wheels on the bus', and not to the production's benefit.

    Warren finds a thematic link between Helicanus jolting Pericles out of melancholy near the beginning of the play (scene 2) and his greater descent into coma at the end, and therefore the two collaborators worked together to give the whole thing a unified shape (pp. 37-38). Warren thinks that Q's omission of some of the questions and answers regarding the imprese has "no obvious dramatic purpose" (p. 39). One purpose could have been found by comparing the play to Middleton's Your Five Gallants written around the same time, in which the heroine Katherine is to choose her husband from the gallants who parade before her in a masque and who reveal their unworthiness by failing to understand the Latin mottoes of the imprese they are carrying. In the quarto of Pericles all six mottoes are read out but Simonides can expound the meaning of only three of them, and quite possibly the dramatic power of Simonides passing over those imprese he cannot understand is considerable; if so, it were better to follow Q in this regard. The tournament itself, according to Warren, should be performed on the stage: Q does not signal that we are to imagine it off stage, and Warren wonders if hobby-horses would do (p. 39). Warren provides an argument for having Pericles play music: it makes him "the Renaissance complete or universal man" just as he says he is ("My education been in arts and arms" 7.77). Warren also gives the reason for extending (from Wilkins's prose novelization) the moment where Thaisa declares her love for Pericles (9.72-96), although admitting that there is a downside: what comes from Wilkins to form their relationship is made to contrast (to its disadvantage) with what Shakespeare does in the shipboard-death scene to break their relationship (p. 41). Because the brothel scenes are about money (Marina buys her way out), not about sexual transgression, an editor is obliged to expand the conversion of Lysimachus using Wilkins's prose novelization because there is just too little in Q to be what Lysimachus calls Marina's having "spoke so well"; what is needed is precisely the eloquence that Wilkins's book gives her and this Warren's text uses, although not precisely as the Oxford Complete works does it (p. 52n1).

    Regarding the status of the text in its own time (pp. 60-71), Warren reckons that the play cannot have been left out of the Folio because Heminges and Condell were unable to get a copy, since it was still in the company's repertory. Warren thinks that Gosson's quarto was "unauthorized" because it was printed from chaotic copy and because there is no evidence of transfer from Blount (who registered it) to Gosson. This is not quite correct: entry in the Stationers' Register showed that one had prior right, but was not mandatory; likewise transfer of right did not have to be recorded: if Blount was happy enough with Gosson's publication of the play, there was no reason to record a transfer. Warren correctly asserts that to explain Pericles's omission from the 1623 Folio by reason of its being a collaboration would require one to account for F's inclusion of other collaborations: Henry 8, Timon of Athens, 1, 2, and 3 Henry 6, and Titus Andronicus. One could, Warren supposes, get around this by saying that Henry 8 was needed to complete the histories, Timon of Athens was included only because of the trouble with Troilus and Cressida, and that the Henry 6 plays and Titus Andronicus were by 1623 so old that Heminges and Condell were unsure about their being collaborations (p. 61n2). Warren concludes that collaboration is the "least implausible" reason for F not including Pericles. Warren provides a standard account of the discovery of the play's joint authorship (pp. 62-71), in which the database Literature Online is attributed to a company called "Chadwyck/Healey" (p. 69). In fact the company (now owned by ProQuest) is named after its founder, Charles Chadwyck-Healey.

    Most important for this review is the section of Warren's introduction that deals with "The Text: Corruption and Reconstruction" (pp. 71-80). Warren gives the familiar arguments for Q being a reported text and what to do about it, and he entirely accepts the Oxford Complete Works's editors views on this, adding only a point about Taylor's supposed reporters, a boy-and-master who played Lychorida and Marina and Fisherman and Pander. The boy might also have appeared in the Fishermen scene, as the third (and clearly junior) Fisherman (p. 79); this would explain why the prose in the Fishermen's scene is well reported. Of course, this memorial-reconstruction explanation has the problem that the meeting of Lysimachus and Marina is so deficient in Q, lacking the speeches for Marina in Wilkins's prose novelization. Taylor's answer was that Q's version shows censorship--Q is  innocuous where Wilkins's book is objectionable--and if that is what happened then it is added reason to put the text back to how it was before it got mangled. Here Warren provides no answer to Gossett's objection (in the essay reviewed here last year) that Shakespeare's Measure for Measure, and Beaumont and Fletcher's Philaster and The Maid's Tragedy depict randy men of authority and seem to have escaped uncensored.

    Warren's statement of his editorial procedure (p. 80) is worth quoting in full:

Where the Quarto makes reasonable sense, it is followed. Where it does not, or where the 'verse-fossils' in Wilkins's narrative offer more plausible readings, the text is reconstructed by re-casting those verse-fossils back into blank verse. This reconstruction is in practice very close to that in the Oxford Shakespeare Complete Works¹, and where there are major differences (e.g. in Sc. 7 and especially Sc. 19), my decisions have been influenced by the practical use of Oxford's reconstruction in rehearsal and performance, so that this edition has been even more influenced than my earlier ones in this series by theatrical considerations. In that respect, the reconstruction carries to its logical conclusion my argument that in Pericles textual and theatrical issues are interdependent.

    ¹ Where this reconstruction differs significantly from Oxford's, this is noted in the commentary or in Appendix C.

Actually, in the reconstruction of the Lysimachus/Marina meeting Warren departs considerably from the Oxford Complete Works and his collation does not note the detail. Instead of arguing each point, word by word, he sends the reader to appendices providing Wilkins's prose novelization version and Q and lets the reader work out for herself why he has done what he has done. Because Q is reproduced in full diplomatic transcript (Appendix A), "collations in the normal sense become superfluous" (p. 81). I disagree: without them one has manually to compare every line to see where the editor has departed from copy. Warren follows the Oxford Complete Works in most things, including the sensible abandonment of the traditional act divisions and numbering the scenes in a single sequence that represent the continuous performance practice of the time. The Oxford Complete Works text introduced elisions in words where this regularized the metre, but Warren chooses instead to spell them out in full because he thinks an actor can say them in such a way as to preserve the rhythm.

    Because Warren's text is highly dependent on the Oxford Complete Works, there is little point going through it with a fine-toothed comb. Rather, I shall refer only to places where Warren has come up with something for himself, departing from Oxford. Warren changes the name of the Second Fisherman to Master to reflect the hierarchy that the dialogue seems to imply, and his 'Attribution of Emendations' mentions the changed speech prefix (5.79 SP) but not the alteration in the preceding entrance direction for the fisherman. Instead of Oxford's "[SIMONIDES] What is the fourth? THAISA A knight of Athens bearing", Warren prints "What is the fourth? THAISA A knight of Athens with" (6.34), the change of bearing to with smoothing the metre slightly. Where Oxford has the metrically incomplete line "[THAISA] The motto, In hac spe vivo", Warren provides "[THAISA] The motto, In hac spe vivo. SIMONIDES 'In that hope I live'", patched from Wilkins and hypermetrical, almost a galloping fourteener. Warren prints "SIMONIDES (aside) By Jove I wonder, that is king of gods" (7.27) whereas Oxford follows Q in having "By Jove I wonder, that is king of thoughts"; as Warren acknowledges, the Oxford editors thought of this but did enact it. In the same scene, Warren has "[THAISA] A gentleman of Tyre, whom sour misfortune, | Bereft of ships and men, cast on this shore" (7.83-4 ) where Q has "A Gentleman of Tyre: who onely by misfortune of the seas, Bereft of Shippes and Men, cast on this shore" and Oxford has "A gentleman of Tyre, who, seeking adventures, | Was solely by misfortune of the seas | Bereft of ships and men, cast on this shore". Certainly Warren has stuck much more closely to Q than Oxford does, and smoothed the metre, but he has not explained how the words he has altered (who onely by -> whom sour) and the words he has dropped (of the seas) came to be corrupted/added in Q.

    Warren has Cerimon say "'Tis by a good constraint of fortune that" (12.56) where Q has "T'is a good constraint of Fortune it belches vpon vs" and Oxford has "'Tis by a good constraint of queasy fortune | It belches upon us". Adding "by" (adopted by Oxford from David Hoeniger's Arden2 edition) and "that" (his own addition), Warren fixes the metre in a new way, having ditched "queasy" that Oxford also got from Hoeniger's edition, which conjectured, but did not enact, its addition. Also fixing of metre is Warren's alteration of Oxford's "Give me your flowers. Come, o'er the sea margin | Walk" to "Give me your flowers. Come o'er the sea-marge walk" (15.78 ). Warren has Lysimachus say "it gives a good report to a member to be chaste" (19.46) where Oxford has "a noble to be chaste" and Q has "a number to be chaste". Warren's idea is that a 'member' is one of a community, or a penis (as in the pun in the opening lines of Marston's The Insatiate Countess). The whole line is of course ironical (being modest suits a bawd about as much as being chaste, flaccid, suits a penis). Warren's reading is possible, although it is a bit awkward to think of even a flaccid penis being chaste. Scene 19 is where major patching from Wilkins's prose novelization takes place, and Warren has Lysimachus say "But pretty one, I do protest to thee" (19.94) where Oxford has "But I protest to thee, | Pretty one, my authority can wink". This is a start of a 6-line insertion from Wilkins, and Warren has simply gone a little further than Oxford in emending to tidy up the metre. I suppose he might as well since he is reconstructing here, and indeed his 19.95-101, 19.103-7, 19.111-48, 19.151-56, and 19.163 are all insertions from Wilkins.

    Towards the end of this stretch of reconstruction, Warren prints "LYSIMACHUS Now to me" (19.160) where Q has "For me be you thoughten, that I came with no ill intent, for to me" and Oxford just cut the line entirely. Warren thinks the line cannot be Lysimachus's speech--it sounds more like Gower with his Middle English verb endings in -en--but also thinks the Oxford cut too abruptly, so, instead of the entire cut he takes just three of the words and changes for to Now. Sensitive to the social implications of you versus thou, Warren has Lysimachus say to Bolt "Thy house, but for this virgin that doth prop it, | Would sink and overwhelm thee" (19.170-1) where Q, and Oxford following it, have "Your" and "you". Warren is sure that Lysimachus would not use polite terms in his rant at Bolt, and he has a point. Finally, Warren gives Lysimachus the lines "She questionless . . . | . . . would alarm | And make a battery through his deafened ports" (21.36) where Q has "would allure" and Oxford has "would alarum". Warren takes the Oxford reading but simply changes it to an equally acceptable spelling of the period that regularizes the metre. Warren provides three appendices: the first reprints the diplomatic transcript of Q from the Original Spelling Edition of the Oxford Complete Works, the second gives the relevant passages from Wilkins's Painful Adventures of Pericles, taken from Geoffrey Bullough's Narrative and Dramatic Sources of Shakespeare and modernized, and the third is the 'Attribution of Emendations'. This last shows the first editor to introduce each emendation of Q that is used in this edition, with the exception that the Oxford Complete Works's emendations are taken as read. However, not even all the departures from Oxford are here recorded: the decision not to include a 7-line insertion from Wilkins in which Simonides bestows gifts on Pericles to cheer him up (Oxford's 7.88-94) is omitted by Warren because not theatrically necessary, but this is recorded only in the explanatory notes.

    The third of this year's substantial Shakespeare editions is John Jowett's majesterial Timon of Athens, which like Pericles is available to us only via a manifestly problematic early printing, in this case the Folio text of 1623. The first page of Jowett's 153-page introduction summarizes Jowett's conception of the problems: some of them are due to its being printed from a two-handed manuscript that "lacked some finishing touches", and as for the division of labour, "Shakespeare concentrated on the opening, the scenes dealing most fully with Timon himself, and the conclusion". Middleton wrote about a third of the play, and Jowett gives a table showing which bits. In places Middleton's and Shakespeare's work slotted together, and elsewhere they were "writing in contestation" (p. 2). The play was entered in the Stationers' Register on 8 November 1623 and no earlier mention survives. The composition seems to have occurred in early 1606: there are no act intervals and no sense that the play was written with intervals in mind, so it was written before August 1608 when the King's men got the Blackfriars and regularized their practices there and at the Globe (p. 4). Stylometric tests against other plays in the canons of Shakespeare and of Middleton place Timon of Athens in 1605-6, and its being clearly influenced by the pamphlet Two Most Unnatural and Bloody Murders (entered in the Stationers' Register on 12 June 1605) and apparently also by the Gunpowder Plot makes the second half of 1605 the earliest likely date (p. 6). Topical material alluding to king James's financial situation and similarities with Jonson's Volpone (first performed mid-March 1606) of the kind that make these plays seem to be in dialogue makes spring 1606 even more likely (p. 7). We cannot be sure that Timon of Athens was ever performed in the early modern period.

    Regarding the play's structure (pp. 9-11), Jowett finds that the action is somewhat in 5 phases (and as someone who wrote for boys in indoor hall playhouses, that would be Middleton's habit) but these do not correspond to the traditional editorially-imposed act intervals. Indeed, the play is "particularly resistant to the editors' act divisions", especially in its dramatically innovative scene 14 that, at 760 lines, is almost an inset play. Act divisions are not used in Jowett's edition. Regarding the staging possibilities (pp. 11-16), Jowett finds no evidence that Timon of Athens was written for anywhere other than the Globe, and, like Gossett and Warren, Jowett thinks that Diana descends from above in Pericles and he notes that Timon of Athens does not call for such things even though the Globe could do them. The masque in Timon of Athens is elaborate, however. The text leaves open just how the unnamed servants of Timon relate to the named ones (same men, now personalized, or different?), and how the guests at his feasting relate to the men who deny him aid (some overlap? same men entirely?). The play needs 13 men (with awkward doubling), or a few more for comfort, and several boys (see note to 2.119-23). Presumably Robert Armin played Apemantus and Richard Burbage played Timon in the first performances. Looking at the sources (pp. 16-23), Jowett finds only patchy borrowing and a tenuous relation to another Timon play known to have existed in the 1580s and to the academic play Timon extant in manuscript (which is not the 1580s play). Regarding genre (pp. 23-29), Jowett suggests that the 1623 Folio ought to have had a section for tragicomedies, including Timon of Athens, Measure for Measure, The Winter's Tale, Troilus and Cressida, and Cymbeline. Unique amongst the tragedies in its F section, it is not called "The Tragedy of . . ." but "The Life of . . ." and although much like Lear in his responses to ingratitude, Timon's death is off-stage and meaningless.

    A substantial part of Jowett's introduction (pp. 29-89) is taken up with thematic matters, which are superbly handled but largely beyond the scope of this review. The subsection on debt (pp. 45-53) is relevant to us because the dramatic material is by Middleton (the specialist in city comedy) and he characteristically contrasts the present, usurious and rapacious age with a notional pre-economic idyll. Jowett wonders whether the friend-soldier that Alcibiades pleads so eloquently for might be meant to be a homoerotic partner: Alcibiades is gay in Plutarch, Spenser, and Marlowe. The senate scene where this pleading happens is the text's "most difficult discontinuity" since it seems unrelated to the rest of the plot (pp. 70-74). Jowett's stage history of the play (pp. 89-120) is not the usual trawl century-by-century but rather is thematised; the play is so rarely produced that most unusually it is possible to list (Appendix D) all the major productions. One error here is worth noting: the Internet URL for further information on an experimental production in Budapest in 2000-1 should be <www.lap.szinhaz.hu> not <www.lap.scinhaz.hu> (p. 103n1). Jowett's account of the text (pp. 120-32) records that this part of the Folio was set by compositor B, and he describes the disruption of signatures and page-numbers around the Romeo and Juliet-Timon of Athens-Julius Caesar sequence. Charlton Hinman's analysis showed that Julius Caesar was printed and the type distributed before composition of Romeo and Juliet was completed and composition of Timon of Athens begun. Thus, Julius Caesar's pagination was set in stone, so to speak, but Timon of Athens proved too short to fill the space allowed for it and the signatures reveal what was done to make up for this. The leaves gg and gg2 in Romeo and Juliet should be the start of a regular 6-leaf quire but analysis of the binding shows that these are in fact the two leaves of an anomalous single-sheet quire. Jowett unfortunately calls it a "single-leaf quire" but he means single-sheet, and this confusion recurs. Thus, when Jowett writes: "To summarize, where one would expect three regular quires of six sheets, signed gg, hh, and ii, one actually finds an anomalous one-leaf quire (gg), two six-leaf quires (Gg and hh), and no quire ii at all" (p. 124) what he ought to have written is "To summarize, where one would expect three regular quires of six leaves, signed gg, hh, and ii, one actually finds an anomalous one-sheet quire (gg), two six-leaf quires (Gg and hh), and no quire ii at all".

    There is one copy of F in the Folger Shakespeare Library in Washington, however, that differs somewhat. It has the last page of Romeo and Juliet with the signature gg3 (so the original plan was the gg would go on as normal instead of being just a one-sheet quire) and this page has been cancelled and on its reverse is the beginning of Troilus and Cressida. This gg3 leaf appears in a separate section between the histories and the tragedies and holds the beginning of Troilus and Cressida that appears there in all copies of F. This shows that when Troilus and Cressida was first printed it was assumed that it would follow Romeo and Juliet, but then Troilus and Cressida was pulled out of this position and Timon of Athens substituted as a replacement. Troilus and Cressida was assigned a new and anomalous position between the histories and the tragedies, but some of the leaves showing the evidence of its original position (by having the last page of Romeo and Juliet on one side) were used in this new position. Jowett summarizes Hinman's account of how, part of the way through the composition of quire gg (the end of Romeo and Juliet and the start of Troilus and Cressida) the printer was made to stop and to end (or at least, to get to the penultimate page of) Romeo and Juliet in such a way (using a single-sheet quire) that it was not attached to Troilus and Cressida. This expedient could reuse standing type set when the end of Romeo and Juliet had been attached (by conjugate leaves) to the start of Troilus and Cressida. It was then decided to print Timon of Athens after Romeo and Juliet, presumably because of rights trouble with Troilus and Cressida; this was lucky for us else Timon of Athens would have been lost altogether. The type set to make the last page of Romeo and Juliet (on the same sheet as the first three pages of Troilus and Cressida) was reimposed as the first page of a new gathering Gg, the other three pages on Gg being the first three pages of Timon of Athens. Quire ii was not needed at all because Timon of Athens is so much shorter than Troilus and Cressida and hence the pagination jumps ahead after Timon. Jowett gives a handy visual summary of the replacement of Troilus and Cressida with Timon of Athens, but here too he writes "one-leaf quire" when he means "one-sheet quire" (p. 126).

    Presumably then, without the trouble over Troilus and Cressida we would not have Timon of Athens at all, and the reason seems to be its collaborative nature: although the Folio is not rigorous on this, ". . . co-authorship is the only consistent ground that can be identified on which plays were excluded" (p. 127). Even with the loss of quire ii, Timon of Athens was still so short that they put in a dramatis personae to fill up space. Compositor B set all of the play except Gg3r (scene 2 lines 10-129), and this change of compositor could be thought significant for identification of Middletonian/Shakespearian authorship because inexperienced compositor E was more likely to be conservative (to follow his copy slavishly) or to regularize in odd ways than was the experienced compositor B. In the event this merely results in some Middletonian spellings (or rather, compositor E's attempts at regularization of unusual Middletonian spellings) breaking through. To judge from the inconsistencies and loose ends, the copy for F was authorial draft. Jowett acknowledges William B. Long's work showing that stage direction inconsistencies often survived into theatrical manuscripts, but Jowett thinks those of the masque in Timon of Athens to be "of another order" from the kind Long writes about. Rather than call the infelicities of an authorial draft manuscript 'corruptions', Jowett points out that this implies a descent from something uncorrupted (which is not the case here) and he instead offers the excellent alternative "pre-completions" (p. 132).

    Relating the two-handed manuscript copy used by the printer to the printing itself (pp. 132-144), Jowett observes that because the play was set (except for Gg3r) by one man, compositor B, whose habits are fairly well known, we can say that shifts of the kind that stylometry can measure are attributable to shifts in the manuscript, probably where the author changes (p. 137). Jowett acknowledges David Lake, MacDonald P. Jackson, and especially R. V. Holdsworth's unpublished PhD thesis as the heroes of Middleton's identification as the other hand in Timon of Athens, and he lists the things they measured: contractions (such as I'm) that Middleton favoured much more than Shakespeare, and archaisms such as doth for does and hath for has that Shakespeare favoured much more than Middleton (pp. 137-38). The distinctive and contrasting spellings of characters' names (such as Ventidius/Ventigius) follow the same divisions of the play into dramatists' stints, as do the variations in the value of the currency unit called the talent, and likewise characteristic Middletonian stage direction phrases that Shakespeare avoids. Jowett briefly surveys claims for there being more than 2 hands in Timon of Athens and rejects them. A section of the introduction called "Shakespeare and Middleton" (pp. 144-53) tries to work out a new approach to play criticism that treats collaboration not as a weakness but a strength. In this regard, Jeffrey Masten's book Textual Intercourse of 1997 is amongst the recent thinking, although here it is misdated to 1977 (p. 144n1). Quite of few of the play's loose ends are begun or anticipated in scene 4: the hopes that Timon places in Ventidius make us expect a big letdown from Ventidius that never emerges, and because Alcibiades enters with Timon in this scene but does not speak or take part in the action we may suppose that there was meant to be here the origins of Alcibiades's unprepared-for plea to the senate about the soldier who has committed manslaughter. Perhaps, writes Jowett, ". . . a planned episode was unwritten, cancelled, or lost" (p. 148). Jowett gets into some speculation about how Middleton and Shakespeare influenced one another--including the evidence the Middleton reorganized material in scene 14, moving the Poet and Painter episode 200 lines later in the action than they originally were--and the mechanisms for it, such as Middleton reading over what Shakespeare wrote in scene 14 including the "plenteous bosom" as a the source of "one poor root" before he, Middleton, wrote scene 2 in which the artificial banquet is ironically referred to with the same phrase "plenteous bosom". Jowett ends this section with a consideration of how the figures of Timon and the Steward that emerge from the collaboration are complicated by Shakespeare and Middleton's different approaches to them.

    In the section called "Editorial Procedures" (pp. 155-64), Jowett provides plenty of detail of certain cases (such as his removal of the one of two entrance directions for the masque of Amazons as repetitious) but not much on his general approach or his answer to question of whether the text needs a lot of fixing. In the list of editions, Jowett names the Oxford Middleton as one that was in proof as this edition was completed, and the way he lists it and the Oxford Complete Works of Shakespeare makes clear that he was the editor of Timon of Athens in both. This raises an unsatisfied curiosity about whether Jowett thinks that the play should be edited differently according to whether it is being done in the context of Shakespeare's or Middleton's other works, or even whether one should try to edit each part of Timon of Athens according to what one thinks are the habits of the man who wrote that part. Since in the collation to this edition the emendations are at times attributed to This edition and at other times attributed to OXFORD MIDDLETON, there must be difference between the two and it would be satisfying to hear something of what that difference consists of.

    So, to the text of the play and those emendations that Jowett is the first to make, whether here or, according to the collation, in the Oxford Middleton. Jowett prints his own "The Persons of the Play" as well as F's, and the opening scene marker (as opposed to an act and scene marker) is attributed to the Oxford Middleton. (The Oxford Complete Works of Shakespeare divided the play into acts even though it did not treat other plays from around the same time such as Pericles and The History of King Lear this way; I wonder why and the Textual Companion does not tell me.) Jowett retains the Mercer as well as the Merchant in the opening entrance direction (whereas other editors have thought it a mistaken repetition) because he "visually introduces the theme of consumption and debt" and because Mercers were known for acting as creditors. In a glossing note to the dialogue mention of the play's main character (1.13), Jowett suggests that Timon rhymes not with Simon but with demon, or even more accurately it should be pronounced as though spelt Teemown. Jowett has the Poet say that when Fortune spurns her favourite, his friends "let him flit down" (1.88) whereas F has "let him sit downe". This is Jowett's emendation, and "flit" means "shift, pass", which suits the meaning; others have gone for "slip", "sink", and "fall". On the basis of a conjecture by R. V. Holdsworth that the Oxford Complete Middleton follows, Jowett prints "APEMANTUS That I had no angry wit but to be a lord" (1.237) where F has "That I had no angry wit to be a Lord". In a glossing note, Jowett identifies a proverb as the source of this--"He has wit at will that with angry heart can hold him still"--although to my ears it does not sound much like the line in question and does not evoke the same self-alienation that Apemantus seems to be speaking of. When Alcibiades enters with his horsemen (1.249 SD) Jowett adds a sentence of "[They greet Timon]" that the ensuing dialogue makes clear, attributing this to the Oxford Middleton. Jowett has Cupid say "There taste, touch, all" (2.122), which is essentially F's reading, but he generously collates as a plausible alternative the opinion of Oxford University Press copy editor Christine Buckley that it perhaps ought to be "There th'ear taste, touch, all". Servilius is given the business of "[presenting a note] to accompany "He's only sent . . ." (6.32), and this is attributed to This edition based on a conjecture by George Steevens.

    Such added business is fairly mundane, but Jowett brilliantly adds a direction when he gives Lucius the line "He cannot want fifty--[reading again] five hundred talents" (6.36-37) where F has "He cannot want fifty fiue hundred Talents". Jowett attributes the dash to H. J. Oliver's Arden2 edition, although the reader will not learn that this is the Arden2 edition from the list of references and abbreviations, where it is given just its title, the editor, and the year (p. 158). However, Oliver meant by it 'Timon cannot want 50, cannot even want 500 talents' in the sense that 'no sum, howsoever large, could add to his wealth significantly'. In Jowett's usage, however, the dash and an added stage direction mean that Lucius misreads and has to correct himself. This means that the note really does ask for 500, whereas Oliver thought that the next line, Servilius's "But in the meantime he wants less, my lord" shows the servant disregarding Lucius's exaggeration and keeping him to the facts of the note (which asks for less than 500), whereas Jowett interprets Servilius's line as a kind of haggling. Timon, he says, will settle for less than is asked for in the note, which is 500. Jowett starts all this off with an observation that fifty-five hundred was not a way of saying 5500 in Shakespeare's time. At 14.433-34 Jowett gives Timon the line "Take wealth and lives together-- | Do, villains, do, since you protest" where F has "Take wealth, and liues together, | Do Villaine do, since you protest". This is Jowett's emendation and is based on the objection to F's singular (villain) that Timon is clearly talking to a group: he calls them "workmen", and moreover were just one villain being spoken to Timon would say 'since thou protest'. This is because 'you' is either plural or respectful, and since Timon can hardly be thought respectful here the 'you' must indicate plurality. Finally, Jowett prints "[SOLDIER] 'Timon is dead, who hath outstretched his span. | Some beast read this; there does not live a man'" (14.3-4) which is essentially what F has. By including this first version of the epitaph of Timon--the second is written on the monument and the soldier says he cannot read it so he takes an impression of it in wax--Jowett is reversing his decision in the Oxford Complete Works of Shakespeare where he defended omitting this first epitaph on the grounds that it was a false start that should not have got printed. Jowett provides four appendices: "Alterations to Lineation", "Narrative Source Materials" giving Plutarch (North's translation), Lucian of Samosata (Frances Hickes translation, 1634) and the Parable of the Unjust Steward from the bible, a "Tabulation of Forms Favoured by Middleton" that show that scenes attributed to Shakespeare generally have few or none of Middleton's preferred forms of spellings and elisions and the ones attributed to Middleton have many of them, and "Major Productions".

    Two book-length collections of essays relevant to this review were published in 2004. The first is Lukas Erne and Margaret Jane Kidnie's Textual Performances: The Modern Reproduction of Shakespeare's Drama that concerns itself with wondering how can (and how should) editors use their knowledge of early-modern theatre, study, and printshop practices to inform their editing, and how they can justify their editorial interventions in relation to the needs of their readers. In the first essay (pp. 21-36) Leah S. Marcus argues that the Folio text of Othello is more racist than the 1622 quarto and that we must concern ourselves with how they got to be different rather than conflate them. The case for splitting Othello is as compelling as the case for splitting King Lear, Marcus argues, yet we have had no two-text editions, and the 160 lines of F that are not in Q are particularly concerned with race. Editors who use Q as their copy nonetheless tend to graft on the racist lines from F. Marcus asserts that close reading is a "rather clumsy, formalist mode" that prevents us seeing "how a given text differs from itself" (p. 23). This is illogical postmodern nonsense: nothing can be said to differ from itself because 'differ' and 'self' are inherently opposite in meaning. On certain textual details Marcus's close reading is in fact quite sound: in F Roderigo echoes Iago's terrifying images of miscegenation in the first scene when they wake Brabantio, but in Q he does not. Thus in F the sexual anxiety is not so much Iago's personal problem but a social norm. Likewise Othello's response to what he thinks is Cassio's confession is more sexually explicit in F. However, Marcus compares Desdemona's claim in Q that she has kept her body for Othello alone and "from any hated foule vnlawfull touch" with F's "From any other foule vnlawfull touch" and claims that the latter means that Desdemona thinks of her marriage as a kind of unlawful touch, a whoredom (p. 26). This is simple misreading of Shakespearian English, for 'other', 'foule', and 'unlawfull' are merely three adjectives modifying 'touch': she has felt no other touch than her husband's, felt no foul touch, felt no unlawful touch.

    F's Willow Scene (4.3) has extra material that is essentially sexually titillating, and although Marcus agrees that this scene gives Emilia more agency (a good thing, most readers will agree) she finds that over against this greater norm of female domesticity and power in F, Othello is made to seem all the more aberrant, all the more of an outsider. In F (but not Q) Othello says that his name that was a fresh as Diana's visage is now as black as his face, but editors usually emend to make him say that her name (that is, Desdemona's) is now black, which is what Q2 has. Only in F is Othello accused of using "foul Charmes" to woo Desdemona and only in F does he liken his icy passion to the Pontic (Black) Sea, which of course makes him more like the Turks. In F, the scene of Desdemona's murder is not just about her: it is also about his alienation from the culture he tried to fit into. In a complex sentence, Marcus suggests (I think) that Q was disliked for its racial tolerance until people started to think it was the earlier version: "Before the quarto version of Othello came to be viewed as Shakespeare's first version of the play, it was reviled in a language of miscegenation that demonstrates the unease textual scholars felt but could not directly express toward the more benign construction of racial difference offered in Q". That is, F was spoken of as having contaminated dirty Q much as Desdemona is contaminated by dirty Othello. Marcus thinks a rising tide of racial intolerance made Shakespeare revise tolerant Q into intolerant F; likewise Folio Titus Andronicus has an additional final 4 lines (over what is in Q) that thrust the blame on Aaron. Here Marcus mentions Scott McMillin's argument that rather than adding to MSQ to make MSF, someone (Shakespeare?) cut from MSF to make MSQ, but omits Pervez Rivzi's argument (reviewed in YWES 80) of the same claim. If so, Marcus speculates, perhaps the most racially-explicit material was cut from MSF for the performance at court (thereby making MSQ) around November 1604 when the Masque of Blackness was performed and when the new peace with Spain with concluded; the Spanish ambassadors did not like anti-Black representations.

    In "'Work of permanent utility': Editors and texts, authorities and originals" (pp. 37-48), H. R. Woudhuysen argues that facsimiles of early drama are not enough: we still need critical editions. Since 1984 the Malone Society reprints have been photo-facsimile rather than type-facsimile diplomatic (or semi-diplomatic) reprints, although W. W. Greg himself was against using photographic reproduction because it exaggerates difficulties. In a photo-facsimile only one state of the text can be shown (and maybe a bad one) whereas an editor might, by comparing several states, easily have worked out the true reading and put it into a type-facsimile. Trevor Howard-Hill has complained that Greg's 1909 Malone Society text of The Second Maiden's Tragedy did not note the marginalia by George Buc in his role as Master of the Revels, but Greg was acting deliberately in this: subsequent theatrical use was not what his edition meant to record, he wanted to treat it as a literary manuscript. Malone Society editors edit the text (the material object) while generally Shakespeare editors edit the play (in the sense of Platonic ideal), and most readers want the latter. An unresolved problem of performance-centered editing of the kind exemplified by the Oxford Complete Works of 1986 and since adopted by the Cambridge and Arden series--aside from the fact that we are only supposing that Shakespeare approved of what the actors did with his scripts--is that the Folio texts of Hamlet and Troilus and Cressida (for example) seem much  too long to have been performed uncut, even though they at the same time seem more theatrical than the corresponding quarto texts. Woudhuysen makes a subtle jibe at the edition's delayed publication when he writes that the Oxford Complete Middleton "may one day" do for that dramatist what has been done for Shakespeare and Jonson, which is provide a plurality of texts. Woudhuysen thinks that all editions are necessarily only approximate to "an unrecoverable original" and yet he is against the principle of 'unediting' since there is always a guiding hand at work, and better a specialist than an amateur reader. In reproducing texts as they appeared in their original, time-bound, material forms (without worrying what the author would have thought about them), the Malone Society has long been at the forefront of literary theory in its stand against the author-function: the facsimile is the socialized text made in the two centres of early modern socialization, the theatre and the printshop. As Joseph A. Dane pointed out, one cannot really make a facsimile of a book: the only thing that is truly reproducible is the text, not its material manifestation. Charlton Hinman's Folio facsimile, Woudhuysen observes, used the page rather than the forme as its unit of reproduction, so it copies no particular object. (One might add that worse still, the object it aims to represent not only does not (to our knowledge) exist and perhaps never did, but further it could not have existed. This is because Hinman put together as though conjugate images of pages from different impressions (that is, from different Folio copies) of the same forme: these pages could never have got together no matter how the bundle of formes was shuffled.)

    Paul Werstine's essay "Housmania: Episodes in twentieth-century 'critical' editing of Shakespeare" (pp. 49-62) is concerned with the practice of 'best text' editing, in the sense of choosing from amongst the various early printings the one that is of overall greatest authority and then reproducing its readings in one's edition except where there is indubitable error. This way of proceeding was advocated by R. B. McKerrow in his Prolegomena for the Oxford Shakespeare (1939) and the alternative, 'critical' editing (advanced in W. W. Greg's The Editorial Problem in Shakespeare, 1942) requires basing an edition on the text closest to the author's foul papers and drawing authority (for example for particular readings) from elsewhere as needed. Greg got from John Dover Wilson's work on Hamlet his confidence that one could tell from a printing whether it was made from foul-paper or promptbook copy, whereas McKerrow thought this distinction nigh-on impossible to make: apparent author's errors that would lead one to infer foul-paper copy look just like other kinds of errors in a printing. Werstine shows that Dover Wilson interpreted a feature of Q2 Hamlet as an authorial false-start (hence Q2's copy was foul papers) where in fact other explanations might be equally reasonable. Explaining certain F/Q differences in Hamlet Werstine writes about "words cut from F" when he means 'words absent in F', because 'cut from' is more than we know (p. 54). Also, this phrasing implies a strange procedure of movement: where did the words cut from get taken to? Gary Taylor knows that inconsistent speech prefixes and stage directions do not show that a printing's copy was foul papers, and said as much in his edition of Henry 5. Yet he still thinks that in this case the copy for F was author's papers because he sees in it a false-start repetition characteristic of this kind of copy: Pistol twice asks, through the Boy, his French prisoner's name (TLN 2390, 2405-6). Yet if that is the explanation, Werstine wonders, why did Taylor include both askings in his edition? Honigmann did the same thing in his edition of Othello: he found what he claimed were false-starts, from them determined the printer's copy to be foul papers, and yet kept them in the edition.

    Characteristically, Werstine overstates his case to make it seem that all New Bibliographical confidence in determining printer's copy is derived from the ability to spot false-starts, just as elsewhere he attributes all the confidence to the ability to spot inconsistent stage directions and speech prefixes. In truth, the confidence usually comes from the coincidence of such evidence, which is mutually corroborating. To show that false-starts can be wrongly inferred from the evidence, Werstine quotes Ralph Crane's transcript of Fletcher and Massinger's play Sir John van Olden Barnavelt which has the line "I know you love the Prince valiant Prince and yet" (line 679). Were this to appear in print it would be seized on as an obvious authorial false-start and hence the printer's copy would be assumed to be foul papers; yet the line appears in a scribal transcript. Moreover, the explanation has nothing to do with false starts. The line read "I know you love the Prince of Orange, yet" and Buc deleted "of Orange," to make "I know you love the Prince of Orange, yet". Crane inserted "valiant Prince and" above the line so it finally reads "I know you love the Prince of Orange, ^{valiant Prince and} yet". Because we now realize that what looks like false starts can be something else, 'critical' editing has had its day. And yet we are not forced back onto McKerrowian 'best text' editing because we no longer believe that we can get back to the author's 'work' at all. Rather, we rightly (according to Werstine) edit not the 'work' but a particular print manifestation of it and hence the next Arden Hamlet will provide all three texts and hence the rationale of the New Folger Shakespeare under Werstine's general editorship. Werstine ends with another example of alleged misdiagnosis: Q2 Romeo and Juliet has a direction that reads "Enter Will Kemp" (K3v). In Dramatic Documents from the Elizabethan Playhouses (1931) Greg acknowledged that in all surviving theatrical documents the appearance of an actor's name instead of a characters is demonstrably not part of the original composition, yet Greg always insisted that Q2 Romeo and Juliet was printed from foul papers. What Werstine obscures here is that Greg changed his mind about many things, and his final word on the subject was the observation in The Shakespeare First Folio (1955, p. 142) that an actor's name sometimes appears in foul papers (especially if the part is being written for a specific actor) and sometimes appears in a promptbook (especially if the part is a minor one), and in the latter case it will probably duplicate the character name. Greg was less dogmatic that his critics would have us believe.

    The best essay in the collection is John Jowett's "Addressing Adaptation: Measure for Measure and Sir Thomas More" (pp. 63-76). The plays in Jowett's title are adaptations of sorts and for that reason will appear in new places: Measure for Measure will be in the Oxford Complete Middleton and Sir Thomas More will appear in the Arden Shakespeare. In both places the editor, Jowett, will attempt to show the pre- and post-adapted states of the text (rather than relegating one to an appendix) by shifting single-text and two-text presentation as the need arises. Whereas many essays in this volume are about the problem of departing from copy text when making a new edition, Jowett's is about the problem of sticking to it. There has been a reluctance to admit adaptation in Measure for Measure because its presence makes it impossible to stick to copy: the copy shows not the original, not the adaptation, but an unplayable mixture of the two. Hence Mistress Overdone's forgetting what she announced in the second scene: that Claudio has been arrested. One can only defend adherence to copy in such a case if one is making a facsimile of an original document and is forbidden from emending error. For Jowett, ". . . the text is by no means necessarily . . . identical to its materialization in a specific document" (p. 64). We are already used to editors disambiguating multiple texts ('texts' in the sense of 'what is represented by the document') where there is duplication, as with Romeo and Friar Laurence being given virtually the same speech about "The grey-eyed morne" in Q2 Romeo and Juliet. For this it matters not whether the underlying manuscript has continuous writing or bits patched over and around: the duplication itself cannot be what was meant. Jowett quotes another case from Q2 Romeo and Juliet: the repetition of lines and ideas around Romeo's death. However, these cases where the revision seems to have happened right away make not for a two-track structure but rather a fold: the first version should really disappear underneath the second. When the revision happened much later, the structure is not a fold but two distinct tracks.

    Jowett does not like disrupting the linearity of the text merely for the sake of disruption (as the book editor Kidnie advocates elsewhere and in this volume) but thinks that where there is adaptation one has to do so, and he presents a sample page of Measure for Measure from the Oxford Complete Middleton to show how it will be done. There is a greyed-out typeface used to represent words by Shakespeare that Middleton cut, and bold-face type is used to show insertions by Middleton. Thus, to read the Shakespeare version one follows the normal text and the greyed-out (ignoring the bold), and to read the Middleton version one follows the normal text and the bold (ignoring greyed-out). This, as Jowett subtly boasts, is a better solution than some two-text typographic arrangements, such as R. A. Foakes's superscripted F...F and Q...Q marks in the Arden3 King Lear. Where there is doubt whether a feature is a result of ordinary linear error or adaptation, Jowett errs on the side of adaptation, just as in a two-text case such as King Lear, where Q and F disagree and we cannot be sure whether one is simply a corruption of the other, we must assume that they are equally valid alternatives.

    Following Taylor's "brilliant" deduction (reviewed below) that Vienna was not the original location of Measure fore Measure but rather Ferrara was, the logic of Jowett's edition means putting Ferrara in grey and Vienna in bold every time. Putting back the profanity in the pre-1606 version is harder, but there must have been more than F records for there are clearly unShakespearian substitutions of inoffensive words. There were also, of course, big adaptational changes (as discussed in Taylor and Jowett's Shakespeare Reshaped): the adding of an opening passage to 1.2, and the alterations around the end of act 3 and start of act 4 that transposed the duke's soliloquies and added the song. There were also some intermediate-scale changes--such as the exchange between Escalus and the Justice at the end of 2.1, the presence of Lucio in 2.2, and Pompey's catalogue of prisoners in 4.3--that were under suspicion in Shakespeare Reshaped and that in the Oxford Complete Middleton will be shown to be Middletonian. All this complexity necessarily fattens the commentary in places, but this is acceptable because the play is so well known that the commentary can be almost entirely devoted to the adaptation. The same is not true of Sir Thomas More that Jowett is editing for The Arden Shakespeare: the aim of that edition is different since it must do much to make the text available to a large readership. Like Measure for Measure, this play is an adaptation and we can tell the Original from the Additional Passages. This makes for binary thinking (before and after), which generalization is indeed accurate, but Jowett will not present Sir Thomas More like he presents the two-state Measure for Measure because in the former the changes run for long stretches whereas in the latter they sent to be short. Also, of course, Sir Thomas More will be in a small, single-column book while Measure for Measure will appear in a large, double-column book that allows one to see more at a glance. Putting Sir Thomas More's additional passages in an appendix would not be ideal either, since all the Shakespeare text would have to go there and this is to be a volume in a Shakespeare series. So instead, where adaptation/revision makes for a two-text situation, the edition will switch to parallel text mode: original on the left, adaptation on the right. Unfortunately, Jowett does not show a sample page but he does describe the difficulties that he anticipates. Sometimes there is no corresponding section in the original to print alongside the addition, and often the additions are of a different length to the passages they replace in the original. The distortions his layout brings will at least be fewer than the distortions of the Revels edition, which is a "synthetic revised version" that "weaves seamlessly between Original Text and Additional Passages". The 'frozen' technology of a paper edition must try, in cases like this, to represent text not as fixed product but as mobile 'process'. Paradoxically, these plays show authoriality to be not isolated but interconnected--by adaptation Sir Thomas More became Shakespearian and Measure for Measure became Middletonian--and yet to see this interconnection ". . . the local markers of authorship . . . must always be clear to the plays' readers".

    In "The New Bibliography and its critics" (pp. 77-93), Ernst Honigmann provides a useful chronology of New Bibliography, showing that it was never univocal and that certain recent criticisms of it (such as attacks on the binarism of 'good' versus 'bad' quartos) were first made long ago. No text is frozen in time, all are variable--even each performance differs from others--yet we are entitled to call most plays just one play. Honigmann thinks that the Oxford Complete Works of Shakespeare was just about entitled to say that there are two King Lears, but the endeavour of editing quartos that are substantially the same as F is "very questionable". (This I take to be a dig at the New Cambridge Shakespeare quarto series in particular.) Like Werstine in his preceding essay, Honigmann quotes A. E. Housman on what is wrong with sticking to one text in all but its manifest errors: it would be amazing, says Housman, that a text's readings were right (authorial) wherever they are possible and wrong (unauthorial) only when impossible. On this basis, Honigmann asserts that the King Lear quarto probably contains many possible but in fact wrong readings, and that any edition that sticks to it in all but error will have "just as many false readings as old-fashioned conflated texts". True, but I suppose there is merit in knowing that one has not introduced certain error, which is what happens when one conflates material that the dramatist removed in revision with the new material with which he replaced it.

    Even unconflated texts conflate to make conflationary emendations (that is, when Q is nonsensical one imports F's reading) so it is not really a matter of whether but of how much to conflate. William B. Long is right that promptbooks need not have been very orderly, but Greg was also right that authorial papers have certain roughnesses (such as variable speech prefixes) that would not purposely be introduced by later copyists. Of course Werstine can find exceptions, but the things that Greg said were characteristic of Shakespearian foul-paper copy for a printing--missing and indeterminate stage directions, variable speech prefixes, false starts, deletions, underpunctuation, unusual and fluid spelling--are found all over the 'good' quartos (Love's Labour's Lost 1598, Romeo and Juliet 1599, The Merchant of Venice 1600, Much Ado About Nothing 1600, Troilus and Cressida 1609) even though the printers were different in each case. What we find common to them are the Shakespearian foul-paper characteristics. There may indeed be an intervening transcript (although we hope not), but the point is that there still are identifiable families of textual origin (foul papers, author's fair copy, scribal copy, promptbook) for us to distinguish. (There is, of course, a logical slip in that last assertion: the wished-away possibility of intervening transcript itself would move the copy from one of the allegedly distinguishable categories to another.)

    In "Scholarly editing and the shift from print to electronic cultures" (pp. 94-108) Sonia Massai makes the now-familiar claim that electronic editions open up postmodern potentialities for the editing of Shakespeare. For example, Michael Best's Internet Shakespeare Editions (ISE) project uses animation to show textual variability such as stop-press variants, which Massai thinks is a good idea. I would argue that this kind of innovation harms the longevity of any textual project: it is inconceivable that the means of producing animation on computer screens used today will still work in 20-years time, so the work will have to either be painstakingly adapted or (more likely) abandoned. The lessons of the BBC's mid-1980s Doomsday LaserDisc project have clearly not been learnt: thousands of people worked together on a digital project that is now entirely unusable. Also, animation destroys the one feature that makes electronic text worth having in the first place: the capacity to be cut-and-pasted for reuse elsewhere. Until 'dancing text' is an international standard for the underlying operating systems of computers, any application that uses it will trap its materials inside a software 'black box' with which no other applications can communicate.

    Massai describes approvingly the ISE's and Shakespeare Electronic Archive's ability to show Q and F texts simultaneously in different windows on the screen. This is of course valuable, but has been available since 1995 from Chadwyck-Healey's CD-ROM product called Editions and Adaptations of Shakespeare and is scarcely an innovation. More interestingly, Massai presents evidence for supposing that the printer Richard Jones, rather than the dramatists, was responsible for improvements when reprinting plays, including Marlowe's Tamburlaine and Edwards's Damon and Pythias. Rather overstating the impact of Lukas Erne's recent book, Massai claims that scholars have "abandoned" the notion that Shakespeare was indifferent to the printing of his plays; I would have thought the debate is still ongoing. Indeed, overstatement (of the kind that Erne and Kidnie as editors of this volume should have saved her from) dominates the remainder of Massai's essay. She claims that editors for the print medium "have no other option but to use the textual apparatus to alert their readers to the presence of editorial variants introduced in early textual authorities" (p. 99), but this is simply not true. They can also (as many Arden3's do) print a photo-facsimile of one of the early textual authorities. Likewise Massai overstates her case about indeterminacy when she claims (on the authority of Werstine and Long) that ". . . it is virtually impossible for editors to establish whether the printer's copy was the author's foul papers, a theatrical manuscript, an authorial fair copy, or a scribal transcript" (p. 101). In fact, we can often distinguish a scribal transcript by such features as Latinate labels (of the kind "Actus primus Scena prima") that would not occur in author's papers nor theatrical manuscript. After all, Massai can scarcely deny that Trevor Howard-Hill has established Ralph Crane's involvement in preparing copy for F1, and yet she asserts that establishing such a thing is "virtually impossible".

    Massai ends with familiar and mistaken assertions about hypertext making the reader Barthesian and about the electronic medium's capacity to disperse authority. (I remain sure that Massai's authority would quickly recoalesce and assert itself were I to download her entire ISE Edward 3 website and republish it as my own edition of the play.) It would appear that contributors to the volume were not able to benefit from reading one another's essays before publication, for Massai would have learnt from John Lavanigno's contribution that print-medium footnotes are essentially hyperlinks and that nothing she describes her edition doing (such as providing contextualizing material) constitutes anything new. All that has changed is how much such material one can include and how quickly readers can get to it: clicking is quicker than page-hunting, but not much. Massai thinks that the electronic medium allows her to capture the "textual plurality and instability which were native to the medium of print in the early modern culture". That in fact is not what she has achieved, and were she really to desire that the best way were to send out a print edition with textual variants between copies and then a subsequent edition with revisions. What her archive tries to do is capture in one place and at one time the variability that was, in the early modern print culture, dispersed over space and time. For that reason, singularity still triumphs over plurality in her work. Massai claims "The notions of Text, Author, and Canon" have been "made . . . obsolete" by the electronic medium, which is an extraordinary assertion that needs substantiation. I am certainly the author/editor of the material on my website, and judging by Massai's citation of an article by Michael Best (p. 106n5) she thinks that he is the author of his text published on his website. Ironically the instability of the medium is exemplified by this citation--it does not work--but that is not because the article cited is not there (it is today, at least) but because the referencing system is inadequate. The URL in question is given as a deep link into the ISE site and as is so often the case, a reorganization of materials has broken the link. The correct current URL for the document is <http://ise.uvic.ca/Annex/Articles/SAA2002/index.html>.  Projects such as ISE should tell people how best to cite their content so that links do not break every time the site owners undertake a spot of tidying-up.

    Ann Thompson and Neil Taylor offer curious insights to the practices of theatrical doubling, as implied by the texts of Hamlet, in "'Your sum of parts': Doubling in Hamlet" (pp. 111-26). All Arden3 editions are supposed to have a doubling chart and discuss casting but few do; the editorial guidelines instructing editors about this arose because apparently theatre directors did not realize that Elizabethan actors doubled and were avoiding Shakespeare for fear of having too few actors. Thompson and Taylor's Arden3 Hamlet will have a separate doubling chart for each of the three texts of the play, and they are reproduced here. Using the rules that Thompson and Taylor explain, all three texts could, they reckon, have been done (at a pinch) by 8 men and 3 boys. This requires doubling Voltemand with Marcellus, as Greg proposed for his famous memorial-reconstruction explanation for the creation of the manuscript underlying Q1. Kathleen Irace proposed that this actor also played Player Prologue and Player Lucianus, in order to rescue the memorial-reconstruction explanation in those moments where the reporting improves and yet neither Voltemand nor Marcellus is on the stage. The trouble is, Thompson and Taylor report, that even with these extra doubles the actor still has not really got enough work to do. It is curious (but Thompson and Taylor say no more than that here) that Q1, Q2, and F Hamlet all allow essentially the same doubling pattern, and that Hamlet, Queen, and especially Horatio are the most unlikely to be doubled. Is this, the editors wonder, a sign that they are more 'stable' characters?

    In "The perception of error: The editing and the performance of the opening of Coriolanus" (pp. 127-42), Michael Warren argues that the traditional reassigning of speeches in the first scene of the play makes for a more conservative reading than the F text (our only authority) actually requires. There are a few speeches assigned to "All" in the opening scene that modern editors reassign because they do not sound right spoken collectively. Also, First Citizen falls silent after announcing Menenius's entrance and Second Citizen does all the talking with him and Martius. Editors often give some of Second Citizen's subsequent lines to First Citizen on the grounds that we can distinguish characters: First Citizen is anti-authoritarian and Second Citizen is conciliatory, and (editors argue) these character traits should be allowed to persist. However, imaginative theatrical thinking that treats the insurgents not as 'types' but as men likely to experience rising and falling emotions and levels of confidence can help us defend the F assignments, and we should not of course emend F without at least explaining why we think it in error regarding these speech prefixes and without explaining how we think the error came about. Warren discusses the theatrical potential of the Second Citizen coming to be the insurrection's spokesperson (reluctantly? by Menenius's selection of him? because the First Citizen has melted into the crowd?), and discusses the potential for this character's views changing during the scene. Thus instead of static characters we have people responding to a developing situation. Altering the speech prefixes has the effect of making the crowd more homogeneous, more like a mob than a group of citizens with diverging opinions but all inspired by a genuine grievance.

    David Bevington's essay "Modern spelling: The hard choices" (pp. 143-57) revisits Stanley Wells's work on modernizing spelling and finds a few borderline cases where the modernization might do more violence to the text than would be done by leaving the original spelling. In fact, several of Bevington's examples (ancient/ensign, travail/travel) were dealt with by Wells, and to the general satisfaction of editors he showed precisely the opposite case: they must be modernized. 'To blame' was, apparently, misunderstood as 'too blameworthy' in the period, so modernizing to 'to blame' rather than 'too blame' loses that sense, and yet to modernize to 'too blame' is to perpetuate an error. Similarly, to modernize 'abhominable' and 'negromancy' is to conceal the period's false etymologies of 'away from man' and 'black magic'. I would have thought those false etymologies were best explained in notes, not perpetuated by body-text spelling. Modernizing 'ancient' to 'ensign' loses the sense of superiority through age and making 'inginer' into 'engineer' is apt to mislead since the wanted sense is close to 'ingenious' (as we might say an inventor is) than to anything to do with an engine.  Bevington admits that preserving unusual spellings because they might reflect original pronunciation is "at the cost of inconsistency", since of course one would be preserving only some, not all, original pronunciations.

    In an article in Shakespeare Quarterly 2000 (reviewed in YWES 81), Margaret Jane Kidnie urged that editors find ways to represent the performance potentialities of plays rather than prescribe too fixedly what should happen on the stage, but she did not suggest how this might be done. In "The staging of Shakespeare's drama in print editions" (pp. 158-77) she offers a solution. As in her article, Kidnie begins with bold assertions that do not bear scrutiny, such as ". . . there is no necessary or transparent link between scripted text and staged performance. Scripts are not comparable to performance, nor can they encode it." (p. 158). If this were true, Kidnie could not object to the Lord Chamberlain's office censoring scripts up to 1968, since this activity would have no effect on the performances. In fact, of course, this censorship was deleterious of theatrical art and its equivalents in tyrannies such as China and Israel today are deleterious. Kidnie cites two modern plays that do not use speech prefixes and stage directions as we would expect, and then refers to this as "the striking diversity exhibited by play-texts of our own time". In truth, her two examples were exceptional: pull down almost any modern play from the shelf in a bookshop and you will find the conventional stage directions and speech prefixes that her two examples have purposely eschewed. Looseness of logic is accompanied by looseness of Kidnie's writing, as in the pleonasm "I myself" whose redundancy is obvious when one considers that there can be no 'you myself' or 'I herself'.

    Relying on William B. Long's observation that theatrical manuscripts of Shakespeare's time do not solve all the staging problems, Kidnie asserts that Shakespeare's scripts only seem deficient to us: they were not so to their original readers. Kidnie thinks that what we do in stage directions in modern editions is simply a matter of "fashions", but since she started with Gary Taylor's notion of a lost para-text provided by the original performance conditions and practices, she ought to acknowledge that adding stage directions can make up for that para-text not being available to the modern reader. The theatre historian's knowledge can supplement the text via knowledge-laden directions. Kidnie accepts that editors might want to specify what the texts leave open because otherwise for the modern reader the script feels incomplete, but she insists that the texts are not incomplete in a transhistorical sense, just incomplete to us. In this she implies that the scripts would not have seemed incomplete to original readers, but she must be forgetting that in many cases the scripts were written for actors not readers and hence they are now being addressed to a readership for which they were never intended. This alteration of address is a source of perceived incompleteness, as one can see from Crane's play transcripts made especially for patrons and Jonson's printed playbooks made for readers; these differ greatly from theatrical manuscripts.

    Kidnie's practical suggestion to editors is to put the stage directions off to one side of the printed page so that they are not aligned with particular lines of dialogue and hence they are drained of their determinacy of timing. Bizarrely, the early modern manuscript examples that Kidnie gets this idea from and which she quotes to support her case show precisely the opposite: an abiding concern with determinacy. The Second Maiden's Tragedy brackets the stage directions that Kidnie uses as her examples, identifying with a single point the moment the action happens rather than allowing the two lines of the direction to sprawl across the paper and thus across stage-time. Thus "Enter | Nobles" is bracketed so that the single point at the centre of the brace is positioned just before the nobles are needed, and this point identifies the exact dialogue occasion for the action: the cry "my lordes treason". The same is true of Kidnie's second example from this manuscript: "Enter | Heluetius" is bracketed so that the point identifies precisely when in the dialogue he enters, in response to "heere comes another". Indeed, although Kidnie could have found theatrical manuscripts that do not do this, The Second Maiden's Tragedy repeatedly connects its marginal stage directions with the dialogue using inked lines to show just where things are to happen. Her second manuscript example Sir Thomas More does the same thing, confining its direction "Enter A messenger" into a wedge-shaped box that seems to press into the dialogue as though concerned to indicate precisely where he is needed. Kidnie could not have chosen more apt examples of why she is wrong to claim that ". . . rarely can they [stage directions] be aligned visually with a precise moment in the dialogue" (p. 165). That Kidnie's third example, Q2 Hamlet, differs from the manuscripts is probably due to the relative trickiness of typesetting brackets and boxes around dialogue.

    Kidnie offers sample text from Troilus and Cressida and Romeo and Juliet laid out in the way she proposes, with the directions in a left-side box whose sides are inked in. Although there are margins used for stage directions in early modern theatrical manuscripts they were made by folding the paper and not by ruling an inked line, so Kidnie has inserted a vertical barrier to disconnect the dialogue from the directions, where the manuscripts she claims to be following did the (sensible) opposite and ran inked lines left-to-right across the page to attach what she boxes apart. Kidnie explains that her layout encourages indeterminacy and seems not to even ponder whether readers want it. After all, theatre practitioners routinely ignore a play's stage directions and editors could simply tell readers that the timings of particular actions are not certain. At least Kidnie admits that one cannot get all the indeterminacy in, else the play disappears. Regarding her sample scene from Romeo and Juliet, Kidnie anachronistically refers to the Capulets' "ball", a word OED records no earlier than about 40 years after the play; clearly she is thinking of later parallel events for what the play itself calls a "supper" and a "feast".

    The Arden editor John D. Cox favours leaving out the stage directions altogether, or at least confining them to the commentary. In "Open stage, open page? Editing stage directions in early dramatic texts" (pp. 178-93) Cox confides that this is what he and Eric Rasmussen wanted for their Arden3 edition of 3 Henry 6 (reviewed in YWES 82) but the house style of the series forced them to compromise. Even entrance and exit stage directions are not straightforward, and Cox justifies his decision that messengers always get off as soon as they have delivered their messages on the grounds of doubling need. (One could argue that such an editorial rule is a perverse closing down of the theatrical options in the name of opening them up, since there can be dramatic power in a messenger hanging loosely about with nothing to do.) What about the vexed issue of implied stage directions given in dialogue? Cox and Rasmussen preferred not to realize them (only to discuss the options in a note) but Arden3 forced them upon the edition. At this point Cox gives a footnote (p. 193n26) about editors who insist on adding a stage direction for York to sit because Henry says "See where the sturdy rebel sits" (Richard Duke of York 1.1.50). Such editors, he claims, are not concerned to help confused readers but rather are "yearning for closure". Having praised and cited Kidnie's perspicacity and respect for theatrical non-closure in this regard, Cox might have observed that in her mocked-up layout of Troilus and Cressida in the previous essay, Kidnie added the direction "[Ajax passes money to trumpeter]" to accompany the line "AJAX Thou trumpet, there's my purse" (4.6.6). What mistake did Kidnie want to save the reader from there, I wonder; perhaps that Ajax might simply point to his purse without using its contents.

    Speech prefixes can give to readers information not available to theatregoers, and Cox uses the example of Don John in Much Ado About Nothing having a stage direction that identifies him as a bastard long before anyone says that about him. This comment rather muddies the water, since it is a stage direction not a speech prefix, and for contrast one could take the case of Q1 King Lear in which, as Blayney showed, Edmund got the speech prefix "Bas[tard]" through type-shortage, not authorial characterization. Cox defends retaining traditional act and scene breaks for the purposes of referencing, pointing out that using Through Line Numbering (the obvious alternative since F 3 Henry 6 is undivided) is awkward. However, although F is undivided, it is pretty obvious (even where the exit directions are imperfect) when there is a clearing of the stage and hence a scene break. Since these breaks would also have been observed in performance, the obvious and most historicized choice would have been to divide the play into a sequence of scenes and use scene and line-numbers for reference. The only pressure against such a decision comes from the inherited editorial tradition, and Cox has been boasting about not respecting that tradition. Importantly, this choice would have alerted ordinary readers who are unaware of when act intervals began to be followed in performance that the play was written for performance without intervals.

    In "Two varieties of digital commentary" (pp. 194-209) John Lavanigno (one of the general editors of the Oxford Complete Middleton) surveys the purposes of commentary in critical editions, from the reader-friendly explanation of unfamiliar words and ideas to the complex tracing of critical opinion in a variorum, and discusses how electronic publication will bear upon these. Essentially, the former are already satisfactory in the print medium and cannot be improved upon, while the latter (which generally involve going beyond the work one is reading to look at other works about it) can be aided by electronic editions' capacity to include longer quotations than are permissible in print. Thus editions of the sonnets have borne longer notes because i) there is more room (the poems are shorter to print than a play), and ii) the reader can spare more time to read the notes since she is not trying to follow the plot. A poem can be consumed in one gulp and then the notes on it can be read; this is not true of a play, for which the reader needs help as she reads. Lavanigno makes the correct but often overlooked point that mouse-clicking to get to a commentary note is not as easy as glancing down the page, and that in general computer interfaces are currently inferior to books for many uses.

    The final essay in the collection--Barbara Hodgdon's "New collaborations with old plays: The (textual) politics of performance commentary" (pp. 210-223)--is among the weakest on matters bibliographical, although the writer is renownedly perceptive regarding performance. Hodgdon begins with the vulgar generalization that the Oxford Complete Works of Shakespeare designated the Folio as a "collection of 'performance-tested' texts" (p. 211). The truth, of course, is that certain Folio texts (but by no means all) seem to be based on manuscripts that have been used in the theatre. Hodgdon is clearly not familiar with the edition she is characterizing, for she includes it amongst those that have an apparatus designed to alert readers to "matters etymological, literary, historical, and socio-cultural". The Oxford Complete Works's apparatus is in fact confined to matters textual. Hodgdon complains about the theatrical uselessness of comments in various editions, but I cannot see what she is objecting to: is it that editors comment at all on performance matters, or that their comments are not the ones that Hodgdon would make, or that they are too prescriptive? Hodgdon is not telling, and quotes the editors as though they were damned out of their own mouths. Concerning editors who add "Exeunt attendants" when Richard 2 asks for Mowbray and Bolingbroke to be brought before him in the opening scene of Richard 2, Hodgdon objects that Gaunt (whom Richard has been addressing) might be the man to do the fetching, just as Gloucester is sent to fetch in the rivals for Cordelia's hand in the opening scene of King Lear, which she calls "A similar exit". It is not similar: Gaunt is the father of one of the two men being fetched in and they are well on the way to trying to kill one another, whereas Burgundy and France are unrelated to Gloucester and he has no stake in the outcome of their relatively friendly rivalry.

    Hodgdon thinks that in the case of Richard 2 no entrance is even necessary, since the actors' 'entrance' could be marked with lighting. (That technical possibility has been available to practitioners for less than half of the play's 400-year stage history, but one gets the sense that Hodgdon is uninterested in theatre practice before the invention of the arc lamp.) Extraordinarily, Hodgdon seems to believe (but gives no reason for it) that Folio Richard 2 was printed from "a manuscript playbook" when in fact it was primarily printed from a copy of Q3 with some promptbook annotations copied in. She claims that the direction "Exit Gaunt" (1.1.195) has "Folio provenance, probably deriving from the manuscript playbook", which is one explanation for its not being in Q; the other is of course that Q omits it by mistake. Hodgdon thinks that Gaunt's leaving is unnecessary since he and the duchess of Gloucester can simply remain on stage to play 1.2 (as in a recent production). That is only possible if one overrules the direction, agreed upon by Q and F, that a new scene is started with an entrance direction for him and her. Even the most postmodern of performers cannot enter twice in succession without an intermediate exit. If Gaunt has to exit in 1.1 (as seems certain), it were better he did it early to avoid breaking what seems to be a general rule (the Law of Re-entry) governing the theatre practice that Shakespeare worked within.

    Hodgdon objects to editors privileging original performance conditions in a Globe-type space, because these conditions are "largely irrecoverable" and we know that they also played in "private theatres" (by which she means indoor hall playhouses) and at court and at touring venues. Actually, of the plays she has discussed so far--A Midsummer Night's Dream, Antony and Cleopatra, Macbeth, The Taming of the Shrew, All's Well that Ends Well, King Lear, Richard 2, and The Winter's Tale--only The Winter's Tale was written when Shakespeare's company could expect to perform in an indoor hall playhouse, and even that play's known stage history begins with a performance at the Globe. When Hodgdon begins to use bibliographical data seriously, she misreads it. Thus she cites the Oxford Textual Companion for her claim that ". . . some editors conjecture that [the Folio text of] Shrew was printed from a transcript based on minor theatrical adaptation" and hence that the entrance of Baptista, Gremio, and Tranio just before Petruccio says "For I am he am born to tame you, Kate" (2.1.270) might have early modern staging authority. What the Textual Companion actually says at the point cited by Hodgdon (p. 170) is that "The Folio text might therefore derive from Shakespearian foul papers, or from a transcript which has undergone some minor theatrical adaptation at a later date. It might also derive from collaborative foul papers . . . ". Wells and Taylor invoke adaptation in the transcript to try to account for F having what they think are features of foul papers and features of a transcript and the "later date" caveat comes from an apparent allusion to a character Soto in John Fletcher's much later play Women Pleas'd. So, Wells and Taylor write of the theatrical adaptation of a manuscript and Hodgdon misreads this as referring to a transcript based on a theatrical adaptation. In a book about performance (Hodgdon's specialism) this might just be excusable, but in a book of essays about matters textual it smacks of editors not reading the contributions carefully nor asking for corrections of palpable error. It is impossible not to wonder if this is why in their introduction Erne and Kidnie "do not here attempt to introduce the essays one by one in the received manner" (p. 6).

    The second book of essays relevant to this review is a festschrift: Theatre Histories: An Introduction. Of its 10 essays, 6 are not about texts of Shakespeare and are not reviewed here although they are all superb. In "The Troublesome Reign, George Peele, and the Date of King John" (pp. 78-116) Brian Vickers argues that Peele wrote Troublesome Reign and since Shakespeare's King John must have been written later, Shakespeare was the borrower. This is contrary to the 'early start' theory of Shakespeare's chronology, and it depends upon Jackson's demonstration (the article "Pause Patterns" reviewed in YWES 83) that Shakespeare's Richard 2 and King John were written around the same time. Since composition of Richard 2 cannot be reassigned to the early 1590s, we can be sure that King John borrows from Peele's play and not the other way around. Perhaps surprisingly, the two central essays of the book undermine an attribution (that "A Lover's Complaint" is by Shakespeare) upon which Jackson's early reputation was built; doubtless the contributors thereby express admiration for their subject's openness to new ideas and evidence. In "Did Shakespeare write A Lover's Complaint? The Jackson ascription revisited" (pp. 117-40) Ward Elliott and Robert J. Valenza give ample reason for thinking the poem is not Shakespeare's. The Elliott/Valenza method of counting showed them that each 1500 word block of a new Shakespeare play would introduce 20-30 new words (new to his vocabulary) and about 320 rare words (words he used fewer that 100 times elsewhere). From the beginning, this method raises questions in the reader's mind that Elliott and Valenza do not answer, most pressingly whether notions of 'new' and 'rare' words are synchronic or diachronic. That is to say, it is not clear from their prose whether a 'rare' word is one that Shakespeare had seldom used before writing the block under consideration (the diachronic view) or do they mean that it is 'rare' across the whole Shakespeare canon? This is worth asking because a word might be 'rare' in early Shakespeare but common later in his writing.

    Within those ranges (20-30 new words and about 320 rare ones), there would be expected variation depending on whether the work was relatively rich or poor in vocabulary. This is measured by the ratio of tokens to types: in any student's 2000-word essay there are 2000 tokens, but a good MA-level essay might have 700 different types (differing words) and a bad essay fewer. If the work was rich in vocabulary (types divided by tokens is high) then it would have a greater number of new and rare words than would be case were it weak in vocabulary. Moreover, the slope of the most-rare to least-rare (but still absolutely rare, occurring fewer than 100 times in the canon) is distinctive too: there is a characteristic Shakespearian steepness to the slope. Jackson's 1965 monograph attributing "A Lover's Complaint" to Shakespeare had discovered all this methodology before, and moreover, Jackson made the proper word/string distinction that Elliott and Valenza's computers could not. That is to say, to a human the three letters making up 'r-o-w' can form different words, being a verb in certain contexts and a noun in others. Unless it is using specialist lexical software to attempt the same kind of distinction, these three letters are to a computer merely a singular alphabetical string of characters. This weakness of computer analysis does not matter as much as one might think, since all the counts performed by the computer will fail to make the distinction, and so all the ratios--say of rare words to common words--will be affected by the same amount. Jackson found that Shakespeare introduced about one new word to his vocabulary every 20 lines in early plays, about 1 in 10 lines in King Lear and Hamlet, and 1 in 11 or 12 lines in Lucrece and Venus and Adonis, and since the rate for "A Lover's Complaint" is 1 in 7 lines, Jackson declared it to be within the Shakespeare range.

    Repeating Jackson's test with the better samples of others' work that we now have, Elliott and Valenza overturn the attribution. We can now detect upper and lower limits of the range of new-word introductions in other writers as well as Shakespeare, and it turns out that the discriminating attribute is the upper bound of new-word introduction. That is to say, what matters is the highest rate of new-word introduction that a given writer can manage. Writing that is not by Shakespeare can be spotted by its having too many new-to-Shakespeare words, although of course those words might well be not new to the person who wrote them. "A Lover's Complaint" ought, by Elliott and Valenza's prediction, to have 55 new strings but has in fact 88: that makes it a little more unlike Shakespeare than the most anomalous Shakespeare. That is, it is like the top 5% of unShakespearian writing that Shakespeare did, such as the French bits of Henry 5, and it is quite like how non-Shakespearian material compares to Shakespeare's norm. (The obvious objection here is that the French bits of Henry 5 are 100% genuine Shakespeare, and a test that rejects one-twentieth of his work as not his own is a poor test.) Elliott and Valenza describe two other tests. One measures the rate at which no is used, expressed as a proportion of how often no or not are used; they call this "no, divided by no plus not" which is rather compressed and cryptic for most readers. The other measures the rate of use of with as the penultimate word in a sentence. Sentence boundaries are often editorial in modern editions and compositorial in early printings, so something ought to be said about how these non-authorial factors might influence the test. Since Shakespeare was habitually light in his punctuation, one would want convincing that this second test would nonetheless be accurate: are sentence boundaries perhaps obvious and the placings of periods (so the machine can catch them) essentially uncontested?

    These tests were shown to be the most sensitive discriminators of Shakespeare in extensive validation of lots of tests, and "A Lover's Complaint" fails them both. The value of no/no+not is 0.184 to 0.536 in Shakespeare's poems and 0.167 to 0.4 in his plays, and "A Lover's Complaint" scores 0.12. Shakespeare's penultimate-with rate is 0.004 to 0.034  but for "A Lover's Complaint" the rate is none. It is worth noting that Elliott and Valenza admit that known Shakespearian works occasionally fail their tests too: their faith in the tests comes from the fact that known Shakespeare writing rarely fails the tests for 'is it Shakespeare?' while known non-Shakespeare writing routinely fails the same tests. Their confidence comes from the relative differences in how often Shakespearian and non-Shakespearian writing fails the tests. Elliott and Valenza ponder some objections to their tests and see how far they could go in adjusting the assumptions to let "A Lover's Complaint" pass a few of the tests it fails: it still fails way more tests than known Shakespeare writing usually does and in order to pass the tests it has to be treated as early Shakespeare, whereas of course everyone who says "A Lover's Complaint" is Shakespeare's says it is late Shakespeare. The authors give a description of a way of visualizing the likelihood of all these statistics misleading the user, using astronomical distances to make the point about how unlikely it is Shakespeare would write material so unlike his usual stuff. This way of presenting the evidence is behind their recent public pronouncements (for example on the email discussion list SHAKSPER) of the kind 'text x is 10,000 times more likely to be by someone else than it is to be by Shakespeare'. By the tests described here, "A Lover's Complaint" is not Chapman's either.

    Reinforcing Elliott and Valenza's conclusion is Marina Tarlinskaja's essay "The verse of A Lover's Complaint: Not Shakespeare" (pp. 141-58). Tarlinskaja's tests cannot be done on a computer: although objective, there are "few formal textual indicators". Tarlinskaja describes her method of counting the actual stresses in the odd and even syllabic positions in lines of poetry, and her making of an average of how often the expected-to-be-stressed positions (S positions) are actually stressed over the average of how often expected-not-to-be-stressed (W positions) are not stressed. In this she seems to be saying that the choice for each syllable is binary (it is or is not stressed) and that the averages expressed as percentages are how often the expectation is fulfilled; not that the amount of stressing on each occasion varies. However, even with binary choices there is surely, at least for monosyllabic words, a choice in the mouth of the speaker and stress does not actually manifest itself until the thing is spoken. I confess that I do not understand how monosyllabic-word stress could encoded in writing and on a binary basis I would expect there to be 64 (that is 2⁶) ways for a speaker to stress the 6 syllables of "The cat sat on the mat". Tarlinskaja charts how Shakespeare's stress habits changed over his career: early on the least frequently stressed midline S position is 6 and late it is 8, with the mid-career habit being 6 and 8 equally often losing their expected stresses. This follows from the mid-line break shifting: early in the career the line usually breaks into 4+6 syllables or 5+5, and later on it is usually 6+4 or 7+3. Likewise the location of syntactic breaks (and speaker-change breaks) moved rightward over the career.

    So, the three tests are: stress profile (averages of fulfilled expectation of stress in each of the 10 positions), syntactic-break location and its opposite the syntactic-link (explained below), and proclitic (forward leaning) and enclitic (backward leaning) microphrases where an monosyllable in a W position gets a bit of stress from the S that follows or precedes it. The result is that "A Lover's Complaint" does not look like mature Shakespeare at all, and is closer to early Shakespeare but even then has marked differences. The most reasonable conclusion is that it is not Shakespeare. Coming after Elliott and Valenza who made much of how they did it, it is noticeable that Tarlinskaja does not validate her tests: that is, she does not describe how well they discriminate a known non-Shakespearian text from a piece of known Shakespearian text. The degree to which her tests are subjective becomes clear in her closing remarks that relate word boundaries to phonetic boundaries, for example adjectives with their long unstressed tails ('ugly' pronounced 'UGlyyyy') create feminine word boundaries while verbs with the long unstressed necks ('permit' pronounced 'perrrrMIT') create masculine word boundaries. The latter, of course, emphasize the foot structure of the line. Tarlinskaja explains the 3 degrees of syntactic link in the second of her tests: she counts as 'strong' such links as those between a modifier and a modified noun and between verb and object, as 'medium' such links as those between a subject and its predicate and between words of no syntactic link, and as 'weak' those where there is a definite syntactic break such as a change of speaker.

    Finally from this collection comes John Jowett's "The Pattern of Collaboration in Timon of Athens" (pp. 180-205). This covers the same ground as his edition of Timon of Athens reviewed above, providing more of the detail than he give there about the reasons for accepting R. V. Holdsworth's division of the shares between Middleton and Shakespeare. Awkwardly, though, Jowett's quotations here are keyed to the Oxford Complete Works text of the play rather than Jowett's new edition. Broadly speaking, Shakespeare did the first work on the play and Middleton did the revising of the text. Jowett attacks Jeffrey Masten's argument on the sociability of dramatic collaboration: claiming that language is socially produced (and so one cannot distinguish individual hands) tends to restore us, by default almost, to the single-author-centered approach that we know is not how dramatists of the period tended to work. To be sure, there are times when the writing is a merger of two men's labours and cannot be unentangled, but just as often it is not so entangled and can be apportioned. Jowett summarizes the shares thus: 1.1 is Shakespeare, 1.2 is Middleton, 2.1 and 2.2 are mixed, 3.1 to 3.6 are Middleton, 3.7 to 4.2 mixed, 4.3 to 5.5 are Shakespeare with a Middletonian insertion of 77 lines.

    The last World Shakespeare Congress produced just one published essay of relevance to this review, and in it Gary Taylor establishes the remarkable fact that Measure for Measure was originally set by Shakespeare in the Italian city of Ferrara ('Shakespeare's Mediterranean Measure for Measure', in Tom Clayton, Susan Brock and Vicente Fores (ed.) Shakespeare and the Mediterranean: The Selected Proceedings of the International Shakespeare Association World Congress at Valencia 2001 (Newark DE: University of Delaware Press [2004])). The location 'Vienna' is said often at the beginning and end of the play, but would have meant almost nothing to the first audiences: this was the only play in the period set anywhere in Austria. In fact, Measure for Measure does not even mention Austria as being where Vienna is located, not does it mention the geographical notion of Germany. Vienna is mentioned in Hamlet ("the image of a murder done in Vienna" 3.2.227), although quite possibly this is a misprint or misrecollection for Urbino, since the source is the murder of the duke of Urbino. Vienna was known for its being under threat from the neighbouring Ottoman Empire (the Turk), but nothing is made of that in Measure for Measure. Claimed allusions to contemporary events (including foreign court visits to London) around the time Measure for Measure was written are all mistaken, and Taylor shows why. The bit of the play that the contemporary (1604) allusions to Hungary are supposed to explain is Lucio's talk about the dukes coming to composition with the king of Hungary (1.2), which the Oxford Complete Middleton has now dated to 1621. Indeed John Jowett has located a precise English newsletter source for Lucio's line in that year, and in 1621 Vienna was hot news: it was the capital of the Holy Roman Empire under Ferdinand II who was hated in England for his campaigns against European Protestants. So it was Middleton who, in adapting the play, set Measure for Measure in Vienna. Obviously from the characters' names it was originally set in Italy, a place known for lechery. The duke of Ferrara is a sexual blackmailer in one of the sources, and Ferrara was a city well-known to Shakespeare from his reading and familiar to his audiences from other plays. This is particularly true of plays in the 'disguised ruler' genre, especially Middleton's The Phoenix and Marston's Parasitaster or the Fawn, both set in Ferrara. In all, a quarter of all early modern plays are set in Italy. Ferrara lost its independence (ceased to be a dukedom) in 1598 and came under the Papal Empire, which made it in the eyes of many a warning of what could happen in England. This event also meant that there was no living duke to take offence at the play. It required just 9 word changes (with no disruption of metre) for Middleton to relocate the play from Ferrara to Vienna. This explains how come there is a Vincentio in the Folio dramatis personae but his name is not mentioned in the play: it was crossed out of the first line during the adaptation as being inappropriately Italian for the new setting; the other names Middleton let stand.

    In a splendid essay within a collection otherwise irrelevant to this review (but highly recommended), Paul Eggert argues that the new materialism in relation to Shakespeare's texts ditches at its peril the notion of authorial agency ('The Way of All Text: The Materialist Shakespeare', in Raimonda Modiano, Leroy F. Searle and Peter Shillingsburg (ed.) Voice, Text, Hypertext: Emerging Practices in Textual Studies (Seattle WA: University of Washington Press [2004])). Eggert's expertise is in editing nineteenth- and twentieth-century texts, and it gives him insights on the recent 'material Shakespeare' shift typified by the work of Peter Stallybrass and Margreta de Grazia and their respondents. The essence of the Stallybrass/De Grazia position is that after 1800 editors sought to regularize the textual variability of the Shakespeare works in order to create the ideal 'Shakespeare' that suited their new ideas about eternal truths of human nature. G. Thomas Tanselle's proposed distinction between the text of a document and the text of a 'work' (someone's, not necessarily the author's, intended text) does not get us off the Platonic hook: this is still idealism, even if not authorial-idealism. Stallybrass and De Grazia accept that their ideas require us to reformulate the notion of 'work' in relation to Shakespeare, but they do not provide the reformulation. Refocussing our attention on the multitude of labours inscribed in an early printing, and the multitude of meanings that early-modern conventions of spelling and punctuation allowed, solves nothing if it leaves out the reader as much as the old-fashioned author-centered view did. Moreover, in the new view (what Eggert calls looking at the material surface of a printing rather than the ideal that lies beyond it) the reader still wants to know whether a particular multiplicity of meanings is Shakespeare's own or simply "passive witness" to the prevailing writing and printing conditions. Indeed, Stallybrass and De Grazia just assume that we cannot attribute the multiplicity to Shakespeare (they assume it is the effect of wider culture) but to make that pronouncement we have to distinguish Shakespeare's input from the other determinants, so we are back where we started with trying to isolate the authorial element. If we just assert that we cannot tell the difference between Shakespeare's and all the other agencies that go into a printing, we just "trade one abstraction (authorship, which, whatever its illusions, has at least spawned finely differentiating analytical methods) for another abstraction ('materiality', as de Grazia and Stallybrass call it, which so far has not)".

    Here Eggert quotes an essay by Graham Holderness, Bryan Loughrey, and Andrew Murphy that objects to De Grazia and Stallybrass on Marxist theoretical grounds but entirely misunderstands the notions of use-value and exchange-value, seeming to think that a one-off manuscript's exchange value is the same as its use-value whereas a printed text's exchange-value is "manifestly more marked" than its use-value. That is, they seem to think that a play becomes a commodity when its printed, which is not true of plays although it is somewhat true of books considered as objects. Eggert insists that we cannot just wish away the notion of a 'work', and notes the irony that Holderness, Loughrey, and Murphy are happy to cite their own works in the footnotes to their essay. The important difference that is being lost here, he argues, is that between unique objects (autographs) and exact-or-variant copies of them (allographs). Eggert observes that Shakespearians are poor in raw materials: they would not dismiss the notion of foul papers as an abstraction or an idealization if they worked on nineteenth-century writers, for whom these documents are in abundance. The problem with the 'material Shakespeare' movement is that it is hindered by a basic mistake about terminology, which makes it fail to distinguish the physical from the mental. Indeed, in the materialists' mouths 'text' seems to mean both or either according to the argument being made.

    Eggert argues that there is nothing wrong with the notion of personal agency in relation to these questions. 'Text' is a poststructuralist term meant to indicate how discourses inscribe texts and people, "who could then be imagined as the provisional sites of discursive traces, rather than as unified or stable entities". This requires that we give up the notion of a 'work' altogether, but if we do that it is hard to see how we would link the two quartos and Folio Hamlet together at all. Bibliography is so well advanced in distinguishing just what changes in the transformation of manuscript to print and what changes between different printings that the notion of 'work' (that which does not change, or at least that which ties these various manuscripts and printings together as a group) "seems to arise naturally from these empirical methods". The alleged transference of authority to the reader in hypertext is illusory: the reader follows paths already laid down, and in any case we have the same links (albeit operating more slowly) in print editions. (This is quite correct: the true hypertext is the library and the point of working in as big and as comprehensive as library as one can visit is that in such a place the links pointing outwards from one's present reading matter to other matter are likely to be followable more quickly than if one were relying on postal services every time one wanted to chase up the reading pointed to by a footnote.) Eggert concludes that to merely gaze at the pages not through them, as the materialists seem to advocate, will not help create editions that answer questions that the reader legitimately asks in the act of reading. Although he does not say so, we might observe that such a pointless edition has already emerged from this postmodern fog of ignorance: the New Folger Shakespeares doggedly refuse to tell the reader most of what she might want to know.

    Finally amongst the chapters-in-books, Graham Holderness argues that we can reconstruct the Macbeth that Simon Forman saw in 1611 and which, had it been printed, would be called Q1-Macbeth (''To be Observed': Cue One Macbeth', in Evelyn Gajowski (ed.) Re-visions of Shakespeare: Essays in Honor of Robert Ornstein (Newark DE: University of Delaware Press [2004])). The Folio text of this play has 'bad' quarto qualities: short, full of cruces, crowd-pleasing in its visuals, and beset with confused stage directions. The topical allusions that date the play to around the Gunpowder Plot are uncertain, and there is no direct evidence of a court performance in 1606. The Padua promptbook of 1625-35 shows that even this short play was cut for performance, and together with the post-Restoration transformations that the play underwent this should make us cautious of assuming that F represents the play as it was performed in Shakespeare's time. Since we know that plays existed in different performed versions, we should treat Forman's notes as a witness to a lost text since it departs from F's action in crucial respects, such as by having Macbeth and Banquo ride through a wood, by saying that this is the "first" thing that happened whereas F starts with the witches and then has the 'bleeding captain' scene, by calling the witches "feiries or Numphes", and by omitting the dancing, the cauldron/apparitions scene, the Porter, the suborning of the assassins, and the discussion of the fate of Scotland by Malcolm and Macduff. Of this list of differences, I would object only to the second: Forman's phrase "ther was to be observed firste howe . . ." means 'these are the things to remember about the play, firstly . . .' not 'the first thing in the play was'. Why, Holderness asks, should we assume that the wood that Forman says he saw Banquo and Macbeth riding through at the beginning of the play was just a memorial contamination from Holinshed? We know, after all, that there was an attempt to show a wood onstage near the end when every man is told to drop his lopped-off bough. (I have an objection here too: those boughs are hand-held properties that viewed close-up would be hard-pressed to connote a real wood; that is why they are discarded as their holders approach Macbeth's castle.)  Holderness describes what should be done to F to make it like the play Forman seems to describe, and wonders if the material Forman omits was written specially for court performance in 1606 and was not staged at the Globe in 1611 and did not make it into the official book of the play until some time between 1611 and 1623. (This is a lot of supposing in order to defend the fidelity of Forman's recollection; might he not just have forgotten some of the play? I do.) That William Davenant's adaptation of the play is preceded by a summary that is very like Forman's account gives, says Holderness, "additional weight to the argument" that there was a kind of lost Q1 Macbeth that Forman's account describes. Holderness ends by presenting the relevant scenes of Macbeth reconstructed and adapted as necessary to suit Forman's account.

    So to the journal articles. Over the last few years the publishers of The Year's Work in English Studies have moved the submission date for copy progressively earlier, and because certain journals are rather behind in their publication schedules--not having produced their volumes for 2003 let alone those for 2004--these cannot be seen in time for inclusion in this review. This year the YWES deadline made it impossible to see the 2004 volumes of the annually-published books TEXT, Studies in Bibliography and The Shakespearean International Yearbook. (It has not escaped this reviewer that there is an irony in his objecting to deadlines creeping forward, since the object must be to solve the very problem--late publication--that he is complaining about.) The longest articles were two in the journal Early Modern Literary Studies. In the first, Sonia Massai describes (as she did in the chapter-in-book reviewed above) what Michael Best's Internet Shakespeare Editions can and will do, via a description of her edition in the series ('Redefining the Role of the Editor for the Electronic Medium: A New Internet Shakespeare Edition of Edward III', Early Modern Literary Studies 9.3 [2004].n. pag.). Essentially this is that she can be bolder in her emendation because the early printings are only a mouse-click away, and she can use animations to show textual differences. Jennifer C. Forsyth's essay addresses the same issues and asks where one is to stop, given e-texts' capacity for copious noting ('Playing with Wench-like Words: Copia and Surplus in the Internet Shakespeare Edition of Cymbeline', Early Modern Literary Studies 9.3 [2004].n. pag.). Forsyth prefers the eighteenth-century editors' endless debates amongst themselves to the majesterial approach of, say, The Oxford Shakespeare that seems univocal and simply tells the reader the meanings of the words. Those old editions also made more plain than do the modern editions just whose Shakespeare the reader was getting. Around the middle of her essay Forsyth rambles off into a dull story about 'what I found on the Internet is an obsession with user feedback' (unworthy even of online journalism), but she pulls the piece back on track by referring this to the reader's agency in new electronic editions. We could, she says, do a Wikipedia Shakespeare in which any reader would be allowed to change the text. The problem Forsyth ignores is obvious: as this review was being written, the former presidential aide John Seigenthaler was objecting most strongly to a Wikipedia entry about John F. Kennedy that implied his involvement in the assassination. The malicious entry, it turns out, was made as a prank.

    Jayne M. Carroll and MacDonald P. Jackson show that Arden of Faversham is at least partly by Shakespeare ('Shakespeare, Arden of Faversham, and 'Literature Online'', ShN 51 [2004].3, 4, 6). The canons of 5 candidates for its authorship were formed from texts in Chadwyck-Healey's Literature Online (LION) database: Robert Greene, Christopher Marlowe, George Peele, Shakespeare, and Miscellaneous (comprising Thomas Kyd, Anthony Munday, Thomas Nashe, and Thomas Lodge). The canons were of roughly equal size and included all the plays of Greene, of Marlowe, and of Peele, but Shakespeare was cut down to The Two Gentlemen of Verona, The Taming of the Shrew, Richard 3, The Comedy of Errors, Romeo and Juliet, Richard 2, and Venus and Adonis. Sample passages from Arden of Faversham were checked against these canons for phrases/collocations shared only with one of the 5, using the method described in Jackson's book Defining Shakespeare (reviewed here last year). The authors make no mention here of the fact that the LION texts of Shakespeare are all taken from F1 ; presumably spelling differences were allowed for in the searching, since the New Mermaid edition was used as the source of Arden of Faversham's text. Overwhelmingly, they report, Arden of Faversham has more links with the Shakespeare plays than the others. Opening up the same searches to include all that Shakespeare wrote up to 1600 (and then scaling the number of links down to reflect how much this inflates the comparative canon) only made the case even stronger. Looking just at the links with the three Henry 6 plays, there were far more with 2 Henry 6 and 3 Henry 6 (the ones thought to be largely Shakespeare's alone) than with 1 Henry 6 (known to be co-authored) and of those to 1 Henry 6 the links were to scenes thought to be Shakespeare's. All this does not make Arden of Faversham solely Shakespeare's work, but it strongly suggests that it is at least partly by him

    What remain are not substantial essays, only notes. Rodney Stenning Edgecombe argues that Mercutio's "Young Abraham Cupid, he that shot so trim" (Q1, "so true" in Q2 and F) should be "abraying Cupid" since this smoothes the metre and has the required sense of waking someone up ('A Take on 'Abraham Cupid' in Romeo and Juliet 2.1', ShN 53 [2004].129). In the 1609 printing, Sonnet 146 reads "Poore soule, the center of my sinfull earth, | My sinfull earth these rebell powers that thee aray", and the problem is the obvious repetition between lines. Like I. A. Shapiro (whose solution was reviewed in YWES 81), Naseeb Shaheen thinks the solution lies in alliteration but also it needs to agree with the structure of the sonnet, which is that the turn-around happens towards the end. Thus Shapiro's "Beat down these rebel powers" is no good. Shaheen proposes "Ruled by these rebel powers" ('Shakespeare's Sonnet 146', English Language Notes 41.4 [2004].15-9). Gretchen E. Minton and Paul B. Harvey Junior attack the problem that Q and F Troilus and Cressida agree on Pandarus teasingly saying to his niece the morning after her night with Troilus: "a poore chipochia, hast not slept to night?" (4.2.34), which editors since Lewis Theobald have emended to capocchia, meaning 'blockhead' but also 'penis-head' (vulgarly, we might say 'dickhead'). Why would Pandarus call Cressida a dickhead? He would not: the unemended word is an Anglicization of che (Italian for that) + poccia (Italian for nipple, meaning clitoris), which suits an address to the newly sexually-active Cressida.

    Arthur John Harris and Frankie Rubinstein think that Jessica's "If on Earth he do not merit it [Portia's Heaven-on-Earth love] in reason he should never come to heaven" (The Merchant of Venice 3.5.72-73) should be restored to Q1's mean it because 'mean' meant 'long for sexually' and 'pander for monetary gain' and 'take the middle ground', all of which apply to this situation and are recurrent themes of the play ('Shakespeare's The Merchant of Venice', Explicator 62 [2004].70-4). Robert B. Hamm Junior looks at Tonson's landmark Shakespeare, Rowe's 1709 book that has been seen as the first modern edition ('Rowe's Shakespear (1709) and the Tonson House Style', College Literature 71 [2004].179-205). It is that, he acknowledges, but it is also merely typical of what Tonson (in concert with Cambridge University Press) was doing with a new layout of text, and with improved standards of editing and printing to produce a standardized-appearance series of volumes across English vernacular literature and aimed at the rich members of the Kit-Cat club. The Folio size connoted importance, but the 2nd, 3rd, and 4th Shakespeare Folios were successively poorer in quality and the rising cost of paper largely killed off the format by the end of the seventeenth century, to be replaced by the multi-volume small-format edition that afforded portability. Paul D. Cannan describes the importance of Charles Gildon's unauthorized 7th volume (comprising the poems plus Gildon's essay on the Shakespeare canon) published in 1710 to cap Rowe's 6-volume Tonson set ('Early Shakespeare Criticism: Charles Gildon, and the Making of Shakespeare the Playwright-poet', MP 102 [2004].35-55).

    Finally to the round-up from Notes and Queries. James D. Carroll argues that Gorboduc being a source for Titus Andronicus explains the things that Brian Vickers has put down to Peele's authorship of Act 1 ('Gorboduc and Titus Andronicus', N&Q 249 [2004].267-9). The plays have themes in common (true, but only general ones such as the young taking over from the old) and Carroll pointlessly lists words and images that they have in common: commonweal, reproachful, lawless, ruthful, wrongful, entrails, sufficeth, aloft, sacrificial smoke, and weaning tigresses. This is pointless because positive likenesses at the level of words and images cannot settle attribution, and moreover Carroll does nothing to dislodge the detailed evidence summarized by Vickers pointing to Peele's authorship; indeed he does not even address Vickers's arguments. Moreover even if the borrowing of words and images from Gorboduc to use in Titus Andronicus were accepted (and the evidence here is far from compelling) that would not tell us anything about who the writer is. Roger Prior thinks that Tasso's pastoral play Aminta (read by Shakespeare in the original Italian) is a source for Love's Labour's Lost and As You Like It ('Tasso's Aminta in Two Shakespearian Comedies', N&Q 249 [2004].269-76). Prior admits that Sidney Lee spotted that Biron's defence of love "has parallels" with the end of Act 2 of Aminta, but Prior thinks there is more borrowing at work. He gives the examples from both Shakespeare's plays and they are not terribly close to the alleged source. Slightly more persuasively, he points out that the lover-fainting-at-the-sight-of-a-bloody-cloth idea is not in Shakespeare's main source, Lodge's Rosalynde, but it is in Aminta. A. B. Taylor finds that Titania's speech about earthly disorder ensuing her quarrel with Oberon (A Midsummer Night's Dream 1.2.82-117) comes from Plato's Symposium ('Plato's Symposium and Titania's Speech on the Universal Effect of Her Quarrel with Oberon', N&Q 249 [2004].276-8). Eryximachus (the doctor) accepts Pausanias's claim that there are two kinds of love, and goes further: the orderly one brings harmonious wealth and agricultural plenitude, and the disorderly one brings the kinds of disorder that Titania's speaks of. I cannot help thinking that it is odd to describe Titania and Oberon as having disorderly love; are they not, in fact, out of love?

    William Lloyd gives good reason for thinking that the copy for Q1 Richard 2 was scribal transcript, not authorial papers, to judge from certain style preferences ('Scribal Copy for Q1 Richard II?', N&Q 249 [2004].280-3). It is a close call whether Q1 was set from authorial papers or some kind of transcript of them, the main piece of evidence being the spelling "Oh" where Shakespeare would have written "O"; this suggests transcript. Lloyd has additional evidence that it was a transcript: he presents a table of the choices of contractions of th' in verse and prose, and of among/amongst, between/betwixt, while/whiles/whilst in Richard 3 (Folio text), Venus and Adonis, Lucrece, Love's Labour's Lost, A Midsummer Night's Dream, Romeo and Juliet (Q2 text), King John, The Merchant of Venice, 2 Henry 4, Henry 5 (Folio text), Sir Thomas More, Richard 2, and 1 Henry 4, all chosen for closeness to Richard 2 in style and date. This table shows that Shakespeare had strong preferences for among over amongst, for between over betwixt, and (somewhat less strongly) for while over whiles or whilst. The editions chosen are, where possible, the ones thought to represent Shakespeare's habits most closely, so for example good quartos made from authorial papers. Q1 Richard 2 bucks the trend: it shows a preference for betwixt over between and for amongst over among and whilst over while. Also, it never contracts the, which is odd. In all these things Q1 Richard 2 is like Q1 1 Henry 4, which is generally accepted to have been printed from scribal copy. Q2 Richard 2 also has other features like the ones in Q1 1 Henry 4 that caused the Oxford Complete Works editors to decide that the latter was printed from scribal copy, such has the phrasing of stage directions that avoids use of and and with, and the use of the Latin plural manent. Lloyd speculates that the same scribe was responsible for the copy that underlay Q1 Richard 2 and Q1 1 Henry 4 and wonders aloud whether there is a connection to their both being censored: the (allegedly, I would say) cut deposition scene and the change of Oldcastle's name. While admitting that this is all speculation, Lloyd ends by noting that an editor who changed all the occurrences of betwixt to between and whilst to while in these two plays might be undoing a couple of genuine Shakespearian choices but would overwhelmingly be putting back Shakespearian preferences that a scribe had undone.

    J. J. M. Tobin thinks that Shakespeare borrowed from Nashe's Christ's Tears Over Jerusalem for Henry 5 2.3 ('Another Psalm for Falstaff', N&Q 249 [2004].283-4). The reason is that the Hostess says "Now I, to comfort him, bid him a should not think of God; I hoped there was no need to trouble himself with any such thoughts yet" (Henry 5 2.3.19-21) and Nashe in Christ's Tears Over Jerusalem quotes Psalm 77 using those words (plus flea, which of course comes into the same scene in Henry 5 regarding Bardolph's nose). Having satisfied himself with those, Tobin gives some more examples of Shakespeare borrowing from Christ's Tears Over Jerusalem in Henry 5. A few pages later in Notes and Queries Tobin gives a string of verbal parallels (not just the howling of his title) between King Lear and Nashe's Christ's Tears Over Jerusalem ('Lear's Howling, again', N&Q 249 [2004].287-91). They are all commonplaces. Robert Ellrodt defends certain emendations in his recent French-language edition of Shakespeare's Sonnets ('The Limits of Interpretation in Shakespeare's Sonnets', N&Q 249 [2004].291-5). I wonder why he did not put them in the book; I cannot imagine most readers will come across them here. R. E. Pritchard thinks that Shakespeare got Prospero's evocation of the dissolving masque from Thomas Coryate's epistle to the reader in his Crudities ('Shakespeare and Thomas Coryate', N&Q 249 [2004].295-6). This epistle refers to "gorgeous Palaces, impregnable Castles and Fortresses, Towers piercing in a manner up to the cloudes", which is quite like Prospero's "The cloud-capped towers, the gorgeous palaces, | The solemn temples" (The Tempest 4.1.152-3). Lastly, Duncan Salkeld defends Folio Henry 5's "Table of greene fields" as an allusion to the points on a backgammon board, and hence it is not in need of emendation ('Falstaff's Nose', N&Q 249 [2004].284-5). The quadrants of a backgammon board are called 'fields' and they are marked with 'points', so when Shakespeare wrote "for his Nose was as sharpe as a Pen, and a Table of greene fields" his mind sprang from a quill to the points on a backgammon table. Shakespeare contracted the comparison, for what is meant is "and [a point on] a Table of greene fields". Salkeld writes that "The Folio capitalizes exclusively proper nouns in Hostess Quickly's speech, and the emended word was therefore itself originally more like to have been a noun (so, 'Table') than a verb (as in 'babeld')". Salkeld is mistaken on two counts. Firstly, he must mean that F capitalizes nouns, not "proper nouns", since looking just at the parts he quotes we see that the common (not proper) nouns 'nose', 'pen', and 'table' are capitalized. Secondly, even confining Salkeld's meaning to nouns, he is mistaken: F capitalizes the adjective "Christome" (meaning 'innocent') in the same speech.

Books Reviewed

Gary Jay Williams, Phillip B. Zarrilli, Bruce McConachie and Carol Fisher Sorgenfrei Theatre Histories: An Introduction. Routledge [2006].

Lukas Erne and Margaret Jane Kidnie (eds.) Textual Performances: The Modern Reproduction of Shakespeare's Drama. Cambridge University Press [2004].

William Shakespeare Pericles, Prince of Tyre, ed. Roger Warren, The Oxford Shakespeare. Oxford Univerrsity Press [2004].

William Shakespeare Pericles, ed. Suzanne Gossett, The Arden Shakespeare. Thomson Learning [2004].

William Shakespeare and Thomas Middleton Timon of Athens, ed. John Jowett, The Oxford Shakespeare. Oxford University Press [2004].