"Shakespeare: Editions and Textual Studies in 2017" Not the Year's Work in English Studies

Four major Shakespeare editions appeared this year, all for the Arden Shakespeare Third Series. Sukanta Chaudhuri edited A Midsummer Night's Dream, Richard Proudfoot and Nicola Bennett edited King Edward III, Kent Cartwright edited The Comedy of Errors, and Valerie Wayne edited Cymbeline. This brings near to completion the Arden Third Series that began in 1995 and plans have been announced for a Fourth Series. In Chaudhuri's edition of A Midsummer Night's Dream, the Preface (pp. xvi-xviii) remarks that there are no serious cruces to worry about, so the editor's approach is largely concerned with other matters. Although he does not get around to mentioning it until page 306, Chaudhuri bases his text on the first quarto (Q1) published in 1600, and in his Preface he supplements the series rules about textual notes with the following principles: when the lemma is not followed by a named textual authority, it is Q1 and Q2 (the 1619 Pavier quarto) and F (the 1623 Folio edition) can be assumed to substantively agree with it. (As we will see, other Arden editions published this year also depart from the series rules but do not alert the reader in this way.) As abbreviations, "QF" means the substantive agreement of Q1, Q2, and F, and "Q" the substantive agreement of Q1 and Q2. This gives the reader a lot to remember when reading the textual notes and spelling out each siglum individually would have made the notes easier to read while adding only a little to their length.

    Chaudhuri's Introduction (pp. 1-115) makes no mention of the textual status of the play nor which early edition this one is based on, but he repeatedly refers to quarto and Folio texts. His first topic is the play on stage, on screen and in art (pp. 3-44). There are only three datable performances before the Civil War: at court in January 1604 (under the title of the "Play of Robin goode-fellow")--which was probably at Hampton Court on 1 January but Chaudhuri does not mention this--and at Hampton Court on 17 October 1630, and before the Bishop of Lincoln on Sunday 27 September 1631. Chaudhuri acknowledges that not everyone thinks the last of these was Shakespeare's play. Chaudhuri mentions the allusion to Thisbe killing herself with Pyramus's scabbard in Edward Sharpham's The Fleir and a few other minor allusions.

    In discussing the staging, Chaudhuri uses rather old terminology, for instance calling the stage balcony "the upper stage" (p. 5), as mid-twentieth century theatre historians did under the influence of John Cranford Adams's idea, now discredited, that an entire second stage was suspended above the main stage. Also anachronistic in relation to the early stage is Chaudhuri's use of the expression "downstage" (p. 5), since early stages did not slope. Chaudhuri supposes that the onstage audience for the play-within-the-play sat in the stage balcony, "mingling with Shakespeare's" audience (p. 6), but this would be most unlike the uses of the stage balcony in other early modern plays, which are generally brief and which draw attention to the elevated characters being elevated. Chaudhuri confidently chooses Will Kemp as the first performer of the role of Bottom and plausibly casts the notoriously thin John Sincklo as Starveling; the other parts he casts more speculatively.

    Samuel Pepys gave his famous negative reaction to A Midsummer Night's Dream ("the most insipid ridiculous play") in 1662. From the Restoration to the mid-nineteenth century the play was most often performed only as comic extracts, with much added music. Felix Mendelssohn's music for the play in the early nineteenth century continued to affect the music accompanying performances of the play right into the late-twentieth century, for example in Michael Hoffman's 1999 film. Lucia Vestris's 1840 London production was the first since the Restoration to use essentially Shakespeare's text, but she added the diorama that became a Victorian standard for productions. Vestris herself played Oberon, which cross-gender casting also became a standard Victoria practice, alongside female Robins. Harley Granville-Barker in 1914 dispensed with the lavish Victorian spectacle and replaced the Mendelssohn music with English folk songs. The next big landmark, for Chaudhuri, is Peter Brook's celebrated 1970 production, the reactions to which he surveys. The European stage tradition starts with Pyramus and Thisbe acted by touring players from England in 1604. This section of Chaudhuri's Introduction is mainly concerned with the German performance histories, with glances at Holland, France and Spain and other countries.

    The first American production of A Midsummer Night's Dream was in New York in 1826. The narrative here is highly compressed, and many productions get only a sentence or two. There is not much to connect the various productions that Chaudhuri sketches and little sense of a overarching theme to this survey, which ends blandly with the reflection that ". . . Shakespeare's material can suit a seemingly endless range of social and cultural situations" (p. 38). The first cinematic version was a silent film of 1909, and the first talkie was Max Reinhardt's in 1935. Chaudhuri's account of other films is brief. Regarding music, he notes that none for the play survives from Shakespeare's time, and there is only one "indubitable song, the fairy lullaby" (p. 40). Then we get a sketch of Benjamin Britten's 1960 opera at the Aldeburgh Festival before turning to ballet. Next comes pictorial art, the highlight of which is William Boydell's Shakespeare Gallery of the late eighteenth century. In all, Chaudhuri's account of the play in performance has the air of a critic itching to move on to the parts of his Introduction he finds more interesting.

    In the section "Sources and Analogues: Fairies and Mortals (pp. 44-71), Chaudhuri begins to offer his critical reading of the play, observing that it is "Shakespeare's first play involving forces beyond the human" (p. 44). (Actually, there are fiends in 1 Henry 6 who refuse to help Joan of Arc when she is captured, and the devil Asnath is conjured from Hell in 2 Henry 6, both written before A Midsummer Night's Dream.) Chaudhuri traces the origins and afterlife of the play's Robin Goodfellow, and of the trick of putting an ass's head on a man. He moves from this to a wider survey of beliefs about the powers and nature of fairies in Shakespeare's time. Chaudhuri detects hitherto unnoticed evidence of Shakespeare's familiarity with a source, the thirteenth-century French courtly romance Huon of Bordeaux via the English translation by John Bourchier, Lord Berner, of around 1530.

    Henslowe's Diary shows a play about this "hewen of burdoche" acted by Sussex's Men on 28 December 1593 and twice more thereafter. In the romance Huon it is said that the storms and floods conjured by Oberon's anger need not affect the humans, who could "ignore them if they wished", and Chaudhuri notices that in the play there is not "even a light shower" (p. 50). True, no shower is depicted, but then scenes set in the rain are rare in early modern drama, presumably because wetting the expensive costumes was to be avoided. Alan C. Dessen and Leslie Thomson's Dictionary of Stage Directions identifies only three plays calling for rain: Thomas Dekker's If It Be Not Good the Devil Is In It, Thomas Heywood's The Brazen Age, and Thomas Drue's The Duchess of Suffolk. Dessen and Thomson note that the stage directions do not make it clear whether rain is seen on stage or only heard, but even if it is only heard Chaudhuri's point stands that Shakespeare could have provided stage directions for the sound of a storm (as he did for other plays including King Lear and The Tempest) but chose not to. On the other hand, the dialogue of A Midsummer Night's Dream insists that the inclement weather caused by the rift of Oberon and Titania does have detrimental affects on humans: ploughing is impossible, crops and sheep are dying, and games are curtailed (2.1.87-101).

    Turning to "Ovid: The Classics and the Fairies" (pp. 52-60), Chaudhuri reports that the most obvious debt is Shakespeare's use of his story of Pyramus and Thisbe, and the borrowing of names such as Titania. Chaudhuri traces the many Ovidian elements in A Midsummer Night's Dream, and then some lesser sources. He thinks that Apuleius' tale of The Golden Ass (translated into English by William Addington in 1566) "may have contributed to the play" but notes a crucial difference reported by Robert H. F. Carver: ". . . the strident bestiality of Apuleius' account is toned down in Shakespeare, perhaps not least because Bottom, an ass only 'from the neck up', lacks the oversized member attributed to that beast" (p. 57). It is probably worth noting that modern productions give Bottom that oversized member: Brook's production apparently did, and Malcolm Storry's Bottom wore one in the 2005 Royal Shakespeare Company production.

    Chaudhuri explores the various versions of the Pyramus and Thisbe tale that Shakespeare might have known (pp. 60-63) and is sure that his "is presented as a bad play by the dramatist of the serious play framing it" (p. 62) and hence ". . . he cannot invest this travesty of serious theatre with serious intrinsic merit" (p. 62). But in fact Chaudhuri's survey of the performance history seemed to mention one production that tried to do this: Aaron Posner's 2016 production at the Folger Theater, Washington DC in which ". . . Bottom [is] a drama teacher whose girls, unprecedentedly, played Pyramus as serious romantic tragedy" (p. 31). The 2005 Royal Shakespeare Company production allowed the mechanicals' play to become serious too. For Chaudhuri, "the trivialization of intrinsically serious material" (p. 62) inherent in Pyramus and Thisbe is the reason that A Midsummer Night's Dream has become so popular with amateur companies, since they "recognize that the piece allows, indeed calls for an undemanding, appropriately flawed rendering" (p. 62).

    The section on sources ends with consideration of "Theseus and Hippolyta: Plutarch, Chaucer and Shakespeare" (pp. 63-71). Shakespeare's Theseus came from Thomas North's translation of Plutarch's account, which North got from the French version by James Amyot, but also more substantially from Geoffrey Chaucer's The Knight's Tale, which also supplied some smaller details. Chaudhuri traces the differences between Plutarch's Theseus and Chaucer's and how they compare to Shakespeare's. Considering how the same characters feature in The Two Noble Kinsmen, Chaudhuri takes into account which parts of The Two Noble Kinsmen are by Shakespeare and which by Fletcher. He also traces the minor sources of A Midsummer Night's Dream, including Anthony Munday's John a Kent and John a Cumber, which as Chaudhuri reports may in fact be influenced by A Midsummer Night's Dream since its date is uncertain.

    The second major division of Chaudhuri's Introduction is "Themes and Designs" (pp. 71-108). The pastoral is a courtly genre about getting away from the stresses of court life, and A Midsummer Night's Dream plays with this by having no humans normally dwell in the wood and by setting it all at night. Chaudhuri traces the three-day time-scheme of the play and notices that the characters are remarkably little explored by Shakespeare: Demetrius and Lysander are almost interchangeable, the women a little less so. Chaudhuri sketches some critical and filmic responses to this. The degree to which patriarchy is "mitigated by the end of the play depends in part on whether Egeus is absent from the wedding revels as in the quartos, or conducting them as in the Folio text" (p. 83). I think Chaudhuri means that if Egeus is present at the end, patriarchy is more visibly subverted since he will be helping to celebrate the wedding he initially tried to prevent. The mere presence of Hippolyta, the defeated Amazon queen, marks the triumph of patriarchy. But there is, Chaudhuri notes, the potential for a lesbian subversion of patriarchy in the childhood bond of Hermia and Helena and in Titania's devotion to her votaress. The Indian boy is probably West Indian not East Indian, and there is a tradition of postcolonial reading around his presence/absence.

    Chaudhuri reads Robin as part classical clever slave and part medieval Vice, whose upsetting of norms and getting things back-to-front is carnivalesque. Chaudhuri explores the other aspects of carnival in the play, and likens Bottom's empowering literalism with Corin's in As You Like It: it makes him unflappable and unawed by his social superiors. The play's links with Elizabeth I are tenuous, although Titania has a whiff of her. Chaudhuri thinks this is Shakespeare's first play about playmaking--I would say The Taming of the Shrew precedes in it that theme--and the interior play resembles the kind of theatricals that went on in aristocratic houses, of which Shakespeare had some experience. Chaudhuri gives a fairly standard reading of the contexts (Plato, Sidney) for the metadrama's comments upon itself. He sees the play raising a series of challenging, even seditious, questions about patriarchy, gender, and class, but never quite dealing with them. In a final section on "Language and Verse" (pp 108-115), Chaudhuri notes that 36% of the verse lines are rhymed, and he explores the varied uses of rhyme in the play, and the use of short lines for the fairies' speeches. Then he turns to the prose comprising 22% of the lines. So ends Chaudhuri's Introduction, rather inconclusively.

    As usual in these reviews, when considering the editorial choices made by Chaudhuri we notice interesting ones that illuminate the editor's attitude towards the text and throw new light on its problems. In the Dramatis Personae, Chaudhuri eschews the label "mechanicals" as "obsolete and class-biased", deriving from Robin's derision, and instead calls the Athenians who put on Pyramus and Thisbe "The Artisans". At 1.1.19.1, Chaudhuri writes that Helena's impossibly early entrance with the other young lovers in Q1 (160 lines before Lysander says "Here comes Helena") might arise from the use of a "'massed entry' for characters appearing at different points in the scene" but that is hard to reconcile with his view (given later) that Q1 is based on authorial papers, which in Shakespeare's case do not show this tendency. The more likely explanation is Shakespearian first thoughts: when he wrote the entrance direction he thought he would use Helena in the row involving Egeus, Hermia, Lysander and Demetrius but as he wrote the dialogue he found that the row developed better in her absence.

    At 1.1.27, Chaudhuri prints "This man hath bewitched the bosom of my child" without mentioning its metrical awkwardness and the possibility that "This man . . ." (referring to Lysander) was accidentally repeated from two lines previously (where it referred to Demetrius) and that the antithesis of Demetrius and Lysander is better made by "This man hath . . . | . . . | This hath". Where Q1 and F agree on "Emptying our bosomes, of their counsell sweld" (1.1.216), Chaudhuri emends to ". . . counsel sweet" to rhyme with ". . . meet" at the end of the next line, since this is a whole passage of rhyming couplets. Likewise he emends ". . . companions" to ". . . companies" four lines later in order to rhyme with ". . . eyes". Chaudhuri retains the early modern spelling "perfit" for 'perfect', explaining it as "the older and commoner (not necessarily uneducated or vulgar) version, used by the artisans" (1.2.88 and 1.2.101. Chaudhuri sticks to Q1 and F to print "But room, fairy. Here comes Oberon" (2.1.56) where other editors have felt the need to emend to "But make room . . ." or "But room now . . ." on account of the "room" being awkward as an imperative on its own, and also to avoid metrical irregularity. At 2.1.61, Chaudhuri emends Q1 "What, Iealous Oberon? Fairy skippe hence" (substantively the same in F) to ". . . fairies skip hence" on the ground of logic: she is making to leave. But he acknowledges that Titania might be speaking only to her fairy who was earlier talking to, and might still be standing beside, Robin.

    Chaudhuri adopts the Q2/F reading "steep" to read "Come from the farthest steep of India" at 2.1.69 where Q1 has ". . . steppe of India". He points out that modern "steppe" was not available in the 1590s and "step" meaning furthest part is not supported by the Oxford English Dictionary (OED). I would say that the aspect of extremity is covered by "farthest" so "step" means simply "walk one might take" as in "shall we step out?" At 2.1.79 Chaudhuri adopts the emendation of Q1/F's "Eagles" to "Aegles" in "And make him with fair Aegles break his faith", deriving ultimately from Nicholas Rowe who used the form "Aegle" for this nymph seduced by Theseus. Thomas North's translation of Plutarch gave her name the form "Aegles" that Shakespeare seems to have intended. Chaudhuri retains the reading of Q1/F to print "The human mortals want their winter here" at 2.1.101, rejecting the common emendation to ". . . winter cheer" on the grounds that ". . . it seems too early for humans to miss the winter . . .". He suggests it means that the summer is now so bad the humans wish it were winter instead. Glossing "the moon, the governess of floods" (2.1.103), Chaudhuri makes the curious remark that "The moon was held to govern tides", which implies that we no longer hold this view when in fact we do. Chaudhuri adopts Thomas Tyrwhitt's emendation to read "And on old Hiems' thin and icy crown | An odorous chaplet" (2.1.109-10) where Q1 and F have "And on old Hyems chinne . . .", on the grounds that a "chaplet" is a wreath for the head and cannot be worn around the chin.

    The Folio prints "I do not, nor I cannot love you" (2.1.201) where Q1 has the impossible "I do not, not I cannot . . ." and Chaudhuri uses the former, thinking the latter's mistake was "no doubt" caused by the preceding two occurrences of "not". Perhaps, but it might also just be foul case since "r" and "t" look alike in Q1's typeface. In another example of reluctance to modernize, Chaudhuri prints "And to speak troth, I have forgot our way" (2.2.48) where we would expect "And to speak truth . . .". At 2.2.50-51, Chaudhuri retains Q1's "Love takes the meaning in love's conference. | I mean that my heart unto yours it knit" instead of adopting Q2 and F's ". . . is knit". He explains that "it" refers back to "love". This interpretation rather overworks 'love' since it has to i) enable Lysander to reproach Hermia (implying that it is something she lacks if she misconstrues him) and then ii) be the cause of their two hearts being one. This seems more strained than the assumption of an error in Q1. Likewise Chaudhuri sticks with Q1 to print "Transparent Helena, nature shows art" (2.2.108), eschewing emendations such as ". . . nature here shows art" and ". . . nature shows her art". He writes that "The line is a syllable short . . ." but this is true only if "Helena" is pronounced disyllabically. Equally likely, "Helena" is trisyllabic, an unstressed syllable is omitted at the caesura, and "art" forms an extra (that is, 11th) stressed syllable at the end of the line; this is within Shakespeare's metrical range.

   Chaudhuri gives to Snout the line "Doth the moon shine that night we play our play?" (3.1.46) where Q1 and F have the ambiguous speech prefix "Sn.". As C. J. Sisson pointed out, doing so gives Snug nothing to say in this scene and even if we suppose that Shakespeare made that mistake it is unlikely that the actor playing Snug failed to object and take the line for himself. At 3.1.77-78, Chaudhuri prints "BOTTOM Thisbe, the flowers of odious savours sweet. | QUINCE Odours, odours" where some editors change "flowers of" to "flowers have" on the assumption that "savours" is a noun and because Bottom-as-Pyramus has to go on to say "So hath thy breath . . .". The problem with the emendation of "of" to "have" is that it seems to require that "odorous" (rather than "odours") is what Bottom-as-Pyramus was meant to say, and yet after being corrected by Quince he still says "odours savours sweet" and Quince does not object. Thus "odours" must be right and hence "savours" must be a single verb with the plural subject "flowers", which is a common disjunction of the period. For Quince's prompt "Odours, odours", Q1 has the misprint "Odours, odorous" and Chaudhuri rightly gets his version of Quince's prompt from the Folio.

    At 3.2.19 Chaudhuri retains Q1's reading of "Minnick" (short for "minikin", meaning a person of "mincing or affected behaviour") so that Robin describes Bottom's entrance with the ass's head with "And forth my minic comes". Most editors switch to F's reading "Mimmick" (meaning actor, impersonator) to read "And forth my mimic comes", but Chaudhuri objects that "mimic" this was a "very new word at the time", which is true: EEBO-TCP's earliest uses are as an adjective not a noun by Edmund Spenser and John Marston in the 1590s. On the other hand, it is not obviously correct to say that Bottom is mincing or affected. Describing the panicked flight of the artisans upon seeing Bottom with the ass's head, Robin says "And at our stamp, here o'er and o'er one falls" (3.2.25), meaning that he stamped the ground and this heightened their alarm. Editors since Samuel Johnson have emended on the grounds that fairies cannot stamp the ground hard enough to scare anyone, but Chaudhuri cites examples of mythical fairies doing so and points out that Oberon says that his dancing with Titania will "rock the ground" (4.1.85).

    Chaudhuri adopts Alexander Pope's insertion of "so" at the end of "And from thy hated presence part I so" (3.2.80), which otherwise not only lacks a final stressed syllable but also fails to rhyme (in the midst of rhyming lines) with ". . . dead or no" in the next line. After praising the whiteness of Helena's hand, Chaudhuri has Demetrius say "O let me kiss | This impress of pure white, this seal of bliss!" (3.2.144), thereby following Howard Staunton's emendation of Q1 and F's "Princesse" to "impress". It is true that "impress"  goes with "seal", but it is not clear what "impress" means in this sentence: Demetrius's kissing would be an impress but the part of Helena he kisses would not. Q1 and F's "Princesse" seems acceptable here. At 3.2.213, Chaudhuri adopts Lewis Theobald's emendation to have Helena say that she and Hermia were "Two of the first, like coats in heraldry" where Q1 and F agree on "Two of the first life . . .", which is a simile from heraldry in which two coats of arms, from two joined families, are equal.

    Chaudhuri follows F to have Hermia say "I am amazed at your passionate words" (3.2.220) where Q1 lacks "passionate". He thinks it only a "possibly authorial" correction in F, which seems slight reason to adopt it. At 3.2.236-237, Chaudhuri prints "HERMIA I understand not what you mean by this. |  HELENA I do. Persevere . . ." where editors usually understand Helena's "I" as "Ay", so she is sarcastically encouraging Hermia: "Yes, do carry on the jest . . .". It seems odd to suppose, as Chaudhuri does, that Helena says that she understands her own meaning. (Perhaps the idea is that Helena does understand Hermia's meaning, so that in reflecting Hermia's "what you [Helena] mean" Helena implies "I do understand what you [Hermia] mean".) Chaudhuri adopts Theobald's emendation to read "Thy threats have no more strength than her weak prayers" (3.2.250) where Q1 and F have ". . . praise".

    In a notorious crux at 3.2.257-259, Q1 has "Dem. No, no: heele | Seeme to breake loose: take on as you would follow; | But yet come not. You are a tame man, go" and F has "Dem. No, no, Sir, seeme to breake loose; | Take on as you would follow, | But yet come not: you are a tame man, go". (Rodney Edgecombe Stenning tackles this crux in a short article reviewed below.) It is unclear whom Demetrius is addressing when he calls Lysander "he" (if indeed "heele" is modern "he'll") nor whom is meant by "you". There have been numerous emendations of words to make sense of this, but Chaudhuri's solution is to adopt Q1 and merely indicate changes of address: "DEMETRIUS [to Hermia] No, no: he'll seem | To break loose, take on as you would follow, | But yet come not. [to Lysander] You are a tame man, go". The explanation is that Hermia is laying hold of Lysander and Demetrius tells her that Lysander is only pretending to try to break free (since he is in fact too scared to fight), and that Demetrius goes on to tell Hermia either i) to pretend that she will follow Lysander if he leaves to fight Demetrius, but not to actually do so, or ii) that Lysander will pretend that he thinks she will follow if he leaves to fight Demetrius and that Lysander will not actually follow Demetrius off to fight. This is ingenious but it makes for a sentence rather heavily condensed in meaning during a piece of physical comedy, as Hermia hangs about Lysander's neck, and it sets a challenge for the actor playing Demetrius to plausibly address Hermia while she is doing that.

    Chaudhuri leaves unemended "The squirrel's hoard, and fetch thee new nuts" (4.1.35) without commenting on its metrical deficiency. Previous editors have emended to ". . . fetch thee thence new nuts" and ". . . fetch thee off new nuts" and ". . . fetch for thee new nuts". At 4.1.39-43, Chaudhuri has Titania say to Bottom that she will "wind him" in her arms and make the horticultural simile: "So doth the woodbine the sweet honeysuckle | Gently entwist; the female ivy so | Enrings the barky fingers of the elm". In a note Chaudhuri deals with the objection that woodbine and honeysuckle are the same plant, so one cannot entwist the other: "The woodbine may here be the bindweed or convolvulus. . .". He also considers the possibility that "bindweed" was what was meant to be set but got corrupted to "woodbine". Another possibility, raised by Edward Capell, is that "woodbine" and "honeysuckle" are in apposition as a pair of plants that entwist some other plant.

    Oberon calls for Titania to summon music to make the four lovers and Bottom sleep: ". . . strike more dead | Than common sleep of all these five the sense" (4.1.80-81). Q1 and F have ". . . fine the sense" and Chaudhuri explains this as a minim error. It could just as easily be a foul case error--an "n" in the "u" sort box when Q1 was printed--which error got replicated in Q2 and thence F. In a rare typo, a note refers to "horns at 127 and 127 SD" (4.1.101.1) but should read "horns at 137 and 137 SD". In printing "And the duke had not given him sixpence a day for playing Pyramus, I'll be hanged" (4.2.21-22), Chaudhuri fails to modernize "and" to "an" in a case where it clearly means "if".

    Being based on Q1, Chaudhuri's edition has Philostrate manage the wedding revels, rather than Egeus as in F. Amidst a flurry of oxymorons, Chaudhuri follows Alexander Dyce at 5.1.59 to print "That is hot ice and wondrous swarthy snow" where Q1 and F agree on "wondrous strange snow". There is clearly a syllable missing in Q1/F and the context demands an oxymoron. Chaudhuri favours Dyce's solution because it is paleographically plausible: "swarthy" could be misread as "strange". At 5.1.177, Chaudhuri comments on Q1 and F's use of the phrase "Jove shield thee" that "Jove for 'God' satisfies the law against stage blasphemy" but Q1 was printed in 1600, before the law against stage blasphemy--the Act to Restrain Abuses of Players--was enacted in 1606.

    Chaudhuri uses his own emendation to print "Now is the more use between the two neighbours" at 5.1.205 where Q1 has "Now is the Moon vsed between the two neighbors" and F has "Now is the morall downe betweene the two Neighbors". Chaudhuri explains how misreading of "mor use" in the manuscript underlying Q1 led to its compositor setting "Moon vsed" and suggests that F is just an attempt to make sense the garbled Q reading. He has little to say on what his emendation makes the line mean: "Use might mean familiarity, customary exchanges . . . or even sexual exchanges or promiscuity . . .". Presumably he thinks that "Now is" means "Now there will be".

    Chaudhuri emends the entrance of Lion and Moonshine agreed upon by Q1 and F to just "Enter Lion" (5.1.216), so he has to explain Theseus's comment "Here come two noble beasts, in a man and a lion" as referring only to the lion. In support of this he cites a late-seventeenth-century Smock Alley promptbook emendation to "a Man and a Lion in one". Chaudhuri prints "Then know that I as Snug the joiner, am | No lion fell, nor else no lions dam" (5.1.211-2) where Q1 has "Then know that I, as Snug the Ioyner am | A Lyon fell, nor else no Lyons damme" and F has "Then know that I, one Snug the Ioyner am | A Lion fell, nor else no Lions dam". Rowe is the source of the emendation of "A lion" to "No lion" in the second line and Chaudhuri records it as "simple and plausible" while noting that the Q1/F reading "is possible". This seems to set the bar for emendation low, since Q1 has its own strong and particular sense: as Snug I am a lion, not otherwise. Rowe's emendation seems to weaken this to a simpler denial of his role and perhaps arose from an anachronistic assumption that "nor" needs an antecedent "no".

    At 5.1.267 Chaudhuri adopts Samuel Weller Singer's emendation to read "For by thy gracious, golden, glittering gleams" where Q1 and F agree on ". . . glittering beames". The reason is that the last word has to rhyme with "beams" two lines earlier and "It is unlikely that beams would be made to rhyme with itself . . .", so that presumably the compositor accidentally repeated "beams" from two lines earlier. But since this is meant to be bad poetry, it is possible that the repetition of "beams" is an intentional anticlimax. Chaudhuri prints "Now the hungry lion roars | And the wolf behowls the moon" at 5.1.361-2 where Q1 and F agree on plural "lions" and have "beholds" where Chaudhuri adopts Theobald's emendation of "behowls". The lion should be singular because the wolf is, Chaudhuri argues, and although Q/F1's "beholds" makes sense it seems "somewhat indirect"; this again sets low the bar for emendation. At 5.1.406-410, Q1 has Robin say "Euery Fairy take his gate | And each seuerall chamber blesse, | Through this palace, with sweete peace, | Euer shall in safety rest,  | And the owner of it blest" and F substantively agrees. Chaudhuri finds that no emendation of words can fix this so he follows Howard Staunton in reversing the order of the last two lines, which "solves the problem".

    In the first of three appendices Chaudhuri offers a casting chart (pp. 279-81). He finds that 12 men and 8 boys can perform the play, which seems like a lot of boys but as he notes earlier "The few words uttered by Cobweb, Peaseblossom, Mote and Mustardseed would not be beyond the powers of . . . casual child players" (p. 9), meaning non-professionals recruited just for this play. True, but we have no evidence that this ever happened in professional plays and it is not clear how such casual actors would fit into the economy of the theatre: who would pay them and who would be in charge of them?

    Appendix Two deals with the play's "Date and Occasion" (pp. 283-94). The play is mentioned in Francis Meres's Palladis Tamia in 1598 so it was in existence by then. There is "no firm evidence" (p. 283) that the play was written for a marriage, but the closing minutes are so formal and elaborate that it is easy to see why this has been suspected. The action is over by the end of Act 4 and there might even been two endings rolled into the texts we have. Chaudhuri tries to figure out which of the weddings around this time it might have been performed at, with the marriage of William Stanley, 6th Earl of Derby, to Elizabeth de Vere, daughter of the Earl of Oxford, on 26 January 1595 the best bet. Everyone accepts that the "fair vestal" remark by Oberon (2.1.158) refers to Elizabeth I, and Chaudhuri considers specific entertainments that it might allude to.

    Regarding other evidence for a date, Chaudhuri considers: the play's bad weather in relation to British weather of the mid-1590s (deciding that there was so much of it that it is little help in dating the play); the problem of bringing a lion in among ladies (which must postdate the same question being raised in connection with a baptismal feast for Prince Henry on 30 August 1595); the Indian Boy (which probably puts the play after Walter Raleigh's return from Guiana with one in August 1595); the "thrice three Muses mourning for the death | Of learning, late deceased in beggary" (5.1.52-53), which some have tried to link to Robert Greene's death or Edmund Spenser's "The Tears of the Muses" but Chaudhuri finds no connection; and finally the date of publication, in the light of Lukas Erne's observation that most of Shakespeare's 1590s plays were published about two years after first performance, but as Chaudhuri notes Erne saw A Midsummer Night's Dream as an exception to this rule. In relation to Shakespeare's 1590s style, Chaudhuri puts the play around 1595-96, which fits the possible topical allusions.

    Chaudhuri thinks it certain that The Merry Wives of Windsor has two endings, one reflected in Q1 and one in F, and that the latter "celebrates the investiture of Lord Hunsdon with the Garter" (p. 292). In fact there is no certainty about this, and Chaudhuri cites no evidence or authorities for his claim. Just why this supports the claim that A Midsummer Night's Dream has two endings is not explained. Chaudhuri pursues the idea of a double ending to suggest that one was for public performance and the other for a subsequent performance at a private wedding. He wonders whether the version adapted for a private wedding was then re-adapted for performance at the Blackfriars, which the company acquired around the time of the play's composition. Or maybe the play was actually written for the Blackfriars and then adapted for the open-air stage. There is no evidence for any of this speculation, although Chaudhuri thinks it would help explain the messy state of papers printed to make Q1 and explain the four-to-six years delay between the play being composed and its being printed.

    Appendix Three is on "The Text" (pp. 295-327). Chaudhuri mentions the features that have made scholars believe that Q1 was printed from foul papers and then acknowledges Paul Werstine's demonstration that for the most part these supposed signs have been overstated: theatrical documents have them too. He turns to distinctly Shakespearian spellings and makes his own count of 127, relying on such things as "ea for e (26), ow for ou (19), oo for o (17), ore for o'er (11 ), to for too ( 11 ), o for oa (10), ai for ei (7), ew for ue (6), plus contractions like heele and weele" (p. 298n4). Also in favour of authorial papers copy for Q1 is the mislining in 5.1 that John Dover Wilson pointed out seems caused by marginally recorded additions to the original composition (since removing the mislined lines leaves the speeches still making complete sense). Chaudhuri considers that Werstine's achievement in Early Modern Playhouse Manuscripts and the Editing of Shakespeare (reviewed in YWES for 2013) was to show that authorial features made it into theatrical documents, not to show that these are not authorial features.

    Chaudhuri surveys recent views of the 1619 Pavier Q2 including whether it was surreptitious, accepting Sonia Massai's in Shakespeare and the Rise of the Editor (reviewed in YWES for 2007) that it was not: "the attempt at deception" in the falsely dated title page "was perfunctory" (p. 302). Despite this rehabilitation, Q2 has no fresh authority, whereas F does in that the exemplar of Q2 it reprints was first annotated by reference to an authoritative manuscript. Chaudhuri examines the features of F that lead to this conclusion before summing up his textual practice: "There seems no reason to depart from Q1 where it provides a viable text, unless F (or rarely Q2) provides a manifestly preferable variant (or, of course, corrects an obvious error)." (pp. 305-6). But as we saw in his actual choices above, Chaudhuri does not always apply this rule: he emends Q1 when it is acceptable and neither Q2 nor F is "manifestly preferrable". Some differences between Q1 and F can be explained by revision, "But there is nothing to suggest repeated stages of general revision" (p. 308) as John Dover Wilson claimed there was.

    Chaudhuri surveys what is known about the printing of Q1, Q2, and F (pp. 309-16), relying rather uncritically on Robert K. Turner's essay in Studies in Bibliography in 1962 on the printing of Q1 that employed methods, such as the analysis of running-title reuse and the apparent recurrence of broken type, that have subsequently been shown to be less reliable than the high New Bibliographers thought. Chaudhuri proceeds to discussions of the printing of Q2 and F, relying on Charlton Hinman's work for the latter. Thus Chaudhuri accepts the division of Folio compositor stints that Pervez Rizvi (in an article reviewed in NYWES for 2016) showed to be unreliable; presumably Chaudhuri's book was completed before Rizvi's article appeared.

    Describing his "Editorial Procedure" Chaudhuri begins by sketching some of his emendation choices, but as he remarks the full justifications are in the commentary notes at the appropriate places in the text so this is unnecessary. He surveys the change in practice regarding intervals at the open-air theatres after the King's men began to use the Blackfriars theatre and attributes to this Q1's not being divided, and F's being divided into acts. One limitation throughout his book is that he refers to the "public" and "private" theatres, where the more appropriate distinction is between indoor hall theatres and open-air amphitheatres (such the public could go to either); this particularly affects his account of act intervals. The action in the forest is particularly hard to divide by the criterion of clearings-of-the-stage, since the fairies remain out of sight somewhere--Oberon and Robin watching the humans, Titania sleeping in her bower--across what otherwise would seem like discrete episodes. Chaudhuri's discussions of where to break scenes and acts, how to reline mislined verse and prose, how to regularized speech prefixes and repair stage directions, and what to do with early modern spelling and punctuation are all conventional and unexceptional.

    So to our second major edition this year: Proudfoot and Bennett's Edward 3. They begin their Introduction (pp. 1-132) with the key facts that the play was published anonymously in 1596 and that Capell, in 1760, was the first to seriously entertain the possibility that it was by Shakespeare. In 1876 F. G. Fleay argued for Shakespeare's hand in the play, writing the scenes concerning the King and the Countess of Salisbury, and E. K. Chambers in 1930 took that idea seriously. If it was partly by Shakespeare, why did the First Folio omit it despite admitting other plays we now know to be collaborative? That the story is not connected to the narratives of the other history plays could be one reason, as could the fact that the rights to Edward 3 were held by Thomas Snodham who was not in the Folio consortium of Isaac Jaggard, Edward Blount, John Smethwick, and William Aspley. And its unflattering depiction of the Scots would not endear the play to England's Scottish king.

    Cuthbert Burby entered the play in the Stationers' Register on 1 December 1595 and when published on 1596 the title-page did not even mention the playing company or the printer, who we now know was Thomas Scarlet who died in July 1596. Judging on internal evidence of quality and resources called for, the play was written by a professional dramatist for a professional company. The depicted events take place in 1337-56, at the start of the Hundred Years War, with subplots about Edward's love for the Countess of Salisbury and the Earl of Salisbury's journey from Brittany to Calais via Poitiers.

    Unlike similar plays of the period, Edward 3 has no thrilling antihero (as in Christopher Marlowe's Tamburlaine) or wider politico-religious subtext (as in Shakespeare's Henry 5) and the audience finds itself watching a war that takes place only because simple oaths and treaties keep getting broken. This connects the love-plot to the military plot via the sacredness of oaths and honour. The play engages a pleasing nostalgia for post-Armada London audiences of the 1590s. After the death of the heirless French king Charles IV, the historical Edward 3 of England had the strongest claim to the French throne via his mother (Isabella, Edward II's widow) who was Charles's sister, but the French invoked the Salic Law to prevent Edward 3's inheritance of the throne of France.

    In the early 1590s, the French succession was a hot topic for London audiences as the Protestant Henry King of Navarre fought his Catholic enemies headed by Henry of Lorraine, the Duke of Guise (who ordered the massacre of Huguenots in Paris in 1572). Henry King of Navarre converted to Catholicism in 1593 and won the civil war to become Henry IV of France. These contemporary events had strong echoes of the Hundred Years War that the play dramatizes, and in any case succession issues generally were topical in 1590s. The play's disparagement of the Scots is probably part of the explanation for its relative neglect after 1603.

    Proudfoot and Bennett notice that the words "love", "oath", and "swear" cluster in Scenes 2 and 3, the love-plot scenes. The struggle is a battle of wits and logic, for example in the Countess of Salisbury's insistence that she cannot simply give her body to the king because it is "souldered to her soul", and in citing scripture (married men have been around since Adam, longer than kings). The wedding-knives that the Countess draws from her belt and threatens to use on herself are "the only weapons drawn onstage with the immediate purpose of inflicting bodily harm" (p. 31) in the whole play. The final scene neatly ties up every plot strand, but the effect is "perfunctory" (p. 31) rather than pleasing. The closing line about "three Kings, two Princes and a Queen" encourages a stage picture in which we see these six, and perhap suggests the importance of Edward's Queen Philippa who may be "armed or visibly pregnant or both" (p. 32). The final moments of the play also stress the importance of youthful, martial princes as the hope of England.

    The play's main source is the English translation by John Bouchier (Lord Berners) of Jean Froissart's French chronicles. Froissart lived shortly after the events portrayed in the play and his account of the history was based on chronicles not otherwise known about until the nineteenth century. Raphael Holinshed's Chronicles are indebted to Froissart but added their own details too, and it is hard to figure out just where the play is drawing directly on Froissart and where it is drawing on Holinshed. The Edward 3 story was also available in the Latin hexameter verse of Christopher Ocland's Anglorum Praelia of 1580 (which is quoted in the 1580 edition of Holinshed's Chronicles) and John Sharrock's English translation of Ocland in 1585; both were used in schools.

    The story also appeared in William Painter's The Palace of Pleasure (published 1566 and 1575) in a version based on Matteo Bandello's Novelle (published 1554). The history play most like Edward 3 is The Famous Victories of Henry 5, of which Proudfoot and Bennett consider the 1598 quarto to reflect only "a reduced and garbled state" of the original Queen's Men's play of 1587-88. All plays of this type were of course influenced by Marlowe's Tamburlaine. Proudfoot and Bennett think that Edward 3 "might have been planned as the counterpart of 1 Henry VI in a diptych of English defeat and victory" (p. 47), and they trace the symmetries of the Henry 6 trilogy and Edward 3.

    Over forty pages of Proudfoot and Bennet's Introduction (pp. 49-89) are taken up with the play's authorship. Four strong candidates died between 1592 and 1596: Greene (in 1592), Marlowe (in 1593), Thomas Kyd (in 1594), and George Peele (in 1596). Two more, Thomas Lodge and Thomas Nashe, stopped writing plays around this time. Most of the candidates have not left us large canons for comparison, and the only writers of the 1590s to later get canon-forming collected works editions were Shakespeare, with the 1623 Folio, and John Lyly. Of the latter, Proudfoot and Bennett give the date of his canon-establishing early edition as 1636 (p. 51) but presumably they mean 1632, when his Six Court Comedies collection was published, since his only book in 1636 was yet another edition of the much-reprinted Euphues: The Anatomy of Wit. Complicating it all further, of course, is the fact that in the 1590s plays were often printed without their writers' names on the title pages. The "verbal links" (p. 51) between Edward 3 and the early Shakespeare canon, of which Proudfoot and Bennett do not at this stage give examples, show that Shakespeare was at least intimately familiar with this play.

    Certainly, Scenes 2, 3, and 12 have markedly more feminine endings than much of the rest of the play and Proudfoot and Bennett tabulate this, cautiously distinguishing those lines with indisputably monosyllabic last words because such counts can be distorted by the fact that certain words--"flower", "heaven", "spirit" and so on--can be monosyllabic or disyllabic. For their table, Proudfoot and Bennett divide Scene 2 into three parts: 2.1-166, 2.167-349, and 2.350-625. What emerges are three distinct areas in the table: one of low incidence of feminine endings, one of medium incidence (including lines 2.1-166 and 2.167-349), and one of high incidence (including lines 2.350-62 and Scene 3 and the marginal case of Scene 12).

    Proudfoot and Bennett try to isolate the evidence of the monosyllabic feminine endings but as there are only 19 of them across the whole play this seems to me rather too little data to go on. Moreover, I am not able to follow in each case their designation of these monosyllabic feminine endings. For example, "Poor sheepskin, how it brawls with him that beateth it!" (3.48) seems to me a regular hexameter ("poor SHEEPskin HOW it BRAWLS with HIM that BEATeth IT"), and "Return? I hope thou wilt not" (11.20) could as well be called an incomplete iambic pentameter as a trimeter with a feminine ending, and "If, then, we hunt for death, why do we fear it? (12.141) seems to me to allow "fear it" to be pronounced monosyllabically to make a regular iambic pentameter.

    Turning to the evidence of n-grams shared across an author's canon, Proudfoot and Bennett report that "It is held that three-word phrases, 'trigrams', offer the best kind of evidence, as occurring with sufficient frequency and yet overall in manageable numbers for collection and analysis" (p. 58). In fact, it is not agreed by investigators that any particular value of n is best for authorship attribution and the "manageable numbers" consideration is only really applicable to investigations relying on the manual collection or processing of data: fully computerized studies tend to work better as the amount of data increases.

    In their survey of authorship attribution scholarship on the play, Proudfoot and Bennett mention the work of Brian Vickers (who favours Kyd's authorship of the parts of Edward 3 that are not by Shakespeare) and the as-yet unrealized ambitions of Martin Mueller, but not the work of Hugh Craig, Arthur Kinney, and Timothy Irish Watt. That is, they consider the searching for n-grams (by Vickers, Mueller, and MacDonald P. Jackson) but not the counting of the most common function words and not the counting of words chosen by computer to be the ones that most distinguish one set of writing from another (that is, John Burrows's Zeta test). This is an incomplete survey omitting some of the strongest evidence about the authorship of Edward 3.

    Proudfoot and Bennett then turn to their own "Internal n-gram examination of Edward III"--that is, testing one part of it against another--to see if it is all of a piece or internally variegated. This investigation, they tell us, is too big to be described in full so their edition will give only the highlights of their findings. Indeed, they do not state their method for finding n-grams, which, for all the reader knows, might be simply the ones that the editors noticed as they worked. With no support other than what seems an appeal to common sense, Proudfoot and Bennett assert that passages sharing one or more longish n-grams were probably written by the same person, but that is more than we know and nowhere do they attempt the necessary negative check to see if the phrase in question is just common in the period. Most importantly of all, the editors do not identify the n-grams they claim to find in the play--they just give the totals found--so that no one else can test to see if these are just common-place phrases of the period. All in all, this study of authorship does not even amount to evidence let alone argument and it may safely be ignored by anyone interested in the topic.

    Proudfoot and Bennett are on firmer ground when they explore the lack of links between the character Lodwick (derived from the Secretary in Painter's version of the story) and the rest of the play, which they diagnose as evidence of revision that added him, or at least that his part was written last. To a lesser extent, Warwick is detachable too and hence "may be an afterthought" (p. 62). In general, the action involving the Countess is not referred to later in the play, suggesting that it may have been added to the play or greatly expanded from an initial hint. In Scene 12, just before the Battle of Poitiers, Lord Audley's advice to Prince Edward not to fear death (12.134-149) is much like the Duke's "Be absolute for death . . ." speech to Claudio in Measure for Measure. This scene also has shades of the pre-battle scenes in Henry 5, including the multiple visits from the enemy's herald proposing a dishonourable way out for the English hero.

    Also in Scene 12, Proudfoot and Bennett diagnose misplacing of lines. The 1596 quarto reads:

Pr: Deathes name is much more mightie then his deeds,
Thy parcelling this power hath made it more,
As many sands as these my hands can hold,
are but my handful of so many sands,
Then all the world, and call it but a power:
Easely tane vp and quickly throwne away,
But if I stand to count them sand by sand     [end of sig. H2r]
_________________________________
The number would confound my memorie    [start of sig. H2v]
And make a thousand millions of a taske,
Which briefelie is no more indeed then one,
These quarters, spuadrons, and these regements,
Before, behinde vs, and on either hand,
Are but a power, when we name a man,
His hand, his foote, his head bath seuerall strengthes,
And being al but one selfe instant strength,
Why all this many, Audely is but one,
And we can call it all but one mans strength:
He that hath farre to goe, tels it by miles,
If he should tell the steps, it kills his hart:
The drops are infinite that make a flood,
And yet thou knowest we call it but a Raine:
There is but one Fraunce, one king of Fraunce,

Proudfoot and Bennett consider the bottom of sig. H2r to be confused here, and think that "Thy parcelling this power hath made it more" and "Then all the world, and call it but a power" are misplaced and belong right before "These quarters, spuadrons, and these regements". The compositor of H2r would have made this change if he could, but H2v was already printed on the other side of the sheet so he could not. What caused the confusion was that some of this speech was written in as interlined or marginal additions to the original writing in the printer's copy. It is not that "Thy parcelling this . . . made it more" and "Then all the . . . but a power" were necessarily the additional writing: they may be the original writing and the bits around them--"As many sands . . . so many sands" and "Easely tane vp . . . indeed then one"--may be the additions. It favours this interpretation that "He that hath . . . but a Raine" is also easily detached without disrupting the sense, so perhaps here we have three counting similes added to the scene: counting sand, counting steps, and counting raindrops.

    The counting steps simile is reminiscent of Love's Labour's Lost at 5.2.183-98, and Proudfoot and Bennett propose that these three counting similes are Shakespeare's subsequent addition, whether to the scene as originally written by him or as originally written by someone else. Proudfoot and Bennett suspect that this addition was made to build up the part of Prince Edward and they identify a series of other moments in the play that establish him as a strong independent character. In their edition they put these lines back into their conjectured proper order. The case of the above additions to Scene 12 is, they reckon, like the cases of revision (detected by mislined verse) in A Midsummer Night's Dream 5.1.5-83 and of course there is in Love's Labour's Lost 4.3 and Romeo and Juliet 2.2 evidence of revised passages printed alongside the lines they were meant to replace.

    Once we suppose that Shakespeare was at least a reviser of an existing Edward 3 play, the echoes of Shakespeare's other works become easier to explain. "Lilies that fester smell far worse than weeds" in the play (2.617) is exactly echoed in, or echoes, Sonnet 94.14, and Proudfoot and Bennett reckon that "the phrase 'far worse than'", which also occurs in the play at 10.16, "may be distinctive of Shakespeare", based on their searching of LION. But searching in Early English Books Online Text Creation Partnership (as of December 2017) I found that this phrase occurs dozens of times in books published before Edward 3, so it is not distinctive of Shakespeare. That Shakespeare knew Edward 3 is shown by his later allusions to it, most certainly in Henry 5's "The King of Scots, whom she did send to France | To fill King Edward's fame with prisoner kings" (1.2.161-2), for which the only known source is Edward 3 18.63-82.

    The words "sovereign" and "liege" abound in Richard 2 (17 occurrences and 20 occurrences respectively) and Proudfoot and Bennett think that they might be Shakespeare marker words so they count them in Edward 3. They find that "sovereign" occurs 20 times in Edward 3, 16 of them in Scenes 2 and 3, and "liege" occurs 29 times in Edward 3, 25 of them in Scenes 2 and 3, and hence these two scenes are particularly Shakespearian and have the preponderance (over 80%) of the play's use of these two words. This certainly suggests that Scene 2 and 3 are unlike the rest of Edward 3, but of course an investigator would have to count these words across all the surviving history plays to see if they are truly author-specific. A quick sample of just a few plays shows that the counts in Richard 2 and Edward 3 are not exceptional: Henry Chettle uses "sovereign" 17 times and "liege" 19 times in The Death of Robert, Earl of Huntington, for example.

    Proudfoot and Bennett trace some loose parallels between the language and themes of Shakespeare's plays and parts of Edward 3. Regarding the authorship, Proudfoot and Bennett conclude that the play as we have it was revised in 1593-4 to extend, using Painter's version, the Countess of Salisbury incident--presumably taking away a now-lost, shorter version of that incident--and that the reviser was Shakespeare. He may, of course, have added tweaks elsewhere in the play at the same time.

    Who else wrote the play? Proudfoot and Bennett go through the well-known dramatists of the late 1580s and early 1590s to see who is plausible. Considering the case for Marlowe, Proudfoot and Bennett turn to Mueller's work on the n-grams common to all the surviving plays of the period, pointing the reader back to their earlier footnote that referred to Mueller's website. As is commonly the case with references to Mueller's project called "Shakespeare His Contemporaries" the weblinks given by Proudfoot and Bennett no longer lead anywhere so this reviewer cannot find the evidence that apparently supports the claim that "In Mueller's list of n-grams across over 500 plays from the mid-sixteenth century to 1663 which show the most links with Edward III, Marlowe figures strongly" (p. 82).

    Regarding Kyd's possible authorship of the play, Proudfoot and Bennett report Vickers's "impressive list of 'rare' n-grams" (p. 84) common to Kyd's work and Edward 3, but withhold judgement until at least Mueller's analysis is completed. At this point in their Introduction it is particularly unfortunate that Proudfoot and Bennett do not engage with the work of Timothy Irish Watt, in Hugh Craig and Arthur Kinney's book Shakespeare, Computers, and the Mystery of Authorship (reviewed in YWES for 2009), showing the unlikelihood of Marlowe or Kyd or Peele writing Edward 3. It is to Peele that Proudfoot and Bennett turn next, but only briefly by noting a few investigations and no serious engagement with them.

    Proudfoot and Bennett mention Nashe's "eulogy of 1 Henry VI in his Piers Penniless", noting that it has long been thought an allusion to Shakespeare, but of course those who consider Nashe a co-author of 1 Henry 6 see this as Nashe praising himself. Proudfoot and Bennett explore the circumstantial evidence for Nashe's hand in Edward 3--principally that he knew his Froissart--and consider the possibility that he plotted the play but contributed no dialogue. Concluding their survey of the authorship, Proudfoot and Bennett are reasonably sure only that Shakespeare contributed the Countess of Salisbury scenes as part of his revision for a revival of an earlier play, and that the date range is 1590-94 with the later years in that range more likely than the earlier.

    Considering "Edward III in Performance" (pp. 90-114), Proudfoot and Bennett note that the staging needs of the play are especially slight: just two doors and perhaps an acting space aloft are essential, plus some ordinary furniture (a chair for a throne, some benches) and the usual handheld properties. The sound directions are somewhat more ambitious, including the sounds of battle but also "A clamour of ravens" (13.18). Proudfoot and Bennett's calculation of the minimum cast size is 13 men and 2 boys. The title page of the 1596 edition is unusually silent about playing company and venue, and the play is nowhere mentioned in Henslowe's Diary so the presumption is that it belonged to the rival Chamberlain's Men at the Theatre in Shoreditch.

    Proudfoot and Bennett survey some possible allusions to the performance of a play about Edward 3, but none is definitely to this play. The earliest known performance is an adaptation by William Poel in 1890. The performance history is thus almost entirely blank until a flurry of productions started in the 1980s, and Proudfoot and Bennett pay most attention to the Royal Shakespeare Company production of 2002 and the National Theatre of Bucharest production of 2008-9.

    Proudfoot and Bennett devote a sizeable part of their Introduction to "The Printing of Q1, 1596" (pp. 114-32). Cuthbert Burby is named on the title page of the 1596 first edition as the publisher but the unnamed printer Thomas Scarlet has been detected by the use of his printer's ornament and his habit of omitting the leaf letter on the second and third sheet of each gathering. Proudfoot and Bennett detail such things as the watermarks and the unusually small typeface of the edition and the increase from 36 to 38 lines per page in the last two gatherings. They also describe the simultaneous typesetting by formes, two compositors taking a half-forme each, but do not reveal, at this point (pp. 118-121), how they arrived at this determination: the compositor identification evidence and argument is deferred to page 124-127. Likewise they report that the outer formes were printed before inner formes for some sheets but not how the running title evidence told them that.

    Proudfoot and Bennett observe some anomalies that are explained by their proposed order of printing. One is the very large typeface used for the stage directions on E1v and E2r, which they put down to space-wasting necessitated by imperfect casting off of the manuscript copy. Another is the apparent jumbling of lines at the bottom of H2r (discussed above) that could not be fixed because the next page, H2v, was printed first.

    Another anomaly is the use of "Lor[raine]" speech prefixes for Lodwick. Setting the pages out of reading order, one compositor wrongly expanded the abbreviation "Lo." as the marker for Lorraine because he had not yet encountered the character Lodwick, while the other avoided the error because he alone had already encountered Lodwick's name in dialogue; the appearance and disappearance of this error matches the differing knowledge of the two compositors. On page 124 Proudfoot and Bennett list the stop-press corrections found across the seven known exemplars of the first edition, which make substantive differences to body-text readings only on one page, H1r.

    Next comes the compositor identification evidence, confined to the final three gatherings, H to K. In earlier gatherings, Proudfoot and Bennett notice that the ends of some long lines are not printed in some exemplars, citing the shortening of "spurre" "spu" on B1v and "Retreate" to "Re" on E3v. It is not clear what Proudfoot and Bennett make of this as they call it "loosening, even to the extent of loss, of the type at the end of long lines" and "loss of type at press" (p. 124). Exactly how looseness could lead to loss of type they do not explain.

    The problem seems to affect the ends of long lines, and Proudfoot and Bennett notice that the final gatherings' adoption of two extra lines to the page (38 instead of 36) is accompanied by the use of a narrower measure, which they think may be connected to this earlier problem. The narrowing of the measure was achieved by indenting the stick--that is, stacking multi-line spacing material against the right edge of the composing stick--resulting in two distinct styles of expedient for cramming long lines into the reduced measure, one style from each of two compositors.

    Proudfoot and Bennett do not explain how they can tell that indenting the stick was used rather than, say, simply shortening the width of an adjustable composing stick or selecting a narrower one. One style of expedient involved the use of lower rather than upper-case letters and "&" instead of "and". The other style retained the capitals and only occasionally used "&" but narrowed the inter-word spacing and turned over long lines. These two styles are accompanied by spelling differences that confirm the compositor stint division as X setting H1r-H2v, I3r-I4v, and K1r-K1v and Y setting the rest of H-K.

    The preceding sheets A-G do not show the same spelling preference divisions and Proudfoot and Bennett leave open the possibility that more than two compositors worked on the book. There are no compositor-specific readings that contribute anything significant to our understanding of the play; stage directions omitted in the edition may or may not have been omitted in the underlying manuscript.

    Proudfoot and Bennett raise a number of questions about particular uncertainties, but they do not answer them. The recurrent misreading as "their" where the context makes it certain that "thy" is intended might be due to Shakespeare's own orthography, since the same error occurs in Sonnets of 1609. Q2 (1599) of Edward 3 corrects its copy Q1 more often than is usual in reprints, suggesting that the publisher was unhappy with the first edition, and many of Q2's alternative readings are adopted by modern editors as correct. Proudfoot and Bennett's edition is of course based on Q1.

    What follows is a selection of Proudfoot and Bennett's emendations. Almost all of them are adoptions of readings from Q2 of 1599 or Capell's edition of 1760. At 1.105, Proudfoot and Bennett have Lorraine call Artois "Degenerate traitor" where Q uses "Regenerate". They give serious consideration to the possibility raised in the OED that "regenerate" was itself a term of abuse linked to "renegade", but find this example and the only other one cited in the OED to be likely errors.

    At 1.368, Q reads "Acquant me with theyr cause of discontent" and Proudfoot and Bennett emend to ". . . with thy cause . . ." and comment that "Q's 'theyr' may originate in misreading of copy 'y^r', but this is too conjectural to recommend Q2's 'your'". This touches on the wider question of just why the Countess is using the familiar "thy" instead of the formal "you", which Proudfoot and Bennett understand as a deliberate character decision: at line 1.410 "The Countess retreats to the formal pronoun". In an article reviewed in YWES for 2003, I. A. Shapiro argued that in fact the Countess's earlier use of the various forms of "thou"--even before she decides to reject his shameful offer--is a serious breach of decorum: in other Elizabethan plays, subjects only ever speak to kings like this in moments of passion. As Shapiro explains, the beginner Shakepeare, newly arrived in London from rural Warwickshire, betrays his ignorance of court protocols.

    Proudfoot and Bennett are prepared to intervene quite extensively when they suspect that corruption in transmission has spoiled the poetic quality of the play even though the text remains perfectly legible. Thus at 2.447, Q reads "So doe her words her bewties, bewtie wordes" and Proudfoot and Bennett emend "bewties" to the singular to avoid breaking the wider chiastic pattern of which this line is part. Two lines later, King Edward imagines himself as a bee labouring "To beare the combe of vertue from his flower" and Proudfoot and Bennett follow John Payne Collier in emending to ". . . this flower" since the flower is the Countess and calling her "his" flower would imply that he possesss her, which he does not.

    Two lines later, King Edward likens himself to a sucking spider trying "To turne the vice I take to deadlie venom", and "vice" seems the wrong word: Proudfoot and Bennett follow Collier's emendation of "vice" to "juice", which makes better sense and supposes a simple minim error. Another two lines later, King Edward decides that religion is "To stricke a gardion for so faire a weed" and again the reading--calling the Countess a "weed"--is acceptable but odd, so Proudfoot and Bennett follow Collier's emendation of "weed" to "ward". This last case is less satisfactory than the others, since the notion of wardship--an underage person having their affairs managed until they come of age--does not clearly apply to the Countess: she could easily choose to give King Edward what he wants. On the other hand, the scene is about to take a new turn with the arrival of the Countess's father Warwick and Edward's attempts to suborn him to assist in her seduction, so the notion of wardship may stand as an analogy for their father/daughter relationship.

    King Edward leaves after telling Warwick what he wants, and Warwick's reaction in soliloquy is "O doting King, or detestable office" (2.513) and Proudfoot and Bennett follow Capell in emending "or" to "O" for the sake of rhetorical balance. Reflecting on his desire, King Edward says in an aside "Lust as a fire, and me like lanthorne show" (3.89) which requires three of Capell's emendations to make "Lust is a fire, and men like lanterns show". But Proudfoot and Bennett do not slavishly follow Capell, and repeatedly defend the retention of a quarto reading he emended. For example, King Edward tells the Countess that what she should give him is no less "Than right for right, and render love for love" (3.124), which is Q's reading, where Capell emended "render" to "tender". Capell's reasoning included the fact that the Countess responds with a contradicting formula: "Than wrong for wrong, and endless hate for hate", so that his emendation puts the adjective "tender" before "love" in King Edward's line to match the adjective "endless" before "hate" in the Countess's response. Proudfoot and Bennett Q's sense--the word "to" is assumed before "render" to make the infinitive--and note that William Painter, a source, uses "render" at the parallel moment in his version. They even wonder whether the Countess's line should be emended to end ". . . render hate for hate" to match King Edward's, noting that the misreading of "render" as "endles" is possible.

    At 4.105 King John asks his son Prince Philip "what is their concept, | Touching the challenge that the English make" and the logic of the conversation demands that he is asking for Philip's own opinion, so Proudfoot and Bennett emend "their" to "thy". They are sceptical that "concept" is merely a spelling variant for "conceit" (the reading they provide) in this period, noting that Literature Online shows it to be "very rare in the drama of the time" and that Vickers in private communication pointed out that the spelling "conceipt" appears in Kyd's The Spanish Tragedy. (Vickers believes that Kyd was Shakespeare's co-author on Edward 3.) In fact, "conceipt" is an extremely common spelling of "conceit", with nearly 500 occurrences in books published before the publication of The Spanish Tragedy in 1592, so this spelling is of little significance.

    At 4.124, Q prints "Like sweete hermonie disgests my cates" and Proudfoot and Bennett find this unacceptably unmetrical so they follow Capell in emending "sweete" to "sweetest". This seems to me unnecessary: the line may simply be headless and stressed quite regularly as "LIKE sweete HERmoNIE disGESTS my CATES". The quarto prints as part of a Frenchman's account of the devastation of war "And as the leaking vapour in the wind, | I tourned but aside I like wise might disserne" (5.5.58-9) and Proudfoot and Bennett delete the initial 'I' in the second line because it "makes little sense, even as affirmative 'ay'". They do not acknowledge or refute the treatment by William Montgomery in the second edition of the Oxford Complete Works (2005) of "I tourned" as a form of "Y-turned", the Middle English form for past participles that is here trisyllabic to form a regular hexameter.

    At 8.10, Q's King Edward asks about his son "what is he prisoner? | Or by violence fell beside his horse" which is metrically irregular. Proudfoot and Bennett find the sense problematic because of the shift in tense from present "is" to past "fell". To fix the tense, they emend "fell" to "felled". This seems to me to leave the line metrically unsatisfactory, since iambic meter would require stress on the "o" in "violence". The Oxford Shakespeare Complete Works emendation was to move "fell" to make "Or fell by violence . . .", which fixes the meter and treats the shift in tense from the previous line as acceptable.

    At the end of scene 8 in Q there are five lines that seem jarringly out-of-place: after the scene has drawn to a natural exit point, King Edward asks what is depicted in a certain picture (on someone's shield?). Prince Edward tells him that it is a pelican feeding her young with blood pouring from a hole she has pecked in her breast and that he, the king, should learn to be like the pelican. This could be understood as a reproach of King Edward for not sending reinforcements to his son, but King Edward makes no answer so the point of the exchange is opaque. Proudfoot and Bennett are sure that these five lines should not appear where they do in Q, and are not sure they belong anywhere--in a Longer Note they explore possible relocations--but they keep them at the end of Scene 8 and place them inside curly braces to mark their special status.

    Proudfoot and Bennett have an entirely new solution to the problem that in describing the enemy forces Audley likens them to a resplendent forest: "Streight trees of gold, the pendant leaues" (12.26). This is metrically short so Proudfoot and Bennett emend to ". . . the pendants pendent leaves", which explains the omission by haplography. They also think the sense is garbled without emendation, and I am not sure it is. The wider context is a description of an enemy's army:

Coting the other hill in such arraie,
That all his guilded vpright pikes do seeme,
Streight trees of gold, the pendant leaues,
And their deuice of Antique heraldry,
Quartred in collours seeming sundy fruits,
Makes it the Orchard of the Hesperides,

Unemended, the "pendant leaues" are clearly the leaf-like pendants flying from the tree-like pikes. Proudfoot and Bennett emend "sundy" to "sundry" in the penultimate line but strangely fail to mark this as an emendation in the collation notes and it is not in their list of stop-press correction affecting only certain exemplars (p. 124n1).

    At 12.146-7, Q's Audley says in his speech on death and fate that "If wee feare not, then no resolued proffer, | Can ouerthrow the limit of our fate" and Proudfoot and Bennett rightly object that this is not what he means. Q's "if . . . then" construction requires that only if we are fearless is our fate unavoidable, but Audley's point is that our fate is unavoidable whether or not we are fearless, so they emend (as no one has before) to "If we fear not, yet no . . .". At 18.216, Prince Edward begins a speech "Now father this petition Edward makes . . ." that refers to "thee whose grace" has been his shield; he says that "thy pleasure" picked him out to be "the instrument to shew thy power", and asks that "thou wilt grant" that many more English princes will enjoy the same success. As Proudfoot and Bennett point out, Prince Edward does not otherwise use "thou"/"thee"/"thy" to his father and they decide that in fact Prince Edward is praying to God; hence, they give him a stage direction to kneel and they capitalize "Thee", "Thy", and "Thy" accordingly. They might also have mentioned that it is hardly within King Edward's power--only within God's--to grant future English princes victories overseas.

    Near the back of the book are "Longer Notes" (pp. 355-86), many concerned with the historical facts underlying the drama and with the versions of these events recorded in the play's sources. The line of dialogue "Lilies that fester smell far worse than weeds" at 2.617 is also the last line of Shakespeare's Sonnet 94. Either Shakespeare wrote both, and repeated himself, or the author of the play knew Shakespeare's sonnet in manuscript (it was not yet in print) and copied it, or Shakespeare knew the play and copied the line. Proudfoot and Bennett explore the line's proverbial and perhaps biblical origins. Regarding the line at 3.169, Proudfoot and Bennett give evidence showing that I. A. Shapiro was wrong to assert (in an article reviewed in YWES for 2003) that there were probably no such things as wedding knives in England.

    Appendix One (pp. 387-98) is concerned with the early reception of the play. There was a ballad derived from the play known from a collection of ballads published in 1608 but which probably existed in an earlier lost edition of the collection and was probably written in 1596-99. Proudfoot and Bennett reprint the ballad. They also reprint Martin Wiggins's conjectured plot of the play of Alice Pierce, mentioned in Henslowe's Diary in the late 1590s, about Edward 3's infatuation with a woman of that name, possibly mounted to compete with a revival of Edward 3 around that time. They also reprint the bits of John Bodenham's Bel-vedere, or the Garden of the Muses (1600) that extract lines from Edward 3, and the bits found in Charles Lamb's Specimens of the English Poets (1808). Lastly, they give what is known of the earliest ownership of the 1596 quarto of Edward 3. Appendix Two is a casting chart (pp. 399-401), in this case a real one that assigns roles to notional actors where some Arden editions merely list which characters are in which scenes. Proudfoot and Bennett manage to get the cast down to 13 men and 2 boys. Appendix Three is a genealogical table (pp. 402-3) giving a who-is-who for the real history.

     The General Editors' Preface (pp. xiv-xviii) to Kent Cartwright's Arden3 edition of The Comedy of Errors has no obvious differences from previous editions. The description of how a textual note is constructed reads "This comprises, first: line reference, reading adopted in the text and closing square bracket; then: abbreviated reference, in italic, to the earliest edition to adopt the accepted reading, italic semicolon and noteworthy alternative reading(s), each with abbreviated italic reference to its source" (p. xvi). In Cartwright's actual practice, however, where the authority for the reading adopted in his edition is the First Folio, he omits the siglum "F" that this description says should indicate the earliest edition to adopt the accepted reading. This sort of contradiction between theory and practice makes critical editions difficult to use, especially for those just beginning their explorations in textual scholarship.

        In his Introduction (pp. 1-132), Cartwright makes a case for the play being a combination of the funniest knockabout humour ever staged in a theatre with a serious undercurrent about identity anxiety. Sketching its themes, Cartwright reckons that "In Errors, Shakespeare is formulating matters that will occupy him for the rest of his career" (p. 7). Early modern thinkers, he argues, were fascinated with error and its elimination from writings (especially about religion), and creative writers used it to drive comic plots. Syracusan Antipholus errs in the sense of wandering the Mediterranean in search of his lost brother. Errors of "misidentification . . . and false inference" (p. 11) multiply one mistake with another in the play, and so they spread virally. Importantly, no one is trying to deceive another: the errors are no one's doing but they are several people's fault in that bad habits of mind encourage them.

    In the play, identity is neither wholly within a person nor on the surface: "it is something of both, in-betweenness" (p. 14). The Syracusan Antipholus is "unstable and transformable" (he takes up the role offered him by his supposed wife) while Ephesian Antipholus is "hierarchical and proprietary" (p. 15). The notion of identity explored in the play is not the Humanist one of a "autonomous, self-unified and coherent" subject (p. 16). "If, in the classical ideal, male friendship consists of two bodies but one mind, as in Two Gentlemen, the reverse occurs in Errors, where the brothers constitute, in effect, one body but two minds" (p. 16). The various ruminations on identity from characters who have doubles suggest a fear that selfhood is not a singular thing, at least in the sense that self-contemplation calls for a kind of splitting of "the 'I' who is the object of self-commentary and the 'I' who comments" (p. 18).

    Identity is tied up with property ownership and the credit reputation it confers, and this social aspect also reaches into the mirroring of self that happens in marriage, as conveyed in Adriana's speech in 2.2 about being polluted by her husband's infidelity. When Antipholus of Syracuse woos Luciana with similar language of two-in-oneness, we have to ask just what it says about her that she seems to respond to what she believes to be a love-suit from her brother-in-law: "The idea of oneself-as-another may have become current by the 1590s, but, regarding love and marriage, Errors parodies as much as affirms it" (p. 22). The play is ambiguous about whether Ephesian Antipholus is guilty of the adultery that Adriana accuses him of.

    Underneath all the anxiety about identity is an anxiety about the human/animal and human/spirit boundaries and the risk of their being traversed. Frequently, people are described as various kinds of beast. Behind all this is Ovid's account of Circe. But the play's actual transformations are all in the mind and almost everyone is affected: "such nativity", the Abbess calls it at 5.1.406 (p. 28). The biblical Ephesus was thought of as a place of magic, and in the play the Syracusans put down their experiences to that effect. There something magical too about the way that images and metaphors pass uncannily from one character's speeches to another.

    The play's use of puns and quibbles gives it a kind of anarchic energy, and characters frequently pick up one another's phrases and re-interpret them in a new way; Cartwright calls this "connective repetition" (p. 33). When language fails in the play, it is generally because the context that grounds it is mistaken: the speaker does not know who the listener really is or what they really want. The chain circulates in the play like a metaphor for characters' connectedness, standing for domestic values in marriage as well as financial ones. It is being made at the start of the play, so it also stands for the intricate plot as it develops.

    The many debts of the play and the circulations of obligation that mercantilism generates are explored by Cartwright. The mercantile system reduces people to things, most literally in the case of the slave Dromios. Cartwright also finds religious overtones in the recurrent language of jugglers, the superstitious fears of possession and especially the anti-Catholic language (such as the Duke's reference to Circe's cup at 5.1.271), but these suggestions are counterbalanced by the Catholic Abbess who is smarter and kinder than the Duke. Cartwright traces the many biblical allusions across the play, including the recurrent use of the notion and word of "redemption" in the religious sense (we all owe God a death) as well as the monetary one.

    The entire action is structured by the implied countdown to Egeon's planned execution at 5pm that day. The passing of time is repeatedly referred to (more than in any other Shakespeare play) and the play adheres to the Unity of Time. Cartwright's linking of time to marriage is a bit strained ("Something lost and partially recovered or newly found through time is marriage", p. 46) but his discussion of marriage is sensitive to the play's nuances. As he points out, there are no words of reconciliation between Ephesian Antipholus and his wife Adriana at the end of the play. Equally, Luciana makes no reply to Syracusan Antipholus's overture of love. At the end the characters do not exit as married couples: the Duke and the women leave together, then the Antipholus brothers leave, and finally the Dromio brothers leave. Thus any "Final questions of heterosexual bonds are elided, and the denouement narrows to celebrate re-established male kinship and fraternal equality" (p. 49).

    What did Ephesus mean for the play's first audiences? The source play Menaechmi by Plautus is set in Epidamnus (on the Adriatic coast of modern Albania), and Shakespeare mentions a place called Epidamium. Ephesus, on the Aegean coast of modern Turkey, was legendary for its wealth from international trade and its pagan sorcery and its conversion to Christianity begun by the Apostle Paul. Cartwright claims that ". . . Falstaff later uses the epithet 'Ephesian' to suggest 'roisterer' . . ." in 2 Henry 4 (p. 51) but in fact it is Falstaff's Page who uses the word (in its plural form) about Falstaff's companions.

    Syracuse was indeed a rival city to Ephesus, and also rich from trade. Cartwright remarks that the Courtesan "turns out to be not the 'devil's dam' (4.3.53) but a kindly businesswoman" (p. 56) and I am not sure where he picks up that idea. Towards the end of the play, Ephesus increasingly comes to resemble Shakespeare's London in its range of urban character types, which is what Cartwright means by "Dark Ephesus" in his title for this section of his Introduction.

    Regarding "Genre and Style" (pp. 58-74), Cartwright notes that "Over a third of Errors's lines are taken up with characters' declaring their versions of prior events . . ." (p. 59-60), as the farce's interwoven and overlaid mistakings become ever-increasingly intricate. Cartwright explores the characteristics and psychological aspects of farce and compares them with its more sedate cousin genre of Romance. He points to a parallel in the way ". . . the metaphorical 'divorce' of the Egeon family [is] replayed in the estrangement of the Ephesian husband and wife . . ." (p. 61). The errors, in the sense of errands that turn into wanderings, of the Dromios are what keeps the plot in motion.

    Cartwright usefully reports that 87% of the play's dialogue lines are verse (and 25% of those rhyme), and 13% are prose. Shakespeare is often criticized for the over-simplified structure of his early verse with its recurrently endstopped lines, but Cartwright makes a convincing case that the endstopping of 1.1.112-117, in which Egeon conveys the story of his family's doubled rescues from shipwreck, is an economical means to convey the key points of a long and convoluted sentence: the last word on each line is a key idea in the tale: "us", "save", "guests", "prey", "sail", and "course" (p. 69). Cartwright is equally erudite on the use of stichomythia, though he uses the rather obscure notion of "tumbling verse" in which there are clearly some stressed syllables (usually four) but between them are an indefinite number of unstressed syllables.

    In the section on "Sources and Influences" (pp. 75-93), Cartwright acknowledges that Shakespeare drew on Plautus for The Taming of the Shrew, The Merry Wives of Windsor, Twelfth Night, and especially The Comedy of Errors (which combines elements of his Menaechmi with those from his Amphitruo). For The Comedy of Errors Shakespeare may have used a manuscript of William Warner's English translation of Menaechmi (published in 1595) as well as a Latin edition. Cartwright details how much Shakespeare keeps from Plautus's play Menaechmi and how much he discards, and how the lock-scene 3.1 was inspired by Plautus's Amphitruo. Cartwright considers the argument for a distinctly sixteenth-century Italian influence on The Comedy of Errors via the commedia grave tradition, and is sceptical that Shakespeare knew much of what was happening on the Italian stage.

    Shakespeare could have got the Apollonius material from John Gower's Confessio Amantis or from Lawrence Twine's The Pattern of Painful Adventures if he saw the latter in manuscript or in a lost edition prior to the earliest surviving one of 1594. The debt to Apollonius is in the theme of a man wandering the ocean after suffering shipwreck, then losing his family and finally regaining it; an especially close detail is the finding of a wife serving as a holy woman in a temple in Ephesus. Cartwright finds Shakespeare drew on the versions in Gower and Twine.

    The debate about a wife's role in marriage (in 2.1) is particularly indebted to Pauline arguments. Images from the Acts of the Apostles and Epistle to the Ephesians pop up across the play. Cartwright considers some earlier English plays based on Roman ones, each of which might have inspired a particular aspect of The Comedy of Errors, including Richard Edwards's Damon and Pithias and John Lyly's Mother Bombie. He finds also an allusive debt to the various pamphlets warning about the dangers of London life, such as debt and arrest, that circulated in the 1580s and 1590s by the likes of Greene.

    In "Staging" (pp. 92-102), Cartwright remarks that modern productions often copy the Roman style of "houses" on the stage representing the home of Ephesian Antipholus (the Phoenix) and the home of the Courtesan (the Porpentine), and perhaps a third for the home of the Abbess. But Cartwright reckons that this need not have been the early modern staging, since the stage doors in the frons scenae will do just as well and there is no need for a permanent association between one door and one location throughout the play.

    In the lock-out scene (3.1), Syracusan Dromio, Luce, and Adriana must somehow speak their lines from inside Ephesian Antipholus's house. Modern productions often wheel a portable door frame onto the stage so that the audience can see outside and inside the house at once. If played using the stage doors and central opening, the characters who speak from inside the house are given entrances that do not actually bring them onto the stage: they are "on" while still behind the frons scenae (as in "Enter within"). But they might still be visible to the audience if the doors had grills in them. The dialogue refers to a "hatch", which might mean that the door has an opening built into it.

    Some of those who speak from within might in fact be on the stage balcony, but this must be an unrealistic convention (rather than, say, suggesting a window that overlooks the street below) since Adriana speaks from within and must not see Ephesian Antipholus below since she is entertaining Syracusan Antipholus in her house. Or rather, as Cartwright points out, she might almost see him, which would add to the fun of the scene. Some productions use one actor for each twin-pair, which makes the confusion more realistic but sets up a problem for the final scene's revelation of twinship by bringing the pairs on stage together. Cartwright does not give examples of how this problem has been resolved in performance and this reviewer has not seen it attempted in the theatre.

    Cartwright's consideration of the "Early Performances" (pp. 102-12) begins with the reopening in June 1594 of London theatres closed by plague for most of the preceding 18 months. It is likely that the newly written The Comedy of Errors was in the first repertory of the newly formed Chamberlain's Men playing at the Theatre in Shoreditch. Cartwright asserts that The Comedy of Errors was unlikely to have been written specially for the Gray's Inn performance in December 1594, but does not say here why he thinks so. (The clarification finally emerges in Appendix One, where the date of composition is discussed in more detail: as Standish Henning points out in his New Variorum edition, the decision to hold the revels at Gray's Inn's was made on 12 December, and that would not have left enough time for Shakespeare to write the play and the company to prepare it before the 28 December performance.)

    Surprisingly, Cartwright repeats on page 103 Donald W. Foster's claim that Shakespeare might have played the part of Egeon, according to Foster's SHAXICON database, which was supposed to track words that Shakespeare used especially frequently on account of having recently learnt them in the roles he performed. The SHAXICON database has never been published and Foster's authorship claims made from it have been comprehensively disproved.

    The Comedy of Errors was performed at court in December 1594, indeed on the same day (28 December) as the Gray's Inn performance according to the Master of the Revels's record, so either the company gave two performances in one day or the court performance was in fact the day before as E. K. Chambers suggested. In his reading of the Gesta Grayorum account, Cartwright ponders whether the play's motif of witchcraft affected the wider celebrations of which it was a part, hence the parodic inquiry into all the night's disorders (extending beyond the play) and the blaming of them on "a Sorceror or Conjurer".

    Cartwright considers what kind of staging would have been possible in Gray's Inn Hall and the evidence that the play was revived in the late 1590s since Edward Guilpin alludes to it in Skialethia in 1598 and John Manningham refers to it in his record of the Middle Temple performance of Twelfth Night in 1602. The Comedy of Errors was performed again at court on 28 December 1604 for the new monarch.

    The last 20 pages of Cartwright's Introduction concern the play's life since Shakespeare's time. Cartwright gives a reading of the picture of the play that Rowe used in his 1709 edition, arguing that it holds in artistic tension all the contrasting aspects of the play that productions tend to pick amongst, emphasizing some and ignoring others. As is commonly the case with Shakespeare's plays, The Comedy of Errors was most often performed in heavily adapted versions in the eighteenth century. In the nineteenth century it became an opera. Essentially Shakespeare's text was restored in an 1855 London production by Samuel Phelps (this was Phelps's regular practice), and in 1895 William Poel directed the play in Gray's Inn using what he considered to be authentic Elizabethan staging and acting.

    For the twentieth century, Cartwright starts with Theodore Komisarjevsky's 1938 Stratford-upon-Avon production, and then moves on to the musical adaptation The Boys from Syracuse opening in New York the same year, which Irene G. Dash pointed out was the first musical featuring music and lyrics that are integral to the plot rather than extraneous reflections upon it. Next comes Clifford Williams's 1962 Stratford-upon-Avon production and then Trevor Nunn's 1976 Stratford-upon-Avon production, much influenced by the American musical comedy. The fourth production Cartwright considers is Tim Supple's sombre Stratford-upon-Avon production of 1996. This whole section is a fast-moving, brief survey of a great variety of productions, which mentions in passing that "black Dromios were first introduced in John Philip Kemble's Comedy of Errors in 1780" (p. 129), which is something this reviewer would like to have heard more about. Regarding the play on screen, Cartwright has not much to say: just one paragraph for this whole section as there are no films, only loose adaptations and borrowings. Cartwright gives the 1984 BBC television version just one sentence.

    As usual, we turn to the play text itself to note the editor's most interesting interventions. At 1.1.14, Cartwright follows Pope's emendation of the Folio's "Syracusian" and "Syracusians" (here and throughout) to "Syracusan", calling it a modernization. It is hard to see why he thinks that and yet remarks that "The -an form was used in the 16th and 17th centuries", which in any case is not borne out by EEBO-TCP: the two forms seem about equally common in that dataset. But as Cartwright points out, metrically they can be the same (so long as "-sian" is spoken as if it were "-shan").

    Cartwright again follows Pope in omitting the "any" in the Folio's "Be seene at any Siracusian Marts and Fayres" (1.1.17), since it was probably picked up by eyeskip from the "any" in the line above or the one in the line below. Likewise at 1.1.22, where the Folio has "To quit the penalty, and to ransome him", Cartwright follows the Second Folio in dropping the second "to". At 1.1.46, Cartwright sticks with the placename "Epidamium" (with four syllables) where editors usually emend to "Epidamnum", for reasons given later in a Long Note.

    Where Folio Egeon say that "A meane woman was deliuered" (1.1.54) of twins in the same inn at the same time as his wife had hers, Cartwright follows Nicolaus Delius in emending to "A meaner woman . . ." to fix the metre. Although Cartwright rightly explains that "meaner" was not always comparative in Shakespeare's usage--it could just mean "mean"--the effect in context here is comparative, and it sounds odd for Egeon to think of his wife that way.

    In the Folio, the Duke allows Egeon one day "To seeke thy helpe by beneficiall helpe" (1.1.151), which makes no sense. Cartwright follows John Payne Collier in emending the first "helpe" to "hope", in the sense of "the thing you hope for". A note on the lemma "maw . . . messenger" (1.2.66-7) is marked with an asterisk, which according to the General Editors' Preface (p. xvi) is supposed to indicate that a note discusses an emendation or variant, but this one does not.

    Cartwright's collation note at 1.2.87 claims that his is the first edition to emend the Folio's "Thy Mistris markes?" to "Thy 'mistress' marks'?" but since the words are merely modernized he must mean that he is the first editor to put the phrase "mistress marks" in quotation marks. It is odd to claim such a small matter of punctuation as an innovation. Cartwright records as an emendation, credited to Thomas Hanmer, the alteration of the Folio's "ore-wrought" (in the sense of over-reached) to "o'er-raught" (1.2.96). Since "raught" and "wrought" are archaic past participles of the verb "to reach" this is really just a modernization and it is not clear why "raught" is preferable to "wrought" unless the unstated principle is that "wrought" should be avoided because it is also the past participle of the verb "to work", which is not the sense wanted here.

    After a catalogue of evildoers whom Syrcusan Antipholus fears will be found in Ephesus ("They say this town is full of . . ."), Cartwright punctuates the end of the sentence as "And many such -- like liberties of sin" where F has "such like" which most editors treat as "suchlike" (1.2.97-102). That is, the evil persons are compared to liberties. Cartwright's dash indicates that he thinks Syracusan Antipholus breaks off his list of persons and finishes his comparison of "this town" with "like liberties of sin", using "liberties" in the sense of independent districts in a town. It is awkward to compare a singular town to a set of liberties, although I suppose that with four intervening lines of dialogue the listener is unlikely to notice the problem. Cartwright acknowledges this but points out that it is also awkward to compare evil persons (a concrete noun) to "liberties" (an abstract noun) and that the emendation to "libertines" adopted by some editors produces the pleonastic "libertines of sin": libertines by definition are sinful. As he concludes in his long note, "No reading answers all objections".

    At 1.4.41, Cartwright retains Folio Egeon's statement that he often traded with "Epidamium" rather than following Pope's emendation to "Epidamnum", the accusative case of the name of the real city Epidamnus. Because "Epidamium" occurs seven times in the Folio text, set independently by two compositors, Cartwright concludes that this cannot be a printing error and that either Shakespeare misremembered the place name or invented it for himself. (The key question, not discussed by Cartwright, is whether Shakespeare would have fixed it if the problem had been pointed out to him.)

     Cartwright follows the Second Folio to print "Sister, you know he promised me a chain: | Would that alone, alone he would detain" (2.1.105-65) where the Folio has "Sister, you know he promis'd me a chaine, | Would that alone, a loue he would detaine", explaining that the second line means "'If only he would withhold just that [i.e. the chain]". Cartwright acknowledges that the repetition of "alone" is awkward and claims that "the second alone refers differently from the first" but without indicating how it is different. He lists various emendations that have been used, including the attractive "Would that alone o' love he would detain", meaning "if only that were the only aspect of love he would keep from me", which supposes just a simple "o" to "a" error in transmission.

    At 2.1.108-12 there is a notorious crux that Cartwright fixes as follow:

Folio	Cartwright
I see the Iewell best enamaled Will loose his beautie: yet the gold bides still That others touch, and often touching will, Where gold and no man that hath a name, By falshood and corruption doth it shame.	I see the jewel best enamelled Will lose his beauty--and though gold bides still That others touch, yet often-touching will Wear gold--and any man that hath a name By falsehood and corruption doth it shame.

Cartwright flips "and" and "yet" in the second and third lines, emends "though" to "the" in the second line, and emends "no" to "any" in the fourth line; this is Cartwright's own solution borrowing bits from various earlier emendations. As he glosses the meaning, the best jewels lose their beauty and although gold is hardwearing ("bides still") it is worn away by frequent touching, and falsehood and corruption will shame the good name a man has.

    The only problem with this emendation, and it is acknowledged by Cartwright, is that there is something of a contradiction between saying that touched gold "bides still" and yet wears away. One solution would be to say that gold is still gold even when touched, but Cartwright points out that "bides" is one of the class of self-contradictory words: it can mean to withstand and to submit, just as "wear" can mean to decay through use but also (used intransitively) to withstand decay.

    Folio Adriana says to Antipholus, about his sexual sins, "I liue distain'd" (2.2.152) but from the context the sense is clearly that she is not stained. Rather than emend the word as other editors do ('undistained', or 'unstained'), Cartwright uses punctuation to read "dis-stained" and argues that it means unstained. As a parallel he offers "dis-horn the spirit" in The Merry Wives of Windsor 4.4, but the parallel is not close since "dis-horn" is already hyphenated in the Folio text of The Merry Wives of Windsor and without the hyphen it would not make another word, whereas the hyphen in "dis-stained" is needed to keep it from conveying the opposite sense. And in any case it is hard to see how the hyphen would help an audience hear Cartwright's proposed sense.

    At 2.2.196, Cartwright retains the Folio's reading to have Dromio of Syracuse say "We talk with goblins, owls and sprites!". The line is short, but tetrameters are not unusual in Shakespeare. Some editors have objected that owls do not belong with the other two kinds of creature, but Cartwright points out that owls are generally associated with the supernatural in Shakespeare and were said to suck breath just as Dromio, two lines later, says he fears will happen.

    In the Folio, Dromio of Ephesus says to Dromio of Syracuse that if the other had become Dromio in his place "Thou wouldst haue chang'd thy face for a name, or thy name for an asse" (3.1.47), which makes little sense. Cartwright emends "face" to "place" and "or" to "and". Thus according to Cartwright the line means that one Dromio accuses the other of changing his place for a name and then changing that name for "ass". Cartwright sounds unconvinced by this when he remarks in a long note that "precise logic may be less important than effect".

    Where the Folio has "Your cake here is warme within: you stand here in the cold" (3.1.72), Cartwright follows Capell in deleting the first "here" since it confuses the contrast of warmth within the house and coldness outside it, and it looks like a simple mistaken repetition in transmission. (The Folio reading is also metrically rough, although Cartwright does not mention that.)

    Persuaded by Balthasar to give up trying to get into his own home, Antipholus of Ephesus says in the Folio "I will depart in quiet | And in despight of mirth meane to be merrie" (3.1.107-8). Cartwright emends "mirth" to "wrath" to restore what he thinks is the necessary contrast, but surely the contrast is already present in the Folio reading: in order to spite merriment at his expense he will be merry rather than reacting angrily. Antipholus has not really been subjected to wrath in being kept out of his house, so in Cartwright's reading the wrath must be his own (in reaction to being kept out) and it is odd to say that he will spite his own wrath. Chiding Antipholus of Syracuse for what she thinks is his neglect of her sister, Luciana asks "Shall loue in buildings grow so ruinate?" (3.2.4) and Cartwright combines emendations by Lewis Theobald and Capell to produce "Shall love, in building, grow so ruinous?". This fixes an obvious nonsense (the point is about the building of love, not a love of buildings) and restores the rhyme of ". . . Antipholus | . . . | . . . ruinous".

    At 4.3.62-3, Cartwright omits a necessary collation note, since he prints "Master, if you do, expect spoon-meat, or bespeak a long spoon" and in fact the Folio has "Master, if do . . ." and the word "you" was an addition first made in the Second Folio. Cartwright retains the Folio's "And much different from the man he was" (5.1.46) where most editors follow the Second Folio to read "And much, much . . .", pointing out that if "different" is trisyllabic there is no missing syllable. True, but it would be helpful to hear how Cartwright thinks its three syllables are stressed. An alternative would be to say that the line is headless (that is, it lacks a first unstressed syllable) so that "And" is, unusually, stressed and "different" is disyllabic and the line is thus a normal pentameter. Cartwright has the Second Merchant say "By this, I think, the dial points at five" (5.1.118). R. B. McKerrow in his unpublished work on the play for his uncompleted Oxford Shakespeare edition (cited in the 1986-87 Oxford Complete Works) remarked that a dial (a circular ring, as on a sundial) cannot point at anything, but the dial's point can, so the punctuation should be "the dial point's at", meaning 'the dial point is at'.

    In the Folio, the reuniting of Emilia and Egeon (5.1.344-52) begins with her saying "Oh if thou bee'st the same Egeon, speake: | And speake vnto the same Aemilia". This calls for Egeon to answer, and the next thing he says is "If I dreame not, thou art Aemilia, | If thou art she, tell me, where is that sonne | . . .". But before Egeon makes this answer, the Duke has a six-line speech  beginning "Why heere begins his Morning storie right . . ." that has seemed to many editors an impossible intrusion on the reunion, so they have moved it out of the way, putting it after Emilia and Egeon have confirmed each other's identity. Cartwright keeps the Duke's intrusion where it is in the Folio, giving in a long note the dramatic justification for doing so: ". . . the father might plausibly stand in awed silence--a pause filled by the Duke--before he can acknowledge a reality beyond his expectations or hopes" and ". . . the Duke's lines might verbalize the hopeful speculation going on in the Merchant's mind that leads him finally to embrace Emilia as his own".

    Cartwright sticks with the Folio to print "And you, the calendars of their nativity. | Go to a gossips' feast, and go with me; | After so long grief, such nativity!" (5.1.404-6). Editors generally treat the line-ending repetition of "nativity" as an mistake and change the second one to "festivity" or "felicity". Cartwright defends the repetition as intentional by comparison with the fourfold occurrence of line-ending "crown" in Richard of Gloucester's soliloquy about deformity and his ambitions in 3 Henry 6 (3.2.168-179). This is also defended in the Introduction at page 67-68, and as here Cartwright claims that a differentiation between the two occurrences of "nativity" could have been introduced if in the first occurrence ("And you, the calendars of their nativity") the last word is spoken trisyllabically (as "nativ'ty").

    This is possible but there is no need for the elision since the line is a perfectly acceptable hexameter with "nativity" having four syllables: "And YOU, the CALenDARS of THEIR naTIviTY". Here Cartwright says that the elision is merely possible ("could have been differentiated with sound") but in the Introduction he insisted on it: "the two occurrences entail different pronunciations" (p. 67) and "The same word pronounced differently: scansion suggests not erroneous typesetting but doubling and varying . . ." (p. 68), despite remarking in a footnote that "full syllabification would make it [line 5.1.404] hexameter" (p. 68n1) and giving no reason to doubt that it gets full syllabification.

    The first appendix concerns "Date of Composition" (pp. 314-25). The terminus ad quem is given by the known performance at Gray's Inn on 28 December 1594. Although the play could not have been written for the Gray's Inn performance it could have received embellishments (such as added bits of legalese) for that show. The terminus a quo is given by the reference to France "making war against her hair", alluding to the opposition to Henri of Navarre taking the throne to which he was declared the heir in 1589.

    To narrow the five-year period 1589-94, Cartwright categorizes several kinds of evidence. The play needs 12 men and 4 boys, which is fewer than Shakespeare's plays typically needed prior to the plague closure of 1592. Also, Egeon is scheduled to be executed at 5pm, and from 8 October 1594 afternoon playing was shifted from a 4pm to a 2pm start, which would make 5pm also coincide with roughly the real-time end of the performance. Dromio of Syracuse's speeches about Nell the spherical woman might allude to the newly fashionable globes of the world being sold with specifically English labelling from 1592.

    Cartwright traces a series of even more minor possible allusions and echoes, including the use of the word "saffron" (4.4.62) alluding to Saffron Walden where Gabriel Harvey was born. Cartwright thinks that "saffron" is "a word that occurs nowhere else in Shakespeare" (p. 320) but in fact Lafeu in All's Well that Ends Well 4.5 calls Paroles "a snipped-taffeta fellow there, whose villainous saffron would have made all the unbaked and doughy youth of a nation in his colour", the Clown in The Winter's Tale 4.3 says he must buy "saffron to colour the warden pies" at the feast, and Ceres in The Tempest 4.1 describes Iris's "saffron wings".

    Cartwright takes seriously Shakespeare's engagement with the rivalry of Nashe and Harvey, to which the play alludes, and also explores the play's possible minor borrowings from Greene and Lyly. Although these seem to point to composition near the end of the 1589-94 window, Cartridge does not find them especially decisive. The play's rhyme is like that of Venus and Adonis and Lucrece, which Shakespeare appears to have written in reaction to the long theatre closure that began in 1592, and its ratio of verse to prose is like that of his other plays around 1594-95. When vocabulary tests are brought in, they too suggest a 1594 composition.

    Appendix Two details "The Text and Editorial Procedures (pp. 326-51). Cartwright describes the orthodox view regarding the typesetting of the Folio text, including without question the attribution of stints to compositors and hence not taking account of Pervez Rizvi's essay "The Use of Spellings for Compositor Attribution in the First Folio" (reviewed in NYWES for 2016), which showed how unreliable compositor attributions are; presumably Rizvi's piece appeared too late to be taken into account. Cartwright surveys the range of textual problems that editors encounter in Folio The Comedy of Errors, which has only one really tricky crux that matters (at 2.1.108-12), discussed above.

    The only problematic stage direction is "Runne all out" at the foot of column 'a' on sig. H6v, the substance of which seems to be repeated just four words later at the top of column b with "Exeunt omnes, as fast as may be, frighted". Cartwright rejects the idea that both directions were followed in the sense that everyone runs out, runs back in again, and then runs out again, which he calls a "Keystone Cops-like chase" even though it is "theatrical fun" (p. 336). Cartwright speculates that the duplication was caused by annotation of the manuscript underlying the printing of the play, or maybe the compositor divided a single stage direction into two parts because the whole of it would not fit at the bottom of H6v column 'a'. The latter idea Cartwright finds unconvincing because the compositor could have put the whole, long stage direction at the top of column 'b' just after the Officer's line "Away, they'l kill vs"; he decides the puzzle is essentially insoluble. In his text, he simply omits the "Runne all out" direction and retains the collective exit after "Away, they'll kill us!" (4.4.146).

    The Folio shows signs suggesting an author in the act of composition, such as changing speech prefixes and stage directions giving detail that no theatrical person would need to know. For example, Ephesian Antipholus is called "Antipholis Sereptus" (meaning the one who was stolen) in the opening stage direction of Act 2, and Syracusan Antipholus is called "Antipholis Erotes" and "Antipholus Errotis" (suggesting the one who wanders?) in the opening stage directions of 1.2 and 2.2, respectively. Likewise there is Luce who is also Nell and three characters all get speech prefixes for a Merchant. That Pinch is also a schoolmaster, as stated in his entrance in 4.4, is knowledge of no use to anyone and of no consequence in the play, but it sounds like a dramatist thinking through his creative options as he writes

     Cartwright acknowledges that although it was once orthodoxy to believe that such things would get ironed out in a document used to run the play in performance, scholars such as William B. Long and Paul Werstine have shown otherwise; the document used to print the Folio text of The Comedy of Errors might, thus, despite these signs of the author in the act of composing, have been used in the theatre. Cartwright claims that "Werstine's perspective has been resisted by [Stanley] Wells and [Gary] Taylor . . ." (p. 341), which is temporally askew since he quotes Wells and Taylor from the Oxford Shakespeare's Textual Companion of 1987, which appeared before the celebrated articles in which Werstine made his claims about the wrong-headedness of the New Bibliographical categories of foul papers and promptbook.

    Regarding the copy for the Folio text of The Comedy of Errors, Cartwright notes that there is no music in the play, which is unusual for an early Shakespeare comedy. The act divisions might come from the Gray's Inn performance, because indoor performance required breaks to tend the candles, but they are not especially closely related to the action of the play, which flows more or less continuously. The only part of the play that requires tricky staging is the both-sides-of-the-door material in 3.1, and it would be better served by the facilities of an open-air amphitheatre (with stage doors and a balcony) than those of Gray's Inn Hall. The play's shortness is no barrier to its having been performed on the professional stage.

    Of his "Editorial Procedures" (pp. 346-51), Cartwright reports the modernizing of the play from the 1623 Folio, with only the occasional retention of an early modern spelling "for its sound" (p. 346). To keep the surprise that the Abbess is Egeon's long lost wife, Cartwright gives her the speech prefix "ABBESS". Punctuation has of course been entirely rethought, and Cartwright finds that he favours longer sentences than other editors. He uses dashes for "a break in a train of thought" and for switches of addressee and interruptions, and laments that "series protocols" do not allow him to use dashes to mark off parenthetical clauses (p. 349).

    Cartwright spends some time explaining his policy of citing first instances of words' usages from the online OED rather than the printed one, seeming not to notice that the dictionary does not claim to be definitive on first usages and that it is often easy to antedate a word in the dictionary by searching for it in EEBO-TCP. Appendix Three concerns "Casting and Doubling" (pp. 352-4), starting from the fact that there are 19 speaking roles. Cartwright reports the play can be performed by 13 men and three boys or 12 men and four boys. He provides a doubling chart showing how the 19 actors could perform the play, which is peculiar since he insists that 16 actors can do it and it is the doubling for that minimal arrangement that would be most interesting to see.

    The last major Shakespeare edition to be considered this year is Valerie Wayne's Arden3 Cymbeline. Like Cartwright--indeed in exactly the same way--Wayne departs from the form for textual notes described in the General Editors' Preface (xvi) in omitting the siglum "F" that this description says should indicate the earliest edition to adopt the accepted reading. The same objection applies as with Cartwright: textual notes are tricky enough for beginners without editions introducing such contradictions between the theory and the practice.

    Wayne's Introduction (pp. 1-136) begins by tracing the origins of the romance genre to twelfth-century France and the language romanz, the vernacular French of the time, as opposed to the Latin that educated people used; she identifies the plays of the 1580s and 1590s that drew on these stories and names a few minor sources for Cymbeline. Wayne finds the origins of what she calls the "calumny plot" in the story of Susanna in the biblical Book of Daniel and follows its development through early modern plays, prose adaptations, and paintings.

    The case of Anne Boleyn made calumny tales topical. In Shakespeare's calumny plots--in The Merry Wives of Windsor, Much Ado About Nothing, Othello, The Winter's Tale, and Cymbeline--the woman in each case is falsely accused and audience's sympathies are firmly against the accusers. Wayne traces the debates about women's roles in general and within marriage in particular across the late medieval and early modern periods, and the various discourses of misogyny and proto-feminism. That Posthumus forgives Innogen in 5.1 when he still thinks she was unfaithful is a remarkable outcome that reflects "a larger cultural shift towards repudiating the most explicit forms of misogyny" (p. 13). Wayne works through the differences between the sexual calumny in Cymbeline and in other Shakespeare plays, most importantly the fact that no one in Cymbeline believes it and that the victim has no female support.

   As a dramatic genre, romance broke the classical rules by leaping over large bounds of time and space, as Philip Sidney amongst others complained. The modern sense of Shakespeare's Romances as a group comes from Edward Dowden's generic classification in the service of biography, in which the writer went through four phases. This perspective detached Shakespeare's plays from a romance tradition on the English stage from the 1570s to 1610s (including Clyomon and Clamydes, The Rare Triumphs of Love and Fortune, and Mucedorus), and their literary precursors. The other Romances in the Folio are placed with the comedies, so why is Cymbeline among the tragedies?

    Wayne considers the classical and early modern attitudes to narratives in which the bad die and the good end happily, which can be consistent with a kind of tragedy. Before the formalization of the category of tragicomedy, it may well have been unremarkable to consider Cymbeline as a tragedy with a happy ending, hence its location in the Folio. Tragicomedies were especially popular on stage because audiences did not know, as they did with comedies and tragedies based on familiar tales, which way the ending would go: deaths along the way did not rule out a happy ending. Part of the delight included seeing a dramatist self-consciously figure out how to bring all the unlikelinesses to a believable resolution.

    Wayne mentions that Cymbeline is like King Lear, Macbeth, and Hamlet in that "it dramatizes events before the Norman Conquest" (p. 27) and while it is obvious that King Lear and Macbeth do, I think the case for Hamlet being pre-Conquest (which Wayne gets from Margreta de Grazia) needs at least some explaining, for example regarding the relevance of a play set in Denmark to events in England. De Grazia's point is that the Danish rulers of England got conquered at Hastings, and that plays about England after 1066 are Histories in the Folio because they concern matters of state and of relevance to the Elizabethan present, while the pre-1066 plays are Tragedies because they concern matters affecting individuals in a dim and distant past. Thus Cymbeline is a Tragedy in the Folio

     Of course, Cymbeline can be read as essentially an historical play. An unusually high number of character types, plot twists, and themes from earlier plays appear in Cymbeline. (It might be speculated that plays near the end of a career are particularly apt to recirculate material if the dramatist is short of fresh ideas.) Wayne repeats here the argument she made in an article reviewed in NYWES for 2015 that this intentionally recapitulatory nature of the play is part of the explanation for its being placed last in the Folio: its very copiousness, its inclusiveness, makes it an apropriate termination since it sums up the dramatist's virtuosity in managing all of its complexities.

    Wayne determines that Cymbeline was probably written between March and November 1610. The terminus ad quem is given by Simon Forman's seeing it at the Globe in 1611. Wayne explores the connections between Cymbeline and The Winter's Tale--including the use of some of the same sources--but that does not establish which came first. Prince Henry was invested as Prince of Wales in June 1610. In January 1610 there was held at the Whitehall Banqueting House a martial contest called Prince Henry's Barriers with speeches by Ben Jonson and quasi-dramatic ceremonies, and Wayne finds in this event some material that inspired the second half of Cymbeline: "a recollection of the nation's Arthurian past through its royal genealogy; evocation of the 'ancient name' of Britain and 'the union of this isle' . . . a celebration of Britain as 'the only name made Caesar fly' . . . anticipation of the nation's future greatness as an empire . . . and martial strength as a confirmation of masculine maturity" (p. 35). (The quotations are from Jonson's text of the event.)

    Prince Henry's militaristic, imperial concerns shaped the Barriers event and, Wayne thinks, "probably influenced the chief playwright for the King's Men, although the play's approach is more politically ambiguous than the blatant royal flattery of productions written for the court" (p. 36). The King's Men performed Mucedorus at court in February 1610 and Wayne sees this play as something of an "antidote to the chivalric fervour" (p. 36) around Prince Henry. Wayne traces connections between details in Cymbeline and various court entertainments (in 1610 and earlier), which makes them at most minor sources for the play. She also lists the newly published books around this time that indicate a surge of readerly interest in Britain as a political entity.

    The theatres were closed for most of the second half of 1610, so if the King's Men wanted to perform the play in public before presenting it at court in the Christmas 1610 season they had limited opportunities. (But in fact there is no evidence that Cymbeline was played at court in the Christmas 1610 season: we do not know the titles of the plays that were performed then.) Wayne gathers a series of topical allusions that she thinks makes composition of Cymbeline in "early spring to late 1610" (p 43) most likely. At that time, Parliament was attempting to abolish the Court of Wards; Posthumus of course is the King's ward, and the Queen and Belarius are further surrogate parents in the play.

    Also at that time, Galileo's Siderius Nuncius was published, announcing the discovery that Jupiter has four moons and hence confirming the Copernican, heliocentric cosmos; this Wayne thinks is alluded to in Jupiter's descent over the four figures of Posthumus's family in 5.5 and in Cymbeline's question "Does the world go round" in 5.6. Even more tenuously, Wayne sees the play's references to oaths of loyalty as allusions to the Oath of Allegiance that James I reinstated in reaction to the assassination of the French king Henry IV in May 1610. Although the direction of borrowing cannot be certain, Wayne also uses the links between Cymbeline and Heywood's plays The Golden Age and The Silver Age for Queen Anne's men to fix the composition in 1610, and likewise Shakespeare's imitation of Francis Beaumont and John Fletcher's Philaster of late 1609 or early 1610.

    James I's attempts to unify England and Scotland made the nature of Ancient Britain topical. James failed to legally unify the kingdoms, but he began to call himself the king of Great Britain. Wayne casually refers to "[Ferdinand de] Saussure's insight that language not only reflects but is constitutive of one's world" (p. 51) to claim that this self-designation by the king carried more than merely symbolic weight. But Saussure never made the claim that language is constitutive, and Wayne cites nothing from his work to support this claim about linguistic determinism, which is commonly known (somewhat inexactly) as the Sapir-Whorf hypothesis and has been thoroughly debunked. Wayne explores James's visions of a united Great Britain and the role in contemporary thinking of the ancient Britons, about whom little was known.

    Wayne considers the ways in which Britain was thought of as a second Rome, especially once its overseas colonial projects got underway, including the writing of Thomas Hariot about Algonkians and the likening of American natives to ancient Britons. When Innogen first encounters Guiderius and Arviragus the same question of who is barbarous and who is civilized emerges, as it does also when Guiderius beheads Cloten. The Italian threat to the British heir's body (by rape) in the first half of the play is mirrored in the Italian threat to the British body politic in the second half. Wayne explores the various Celtic aspects of the play, noting that the Welsh were (in Shakespeare's time) thought of as the quintessential early Britons.

    Innogen's part is the third-largest female role in the Shakespeare canon, after Rosalind and Cleopatra. Wayne considers each of Cymbeline's main characters in turn, attending to their lineages and their engagement with the sources' attitudes towards, for example, social class and parentage. There appears to be some aspect of Rome-resister Boudica in Cymbeline's queen, although her interest in poisons makes her a bit Italian too. And of course Iachimo is a stock figure of Italian degeneracy. And so on through the character list, the main point being that many of the characters are hybrids: "The complexity of these identities suggests that the play actively mixes affinities within each of its characters to convey the Britons' multiple origins and subsequent history" (p. 80).

    There has been debate about whether Posthumus and Innogen are legally married in the eyes of the first audiences, but Wayne points out that Jupiter says that their marriage took place in his "temple" (which makes it formal) and the First Gentleman says Posthumus is "wedded" at the start of the play. Claims of Innogen's possible virginity are based on contestable readings of verbal details in the play. The Soothsayer explains the prophecy's reference to "a piece of tender air" as "mollis aer" (Latin for "tender air") and hence "mulier" (Latin for "woman" or "wife"). Thus etymology seems to embody the cultural assumption of female softness and weakness. But this particular false etymology was known in Shakespeare's time to be false, Wayne points out, and it was cited in debates about anti-feminism.

    Posthumus and Cloten are alike in a number of ways--desiring Innogen, resenting her, given to wagers--and the roles can be doubled. Posthumus's misogyny is straight out of Juneval's Satire Six, according to Wayne. She considers the play's treatment of deviant sexualities, including incest, and its disrupting of gender boundaries: in the end, binary thinking loses out to hybridity and multiplicity in desire. For the historical plot, Shakespeare used Holinshed's Chronicles, Geoffrey of Monmouth's History of the Kings of Britain, and the multi-authored The Mirror for Magistrates. Wayne goes through some of the more minor historical sources. For the wager plot Shakespeare used Giovanni Boccaccio's Decameron (Day 2, Number 9), which he perhaps read in the French translation or the Italian original (the first English version was not available until 1620), and the anonymous story Frederyke of Jennen that he could have encountered in a variety of places.

    Wayne considers European versions of the wager plot that precede Boccaccio's and then turns to classical motifs. Then she moves on to Cymbeline's parallels with the play The Rare Triumphs of Love and Fortune, and then the debt to folktales, including Snow White. There might also be a link to Chaucer's The Franklin's Tale about Arveragus's wife Dorigen who agrees to sleep with her suitor Aurelius if he can remove the rocks that might wreck her husband's ship on its return, which Aurelius seems to achieve by an illusion. In the play Innogen refers to being dressed like a "franklin's housewife" in 3.2 and to Posthumus being "upon a rock" in 5.5, which might be allusions to Chaucer's tale. Wayne next traces even more minor borrowings from other works, including Miguel de Cervantes's Don Quixote.

    As is conventional, the last part of Wayne's Introduction concerns the play's "Afterlives" (pp. 109-36). Until the twentieth century, the play's penultimate scene, including the descent of Jupiter, was largely cut in performance. After the first performances in 1610 and 1611, the next sighting of the play is a court performance on 1 January 1634. In the Restoration, it reappeared in Thomas D'Urfey's heavily adapted The Injured Princess, or the Fatal Wager. After D'Urfey, Charles Marsh and David Garrick returned to Shakespeare's text, but with large cuts, in particular the descent of Jupiter. Garrick made the play markedly more realistic and nationalistic, evoking the frightening context of Napoleonic invasion. Cutting Posthumus's dream reduces the sense of Posthumus's noble ancestry, which necessarily diminishes his status at the end of the play, and indeed in some productions it seemed as if the male lead was really Iachimo. The play was first performed in America in 1767. (In describing productions, Wayne switches between the forms "Innogen" and "Imogen", apparently using the former for discussions of the text and its meanings and the latter for discussions of actual performances of the role by specific actors, presumably because performers spoke the latter form of the name.)

    Wayne gives sketches of various twentieth-century productions. The play's mixing of genres was to some extent more acceptable to late twentieth-century audiences than earlier ones because cinema had got people used to such mixing. Wayne remarks in passing that using the name Innogen in performance began with Peter Hall's 1988 production. She describes a few recent productions that have set the play in a colonial context, and others that were performed in actual colonial contexts such as Jamaica in the eighteenth century and Sri Lanka in the nineteenth. Then she sketches a few other productions from across the world.

    The first screen Cymbeline was a silent film from 1913. Wayne gets wrong a detail of the Cymbeline segment of Douglas Hickox's 1973 film Theatre of Blood: the wife who wakes up next to her decapitated husband was played by Joan Hickson not Diana Rigg, the latter taking the much larger role of the protagonist's daughter Edwina Lionheart. Here the convention of putting an actor's name in brackets after a role conflicts with the convention of putting a writer's name in brackets after a quotation or summary of their ideas, since it is only by looking at the list of References near the end of the book that one can find that Wayne's phrase "her husband's headless corpse (Díaz Fernández)" does not mean that Fernández played the decapitated man but rather that he is the cited source of Wayne's inaccurate information about the film. Poets, playwrights, and novelists have liked the play and Wayne finds echoes of it in their works, including those by Charlotte Brontë, Samuel Beckett, James Joyce, and Virginia Woolf.

    Turning to Wayne's text of the play, we find that she lists many emendations that she clearly thinks have some plausibility but has not adopted, preferring to seek explanations for what the Folio has rather than to change it. The following are some of her more interesting editorial decisions. At 1.3.8-10 the Folio has Pisanio say that Posthumus stayed on deck, waving "for so long | As he could make me with his eye, or eare, | Distinguish him from others". Editors since Lewis Theobald have emended "his eye" to "this eye" so that Pisanio is referring to his own ability to make out Posthumus. Wayne does so too. The problem with this emendation is that it seems odd for Pisanio to use his ear to distinguish someone waving at him. C. J. Sisson proposed that the Folio is correct, and that Pisanio means that Posthumus's eyes and ears were telling him that Pisanio could distinguish him, presumably because Pisanio was waving and hollering back. The trouble with that idea is that Pisanio says nothing about waving and hollering back, and the ensuing discussion is about Pisanio's (and then Innogen's) ability to see the diminishing Posthumus.

    Folio Iachimo say of Innogen "I could not beleeue she excelled many" (1.4.76-7) in a context where he must mean that he believes it (because she excels many but not necessarily all), so Wayne adopts Edmond Malone's emendation "I could not but believe she excelled many". As explained in her essay "The Gendered Text and its Labour" (reviewed in NYWES for 2016), Wayne thinks that the part of Posthumus's letter delivered by Iachimo that Innogen reads aloud should be as in the Folio (1.6.23-5), to give in her edition "Reflect upon him accordingly, as you value your trust. Leonatus" rather than being emended to ". . . your truest, Leonato". Wayne's argument is that this cannot be the sign-off of the letter since Innogen goes on to say "So far I read aloud | . . .  my heart | Is warmed by th'rest", meaning there is more.

    The problem with that argument is the word "Leonatus" at the end of the part that Innogen reads aloud: why would he put his own name anywhere but at the end of his letter? On the other hand, it is a strain to understand Innogen's "So far" to mean that she has read aloud only this end-part of the letter and that "th'rest" is the preceding matter. Wayne prefers Posthumus to be telling Innogen to reflect on Iachimo to the extent that she values her trust, in the sense of her ability to give trust where it is deserved. That is, according to Wayne, Posthumus is not recommending Iachimo to Innogen by reference to the value she places on Posthumus ("your truest, Leonato") but by reference to the value she places on her own powers of judgement. That is an attractive critical argument.

    At 1.6.35-36, Folio Iachimo says that men have eyes that "can distinguish 'twixt | The firie Orbes aboue, and the twinn'd Stones | Vpon the number'd Beach", which Theobald emended to ". . . unnumbered beach". Wayne adopts this emendation. The problem is that it is not clear if Iachimo means that men can distinguish the similar-looking stars from one another and the similar-looking stones from one another, or that men can distinguish the stars from the stones. The former seems more likely since it is scarcely a fine distinction to tell stars from stones, although Iachimo may be deliberately speaking ambiguously. Sisson thought the latter was the point and defended "number'd" as an extension of the image: men can not only tell the stars from the stones but they can count (number) them. In King Lear, Edgar refers to "unnumbered . . . pebbles" on the beach (Scene 20/4.5), which lends weight to Theobald's emendation.

    Where in the Folio Iachimo looks at Innogen as he says "Had I . . . | . . . | . . . | This obiect, which | Takes prisoner the wild motion of mine eye, | Fiering it onely heere" (1.6.101-3) the Second Folio substitutes "Fixing" for "Fiering". Wayne adopts the Second Folio reading, but Sisson's defence of "Fiering" is possible. Shakespeare recurrently refers to eyeball movements as an index of emotional intensity, as in "Move these eyes? | Or whether, riding on the balls of mine, | Seem they in motion? (The Merchant of Venice, 3.2), "Rolling his greedy eye-balls in his head" (Lucrece 368), and "And make his eyeballs roll with wonted sight" (A Midsummer Night's Dream 3.2). These parallels make attractive Iachimo's image of the wild motion of his eyes being taken prisoner and hence fixed.

    Wayne also follows the Second Folio to print "I give him satisfaction?" (2.1.14) where the Folio has "I gaue . . .". Cloten is talking about a man who beat him at bowls and criticized him for his swearing, which caused Cloten to strike him, and Wayne rightly points out that Cloten says that the man was beneath him in rank. Thus the stronger reading is that Cloten pretends that he could not give the man satisfaction in the aristocratic-honour sense of fighting a duel when in fact he was scared to--he also goes on to say that men "dare not fight" with him--and hence the Second Folio is right.

    At 2.3.42-43, the Folio Cymbeline advises Cloten that Innogen will need time to forget about Posthumus: "some more time | Must weare the print of his remembrance on't". Wayne accepts the Second Folio's emendation of ". . . out", in which the print is worn away, but the Folio's reading is defensible: more time must bear (wear upon itself) his remembrance. Inexplicably, Wayne's note refers to "Theobald's emendation" when in fact Theobald simply accepted the Second Folio's reading as his predecessors did. Wayne retains the Folio reading so that the Queen advises her son: "Frame your selfe | To orderly solicity" (2.3.47-7). The problem is in that final noun, which editors often emend, usually adopting the Second Folio's "solicits". Wayne shows that although OED does not record it, "solicity" was established as a noun (meaning "solicitude") before Cymbeline was published; presumably she searched EEBO-TCP to discover this: she does not say.

    Wayne has Posthumus say of the British army that their discipline is "Now wing-led with their courages" (2.4.24), which is the Folio's reading, rather than following the Second Folio's "Now mingled . . ." as many other editors do. She notes the usual arguments on each side, including the implausibility of supposing that the Folio's compositor had trouble making out whether the minims represented "m" or "w" and settled on the latter to make a most unusual and incorrect reading. She also brings new evidence from EEBO-TCP about the use of "wing" in relation to an army, including Fynes Moryson on "the forlorne Hope and Wing led by Captaine Billings". The trouble is, Moryson here uses "wing led" as a noun and its verb, so two words, and not a compound verb as in Folio Cymbeline.

    Describing Innogen's body, Iachimo in the Folio says that "vnder her Breast | (Worthy her pressing) lyes a Mole" (2.4.134-5) and Wayne follows Rowe's emendation of the second "her" to "the". Otherwise, it sounds as if Innogen presses her own breast or the mole or else the breast (figured as female) presses the mole, when the point of Iachimo's description is that someone/something other than Innogen--by implication, Iachimo--does the pressing. Wayne glosses "cuckold" as "a man dishonoured by having an unfaithful wife in a culture that believed husbands should control their wives' sexuality" (2.4.146). I would have thought that qualification (". . . in a culture . . .") overstated, since the term does not require a norm of asymmetrical gender relations: egalitarians might also think that an unfaithful wife dishonours her husband and vice versa.

    Where the Folio Pisanio exclaims, upon reading the letter from Posthumus saying that Innogen is adulterous, "What Monsters her accuse?" (3.2.3), Wayne adopts Capell's emendation to "What monster's her accuser?" Wayne's logic is that two lines later Pisanio refers in the singular to some "false Italian" who must have belied her, but that does not require him to use the singular here. At 3.2.67, the Folio Innogen asks Pisanio "How many store of Miles may we well rid" in an hour, and Wayne adopts the Second Folio emendations of "store" to "score" and "rid" to "ride". The former is ratified by Pisanio's reply "One score 'twixt sun and sun". Although it is grammatically acceptable for her to ask "How many store of miles. . .?" and him to reply "One score", the joke about her optimistically thinking in terms of multiple scores per hour and his realistic answer of one score a day would thereby be lost. Wayne's defence of "rid" in the sense of "get through" is more strained.

    Reflecting on the country life as opposed to the court life, Belarius in the Folio says that it is "Richer, then doing nothing for a Babe" (3.3.23). Editors usually emend the last word, for example to Thomas Hanmer's "bribe", but Wayne sticks with the Folio reading. Her argument is that "babe" refers to negligently-cared-for wards of court, which makes a topical allusion (as discussed in her Introduction) and links to other examples of surrogate parenthood in the play. Against it I would argue that it is a decidedly odd thing for Belarius to say to Arviragus and Guiderius, since they do not know that they are being raised by a surrogate, and the audience is unlikely to catch the allusion unless they were especially attentive to what the First Gentleman said in the first scene and guess that these boys are really Cymbeline's.

    In the Folio, when Innogen tells Pisanio to get on with murdering her and then get some sleep, he replies "Ile wake mine eye-balles first" (3.4.101) which does not make much sense and does not form a metrical line with her response "Wherefore then". Wayne adopts Samuel Johnson's suggestion that Pisanio says ". . . out first", meaning that rather than kill her he would choose to stay awake until he lost his sight.

    Contemplating returning to the court, Innogen in the Folio says that she can have nothing to do "With that harsh, noble, simple nothing: | That Clotten" (3.4.132-3). The first line is metrically short and it is perhaps odd for her to call Cloten "noble" without qualifying it so editors sometimes add an adjective before "noble" that reshapes it to make "churlish noble" or "feeble noble" and regularizes the meter. Wayne retains the Folio reading, arguing that rough meter is acceptable given Innogen's anger and that the three adjectives in this line match the three negatives in the preceding line "No court, no father, nor no more ado". It might also be noted that metrically "With that harsh, noble, simple nothing" lacks only a final stressed syllable and that the omission might coincide with Innogen's struggling to find a suitable word for the nothingness of Cloten (so the metrical absence expresses his nothingness), which tongue-tiedness she breaks by just naming him "That Clotten".

    After his catalogue of Posthumus's faults, Cloten in the Folio refers to Innogen as "this imperseuerant Thing" (4.1.14) who nonetheless loves Posthumus. The problem is how to modernize "imperseuerant" and Wayne goes for "imperseverant" meaning "lacking in perseverance", which she acknowledges is untrue of the steadfast Innogen. Wayne rejects the modernization to "imperceiverant", meaning that Innogen cannot perceive Posthumus's faults that Cloten has just listed, on the grounds that this sense of "imperceiverant" was not available at the time. She writes that ". . . 'perceiverant' was a noun meaning 'perceiver' (OED 'perceiverant' n., one instance), not an adjective meaning 'perceptive'". True, but Cloten's sense does not require an adjective since a noun could stand in for an adjective in Shakespeare, or else the noun could be in apposition to "thing" so Cloten is saying that Innogen is both. The problem is not that Cloten is wrong in saying that Innogen lacks perseverance but rather that saying so does not fit his wider complaint that she sticks with such a man in preference to himself.

    After Arviragus has referred to Fidele's sighs and smiles, Guiderius says in the Folio that "greefe and patience rooted in them both, | Mingle their spurres together" (4.2.57-8). Wayne follows Pope's emendation of "them both" to "him both", arguing that "them" cannot refer back to the sighs and smiles because these are "manifestations of those emotions rather than grounds for them". She might also have mentioned that it would be hard for the audience to find the antecedents of the pronoun "them" by recalling the sighs and smiles from 30 words earlier. But her claim about the relationship between sighs and smiles on one hand and grief and patience on the other seems back-to-front: Arviragus can quite sensibly say that sighs and smiles (the manifestations) are "rooted" in their causes of grief and patience. Wayne seems to take "rooted" to mean "are caused by these origins" but it can just as easily mean "form the origins of".

    The Folio Belarius marvels at his two adopted sons, saying "Oh thou Goddesse, | Thou diuine Nature: thou thy selfe thou blazon'st | In these two Princely Boyes" (4.2.168-70), and editors have worried about these four occurrences of "thou" in two lines. Wayne adopts the Second Folio reading in which the "thou" after "Nature" is omitted, but does not explain how this is acceptable metrically; presumably she assumes a missing stressed syllable at the caesura and a feminine ending. I would have thought the Folio reading acceptable.

    At 4.2.261-62 Wayne's glossing note reports that "'Chimney sweepers' was a Midlands colloquialism for dandelion clocks, the grey heads of dandelions gone to seed . . . apparently because they resembled chimney sweepers' brooms". As Charles E. Nelson showed in an article reviewed in NYWES for 2016, there is no evidence for this claim and it may well be a twentieth-century invention.

    After the Folio's stage direction "Enter Lucius, Captaines, and a Soothsayer" its next lines are "Cap. To them, the Legions garrison'd in Gallia | After your will, haue crost the Sea" (4.2.331-3). Wayne thinks that "To them" cannot be what the Captain says because "the legions from Gallia arrived in Britain first and the forces from Rome are supplemental to them", as the dialogue elsewhere makes clear. "To them" must be a displaced part of the stage direction, and Wayne thinks that Lucius and the Captains enter and then the Soothsayer enters "to them", which as she points out is a common idiom on stage directions. Instead of "them" being Lucius and the Captains, as an alternative she acknowledges that "them" could be Innogen and Cloten who are already on stage. This she rejects, arguing that it is more likely to be the Soothsayer who enters "to" his fellow Romans "given the character's peripheral relation to the military figures and his silence for 13 lines following his entrance, which may place him at some distance from those on stage". The trouble with this suggestion is that when one person or party enters "to" another, they usually interact right away: this locution is not commonly used for persons who enter and then stay aloof.

    At 4.4.27, Wayne renders the Folio's "this heard life" as "this hard life". She derives "hard" from the Second Folio, marking it as an emendation, but of course "heard" is just an early modern spelling of "hard" so this is modernization not emendation. Wayne offers a gloss to the word "egregious" (as "outrageous, flagrant", 5.5.211) which I think is still in common usage, but perhaps her felt need to gloss reflects the particular market that modern editions must aim at. At 5.5.229, Wayne uses the substance of Rowe's stage direction to interpolate "He strikes her and she falls" but as Sisson pointed out this is not quite what the subsequent dialogue seems to suggest. Innogen's subsequent question to Posthumus "Why did you throw your wedded lady from you?" (5.5.260) suggests that he pushes her away so she falls but he does not strike her as such.

    Embracing Posthumus, Innogen in the Folio says to him "Thinke that you are vpon a Rocke, and now | Throw me againe" (5.5.261-2) and Wayne retains "rock" where other editors often emend to "lock", in the sense of a hold in wrestling, as suggested by Edward Dowden. The idea of the emendation is that Posthumus threw Innogen down and she now rises and embraces him, so that it looks somewhat as if they are wrestling and she invites him to throw her down again. Wayne finds the emendation "less appropriate" than the Folio's reading, which she glosses as "Imagine we are standing on a precipice: if you throw me again, we will both go down".

    Wayne's first appendix is a substantial account of "The Texts of Cymbeline" (pp. 378-401). First she revisits some of the textual choices detailed above, remarking that the existence of digital databases now shows us that words such as "solicity" that were formerly thought to have been created by textual corruption in Cymbeline are in fact witnessed in other early books. Wayne revisits the layout of the letters from Posthumus to Innogen in scenes 1.6 and 3.2, discussed above and in Wayne's essay "Gendered Text". She explains her decision--deriving from work by Maurice Hunt and George Walton Williams--to set the complaints to Jupiter in scene 5.5 as heptameters (true fourteeners) rather than as alternating tetrameters and trimeters (ballad meter), as most editors previously lined them following the Folio's intermittent practice. Until recently editors seem not to have accepted that Shakespeare did occasionally write in fourteeners.

    Next Wayne tackles the question of whether metrical irregularity should be used to diagnose textual corruption and the problem that the late plays are in any case quite metrically irregular. Although Wayne does not mention it, there is a danger of circular reasoning here, since some of what we think of as the characteristic metrical irregularity of the late plays might in fact be textual corruption: that is, the late plays might be more regular than we have thought. Paul Werstine argued this case in his essay "Line Division in Shakespeare's Dramatic Verse: An Editorial Problem" which Wayne includes in her bibliography but does not cite at this point. Wayne's policy is that it is "inappropriate to emend primarily on the basis of metre" (p. 381).

    Wayne reports that the only part of the Folio printed after Cymbeline was completed was "the first three pages of Troilus and Cressida" (p. 382) but in fact Charlton Hinman's reconstruction shows that almost all of Troilus and Cressida was printed (on gatherings ¶, ¶¶, and ¶¶¶) after Cymbeline. Wayne gets the mechanics of Folio printing wrong too, thinking that a gathering comprises six sheets when in fact it consists of three sheets. Wayne's account of the division of the typesetting among the Folio compositors represents where matter stood before Pervez Rizvi's recent article (reviewed in NYWES for 2016) threw doubt on most of the identifications.

    Because the pages attributed to Compositor E show more proof corrections than pages attributed to other compositors, Wayne thinks that this "means E's pages were carefully proofed" (p. 383). Actually, as D. F. McKenzie first pointed out in his 1969 essay "Printers of the Mind" and as Peter W. M. Blayney repeated in his introduction to the 1996 second edition of the Norton Facsimile of the First Folio, the amount of stop-press correction evident in the surviving sheets of the Folio does not indicate the overall amount of proofing since corrections made to the type before the print run started in earnest leave no evidence.

    Wayne lists errors in the play that she attributes to its compositors but it is not clear why she thinks they must all have been introduced during printing. Some of these errors, such as substituting "f" for long "s" and certain minim errors, could have existed in the manuscript that was printed to make the Folio, having got there by prior scribal error. Wayne surveys the growing scholarly consensus that the scribe Ralph Crane probably created the manuscript used to print Cymbeline in the Folio, and the reasons for this arising from the presence of his characteristic habits of punctuation and spelling in the printed book.

    In this light of this consensus, Wayne surveys some of the text's cruxes but these do not really make any difference to her choices. It is hard to say how far removed from Shakespeare's papers Crane's manuscript was, but Wayne thinks that the distinctly Shakespearian spelling "Iarmen" (where Shakespeare wrote "Iarman" for modern 'German') got through, as perhaps did "one" for "on". She accepts the (slight) evidence that Crane was copying from a manuscript in two hands (one preferring the spellings "O" and "Clotten" and the other preferring "Oh" and "Cloten"), or from a manuscript that was itself a copy of such a manuscript in two hands.

    Wayne adds to the evidence for the name of the play's heroine being Innogen not Imogen, pointing to a forthcoming essay of hers and mentioning here that the form "Gennogen" appears in two history books published in the late fifteenth century and has not been previously noticed (p. 395). The long-established evidence includes the wife of Leonato being called Innogen in Much Ado About Nothing and the name Innogen being in Holinshed's Chronicles, including an occurrence where the compositor set it as "Imogen" but its appearance within an alphabetized lists proves that his manuscript copy read "Innogen". There are also many other printed occurrences that show that Innogen not Imogen was the standard form of the name. And there is the evidence of Forman's eyewitness account of the play in performance with the name Innogen. Since Crane wrote clearly, it is unlikely that the Folio compositors repeatedly misread his "Innogen" as "Imogen": rather, the mistake was Crane's in misreading his copy, or it existed in that copy.

    There is some slight evidence that could point to authorial revision, such as interruptions from Cloten that when removed would leave the meter intact and Belarius's repetition "Hark, the game is roused" (3.3.98) and "The game is up" (3.3.107) between which he gives a standalone speech that fills the audience in on the backstory. And of course the dream sequence in 5.4 has long been suspected as an interpolation of someone else's writing. Wayne also suspects revision because the play is so strongly divided between the first two acts concerning Innogen and Posthumus and the last three concerning the break of Britain from Rome and her brothers, for which quite different sources are used. Maybe revision explains the evidence for two hands in the manuscript Crane copied.

    In an appendix on music (pp. 402-6), Wayne reproduces an early seventeenth-century setting for "Hark, hark, the lark" (possibly by Robert Johnson), one of the two sections of the play marked as "SONG" in the Folio. The other is "Fear no more the heat o' th' sun", which the dialogue insists was spoken and for which no early setting exists although contemporary tunes would fit it. Wayne then briefly considers music later written for the play, and operas and other musical spinoffs based on it. In the third appendix, on "Casting and Doubling" (pp. 407-12), Wayne finds that the final scene is the most populous, needing 19 actors of whom three play women. Doubling Cloten and Posthumus is attractive, and Wayne considers some other possibilities too, including those that make conceptual links between the characters, such as Posthumus's dead brothers in 5.4 being played by the actors of Guiderius and Arviragus who unbeknownst are his brothers-in-law. Wayne ends with what she calls a "Casting Chart" but is only a list of which characters are in which scenes.

    This year only one monograph was central to our concerns: Hugh Craig and Brett Greatley-Hirsch's Style, Computers, and Early Modern Drama: Beyond Authorship. The present reviewer gave detailed feedback on a draft of this book and provided an endorsement printed on its cover and is currently engaged in separate research projects with each of its authors, so this review will necessarily be circumspect and stick to reporting what is claimed by the authors. To the texts of 39 plays from across the theatre companies of the 1580s and 1590s, Greatley-Hirsch and Craig apply the techniques of computational stylistics that have had great success in distinguishing authorial style, genres, and chronology. Their aim is to divide the plays by means other than authorship, for example to see if there really was such a thing as playing-company style. Craig and Greatley-Hirsch's comprehensive study takes in the differences in dramatic style that emerge between prose and verse dialogue, the variety of stage properties mentioned in the drama, and the appearance of the drama if we break it into individual speeches and compare these across plays and across canons.

    Craig and Greatley-Hirsch find that when plays are wholly prose or verse there are no measurable stylistic consequences of these choices--each writer's prose plays test like their verse plays--but when an author substantially mixes the two forms in one play the consequences for word choice are apparent across that play. Regarding properties, there are signs that genre affects usage. Naturally, thrones go with histories, and hearses, coffins, tombs and beds go with tragedies, and tables, seats, and stools go with comedies. Naturally, horns, trumpets, and drums go with histories, and lutes with comedies. Daggers go with tragedies, halters with tragedies and comedies, while cudgels, clubs, rapiers, and pistols go with (especially city) comedies, and halberds and bows and shields go with histories. Most unexpectedly, babies go with histories and the authors explain this as a result of babies appearing in comedies that in fact test more like histories in their other property types.

    When Craig and Greatley-Hirsch compare characters from different plays with one another, some expected and some curious affinities emerge. In Julius Caesar, Cassius calls himself a mirror to Brutus and indeed their language is alike. A character who appears in multiple plays can be similar from play to play, as are the Tamburlaines in the two parts of Marlowe's Tamburlaine. But in other cases--the two Byrons of George Chapman's Byron plays, the two Falstaffs of 2 Henry 4 and The Merry Wives of Windsor--the multi-play selves are less alike. Being exaggeratedly Welsh, Fluellen in Henry 5 is a most exceptional character (especially in his high count for "is"), and Craig and Greatley-Hirsch find a pattern here. It is the monomaniacal types, such as the obsessed-with-war Fluellen, who have the most exceptional language.

    Most importantly for theatre history and literary criticism, Craig and Greatley-Hirsch find that there is no such thing as company style, at least not in the kinds of things that distinguish authorial style, genre style, and period style. What is detectable is that dramatic language changes over time. Craig and Greatley-Hirsch quantify "a marked and consistent change in play dialogue over forty years between 1585 and 1624" as ". . . early modern English tragedy becomes steadily more comic, and comedy more farcical", and as ". . . elaboration and picture-painting in speeches give way to categorical thinking, and more attention is paid to attitudes and interactions" (p. 162).

    No one writer is leading this change which, rather, occurs between authors as they come and go. In general, dramatists got better at realistic representation of human speech and thought. This period, the Shakespearian period, was a special time; results are different later on. Craig and Greatley-Hirsch find "continuity rather than rupture as a literary history connecting pre-1642 and 1660s English drama" (p. 223). Richard Brome was the forerunner of Restoration comedy and James Shirley the forerunner of Restoration tragicomedy. The new writers of the 1660s did not revert to an older style, such as Shakespeare's.

    Only three chapters in collections of essays are relevant to this review. In the first, Paul Werstine argues that we have no reason to believe that Shakespeare revised his plays and hence where early editions vary we should conflate them ('Authorial Revision in the Tragedies', in Neill & Schalkwyk, eds. The Oxford Handbook of Shakespearean Tragedy, pp. 301-15). Werstine traces Shakespeare's editors' belief in Shakespearian revision, especially in Romeo and Juliet, from Alexander Pope on to the twentieth century, including the complex theories of A. W. Pollard and John Dover Wilson.

    Werstine then turns to the theory of Memorial Reconstruction, now widely rejected. It is, according to Werstine, clear that the 1597 first quarto (Q1) of  Romeo and Juliet derives in some way from the play underlying the 1599 second quarto (Q2). He gives as an example of this dependence the fact that in Q2 Capulet gives a courteous farewell to the maskers whom in Q1 "he suddenly selfishly blames for keeping him from his bed by their not knowing when to leave" (p. 304). The supposedly "churlish" lines are "but for your company, | I would haue bin a bed an houre agoe". I cannot see why this has to be performed as churlishness since it may be a compliment: you have not bored me, he says, you are interesting enough to stay awake for.

    The quoted lines appear later, only in Q2, said by Capulet to Paris after explaining that Juliet has taken Tybalt's death hard, and again Werstine treats them as suppressed irritation ("Capulet is trying to get away from Paris after that persistent suitor has intruded", p. 305). Because these lines are, according to Werstine, wrongly placed in Q1 and rightly placed in Q2, Q1 must depend in some way on the script behind Q2. I would say that this is just as easily an example of lines transposed in revision, and that if they are complimentary then intentionally moving them from the maskers, about whom Capulet cares little, to Paris, whose favour Capulet courts, has a dramatic logic to it.

    Werstine reckons that the fact that Q1 Hamlet becomes verbally close to Q2/F with first the entrance of Valtemand and Cornelius and then again with the entrance of Marcellus and others (all in the second scene) shows Q1's dependence on the script(s) underling Q2 and F, at least at the beginning of the play. (It is odd to hear this argument coming from Werstine, since it is, as he acknowledges, the essence of the Memorial Reconstruction hypothesis and he was perhaps the most influential voice in the widespread rejection of that idea in the 1980s and 1990s.) Thus, argues Werstine, if the long versions of Romeo and Juliet and Hamlet were in existence before the making of the short versions represented in their first quartos, we cannot use authorial revision to explain the differences.

    Werstine turns next to the differences between the 1604-5 quarto and 1623 Folio Hamlet, the 1622 quarto and Folio Othello, and 1608 quarto and Folio King Lear, where in each case there are more than a hundred lines present in one version and absent in the other. He argues that these can be explained as cuts made in one or the other or both (that is, different cuts in the different versions) of the scripts underlying the different editions. This argument directly contradicts Werstine's demonstration in his 1988 essay "The Textual Mystery of Hamlet" that such an explanation will not work for Hamlet because if we conflate the two versions we get contradiction or doubling in the form of two competing explanations for one event in the story. Theatrical cutting accounts for whole passages absent from particular editions, according to Werstine, and to account for small verbal differences he blames the scribes that copied the underlying manuscripts. Werstine here cites in a footnote his 1988 essay that his present argument is contradicting, as if it were supporting what he now claims.

    Werstine returns to sketching the various editorial views on Shakespearian revision, lauding those who conflate even when they suspect revision. He thinks that the 1986-87 Oxford Complete Works project was misled by E. A. J. Honigmann's mistaken view, in The Stability of Shakespeare's Texts of 1965, that Shakespeare's first manuscript of each play was too messy to be used in the theatre so he had to copy it out fairly. Honigmann supposed that during this copying Shakespeare made small verbal changes and that the first manuscript went on to supply the printer's copy for one early printed edition and the fair copy went on to be the basis for another. Werstine considers that his own book Early Modern Playhouse Manuscripts (reviewed in YWES for 2013) showed that dramatists' own untidy manuscripts were in fact used in the theatre, and therefore that no copying out was necessary.

    I think it is unfair to suggest that Honigmann's argument was entirely dependent on the idea that authorial manuscripts were too untidy for use in the theatre, since Shakespeare may have copied his own manuscripts fairly even if they were strictly usable. And it is also unfair to suggest that the 1986-87 Oxford Complete Works claim for authorial revision rested largely on Honingmann's work: it gave other reasons for believing that revision happened. Werstine suggests that the stage-centered thinking that dominated 1970s and 1980s Shakespeare studies made the Oxford Shakespeare prefer the Folio versions of plays even where, as in Hamlet, these lack lines present in preceding quartos: it was not so much that the cuts were authorial as that the cuts were theatrical that persuaded them, Werstine argues. This is essentially true, but for the Oxford Shakespeare editors the fact that the revisions were theatrical itself implied authorial warrant for them, since Shakespeare was a senior member of the theatre company and its interests and his interests were aligned.

    Werstine observes that almost all the surviving play manuscripts that show signs of being used in the theatre have deletions marked in them, and hence it is reasonable to assume that the cuts evidenced in Folio Hamlet and King Lear and in Q1 Othello are theatrical not authorial. We have no evidence that authors would be involved in such cutting, he reckons. Again, the unique position of Shakespeare ought to be considered here: he was a sharer in the playing company and hence did not lose control of his scripts as other writers might have. The sample Werstine draws upon, fewer than two dozen manuscripts, need not be representative of the 2,000 plays that existed and may not reflect Shakespeare's unusual position as actor-writer-sharer in a playing company. Werstine ends by quoting the journalist Ron Rosenbaum on Richard Knowles's role in "calming the revision craze" (Werstine's words, p. 314), as will apparently be seen in his New Variorum Edition of King Lear.

    Immediately following Werstine's essay in the same collection, Michael Witmore, Jonathan Hope, and Michael Gleicher show that Shakespeare's tragedies and comedies can be distinguished using the counts of features made by their software, called DocuScope ('Digital Approaches to the Language of Shakespearean Tragedy', in Neill & Schalkwyk, eds. The Oxford Handbook of Shakespearean Tragedy, pp. 316-35). The freely available Phase One of EEBO-TCP gives us 554 early modern plays. Can we tell from these if there is a distinctive language of tragedy? And a distinctively Shakespearian language of tragedy? The investigators first removed all the stage directions and speech prefixes and anything else not spoken on the stage, then they used an automated spelling modernizer to normalize the spelling. They then counted the occurrences of 113 linguistic features and applied various statistical tools to process the counts and visualize them. In a footnote they explain the specific technical processes that they applied.

    The essay's first visualization plot shows the first and second principal components (PC1 versus PC2) for all the plays. The tragedies are not evenly dispersed among the plays: as a group they cluster in the top left corner, having low PC1 and high PC2. The comedies go the other way: clustering in the bottom right corner, with high PC1 and low PC2. Looking at which words are causing this distinction, the investigators find that ". . . tragedies (perhaps not surprisingly) favour a set of linguistic features used to communicate negative emotion and affect, while comedies, less predictably, favour a set associated with the representation of rapid, highly interactive speech, including first and second person pronouns, questions, discourse markers, words for social roles and relationships, and imperatives" (p. 320). Looking just at Chapman's tragedies and comedies, they show the north-west versus south-east distinction most clearly.

    In a highly quotable image for how quantitative methods differ from traditional literary scholarship, Hope, Witmore and Gleicher remark that "When we present you with quotations from plays, they are intended to be representative of what goes on across very large amounts of text: they are not 'plums' plucked from a literary pie, but are the stodge that makes up the vast majority of what is going on in the text" (pp. 312-2). They quote some parts of Chapman's works to illustrate the words that turn out to be distinctive of genre by this method. Witmore, Hope, and Gleicher find that another group of words about selfhood--involving terms featuring Direct Address, First Persons, Questions, Oral Cues, Person Properties, and Imperatives, Self Disclosure, Autobiography, and Metadiscourse, all of which they illustrate--are more common in comedy than tragedy. This is surprising since tragedy is traditionally thought to be the genre where the self gets defined and explored.

    Turning to Shakespeare, the sharp tragedies/comedies distinction visible in the full set of 554 plays and in Chapman's canon is not present and overall the tragedies cluster a bit more south-westerly than the norm: lower than the comedies on PC1 but also lower than the tragedies on PC2. The impression we get is that Shakespeare mixes genres more than others do. What is it in Shakespeare's tragedies that accounts for this? His tragedies have more of the words typical of comedies, but also they have unusually high counts for a group of words for describing the world and mapping spaces. His tragedies also avoid the words about the self that characterize others' tragedies. Between Shakespeare's comedies and tragedies there is little distinction, although his comedies do have more of the words about the self than his tragedies, and this distinguishes him from other writers. Having sketched their results, Witmore, Hope, and Gleicher describe in technical detail their method.

    Docuscope contains strings of characters classified into 113 rhetorical-linguistic human-crafted categories called Language Action Types (LATs), and when it finds an occurrence of one of these strings in a play the tally for the corresponding LAT is raised by one. Each string can belong to only one LAT and the method is greedy in the sense that where a longer string would match one LAT and a shorter component string within that longer string would match another LAT, the longer string is counted and the shorter one discarded. Witmore, Hope and Gleicher freely acknowledge that this is somewhat crude. The occurrences in the plays produce the data points that are fed into the Principal Component Analysis. In their conclusion, Witmore, Hope, and Gleicher make a point that others in this field have recently made, which is that empirical methods allow us to consider the typical cases--say, the average tragedies--where traditional literary methods have inherently priviliged the exceptional cases because they most stand out to the human reader.

    In our last chapter-in-book, Lynne Magnusson argues that the dedication of the 1609 edition of Sonnets is addressed to Shakespeare himself ('Thomas Thorpe's Shakespeare: 'The Only Begetter'', in Crawforth, Scott-Beaumann & Whitehead, eds. The Sonnets: The State of Play, pp. 33-54). For her edition of Sonnets in the Norton Complete Works Third Edition, Magnusson emended the title page dedication to read "TO. THE. ONLY. BEGETTER. OF. THESE. ENSUING. SONNETS. Mr. W.[S]. ALL. HAPPINESS", where the original has "W.H." for "W.[S]." In fact, Magnusson reports her emendation as "W.[S.]" but that cannot be right since the second full-stop is present in the original.

    Magnusson is convinced by Donald Foster's argument, made in 1987, for Shakespeare being the addressee, and she sketches its lineaments. What difference does it make if Shakespeare is the dedicatee rather than Henry Wriothesley (Earl of Southampton) or William Herbert (Earl of Pembroke)? It is hard to say, Magnusson observes, but major editions produced since Foster's article appeared have not accepted his claim. One obstacle is that Foster's argument requires that the phrase "ever-living poet" in the dedication means not Shakespeare--since he, in Foster's view, is mentioned earlier--but rather God.

    For Magnusson, the dedication is more complex and subtle than it might at first seem. Using all capitals it imitates monumental inscription, which is a theme in Sonnets. The phrase "ever-living", referring to God, was familiar to readers from the Book of Common Prayer. Also, "only begetter" was bound to have Biblical overtones, as in "only begotten son", which the Book of Common Prayer also emphasized; this makes the poet a kind of miniature God. The phrasing about a begetter also engages with the theme of procreation in Sonnets, showing that Thomas Thorpe read them carefully. Likewise the dedication's use of "only" to mean "preeminent", which echoes "only herald" in Sonnet 1, and also likewise the "these"/"that" of the dedication echoes the recurrent deixis of the poems, which Magnusson illustrates at length. In the dedication and the poems, the term "this"/"these" refers to the temporal world (even as it offers a kind of immortality by textual transmission) while "that" refers to the truly everlasting, the afterlife.

    But why dedicate a book to its author not a patron? Magnusson shows that the publisher Thorpe had a track record of playful, innovative paratexts. Why is the dedication just a greeting without the extended praise that usually follows one? Because it is trying to be neither deferential (as one must be to an aristocrat) nor overly familiar (as if to a close friend); the rhetorical mode is "sweet respect" (Sonnet 26.12). This mode was perhaps meant to mitigate the harm the publication would do in disclosing Shakespeare's love life to his acquaintances and family.

    This year 44 journal articles on our topic were published. We will start with those in Shakespeare Quarterly, taking first Sonia Massai's thoughts on how editions could be hyperlinked to useful additional resources ('Editing Shakespeare in Parts', SQ 68[2017] 56-79). According to Massai we now know, thanks mainly to Tiffany Stern's work, that plays were essentially fragmentary, made of multiple documents. "How differently might editors re-present Shakespeare", Massai asks, "if they were to acknowledge the primacy of the part over the whole?" (p. 59). The premise is contentious: some scholars (including the present reviewer) think that Stern's claims for the fragmentary nature of plays are substantially overstated. Massai accepts as true Richard Preiss's remarkable idea that early moderns went to the theatre mainly for entertainments other than the plays: ". . . the play was what interrupted [the day's offerings]: it was an afterthought, and the clown, the ringmaster who transcended it, was the main attraction" (Preiss quoted p. 60).

    Massai takes the play-within-the-play in Hamlet as evidence of how fragmentary early modern plays were, but one could just as well argue that Shakespeare is there showing something utterly unlike professional, London theatre: these are itinerant players performing before aristocrats. She also repeatedly refers to early moderns "hearing" a play, when in fact the evidence shows that like us almost all of them almost always called it "seeing" a play and that Shakespeare was an exception to this rule. Massai repeats Stern's mistaken claim that "Sometimes parts were copied when the play was still being written, in order to maximize the time actors had to memorize their parts . . ." (p. 62). This claim arose from Stern's misreading of an entry in the office book of Henry Herbert, the Master of the Revels.

    In general, by overstating the effect of all the other bits that might get included in the printing of a play (prologues, epilogues, songs, and so on), Massai makes the claim that the act of publication "divided the text of the play from itself and connected it to other texts and textual networks" (p. 64). I would agree with the latter but not the former. What follows are Massai's reflections on how everything connected with early modern drama is in parts.

    So what does any of this mean for editing Shakespeare's works? Massai thinks that the "electronic medium . . . also privileges the networking of parts over self-contained and sealed-off units of information" (p. 67). This seems to be a revival of the claim often heard in the 1980s that the digital medium (and especially hyperlinking) would thoroughly transform the nature of writing, which claim turned out to be untrue: the codex form endures and computers just help us disseminate and search codices more conveniently than we could before.

    Massai fantasizes about what a digital edition built along the notion of "parts" would look like, and the picture is fuzzy. She imagines all the opening sequences (incipits) of plays being tagged together so that they can be compared, using the beginnings of King John, King Lear, and Cymbeline as her examples. It is not clear why this would be any more useful than just a glossarial note in one play advising the reader to compare the opening to those of other named plays.

    Massai thinks that tagging could also provide links to other available databases that provide interesting sidelights on the text, such as which bits got written into their commonplace books by which known readers. This has been claimed as a desideratum for decades, but there is a mountain of work needed to make such links stay live for any substantial period of time. For example, there has been a mass migration of websites from using the unencrypted HTTP protocol to using the encrypted version HTTPS. Whose job would it be to update all the affected links in the kind of edition Massai envisages?

    Illustrating this point is the fact that this very essay's deep links into online databases that Massai uses to make her argument are already (two years after publication) broken, for example because the Database of Dramatic Extracts (DEx) has moved servers. Massai goes on thinking up other useful links that could be made between plays, such as to all the plays that belonged to one company or to their entries in the Database of Early English Playbooks (DEEP) or the Lost Plays Database.

    The remaining three articles from Shakespeare Quarterly need little attention. Paul D. Cannan gives an account of how Edmond Malone treated The Passionate Pilgrim in his Supplement to the Edition of Shakespeare's Plays Published in 1778 (itself published in 1780) and his The Plays and Poems of William Shakespeare published in 1790 ('Edmond Malone, The Passionate Pilgrim, and the Fiction of Shakespearean Authorship', SQ 68[2017] 139-71). Malone added to the collection the song "Take, oh, take those lips away" using the full two stanzas from Rollo, Duke of Normandy and this implicit validation contributed to a longstanding reluctance to challenge the attribution to Shakespeare of the first stanza of this song as it appears in Measure for Measure. Malone also appended "Let the bird of loudest lay" to his edition of The Passionate Pilgrim.

    Like Massai, but with a firmer foundation in the practicalities of digital methods, Alan Galey and Rebecca Niles explore the notion of books-as-interfaces--who first used which kind of notes, who first gave each line a number, and so on--and how this continues in the digital world. They write that "If post-New Bibliographical editorial theory and practice mean letting go of hypothetically reconstructed Shakespeare manuscripts, then we must learn to live with the fundamental printedness of Shakespeare texts . . ." (p. 22).

    I am unclear what this "letting go" is meant to consist of, since when editing a play for which we believe the printer's copy was a manuscript, or was a prior printed edition annotated by reference to a manuscript, we have to take into account the particularities of manuscript transmission, such as the kinds of mistakes that are easy and those that are hard to make in handwriting and the reading of it. Lastly, the late David Bevington offers some reflections on how we should gloss 400-year old language, making the undeniable case that editors should try to explain the range of possibilities without being overly prescriptive ('Confessions of an Annotation-note Writer', SQ 68[2017] 7-20).

    This review does not normally notice other reviews, but Peter W. M. Blayney's response to Brian Vickers's recent book on King Lear conveys substantial new knowledge about the early editions, including the fact that the indenting of the composing stick for the 1608 (Q1) edition was achieved using quotation quads normally reserved for making columns to contain marginal notes ('"Quadrat Demonstrandum": Review of Brian Vickers The One "King Lear" (Cambridge MA: Harvard University Press, 2016)', PBSA 111[2017] 61-101). The big editorial question with King Lear is whether to conflate the Q1 text with the substantially different 1623 Folio (F) text or to treat these as two distinct versions of the play.

    Vickers claims that conflation was the common practice until the 1980s, but Blayney points out that in fact the first completely conflated edition did not appear until 1770, from Charles Jennens. Blayney suspects that it is difficult for anyone who grew up with the conflated King Lear, as Vickers did, to read either Q1 or F as anything but an incomplete text. The landmark collection of essays called The Division of the Kingdoms, published in 1983, largely assumed that only Shakespeare could have revised the play, but Blayney's view is that someone else adapted it to make the version underlying the Folio, which he considers markedly inferior to the quarto.

    Blayney surveys Michael Warren's argument that F makes Albany less ineffectual than he is in Q1 and disagrees: in F he is given two additional short speeches trying to placate Lear and both are ignored by Lear. (At this point Blayney silently assumes that the line "Alb. Cor. Deare Sir forbeare." in the first scene is spoken by Albany and  Cornwall but in fact it could be Albany and Cordelia, as some editions have it.) Likewise, F gives Goneril a long self-justificatory speech to Albany in 1.4 that Q1 lacks, which again (according to Blayney) makes her seem stronger than him. The reassignment of Albany's lines to Edgar in the final scene--once to hasten the rescue of Lear and Cordelia and once at the end of the play--diminishes Albany in F. All this does indeed look to Blayney like conscious artistic reshaping of Albany's character.

    Blayney finds untrue Vickers's claim that "If you were to complete either version by adding the passages preserved by the other, you would have, in terms of characters and events, two identical plays" (Vickers's p. ix), since this claim takes no account of the artistic effect of the reassignment of lines from one character to another. There are "significant differences that are neither omissions nor additions" (Blayney's p. 66). I would say that Vickers's claim is in any case meaningless since it relies on an unstated threshold for the notion of textual difference and on an unstated size of passage: if for every difference in wording you used both words rather than one or other, the result would be gibberish. Vickers's explanation for the existence of the Folio-only lines is that they were already in existence when Q1 was printed, but were omitted from Q1 by the printer Nicholas Okes. This argument Blayney comprehensively demolishes, first by highlighting a long list of Vickers's elementary mistakes regarding bibliographical terminology, printing practice, and publishing history (pp. 66-70).

    The most important error is that Vickers thinks that a printer was always paid by the publisher for the paper, but in fact the publisher could, and in the case of Q1 King Lear Blayney thinks he did, provide the paper himself. Vickers thinks that what he takes to be the regular absence of space after punctuation in Q1 King Lear is significant, but it is typical. Likewise unspaced words: often there is in fact a hair space though hard to see. Vickers's misunderstanding on this point arises from misunderstanding what Blayney elsewhere meant by "space-metal", which was not the ordinary inter-word spaces--called thick, mid, thin, and hair in modern founts--but quadrats placed beside the text to span multiple rows of type.

     Blayney admits that he claimed in his 1982 book on the quarto that the copy for Q1 King Lear was cast-off to determine how many sheets it would fill; this mistake has "unfortunately become the core of Vickers's theory" (p. 71), which is likewise wrong. Blayney realized as his 1982 book was being finished that in fact certainly from sheet C and almost certainly from the beginning, Q1 King Lear was set seriatim, most likely because the copy was hard to cast off. Thus Okes could not have known how many sheets it would fill.

    Okes could not have intended to end Q1 on L4r and did so only because he was short of pica quads. Blayney details Vickers's excursus into the casting off of the Folio copy, which is not relevant to his argument about Q1 where the number of lines skipped over in order to start typesetting pages out of reading sequence is much shorter: dozens in a quarto rather than hundreds as in a folio. Vickers wrongly believes that play quartos were typically 8-9 sheets, and hence that Okes was trying to keep to about that size. Vickers's first example of what he calls space-saving, the use of some barred vowels on C4r, does not save any space at all since the speech is a prose paragraph and has plenty of room on its last line.

    Vickers's example from Edgar's speech about dogs on G4r--verse set as prose--wrongly attributes the use of barred vowels and commas not followed by spaces to the saving of space, but again it is a prose paragraph (although its contents are verse) and the last line is not full so even without those space-saving measures the speech would still have fitted into the same seven type lines. (This point stands even if one disagrees with Blayney, as I do, about how much of Edgar's speech is verse and how much prose.)

    Next comes Blayney's discussion of the indenting of the compositing stick first raised by D. F. McKenzie in 1973, in which Blayney finds substantial errors of fact. The difference between the narrow measure and the wide measure in Q1 King Lear is not exactly three ems, as McKenzie claimed. We want to know how wide the narrow measure was in terms of the type dimensions, and McKenzie found that the width of the narrow measure as measured directly off the page with a ruler, 80-81 mm, is also exactly the vertical height of 20 lines of type measured directly off the page, so (since an em is square) this vertical dimension tells us that the narrow measure was 20 ems. By the same calculation, the fact that 23 lines of type vertically equals 93-94mm gives the width of the wide measure as 23 ems.

    Blayney has measured the same printed width in Q1 King Lear and comes up with 82.5 mm as the narrow meausure as taken off the paper. The paper will have expanded by 1% horizontally and 2% vertically when wetted, according to Blayney. Presumably he means that it shrank by this much after it was printed, since it was printed wet and then it dried. This means that the type that impressed an inked image of a given size on the paper while the paper was in its wet-and-expanded state would have been bigger than the size that this image now occupies on the paper because the paper has, since being impressed, dried and contracted.

    Blayney believes that the paper expanded by 2% when wet and that it then contracted by 2% when it dried, but I find this rather too precise an assumption. Measuring the impressions left by the same page of type on different copies of Q2 Hamlet, I found that the impressions in one exemplar (at the Huntington Library) were consistently wider than those in other copies (by about 0.5%) so we should not assume too much consistency and exact reversibility in the reactions of paper to water.

    Moreover, Blayney gets his calculation of the effect of shrinkage wrong. Since the impression on the printed page is, supposedly, 2% smaller than the type that made the impression (since the printed page has shrunk by 2% since it was impressed), the correct calculation is that the type that left an image 82.5mm across must have been 82.5 ÷ 0.98 millimeters across, which is 84.18mm. But Blayney instead uses the calculation 82.5 × 1.02, which is 84.15mm. This difference is trivial because the percentages in this case are small but the fact that Blayney uses the wrong calculation is disturbing.

    So much for the vertical dimension. In the horizontal dimension the shrinkage of the paper was, according to Blayney, about 1%. So, if 84.15mm is the true size of 20ems in this type then type that was 84.15mm wide would leave an impression that is 84.15 × 0.99 millimeters wide, which is 83.3mm. Again, Blayney uses the wrong calculation (84.15 ÷ 1.01) and gets the slightly different width of 83.32mm as what we should find on the page, when in fact McKenzie found a width across the page of the narrow measure as 80-81mm. The discepancy is not as bad as Blayney makes it seem: the expected 83.3mm width of the narrow measure as found on the page is less than a millimeter above Blayney's own measurement of the narrow measure on the page as 82.5mm. It seems to me that to expect closer tolerances than this in calculations about supposed expansion and shrinkage of paper is unrealistic.

    Blayney explains McKenzie's hypothesis that quads from a different size of type were used to give a multi-line right-edge indenting of the stick, which is the key to saving quads that are the same size as the type being set. Blayney objects that McKenzie was being anachronistic: the heights of different sizes of type had not yet been standardized, and finding quads from a different size of type that were an exact multiple of the height one was setting in (to give exactly two, three, or four line indents at once) would have been virtually impossible. In fact, reckons Blayney, the indenting of the stick for Q1 King Lear was done not with oversized quads from a different size of type, but rather by using "quotation quads" that were made specially to support marginal notes used in some books.

    Blayney reports that the standard size for these quotation quads was 12.5mm by "a little short of" (p. 78) 18mm and notices that the latter figure is the former multiplied by the square root of 2. (Blayney does not define "a little short of" but 12.5mm times the square root of 2 is 17.68mm, or about one-third of a millimeter short of 18mm.). These proportions involving the square root of 2 are used in paper sizes so that folding a sheet, or joining two sheets together, does not change the sheet's shape, but Blayney cannot see (and nor can I) why these proportions should govern metal spaces.

    Although there was no intention to set a column for marginal notes in Q1 King Lear, on three occasions (on C2v, D3v, and F3r) it was done during imposition to accommodate stage directions that had been omitted in the first setting. Blayney thinks that a printer who went to such lengths to avoid losing three small directions would not drastically cut lines for expediency as Vickers believes Okes did. Actually, we do not know that these stage directions were originally omitted and then added later: the first two instances are entrances by one character in a middle of a long prose speech by another character and Okes might genuinely have not known how to set such a thing and been trying out the marginal setting as an experiment.

    Okes often did use quotation quads (usually reserved for setting marginal notes) inside the margin to indent the stick, according to Blayney. Unfortunately, in their heights these were not any integer multiple of the line-height in use, and Blayney describes the dodges that could be employed to make up for this discrepancy. Blayney lists the visual evidence in certain examplars that quotation quads were indeed used in Q1 King Lear in this way, and points to evidence on I1v, visible in all exemplars, of the lines of type being bent upwards and downwards around the inserted quotation quads because these did not exactly match multiples of the line-height in use.

    Returning to Vickers's account of space-saving in Q1, Blayney finds his bibliographical narrative amateurish: Vickers never explains why the compositors reverted to a narrow measure if, as he believes, they were forever trying to cram in as much text as they could. And nowhere does Vickers manage to substantiate his claim that Okes ordered too little paper for the job. Elsewhere, as Blayney notes, Vickers is mistaken about the rearrangement of type on the stone which, as Blayney points out, was confined to at most the moving of whole lines from the bottom of one page to the top of another, and not the kind of within-line alteration that Vickers (misled by William Proctor Williams) imagines happening. Blayney acknowledges that in 1982 he suggested that Okes was short of paper, but he now thinks he was wrong about that.

    Blayney points out that the first dialogue page of Q1, B1r, wastes rather than saves space. The next page uses a turn-down, but it saves no type-lines. What followed across sheet B was almost entirely narrow verse lines and even with quotation quads used to indent the stick, Okes must have been severely short of ordinary quads. On sheets C and D, Okes saved many quads by setting verse as prose, but even if he had not done so the book would have run to only one extra sheet (ending on M rather than L) and if Okes had kept up the rate of saving he achieved on sheets C and D he could have got the whole thing into nine sheets rather than the 11 he used. Therefore, Blayney assumes that Okes's motivation was saving quads not paper. Vickers's idea that Okes started radically cutting the text--removing what we now know as the F-only lines--is in fact absurd: printers just did not do that.

    Blayney finds implausible Vickers's account of the omission in Q1 of F-only words and lines, as many of the supposed cuts do not save type lines, which were the ones that mattered. Blayney agrees with Vickers that Madeleine Doran was right to suggest that the copy for Q1 was the author's own papers, which the playing company did not need because these had been copied to make the promptbook. Because the revision of the version underlying Q1 to make the version underlying F was in Blayney's words "both radical and damaging" (p. 95) to the play, Blayney cannot believe that Shakespeare was responsible. Blayney agrees with Vickers that the copy for F was the manuscript promptbook of the 1620s performances but disagrees with him (and Doran and René Weis) that even if the promptbook were recopied between 1606 and the 1620s its readings would have been largely preserved intact. Rather, Blayney believes that over those 15 years the actors and company writers would inevitably have altered the play.

    Certainly, Q1 and Q2 had some influence on F even though F was (in Blayney's view) printed from a manuscript, but just how this occurred is unknown. Blayney thinks that Folio Compositor E was given Q2 as a crib to help him make sense of his manuscript copy. But the presence in F of miscorrections of Q1 readings--and, I would say, the presence in F of press variants found only in Q1--means that the manuscript underlying F cannot have been created before the printing of Q1 in 1608. Blayney considers a series of cases in which Q's reading has traditionally been taken as a garbling of the true reading found in F, and argues in each case that Q's reading makes sense on its own, or else is (by a graphically plausible argument) a misprinting in Q1 of a good reading that is different from F's reading at that point. These are not simply garblings in Q1 but attempts in F to correct a garbled reading inherited from Q1, which is why in retrospect Q1 reads like a garbling of F. To Blayney it seems that ". . . somewhere in the chain of transmission from Q to F, someone tried to emend the Quarto readings without having properly understood what was wrong with them" (p. 98).

    Blayney finds what he considers a clear Q1 error in Kent's being given four days to prepare for exile, being made to leave on the fifth day, and threatened with death if found in the kingdom on the tenth day rather than the sixth. (I would say that the extra five days--day 6 to day 10--are allowed for Kent to cross the kingdom on his journey out of it.) Blayney's explanation is that the copy did not read "tenth day" but "next day" and "next" was misread by the compositor as "xth". Since the Folio also gives Kent until the tenth day, it must have got this error from Q1, since two independent misreadings of the correct word "next" is implausible, so again the Folio depends on Q1 rather than being wholly independent. Thus Vickers must be wrong that ". . . the Folio’s copy was either the 1606 promptbook itself or a near-perfect clone" (p. 100).

    Shakespeare Survey published four articles relevant to this review. John Jowett argues that the surviving theatrical manuscripts that were used by Werstine to overturn a key New Bibliographical assumption--that authorial papers served as the printer's copy for early editions of Shakespeare--are not themselves good evidence for what the printer's copy looked like ('Exit Manuscripts: The Archive of Theatre and the Archive of Print', ShSurv 70[2017] 113-22). Theatrical manuscripts are often not so much snapshots in a text's life as the embodiments of complex changes over time. Jowett illustrates this from the manuscript of Sir Thomas More and observes that it is also true of Thomas Middleton's The Lady's Tragedy and the Thomas of Woodstock manuscript. Thus ". . . the variants between an early and a late manifestation of a play are much more likely to arise within a manuscript than they are to arise between manuscripts when the source manuscript is copied" (p. 115).

    Paper was expensive, so recopying was not undertaken when overwriting an existing manuscript would do just as well. The majority of the surviving theatrical manuscripts are transcripts of lost originals and Jowett reckons that ". . . transcription can safely be taken to imply prior foulness . . ." but he accepts Werstine's point that a document only became foul papers by virtue of its being transcribed. As a neutral term for the lost antecedents of the transcripts, Jowett prefers "author's papers" (or "author's draft"), some of which may not have had to be copied to make a document useable in the theatre.

    Because foulness comes from the act of copying, any document (not just the author's papers) may become foul, for instance a document used for theatrical adaptation that thereby becomes messy. Werstine is right that theatrical documents could be messy so the New Bibliographers were wrong to simply assume that inconsistencies in early printed texts were a sign that the printer's copy was authorial papers not theatrical documents, but Jowett observes that "most of these inconsistencies and ambiguities" found in surviving theatrical manuscripts "are features of the initial authorial script" (p. 117).

    In the case of Sir Thomas More it is likely that at least for Addition V a fresh transcript (now lost) was in fact made. Noticeably, Hand C (the theatrical functionary) went to some trouble to ensure that although the document was becoming messy from revisions, the text of the play it contained "was properly joined up" (p. 117), and it is also noticeable that the parts of the manuscript that are transcripts are "more orderly" (p. 117) than the parts that seem to be uncopied drafts. That is, authors are messier in composition than people trying to put together a theatrical document. For instance, the Hand D part by Shakespeare is unusually vague and variable in speech prefixes, and so is Heywood's Addition IV. In fact, Sir Thomas More is more complex than is supposed by either the New Bibliographical model (of texts getting increasingly consistent on their way to the theatre) or Werstine's alternative model (that a playhouse manuscript is always annotated), since the manuscript appears to have been unannotated at the point of submission to the Master of the Revels.

    What does all this mean for our interpretation of printed books? What if the manuscript of Sir Thomas More had been the basis for a print edition? Jowett points out that Fredson Bowers overstated his claim that printing a manuscript generally destroyed it, noting that J. K. Moore in his book Primary Materials Relating to Copy and Print in English Books of the Sixteenth and Seventeenth Centuries in 1992 showed manuscripts that had survived being used for printing. Nonetheless, Jowett observes, the trip to a printshop was generally one-way for a play manuscript: it did not go back to the playing company that had sold it. The extant theatrical manuscripts are, then, not ones used for printing books but ones that survived in theatre archives, and we should be cautious in inferring from those the characteristics of manuscripts used to print plays. Moreover, unless the company was willing to stop performing the play, what went to the printer was not a theatrical document they relied upon for performance.

    The New Bibliographers assumed that the authorial papers were the spare copy sent to the printer, but Jowett points out that Q2 Hamlet has at least one repeated stage direction that looks like the work of a prompter. The assumption that good quartos were printed from authorial papers rather than theatrical ones is insecure since, as Werstine shows, we cannot assume that the good quartos' inconsistencies prove that their underlying manuscripts were unperformable. Jowett concludes that the evidence of the author's hand in the good Shakespearian quartos is weaker than previously thought and that the hypothesis that these good quartos were printed from non-authorial transcripts solves a number of problems.

    Gary Taylor, John V. Nance and Keegan Cooper show that Kyd did not write the Mariner speech from Edward 3 (4.137-184) and that we have no reason to agree with Brian Vickers's claim that he wrote the rest of the play ('Shakespeare and Who? Aeschylus, Edward III and Thomas Kyd', ShSurv 70[2017] 146-53). They start with a critique of Vickers's method of attribution and describe their own approach, which allows all candidates an equal chance to come out as the author. They take a passage from Edward 3 that Vickers has insisted is by Kyd, the Mariner's speech at 4.137-184, and look within it for 2-grams, 3-grams, and 4-grams and collocations spanning up to 10 words and count how often these also appear in LION and EEBO-TCP.

    As a validation step Taylor, Nance, and Cooper check if their method will correctly attribute a portion of the play Cornelia (widely believed to be by Kyd and attributed to him in his lifetime), using the first 173 words from the messenger's speech at 5.47-79. This sample size of 173 words is one they have used in previous micro-attribution studies of this kind. They confine their searching elsewhere to plays from 1585 (the earliest date for The Spanish Tragedy) to 1594 (the year of Kyd's death), and in that set they find 24 n-grams or collocations unique to just this part of Cornelia and one other play. Five of those 24 are links to Kyd plays, and Peele comes second with three links, and then come Lodge, Shakespeare, and Marlowe.

    For the Edward 3 test the authors take various 173-word samples from the Mariner's speech at 4.137-184 and whichever way they slice it Marlowe has the most links, closely followed by Shakespeare. Some of the Marlowe links are within the parts of the putative Marlowe canon that not everyone agrees about, such as his contributions to 1 Henry 6 and 2 Henry 6, but even excluding the contentious parts does not change the overall picture. Importantly, the Marlowe links (19 certain, 11 probable) and the Shakespeare links (16 certain) to this part of Edward 3 greatly outnumber the Kyd links (9), which number is roughly as many as Peele or Greene have.

    Looking at particular plays that have a lot of links, we can determine that Shakespeare's Richard 3 on its own has as many links as all Kyd's plays, and each part of Marlowe's Tamburlaine has twice as many links as any play by Kyd, Peele, Nashe, or Greene. "Like the total number of parallels overall, the evidence of parallels clustered in a single play strongly points to either Shakespeare or Marlowe as author of the Mariner episode" (p. 149) in Edward 3. When the investigators exclude links that appear in non-dramatic works in the relevant decade, Marlowe moves further ahead of other candidates. When they include links that appear in more than one play in the decade but require that the plays are by the same author (that is, the links have to be unique to an authorial canon), then Marlowe again stands out even further. Lastly, Taylor, Nance, and Cooper look at links with the non-dramatic works written by dramatists in this decade, and this raises the count for Lodge.

    Considering links to the non-dramatic canon reveals that the collocation "were . . . to give them way" from this part of Edward 3 appears also (uniquely) in Petruccio Ubaldini's pamphlet about the defeat of the Spanish Armada, which in the same place uses the phrase "enforced to giue them waie", which appears in the Mariner's speech in Edward 3 as "perforce were fain to give them way". Thus Ulbadini's 1590 pamphlet is confirmed as a likely source for this part of Edward 3. Vickers argued that the Senecan aspects of the Mariner's speech point to Kyd, but Martin Wiggins has shown that the influence of Seneca on plays from this decade was widespread.

    Taylor, Nance, and Cooper think the author of the Mariner's speech in Edward 3 was influenced by Aeschylus's The Persians, which not only has a similar messenger speech but also is about the king Xerxes (mentioned in Edward 3 just before the Mariner's speech) and mentions a fleet of 1,000 ships, which is also the number given to the French fleet (contrary to the historical sources) in Edward 3. Taylor, Nance, and Cooper find other more qualitative and circumstantial parallels between the scene in Aeschylus and the scene in Edward 3. Indebtedness to Aeschylus's Greek play, unavailable in English or Latin at the time, would rule out Kyd or Shakespeare as author of the Mariner speech. Thus Kyd did not write the Mariner speech in Edward 3, on which rests much of Vickers's claim for Kyd's authorship of the whole play. We do not know who wrote the play, but Marlowe is a strong contender.

    Hugh Craig finds that authorship is the category that most affects style when we compare early modern plays, but genre and date are also significant ('Authorial Attribution and Shakespearean Variety: Genre, Form and Chronology', ShSurv 70[2017] 154-64). Critics often comment on the great variation in Shakespeare's writing style from work to work, which if true should make it harder to do authorship attribution investigations. But in compensation, statistical methods are well developed for telling us how much variety there actually is in a dataset, in this case the variety of word use. Moreover, quantitative methods are relatively immune to the biases we find in qualitative methods.

    Joseph Rudman, in an essay reviewed in NYWES for 2016, claimed that genre trumps authorship: the variation between one author's work in different genres is greater than the variation between two authors in one genre. Craig sets out to see if this is true. He takes Shakespeare's eight sole-authored well-attributed tragedies (Antony and Cleopatra, Coriolanus, Hamlet, Julius Caesar, King Lear, Othello, Romeo and Juliet, and Troilus and Cressida) and a random sample of eight plays from Shakespeare's twelve sole-authored well-attributed comedies (All's Well that Ends Well, As You Like It, The Comedy of Errors, Love's Labour's Lost, The Merchant of Venice, The Merry Wives of Windsor, A Midsummer Night's Dream, Much Ado About Nothing, The Taming of the Shrew, The Tempest, Twelfth Night, and The Two Gentlemen of Verona).

    For comparison with the Shakespeare tragedy and comedy sets, Craig selects the set of eight Chapman comedies (All Fools, The Blind Beggar of Alexandria, The Gentleman Usher, A Humorous Day's Mirth, May Day, Monsieur d’Olive, Sir Giles Goosecap and The Widow's Tears), the set of eight John Fletcher comedies (The Captain, The Chances, Monsieur Thomas, The Pilgrim, Rule a Wife and Have a Wife, The Wild Goose Chase, Wit without Money and The Woman’s Prize), the set of eight Jonson comedies (The Alchemist, The Case is Altered, Cynthia’s Revels, Epicoene, Every Man in his Humour, Every Man out of his Humour, The New Inn and A Tale of a Tub), and the set of eight Middleton comedies (A Mad World My Masters, More Dissemblers Beside Women, The Nice Valour, No Wit No Help Like a Woman’s, A Trick to Catch the Old One, The Phoenix, The Widow and Your Five Gallants).

    These sets enable Craig to undertake exactly the comparison set up by Rudman and ask "Are Shakespeare's comedies and tragedies more different than Shakespeare's comedies and others' comedies?" He uses a set of 100 very common words and counts their frequency of occurrence in each eight-play group, using Welch's t-test to see if the resulting differences are statistically significant. Craig lists his 100 words, which distinguish between parts of speech so that "no" as an adjective is counted separately from "no" as an exclamation. These 100 words account for over half Craig's entire corpus.

    It turns out that the words "a" and "I" are in fact consistently used more often in Shakespeare's comedies than his tragedies and "for" (as a preposition) is consistently used more often in Shakespeare's tragedies than his comedies. But when we compare the Shakespeare comedies to i) the Shakespeare tragedies, ii) the Chapman comedies, iii) the Jonson comedies, iv) the Middleton comedies, and v) the Fletcher comedies the differences (regarding the frequencies of these 100 words) are least for (i) and rise steadily for (ii) through (v). The measure is how many of the 100 words are used significantly differently in the two sets being compared.

    Next Craig takes just the speeches of eight characters from each set and reruns the test, getting the same result. Indeed, the pattern of increasing difference is identical: the Shakespeare comedy characters are most like the Shakespeare tragedy characters, and then (in order of increasing unlikeness) come the Chapman comedies, the Jonson comedies, the Middleton comedies and the Fletcher comedies. (It is interesting that in these tests Chapman's style is the one most similar to Shakespeare's; the present reviewer's article on Word Adjacency Network's, briefly mentioned in NYWES for 2016, found the same thing.)

    To move to larger datasets, Craig constructs a 14-play, multi-author collection of comedies, and another of tragedies, and another of histories. (All the familiar names are represented, including Chapman, Field, Dekker and Henry Porter for comedy, Chettle, Marlowe, Marston, and John Webster for tragedy, and John Ford, Greene, Peele, and Anonymous for history.) Do these three genre sets show any differences from one another? Using the same test as before--asking how many of the 100 words are used significantly differently in the two sets being compared--Craig finds that history versus tragedy has the smallest difference and comedy versus history the largest, with comedy versus tragedy in the middle.

    Craig repeats the process using sets of 14 plays of varying genres for the authors Shakespeare, Jonson, Fletcher, and Middleton and compares pairs of sets. He finds that the Shakespeare-Jonson comparison produces a much larger number of words used at a significantly different rate than the genre comparisons did. (He puts them together on the same histogram to make this point.) And the results of the Jonson-Middleton and Shakespeare-Middleton comparisons are greater still.

    Next Craig tries sets of 14 plays of one author and one genre only--or at least the minimally hybrid sets of 12 comedies plus two tragicomedies (which he calls the "comedy-plus" genre)--and finds that this produces far fewer differences. "Here authorship does not trump genre", he writes, since "Shakespeare comedy-plus is close to a general comedy-plus grouping of his time" (p. 161). Craig tries making up sets of mixed authors and genre but all from the same decade. He finds that adjacent decades are fairly alike but across wider gaps the differences are noticeable.

    What about the difference between verse and prose? Craig describes tests that are detailed in his book with Brett Greatley-Hirsch (reviewed above), and the upshot is that when looking at whole plays no significant differences emerge, but when looking at plays that mix verse and prose significant stylistic differences emerge between the verse and prose parts. "It seems that when playwrights are working in prose and verse in the same play, they specialise--clowns and the middling sort in prose, for instance, and kings and lovers in verse" (p. 163).

    Craig's conclusion is that authorship is the dominant factor in style, but that the other factors are not trivial: ". . . genre, date, and mode also are present statistically" (p. 163). Thus "When looked at as discourse, one insistent thing a passage from a Shakespeare play says to us is 'this is by Shakespeare'. It also says 'this is a comedy or a tragedy'. After that it says, a little less loudly, 'this is early or late' and in some cases it may also say 'this is prose or verse'" (p. 163). In one last article from Shakespeare Survey, Jesús Tronch finds that Paul Werstine's critique of W. W. Greg's binary categories for English play manuscripts of "foul papers" and "promptbooks" holds true for Spanish plays of Shakespeare's time too (Jesús Tronch 'What If Greg and Werstine Had Examined Early Modern Spanish Dramatic Manuscripts?', ShSurv 70[2017] 99-112).

    A special issue of the journal Shakespeare, of which this reviewer is one of the five main editors, was guest-edited by Emma Smith on the topic of "Shakespeare's Changing Canon" and essays from it are relevent to this review. The present reviewer played no role in the selection or judging of the contributions to the issue. The first is by James Purkis and sketches the contested status and publication history of the Apocryphal books of the Bible in the sixteenth and seventeenth centuries ('Apocryphal Thinking, Or, Deuterocanonical Shakespeare', Shakespeare 13[2017] 292-308). Purkis surveys the recent changes to the Shakespeare canon as formerly apocryphal plays are admitted, and as alternative versions (even ones formerly discounted as bad editions) get published in adjunct volumes or series alongside their better-known companions.

    Purkis considers how C. F. Tucker Brooke characterized the plays in his influential The Shakespeare Apocrypha of 1908, ending with Brooke's distinction, which Purkis finds newly pertinent, between what Shakespeare wrote and what deserves to be called (is good enough to be) Shakespearian. Purkis summarizes that the ways that the Royal Shakespeare Company Complete Works, the Third Norton Shakespeare, and the New Oxford Shakespeare handle the Apocrypha, with special attention to Gary Taylor's essay on "Artiganality" in the New Oxford Shakespeare Authorship Companion (edited by Taylor and the present reviewer). Purkis's main point seems to be that an examination of playhouse manuscripts--Purkis's area of expertise--makes one less certain than Taylor is about just whose agency is present where.

    Jacob Boguszak also shares some reflections on the New Oxford Shakespeare, of which the present reviewer is a general editor (''Scarlet Experiment':! The New Oxford Shakespeare and the Importance of Authorship', Shakespeare 13[2017] 309-12). He objects that the volumes published so far give only the Shakespearian parts of The Spanish Tragedy and Sir Thomas More but give all of 1 Henry 6 and he acknowledges (but implicitly does not accept) the editors' excuse that for 1 Henry 6 it is difficult to say which bits are by Shakespeare. Boguszak also does not accept the notion of "alternative" in the design of the New Oxford Shakespeare Complete Alternative Versions volumes currently in progress. "How can a complete text [of, say, The Spanish Tragedy] be considered an alternative version of an excerpt?" (p. 311)  And he especially dislikes the fragmentary representation of Cardenio.

    Boguszak is not well informed on the state of authorship attribution: "To assume that one can locate Shakespeare's 'bits' in plays of uncertain authorship is to assume that Shakespeare was not very good at collaborating with others" (p. 311). In fact, in some cases we can distinguish the different authors' bits quite readily: a lot depends on how different their styles are, and the differences may well not be apparent to the naked eye or ear of the reader or playgoer so that superficially the work is a seamless whole. Boguszak thinks that maybe editions should be organized company-wise rather than by author. Craig and Greatley-Hirsch's book reviewed above tells us why this would not in fact be helpful.

    Also on the topic of the New Oxford Shakespeare's deficiencies, Harry Newman objects to its omission of some of the paratexts, such as John Milton's contribution to the 1632 Second Folio preliminaries and Leonard Digges's prefatory poem to John Benson's 1640 edition of the Poems ('Paratexts and Canonical Thresholds', Shakespeare 13[2017] 313-7). He also objects to the project's Modern Critical Edition borrowing from the Folio lists of characters that appear in full only in the original-spelling Critical Reference Edition. If Milton's and Digges's works were to be included in an edition of Shakespeare's works that is not based on the editions to which these form the paratexts it is hard to know where such a process might stop since Newman does not indicate what makes these particular paratexts more pertinent than the thousands of others that have been produced across the centuries.

    It is also unclear why Newman objects to character descriptions in a modernized edition drawing on those given in the early editions of Shakespeare; he appears to think the problem with this procedure is self-evident. Much the best thing in this special issue is John Jowett's essentially literary-critical reading of Shakespeare's attitude towards collaboration ('Shakespeare's Metamorphosis', Shakespeare 13[2017] 318-32). According to Jowett, Shakespeare used collaboration in his early years to develop his writing skills, and then he firmly renounced it; the break happened with Richard 3, Love's Labour's Lost and Venus and Adonis. Lastly from this special issue, Claire Bourne considers the reception of Shakespeare's works as witnessed in early readers' annotations, which is not directly relevant to this review ('Marking Shakespeare', Shakespeare 13[2017] 367-86).

    Outside of Emma Smith's special issue, Shakespeare published one other article of interest here: John V. Nance shows that scene 4.2 of  2 Henry 6, a crucial part of the Jack Cade rebellion, is by Shakespeare and Marlowe (''We, John Cade': Shakespeare, Marlowe, and the Authorship of 4.2.33-189 2 Henry VI', Shakespeare 13[2017] 30-51). We first see Jack Cade in scene 4.2, which is also where we first find the play's "sustained use of dramatic prose" (p. 30) that is so unlike 1 Henry 6 or 3 Henry 6. Hugh Craig's investigation--in his book with Arthur Kinney reviewed in YWES for 2009--was inconclusive about the authorship of at least part of scene 4.2 (lines 1 to 159) and Nance wonders if this is because of its preponderance of prose, which form makes writers write differently. (This last point is confirmed by Hugh Craig's work with Brett Greatley-Hirsch reviewed above.) Nance traces Cade's transitions between prose and verse and persuasively argues that he slips into verse when the verse-speaking Staffords enter in order to mock them: "Cade speaks verse to devalue verse" (p. 32).

    Nance attends to 4.2.33-189, looking for every 2-gram, 3-gram, and 4-gram and collocation in that passage and in EEBO-TCP and LION for i) all plays from 1576 to 1642, and ii) all books in that period. For additional filters, Nance restricts himself to authors whose canons have more than one such unique parallel with this part of 2 Henry 6, and to authors born before 1595; he does not count hits solely to religious pamphlets because this genre's preponderance in the extant books distorts the picture without bringing in plausible candidate dramatists. Using EEBO-TCP and LION together helps overcome the problem that the former classifies as drama any work structured as a dialogue, and cannot find 2-grams that are split by a line break.

    As a validation step, Nance finds other parts of plays of known authorship (confining himself to Shakespeare and Marlowe) that he attempts to attribute with his method, requiring also that these must mix prose and verse and mix comic and tragic elements in the way that 4.2.33-189 of 2 Henry 6 does. Finding a suitable passage for Marlowe is tricky because of the possible presence of other hands in his plays, as Nance explains before settling on scene 4.2 of The Jew of Malta. In a validation run looking at 173 words of prose from that scene, Nance gets more hits with Marlowe's canon than anyone else's, whether we look amongst just the plays and or at all books published. In the same kind of validation run with 173 words of verse from scene 4.2 of The Jew of Malta the hits with Marlowe's canon again predominate, although if we exclude Dido, Queen of Carthage from Marlowe's canon (on account of its co-authorship) then Samuel Purchas has just as many hits, although of course he is an unlikely candidate for authorship of The Jew of Malta.

    For a Shakespeare validation, Nance picks part of scene 1.4 from Richard 3. Again, in the separate verse and prose validations his method correctly points to Shakespeare as the author, whether we look at dramatic canons or everything published in the period. Noticeably, Marlowe's canon being one-sixth the size of Shakespeare's did not prevent Marlowe being picked out as the author of the sample from The Jew of Malta. Nance believes that because he counts only unique parallels his test "dilutes the statistical advantage gained by large canons" (p. 39).

    Because he used 173-word segments of plays in his validation, Nance divides the 1,166-word section  of scene 4.2 of 2 Henry 6 that he wants to test into six 173-word segments and one 128-word segment. He runs the same set of tests as he did in his validation phase and for comparison with dramatic canons the results are:

Segment One: Shakespeare
Segment Two: Shakespeare
Segment Three: Shakespeare
Segment Four: Shakespeare tied with Marlowe tied with Heywood
Segment Five: Marlowe
Segment Six: Marlowe
Segment Seven: Marlowe

When comparing the passages with all books rather than just plays, a couple of extremely unlikely candidates creep in, but overall Shakespeare and Marlowe dominate the results and the above division is well-supported: "The data indicate that 4.2 is collaborative" and "Shakespeare seems responsible for 4.2.33-120 and Marlowe for 4.2.121-89" (p. 46).

    The mocking prose asides to the audience in 4.2, which Vickers says are a uniquely Shakespearian feature, do in this analysis fall in the Shakespearian half of the section. Shakespeare is also responsible for nearly all the comic prose in the section, including the killing of the Clerk of Chatham. Nance finishes with fine literary-critical remarks on the transitions from comic to serious matters in the scene, especially in the context of Marlowe's concerns and habits of writing.

    Two articles by Darren Freebury-Jones may usefully be taken together. In the first he argues that MacDonald P. Jackson's use of shared phrases to attribute part of Arden of Faversham and to dismiss Thomas Kyd's authorship of other works is faulty since Freebury-Jones is able to find shared phrases that Jackson overlooked ('Kyd and Shakespeare: Authorship Versus Influence', Authorship 6.i[2017] 1-24). Freebury-Jones starts by summarizing what we know of Shakespeare's career as an actor, but overlooks the newest discovery: in an article in Shakespeare Quarterly in 2009, Alan H. Nelson and Paul Altrocchi report their finding of an early seventeenth century book annotation describing Shakespeare as "our Roscius". Freebury-Jones then  surveys some out-of-date scholarship on the 1590s editions of The Contention of York and Lancaster, and Richard Duke of York, declaring himself convinced that these are based on memorial reconstructions of Shakespeare's plays. Freebury-Jones is sure that Shakespeare began his theatrical career in Pembroke's men.

    Freebury-Jones thinks that being an actor gave Shakespeare his extraordinary memory for other people's plays. To trace Shakespeare's recall and reuse of others' phrases in his plays, he turns to Martin Mueller's dataset called "Shakespeare His Contemporaries" for which Freebury-Jones gives an URL that at the time of review (December 2019) returns a "page not found error". Even searching for "Shakespeare His Contemporaries" (with and without Mueller's name) in the popular search engines Google and Bing does not turn up the material Freebury-Jones depends upon, so there is no way to replicate his work although one can use EEBO-TCP to check the parallel phrasings he claims.

    Freebury-Jones finds phrases that are unique to Kyd's Soliman and Perseda and Shakespeare's 3 Henry 6. Some of these phrases are indeed unique to Kyd's play and Shakespeare's, but "I [would|will] not hence till" that Freebury-Jones claims is unique also turns up in Edward 3: "Ye will not hence, till" (STC 7501, B2r); I assume that if "would"/"will" are allowed to stand for each other, then "I"/"ye" are too. It is not clear why Freebury-Jones's use of Mueller's dataset did not turn up this usage, nor why it missed the same phrase in Dekker's Match Me in London (STC 6529, D4r) which although somewhat later should still be in Mueller's dataset. In claiming that he is tracing unique parallels Freebury-Jones does not indicate that he is applying chronological limits.

    It is Freebury-Jones's and his reader's bad luck that some of the URLs he cites become broken after the time of his writing, but Freebury-Jones brazenly cites an URL he knows was broken even before his paper was published: "Vickers's paper is no longer available online. It was previously available at http . . ." (p. 8n50). Freebury-Jones turns to phrases shared by Kyd's Soliman and Perseda and The Two Gentlemen of Verona, and one of them, "sweet glances", seems to be such a commonplace--Greene used it in Arbasto, the Anatomy of Fortune, Heywood used it in Troia Britannica, and so did more than half a dozen other writers--that although Kyd's and Shakespeare's are the only two plays it occurs in, we probably should not read too much into that. Two writers may, of course, independently use a common phrase.

    Freebury-Jones revisits Vickers's attempt to expand the Kyd canon and Jackson's demonstration (in an essay reviewed in YWES for 2009) that this method is a "one-horse race". Freebury-Jones tries to undermine Jackson's work by pointing out that Jackson finds more phrases shared by the non-Shakespearian (according to Jackson) rather than the Shakespearian parts of Arden of Faversham and 2 Henry 6 and shared by the non-Shakespearian (according to Jackson) rather than the Shakespearian parts of Arden of Faversham and The Taming of the Shrew.

    Obvious objections to Freebury-Jones's observation are that the authorship of 2 Henry 6 is itself disputed (many investigators think it co-authored) and that the sole-authorship of The Taming of the Shrew is under suspicion too. Although Freebury-Jones refers to his adjustment of the counts of shared phrases to compensate for the texts being different lengths, it is not clear from his phrasing that he properly allowed for the non-Shakespearian parts of Arden of Faversham being four-fifths of the play and having, as it were, more opportunity to match with other plays.

    Freebury-Jones considers the influence of King Leir on Shakespeare, citing a series of phrases-in-common that Mueller's dataset tells him are unique. It is not clear what Freebury-Jones thinks this indicates, since if he went looking for phrases common to only King Leir and the Shakespeare canon he was bound to find them. The clincher for significance would be to show that he went looking for phrases common to King Leir and a series of other writers' canons and was unable to find them, so that the finding for Shakespeare stands out. Freebury-Jones does not make this claim at this point, but he goes on to do so for other phrases-in-common and as we shall see they turn out to in fact not be, as Freebury-Jones believes, unique.

    At this point Freebury-Jones tabulates his data, reproducing from Mueller's work a table showing how many 4-grams are uniquely common to The Spanish Tragedy, Soliman and Perseda, King Leir, and Arden of Faversham on the one hand and on the other a selection of Shakespeare plays (3 Henry 6, Titus Andronicus, Richard 3, The Two Gentlemen of Verona, 1 Henry 4, King John, The Merchant of Venice, Much Ado About Nothing, Troilus and Cressida, Cymbeline, and Henry 8).

    The ones that stand out are The Spanish Tragedy with 45 matches and King Leir with 41. Freebury-Jones goes through the table, pointing out that quite a few of the matches between Arden of Faversham and the Shakespeare plays are found in parts of Arden of Faversham that Jackson does not think are by Shakespeare, but again he gives no sign of having adjusted for the fact that those parts of Arden of Faversham (Acts 1, 2, 4, and 5) are four-fifths of the whole play so we would expect to find more hits to those parts than to Act 3, all other things being equal.

    Looking for 4-grams common to Arden of Faversham and all other plays, Freebury-Jones reports that in Mueller's dataset "the play with the most unique matches is Kyd's Soliman and Perseda, with a total of eighteen" (p. 18). Mueller puts the odds of so many shared 4-grams coming up by chance as smaller than one in 10,000. Freebury-Jones does not mention by how much Soliman and Perseda has "the most" matches, and it would make a great difference to the significance of his discovery if the next highest play had, say, only half as many matches. On the other hand if Soliman and Perseda has 18 matches and five other plays each have 17 matches, the significance of the finding is much less. Without such details, a mere rank order ("the play with the most") does not really tell us anything and Freebury-Jones is overplaying his hand in claiming to have refuted the evidence that Arden of Faversham is not by Kyd.

    Freebury-Jones finds parallels between Arden of Faversham (outside of Act 3) and Richard 3, but again the problem is in the data's presentation since to see the significance one would require comparison between plays from multiple canons and Arden of Faversham. Moreover, some of Freebury-Jones's claimed parallels are not in fact unique to the pair of plays he claims. For example, "that loves not me" also appears in William Davenant's play Albovine (STC 6307, G3v) and it is not clear why Freebury-Jones did not find it in Mueller's dataset of "over 500 plays dated between 1552 and 1662" (p. 5). When arguments are built upon the absence of phrases from all but the texts in question, it matters greatly that we know why certain examples of those phrases in other texts have been overlooked by the investigator.

    Another example is Freebury-Jones's claim that the collocation of "in his bed" with "he is"--a collocation rather than an n-gram because the phrases are in different speeches separated by a speech prefix--is unique to Arden of Faversham and Richard 3, whereas in fact this collocation also appears in the plays Eastward Ho! by Jonson, Chapman, and Marston (STC 4971, D4r) and Anonymous's 2 Newmarket Fair (Wing S2318, A3v), both of which ought to be in Mueller's dataset. It takes little time to check on Freebury-Jones's claims and find their inaccuracies; the interesting question is why Mueller's dataset is misleading him.

    In his second article this year, Freebury-Jones concludes that 2 Henry 6 was not co-authored by Christopher Marlowe ('Did Shakespeare Really Co-write 2 Henry VI with Marlowe?', ANQ: American Notes and Queries 30.iii[2017] 137-41). He starts by summarizing his 2016 essay "Exploring Co-Authorship in 2 Henry VI" (reviewed in NYWES for 2016), reporting that his use of anti-plagiarism software turned up the extraordinary "parallelism of thought" (p. 137) that occurs when the Duke of York says "Cold news for me--for I had hope of France, | Even as I have of fertile England's soil" (1.1.237-8) and later says "Cold news for me, for I had hope of France, | As firmly as I hope for fertile England" (3.1.87-8). It is hard to see why Freebury-Jones would need software to detect this repetition, since it has long been noticed by readers and is frequently marked in scholarly editions as such. In response to the New Oxford Shakespeare attributing parts of 2 Henry 6 to Marlowe, Freebury-Jones quotes others including Vickers who disagree with this attribution.

    Freebury-Jones returns to the work of Philip Timberlake on feminine endings to verse lines, pointing out that early Shakespeare's rate of feminine endings is much higher than the rate in Marlowe's plays, and by Freebury-Jones's counts the parts of 2 Henry 6 that have recently been attributed to Marlowe have feminine ending rates much too high to be his. Freebury-Jones quotes Marlowe's mean average rate of feminine-endings of under 4% (of verse lines, by Timberlake's "strict" count), and Shakespeare's mean average of over 10%.

    But Marlowe's low mean conceals a wide variation in short sections, hitting 12.5% in Doctor Faustus, 9.6% in The Jew of Malta, 11.1% in Edward II, and 12.5% in The Massacre at Paris, as tabulated in the pages of Timberlake's book that Freebury-Jones refers his reader to. The problem with short sections of plays is in deciding where to draw the boundaries: a single scene of 8 lines containing one feminine ending comes out as a 12.5% scene, and a better measure of variation across a play would be to ignore dramatic structure and use a rolling window taking in a given number of lines.

    Freebury-Jones shows that if we use Ants Oras's pause-pattern measurements, the parts of 2 Henry 6 supposed to be Marlowe's and the parts supposed to be Shakespeare's test alike. This tells us nothing since Freebury-Jones has not established that Shakespeare and Marlowe (or anyone else for that matter) were distinctive in their use of these pause patterns. Freebury-Jones finishes by returning to those who do not accept Marlowe's contribution to 2 Henry 6, quoting from private correspondence from Marcus Dahl and an essay by Mueller that, as is unfortunately usual in Freebury-Jones's publications, is not at the URL he provides, which points to a location within Vickers's personal website.

    Poorly performed computational work on Shakespeare's writing is apparent in two more articles. In the first, Maria Ryskina, Hannah Alpert-Abrams, Dan Garrette, and Taylor Berg-Kirkpatrick train a machine-learning system on the Folio compositor discriminations made by Charlton Hinman and find that the system is then capable of making Hinman-like attributions of Folio compositors ('Automatic Compositor Attribution in the First Folio of Shakespeare', arXiv 704.07875v1[2017] n. pag.). The futility of this work ought to have been apparent at least to Garrette, whose affiliation is given as the Google corporation; in computer science the acronym GIGO (for garbage-in, garbage-out) has long been axiomatic.

    To illustrate that spellings and spacing around punctuation are chosen by compositors, the authors photo-quote a Folio page from The Comedy of Errors and another from Henry 5, apparently not noticing that the latter is prose. This matters because when setting prose compositors would routinely adjust spacing and spelling to help justify a line and no one would use the spacing and spelling in such lines as part of a compositor attribution.

    The investigators have developed a machine-learning system that learns the different spellings and spacings in the Folio pages set by known compositors and then, using this knowledge, makes its own predictions about which compositor set the other pages. On this task, the method agrees 87% of the time with the known attributions. We now know from Pervez Rizvi's article on the topic (reviewed in NYWES for 2016) that those supposedly secured attributions are insecure, so there are no solid ground-truth attributions for this method to work from.

    The investigators used as their ground truth Hinman's 1963 book on The Printing and Proof-Reading of the First Folio of Shakespeare, Trevor Howard-Hill's 1973 essay in Studies in Bibliography on "The Compositors of Shakespeare’s Folio Comedies", and Gary Taylor's 1981 essay in Studies in Bibliography on "The Shrinking Compositor A of the Shakespeare First Folio". Those three sources represent only a fraction of the work done on this topic and the investigators ignore subsequent work.

    The computer model works from images of the Folio that have been through the process of Optical Character Recognition (OCR) to get at the spellings and uses the images themselves to judge spacing around punctuation. (Nothing is revealed about how the process distinguishes prose from verse and short lines from long lines, so presumably this was not done and hence the process includes many lines that are irrelevant.) I do not understand what the authors mean when they write that "To separate the effect of the compositor from the choices made by the author or editor, we condition on a modernized (collated) version of Shakespeare's text as was done by scholars" (p. 2). What does "condition on" mean and what is this "modernized (collated)" Shakespeare text they used? Their Figure 2 is some kind of picture of how their model works, but it is too condensed and packed with unexplained abbreviations to convey anything.

    The failure to explain things gets worse as the abbreviation "EM" goes unexpanded in the sentence "In order to fit the model to an input document we estimate the orthographic preference parameters, w_c, and spacing preference parameters, θ_c, for each compositor using EM" (p. 3). The question of which modernized Shakespeare text they used gets answered on page 4: it was the Moby digital text based on the 1864 Globe edition. For the evaluation of how well their model makes attributions on its own they compare its determinations with those by Blayney in his 1996 Norton Facsimile of the Folio based, as they acknowledge, on work by Hinman, Howard-Hill, Taylor, John O'Connor and Werstine.

    It seems that the investigators took Hinman as their ground truth, trained a model based on the Hinman attributions, and found that this model made distinctions of the kind Hinman made. Then they took the Blayney attributions as their ground truth, and found that a model trained on those made distinctions of the kind Blayney made. It is not at all clear what criteria determined whether the model was judged to have independently made a correct attribution. Perhaps they trained the model on some parts of the Folio for which it was given the attributions claimed by Hinman and Blayney and then used the trained model to make predictions about other parts of the Folio and compared the results with what Hinman and Blayney claim about those other parts. This at least would be of some value, but the authors fail to coherently claim anything as straightforward as such a test.

    The same problem of inadequate description of a technical method vitiates David Kernot, Terry Bossomaier, and Roger Bradbury's claim that a series of tests that count various features of the language is able to differentiate the thematic groupings of Shakespeare's Sonnets identified by traditional scholarship ('Novel Text Analysis for Investigating Personality: Identifying the Dark Lady in Shakespeare's Sonnets', Journal of Quantitative Linguistics 24[2017] 255-72). Their approach was developed for forensic analysis of blog postings and applying it to Shakespeare's Sonnets they claim that it "uses word semantics that reflect personality or characteristics of self [and] can provide a more accurate profile of a person" (p. 255). Presumably they mean "more" than hitherto but who the "person" is in this case never becomes clear. The Sonnets divide into numbers 1-126 addressed to the Fair Youth and 127-154 addressed to the Dark Lady, with the Procreation (1-17) and Rival Poet material (78-86) falling in the Fair Youth group.

    The investigators call their method "RPAS", from (R)ichness (number of words used), Personal (P)ronouns, Referential (A)ctivity Power (word particulars and function words associated with clinical depression), and (S)ensory (words related to the sense of sight, sound, touch, smell, and taste). Applying their tests to each of the sonnets, the investigators want to see if they can objectively classify them into the groups previously identified by criticism (Dark Lady, Fair Youth, and Procreation).

    (R)ichness is just the type-token ratio. Personal (P)ronouns counts the occurrences of "my", "her", and "its" and puts them into an equation that multiplies them by arbitrary constants and sums them before raising the mathematical constant e (Euler's number) to the power of one such set of multiplications and sums divided by another such set of multiplications and sums. No defence of this equation is given: the investigators just point to prior studies that show that it produces a final number between 0 and 1 that, to a high degree of confidence, is below 0.5 when the writer's style is "feminine" and 0.5 or above when the writer's style is "masculine". No room here for gender that is not binary.

    The investigators acknowledge that "its" occurs nowhere in Sonnets but insist that ". . . the algorithm is effective at comparing data from within the Early Modern English period" (p. 259), without explaining why this is so. The Referential (A)ctivity Power measure counts 117 "highly concrete and imageability function words" (p. 259), and "imageability" seems to be the wrong part of speech here (a noun instead of an adjective). The description of how these counts gets processed is even less self-contained and well-explained than the one for Personal (P)ronouns, not least because a term "ɛ_i" is used before it is defined as "the weight for each category" (p. 259) of articles, conjunctives, prepositions, and pronouns.

    The reader has to figure out for herself that the i in ɛ_i takes on the values 1, 2, 3, and 4 to iterate through these categories. (At least, I think it does: the explanation is poorly phrased and infelicities in typography and layout make it hard to separate the symbols being glossed from the words used to gloss them.) An equation is given for how the various bits of data are processed but it is not defended; it seems to provide as its output four numbers, one for each of the four categories of function words as represented in the text.

    Lastly, the (S)ensory adjectives test counts the occurrences of 774 words "in two different contexts" (p. 260), sorted into five categories to match the five senses. It is not explained what the two different contexts are and the idea of there being a context seems to contradict the previous sentence's claim that adjectives are used because "their context is not necessary". As before, there is an equation that processes the various counts used for this test and produces a set of numbers, five in this case (one for each sense); the defence of the equation is again merely a reference to prior work.

    Bringing the various tests together, Kernot, Bossomaier, and Bradbury refer to "different configurations of the 14 elements from above" (p. 260) but I cannot see where this number 14 comes from. (R)ichness yields one number, Personal (P)ronouns yields one number, Referential (A)ctivity Power yields four numbers, and (S)ensory adjectives yields five numbers; that makes 11 elements so what are the other three?

    Moving on from this mysterious counting, the authors claim that only nine values are used because the four Referential (A)ctivity Power numbers are summed to make one and likewise the five (S) sensory adjective numbers are summed to make one. But that would seem to leave us with one value for each of R, P, A, and S, making a total of four dimensions not nine. The confusion mounts as the investigators refer to the Referential (A)ctivity Power using not the abbreviation "(A)" but rather the abbreviation "(P)", which also seems to be the abbreviation for Personal (P)ronouns.

    I am unable to make sense of how the investigators processed the data for this work. Just what they used as their source text for Shakespeare's writing is also uncertain as their reference to an online resource called "Farrow, 1993" makes one's browser attempt to access materials within a computer user's account on a server at the University of Sydney for which the reader would need log-in credentials that are not here provided. The investigators report that they used a Parts of Speech Tagger (POS) to "remove all stop words" (p. 260) but they do not say what those stop words are nor why a POS tagger was the way to remove them.

    Next a new technical term appears: "seriation" is a technique for placing a set of multi-dimensional quantities into an ordered sequence. This involves the use of the mathematical factorial function (conventionally represented by the symbol "!"), which the investigators several times pointlessly exemplify, such as "10! = 3.6 million", without actually mentioning how the factorial function is calculated. They introduce another new technical term, "the shortest path" (p. 261), without mentioning what the path travels through or how it might be calculated. Some of this becomes apparent later when they imply that they treat their nine pieces of data for each sonnet as coordinates in nine-dimensional space and then use software to find the shortest nine-dimensional line that touches each of these points.

    The quality of the explanatory prose is low. For example, it is not explained what is meant by "consistency" in the observation that ". . . it might be necessary to try a range of different estimators and look for consistency among them" (p. 261). The investigators appear to have simply entered their numbers into the "seriation" function of the software package R and accepted the results, which is an ordering of the sonnets so that ". . . the sonnets on the far left-hand side are most dissimilar to those on the far right-hand side" (p. 261).

    Having produced a nine-dimensional line that touches all the points in space represented by the sonnets, the investigators remove, one at a time, a single RPAS element--presumably one of the four from R, P, A, and S--to see if this shortens the path "without fragmenting the sonnet group clusters" (p. 262). They do not positively assert that the sonnet groups do indeed form clusters in nine-dimensional space--which if true would be the big news of this article--so we have to take this as implied.

    "We do this six times", they write (p. 262), but I cannot see where the six comes from: R, P, A, and S seem to me to present four opportunities to remove a single RPAS element. The method gets even mysterious when the investigators refer to the "PR(G)AS" elements performing "equally well" (p. 262) as the RAS elements (presumably that is RPAS without the P): it is unclear what "(G)" refer to and exactly what that performance consisted of.

    After some long tables consisting mainly of numbers that the investigators do not explain, and a brief discussion of adding "noise" to the data to determine its effect, the investigators conclude that the method is able to differentiate empirically between the Procreation sonnets, the Rival Poet sonnets, and the Dark Lady sonnets. Nowhere do they point the reader to the results that establish this. They do discuss the various permutations of their test's conditions and how these affect the tightness of the clustering within each of these groups, but I cannot see where their results show clear differentiation between the groups.

    Instead of their Table 3, which lists 1,540 numbers, they should have generated a visualization so the reader can see the claimed clustering and separation. The section headed "Conclusion" is an inadequate 63 words in all. For all one can tell this method may work perfectly well and do what is claimed for it, but the low standard of description offered here gives no reason to believe that this is so.

    Yet another flawed computational study is Hartmut Ilsemann's attempt to show that Sir John Oldcastle, attributed to Shakespeare in its second edition (published 1619), really is by Shakespeare ('The Two Oldcastles of London', DSH 32[2017] 788-96). Henslowe's Diary tells us that he bought a play about Oldcastle from Anthony Munday, Michael Drayton, Robert Wilson, and Richard Hathwaye in October 1599 and that it was performed at the Rose theatre in November 1599. A quarto appeared in 1600, and its prologue indicates that it is meant to correct the false image of Oldcastle as a glutton and bad counsellor to youth.

    Ilsemann sketches the well-known story of Shakespeare's being compelled to rename his Oldcastle character as Falstaff and of Shakespeare's 1 Henry 4 being referred to as Sir John Old Castell when the Chamberlain's men performed it in 1600. Ilsemann thinks that perhaps the Chamberlain's men really did perform the play we now know as Sir John Oldcastle and that it really is by Shakespeare. At least I think that is what he is suggesting, namely that rather than just changing the character's name to placate the Cobham family, Shakespeare wrote the play we know as Sir John Oldcastle. The point is not entirely clear from Ilsemann's phrasing.

    Ilsemann used the R Stylo software package to run Rolling Delta tests on a set of "reference texts"--Dekker's Satiromastix, Middleton's The Phoenix, Munday's Death of Robert Earl of Huntingdon, Samuel Rowley's When You See Me You Know Me, and Shakespeare's 1 Henry 4, Hamlet, and Richard 3--and the anonymous first edition of Sir John Oldcastle, looking at character 3-grams and using a window of 4,000 words that moved forward in steps of 800 words. Ilsemann is unhelpfully vague in describing the results of these tests, for example commenting that "The information that remains stable is Shakespeare’s predominance, and in Act IV, 2 and 3, it is the reference text Richard III that comes to the foreground" (p. 789). I cannot tell which play's fourth act is being referred to there and I presume but am not sure that the arabic numerals refer to scenes within that fourth act.

    The confusion gets worse when Ilsemann appears to describe the experiments out of chronological order: "Munday's texts also underwent elimination in previous runs of the program" (p. 789). Previous to what, and what does it take for a play or canon to be eliminated? Ilsemann uses R Stylo as a black box--a system whose inner workings he does not explain and perhaps does not understand--and his descriptions of his results indicate that it gave him conflicting results depending on the starting parameters he chose. For example, he remarks that "The general outcome is certainly that smaller windows tend to bring up collaborative scenarios" (p. 789).

    Collaborative authorship either did or did not happen in certain plays, so if changing the size of the rolling window in his R Stylo software makes it point towards or away from collaboration in those plays then there is nothing scientific about his method. The proper way to proceed is to first set up the software so that for a given collection of initial settings it correctly discriminates between collaborative and non-collaborative plays, or makes whatever discrimination one wants it to perform, and then to apply the software to the problem at hand. Altering the parameters during the experiments--twiddling the knobs, as it were--invalidates the results.

    These two problems--Ilsemann's inability to give a clear description of the experiments and his apparent twiddling of the knobs during them--undermine the remainder of this article. The tables and visualizations in Ilsemann's article are impossible for even a specialist to make sense of since he does not take the time to talk the reader through the interpretation of any of them: they are presented as if they are self-explanatory, and they are not. For these reasons, none of the conclusions drawn by Ilsemann should be credited.

    An additional problem occurs in Ilsemann's discussion of the date of the letter from Rowland Whyte to Sir Robert Sidney. The letter reports that the performance of the Oldcastle play was on Thursday 6 March 1599, but 6 March was a Tuesday not a Thursday in 1599 and in 1600 it was a Thursday. Ilsemann decides that Whyte got the day wrong and got the year right, but of course the true explanation is that, like many devout Christians, Whyte deferred rolling over to a new year until Lady Day, 25 March, so that what he called 6 March 1599 we would call 6 March 1600. This simple error takes Ilsemann on a trail of theatre-historical speculation leading nowhere.

    Much better work in this field is found in Thomas Merriam's presentation of evidence that Shakespeare wrote Acts 1 and 4 of King John and Peele Acts 2 and 5 ('Is it Time to Re-think King John?', DSH 32[2017] 591-601). After surveying some of his past work on the co-authorship of King John, Merriam's latest contribution begins with his division of the play into 204 blocks of 100 words each, and applying Principal Component Analysis (PCA) and cusum graphs, both of which techniques are repeatedly used in his previous research spanning 50 articles that have been reviewed and explained in YWES and NYWES during the past two decades.

    What is unclear at first is exactly what Merriam counted to provide raw data for his PCA and cusum processes. The cusum graph has positive slopes approximately aligned with Act 1 and 4, and negative slopes with Acts 2 and 5, suggesting different authors for these two parts. The play's treatment of the break with Rome is likewise two-sided, positive in places and negative in others. Merriam reports counting the lengths of words (measured in letters) in the First Folio, The Troublesome Reign of King John (possibly by Peele), five plays securely attributed to Peele, and the parts of King John that he thinks are by Shakespeare and the parts that he thinks are not.

    The result is that the First Folio and the parts of King John that Merriam thinks are by Shakespeare have a median word length of 4.07 and 4.04 respectively, while the others are all 4.20 or above. Next Merriam turns to tests he made using the software package R Stylo to apply to corpora of Shakespeare's and of Peele's words the Delta method invented by John Burrows, for which he describes the parameters selected. R Stylo was used to draw two dendrograms from the Delta results--Merriam's Figure 4 using the 131 most-frequent words and Figure 5 using the 864 most-frequent character 3-grams--and in both trees we find that Acts 2 and 5 of King John are on the Peelean branch of the tree and Acts 1 and 4 are on the Shakespearian branch.

    As Merriam points out, the agreement of a most-frequent words approach with a most-frequent character trigrams approach is remarkable since the latter, which includes such trigrams as "e t" because a space is a character, "defy conscious preference" (p. 595). Merriam finishes with literary critical remarks on the contradictory presentation of King John in King John--in some places he is to be admired, in others reviled--and suggests that co-authorship would explain this.

    A model of traditional methods of editorial scholarship is Rory Loughnane's study of the peculiar role of the character Mutius in Titus Andronicus ('Re-editing Non-Shakespeare for the Modern Reader: The Murder of Mutius in Titus Andronicus', RES New Series[2017] 268-95). Loughnane finds that the theory that the murder of Mutius was added to the play after Shakespeare had finished it is not entirely consistent with the textual evidence, since there remain logical and staging problems, which also extend to the sacrificing of Tamora's son Alarbus. The stage directions of the opening scene that call for the action to be split between the main stage and the stage balcony are Shakespearian not Peelean, so perhaps that split was an alteration introduced by Shakespeare after Peele had finished his work on the play.

    Loughnane remarks that we have long known that Peele wrote the beginning of Titus Andronicus, but we have not edited the play using this fact to help establish the text. He atttempts to give an account of the editing of the first scene "with Peele’s linguistic preferences and dramaturgical practices in mind" (p. 270) rather than Shakespeare's, and explains why he does not think the murder of Mutius was added by Peele after Shakespeare completed the rest of the play. The 1594 quarto (Q1) has "Though change of war hath wrought this change of chear" and the 1600 second quarto (Q2) alters the first "change" to "chance". Since Peele elsewhere uses "chance of war", Loughnane adopts "chance".

    Brian Boyd argued that Titus's murder of his son Mutius is anomalous because it is never referred to again in the play, but Loughnane finds the staging hard to explain if we cut from Q1 the parts that Boyd thinks were not present in the script before Peele interpolated this episode. In particular, in this cut version, Titus tells Saturninus to follow him and that he will bring Lavinia back after her brothers whisk her away to save her from marrying Saturninus, but Saturninus does not answer Titus until he (Saturninus) has gone back into the stage balcony ("Enter aloft . . ."). That ascent would take Saturninus some time.

    Loughnane shows that the same problem of implausible dialogue recurs if we cut the lines about Mutius's burial. The point is that "The Mutius murder and burial may jar, may seem unanticipated, may not be mentioned again elsewhere in the play, but once removed, it is difficult to reconcile what is left behind into a coherent script" (p. 276). The Mutius episode does not lift out cleanly from the script we have, therefore if it is a late interpolation other material around it must have been altered too, including perhaps some cuts that we know nothing of.

    Perhaps the sacrificing of Alarbus was also not originally in the play, since he is not mentioned in the stage direction that first brings on Tamora, "and her two sonnes" Chiron, and Demetrius. That is two sons not three, and before this entrance Marcus says that "the Noblest prisoner of the Gothes" has already been sacrificed at the monument of the Andronici. Even on its own terms, aside from its incompatibility with the ensuing sacrifice of Alarbus, Marcus's report that the Goths' noblest prisoner has already been sacrificed at the monument of the Andronici is problematic: why did Titus not bury his dead sons there while making this sacrifice?

    Marcus's account of the sacrifice cannot be a simple alternative to the later-interpolated scene of Alarbus's execution, because it is also incompatible with the entrance of Titus accompanied by his dead sons. If we try to explain all these things by supposing that there were shifting authorial intentions during the creation of the scene, it is very hard to know which parts to keep and which to omit because they are incompatible with where the intentions ended up. Loughnane offers no easy solutions to this conundrum.

    Loughnane works his way through the staging problems in this opening action of the play, looking at how Stanley Wells and Jonathan Bate addressed them in their different ways. A key question is what Saturninus is doing while the whole dispute of Titus with his sons about Lavinia takes place. If all this time Saturninus is busy wooing Tamora and does not notice the dispute that culminates in Mutius's murder, it is odd that he seems to know of the dispute about Lavinia when he announces that Tamora will be his wife. Yet if he does take notice of the dispute about Lavinia he seems to choose Tamora without having paid any attention to her.

    So far the discussion has been about the 1594 Q1, which is the basis of all modern editions, but there is also the 1623 Folio text, which was printed from an exemplar of the third quarto (Q3) published in 1611 that was first annotated from a theatrical document. The Folio-only stage directions are an attempt to make staging sense out of the dialogue inherited from Q1 (via Q3) and Loughnane explains how he handled it, quite conservatively he says, in the New Oxford Shakespeare.

    Peele does not use an entrance "aloft" in any of his plays, but Shakespeare does, so the stage directions in Titus Andronicus that do so are already immediately suspect. Indeed, the opening of Titus Andronicus also has the stage direction "They go vp into the Senate House" (Q1 and F), which phrase "go up into" is vanishingly rare in the drama but does appear in dialogue in Julius Caesar: "Let him go up into the public chair". We seem to have some Shakespearian material in a scene written by Peele. Perhaps Shakespeare imposed the upper/lower staging we find in Q1 and F, after Peele had written the scene for performance entirely on the main stage; that would explain quite a few of the opening scene's conundrums.

    Henslowe's Diary records a performance of Titus Andronicus at the Rose on 24 January 1594 and the New Oxford Shakespeare gives reasons for dating its composition to the late 1580s. Peele was eight years older than Shakespeare, but the latter had a string of stage hits in the late 1580s and early 1590s so Loughnane thinks that we should not assume that this was a master/apprentice relationship. If, as Anna Pruitt and William W. Weber argue (the latter in an article reviewed in YWES for 2014), Titus Andronicus 4.1 is by Shakespeare not Peele then rather than the two men actively collaborating we should should consider the possibility that Shakespeare took over a play that Peele had merely begun. Perhaps that is what Greene's Groatworth of Wit alludes to as Shakespeare beautifying himself with others' feathers.

    The remaining articles outside of Notes and Queries do not call for extensive review. Eric Rasmussen suggests that perhaps Leonard Digges or John Florio served as editor of the 1623 Shakspeare Folio ('Who Edited the Shakespeare First Folio?', Cahiers Élisabéthains 93[2017] 70-6). He starts with the assertion that it is obvious that "someone regularized the metrical irregularities and subject-verb disagreements of the quarto texts" and "introduced thousands of verbal changes" (p. 71), but those are claims that I would want unpacked with some discussion of the evidence that quartos were the sources for the Folio texts. They clearly were not for those plays that had not been previously printed, and for the others I would expect some discussion of the evidence for dependence on the quarto texts rather than, say, independent manuscripts of the plays. Moreover, why "someone" for the whole of the Folio rather than, say, a different person for each play? There are a lot of assumptions built into Rasmussen's first two sentences.

    After mentioning some previously suggested candidates for the editorship of the 1623 Folio, Rasmussen announces his as Leonard Digges, who contributed commendatory verse to the 1623 and 1632 Folios and the 1640 edition of Shakespeare's Poems, and who had his own works published by Edward Blount of the First Folio consortium. Digges lived near John Heminges and Henry Condell and Shakespeare, and his brother married Will Kemp's niece and his widowed mother married Thomas Russell, executor of Shakespeare's will. Rasmussen lists some slighter Shakespeare connections too.

    Rasmussen also considers the possibility that John Florio did the editing, with rather a lot of speculative phrasing: "may have tutored Ben Jonson . . . . might also have had a connection to Shakespeare through his patron, Henry Wriothesley . . . . Shakespeare may have modeled the pedant Holofernes in Love's Labour's Lost on Florio" (p. 72). Rasmussen surveys the rare-word evidence underpinning Saul Frampton's case for Florio being the Folio editor, pointing out that some of it is wrong, finding misquotations and an overlooked occurrence of a rare word.

    Rasmussen takes the example of "my" before a vowel in Q2 Hamlet being rendered as "mine" in Folio Hamlet, tabulating 16 examples, and showing that Digges and Florio also prefer "mine" in such cases. The obvious objection is that many people might have preferred "mine" in these cases so this is not strong evidence for the handiwork of these two men in particular. Nor does Rasmussen address the question of provenance of Folio Hamlet, referring to how the "Folio . . . changed" (p. 73) Q2 readings without discussing the considerable evidence that in fact Folio Hamlet was printed from a transcript of a theatrical document and not a preceding quarto. Rasmussen mentions some other words for which the Folio preference appears to agree with the Digges or Florio preference, but is explicit in making no claim to being comprehensive or systematic.

    In the 1609 Sonnets, Sonnet 76 reads "That euery word doth almost fel my name" and "fel" is usually emended to "tell" but Mingqiang Li thinks it should be "sell", since the long-s is more easily misread as "f" than a "t" is ('William Shakespeare's Sonnet 76', Explicator 75[2017] 185-7). Li defends "sell" as meaning "betray" in this context, using the analogy of Exeter's "That he should for a foreign purse so sell | His sovereign's life to death and treachery" in Henry 5, to which the obvious retort is that in Exeter's line the word "purse" accompanies "sell" and the sense is obviously not betrayal in general but taking money to betray someone. There is no such sense in Sonnet 76, and there "sell" meaning "betray" would have to be doubly metaphorical since the betrayal is only a revealing (of the poet's name) not a giving-up-for-money as in the play.

    Rodney Stenning Edgecombe tackles a crux (''He'll | Seem to Break Loose' in A Midsummer Night's Dream', Explicator 75[2017] 140-2). The 1600 first quarto of A Midsummer Night's Dream has: "Lys. Away, you Ethiop. | Dem. No, no: heele | Seeme to breake loose: take on as you would follow; | But yet come not". The 1623 Folio has Demetrius say "No, no, Sir, seeme to breake loose . . .". Edgecombe's suggestion is that Demetrius says "No, no--a Helen! | (To Lysander) You seem to break loose . . .", which response perfectly completes the meter of Lysander's "Away, you Ethiope".

    The reason that Demetrius defends Hermia at this point (likening her to another Helen) is that "Love in idleness might have revoked his attraction to Hermia, but Demetrius is still besotted, and honor-bound by the code of chivalry to protect his beloved from insult" (p. 142). Edgecombe suggests the quarto's compositor was confused by the use of the name Helen at this point and independently changed it to "heele".

    Where Sonnet 18 in the 1609 Sonnets has "Sommers lease hath all too shorte a date", Charles Llewelyn Davies thought that the word "lease" was really "leafe" (that is, he mistook what is in the quarto) and he defended "leafe" as the better reading. Richard J. Moll wonders if Shakespeare meant this as a kind of visual pun: readers were supposed to see both possibilities at once, and both make sense ('A New Look at an Old Reading: William Shakespeare's Sonnet 18', Explicator 75[2017] 137-9). If it were intended that we see "leafe", Moll speculates, then there might also be a pun on leaves of paper of the kind upon which the sonnets are being written. If so, "leafe" becomes the antecedent of "this" in the final line (the subject of "lives" and "gives life"), making a better antecedent than "eternal lines", for which the pronoun should be "these" not this".

    Jason Scott-Warren finds that much of what Francis Meres wrote in his Palladis Tamia of 1598 about various writers was plagiarized from his sources, including the phrases used for comparison taking the form "as x said of y, I say a of b" ('Commonplacing and Originality: Reading Francis Meres', RES 68[2017] 902-23). Martin Dodsworth reports that anagrams were popular in the French literary tradition of Shakespeare's time but not, as is sometimes claimed, in the English tradition and that there are no anagrams in Shakespeare's Sonnets (Martin Dodsworth 'The Elizabethan Anagram and Shakespeare's Sonnets', RES 68[2017] 666-88).

    And finally to the round-up from Notes and Queries. Thomas Merriam offers new evidence that Hand D of Sir Thomas More contains dramatic material written by Shakespeare ('The Proximity of Hand D of Sir Thomas More and Othello: A Detail', N&Q 262[2017] 283). The word "trash" is repeated within a line in Sir Thomas More (in Addition II) and in Othello (in 2.1) and shortly thereafter in both cases comes the word "hip"; this collocation is unique in LION and EEBO-TCP for the period. Actually, the repetition of the word "trash" in Othello is not in the 1622 quarto, where the second word is "crush", nor in the Folio where the second word is "trace": it exists only in editions that follow George Steevens's emendation of "crush" or "trace" to "trash". Merriam observes that Sir Thomas More and Othello are usually dated around 1604 and concludes that both having this otherwise never seen collocation is more evidence that Hand D is Shakespeare.

        Michael L. Hays has a theory that Iago did not know that Cassio served as Othello's intermediary in the wooing of Desdemona until she mentions it in 3.3 and that he seizes on this fact as a way to trigger Othello's sexual jealousy ('Who Wooed Desdemona?: The Crux at Othello, III.iii.94', N&Q 262[2017] 284-7). In the 1622 quarto, Iago in 3.3 asks Othello "Did Michael Cassio when you wooed my Lady, | Know of your loue?" and in the Folio he asks "Did Michael Cassio | When he woo'd my Lady, know of your loue?" Shortly before this in 3.3, Desdemona mentions that Cassio came with Othello when Othello wooed Desdemona ("Michael Cassio, | That came a-wooing with you").

    In F, Iago insinuates that when Cassio served as Othello's intermediary in his (Othello's) wooing of Desdemona, Cassio wooed her for himself. Hays sketches the literary history of unfaithful intermediaries in love, but in illustrating it from Shakespeare's works he leaves out Suffolk wooing Margaret for King Henry 6 at the end of 1 Henry 6 and Claudio's fear that Don Pedro woos Hero for himself during the mask in Much Ado About Nothing 2.1. The many couplings that Iago says Cassio had with Desdemona thus took place in Venice before her wedding, not (implausibly) since the wedding. Thus F's "When he woo'd" is the better reading.

    Francesco Guicciardini's A History of Italy--first English translation 1579, republished by Richard Field in 1599--contains a prophetic dream of ghosts that Richard M. Waugaman thinks provided the inspiration for Richard's dream before the Battle of Bosworth at the end of Richard 3 ('A Possible Source for Richard III's Prophetic Nightmare in Guicciardini', N&Q 262[2017] 265-6). It is probable that Shakespeare knew Guicciardini's book since it is a likely source for The Tempest.

    Christopher Baker finds that Naseeb Shaheen in his study of biblical allusions in Shakespeare overlooked the fact that just after the St Crispin's Day speech in Henry 5, Henry alludes to Mark's gospel ('Henry V and Mark 923:', N&Q 262[2017] 266-7). Henry speaks of his hope and confidence of victory and remarks that "All things are ready if our minds be so" which echoes "All things are possible to him that believeth" (Mark 9.23).

    Thomas Merriam has more evidence for his previously advanced claim that Henry 5 is co-authored, perhaps with Christopher Marlowe ('Verse and Prose in Henry V', N&Q 262[2017] 266-9). His previous investigations led him to this conclusion because the prose parts test quite positively Shakespearian while the verse parts do not and seem more like Marlowe's style. For the present analysis he divided the play into prose and verse sections and put them with Romeo and Juliet, Hamlet, Othello, King Lear, The Winter's Tale, 1 Tamburlaine, and 2 Tamburlaine in a test performed using the software package R Stylo that counts the most-frequent bigrams common to all these plays (278 bigrams in all). Merriam used modernized texts and omitted stage directions and speech prefixes.

    The resulting 278 counts for each play were reduced to just the first two principal components and these Merriam plots on an x/y graph. On PC1, the verse part of Henry 5 clusters with the two Marlowe plays, while the prose part clusters with Shakespeare's. (Clustering of the two authorial groups was also evident on PC2, but both parts of Henry 5 score significantly higher on PC2 than the Shakespeare and Marlowe plays.)

   Referring to the Gravedigger's evasive answers, Hamlet says "We must speak by the card, or equivocation will undo us", and Steven Doloff believes the card refers to the catechism ('Speaking 'By the Card': Hamlet Remembers His Catechism', N&Q 262[2017] 270-3). Previously this "card" has been taken to mean a 32-point compass card used in navigation, so that Hamlet is saying that they must precisely steer the conversation, choosing their words carefully, to get what they want from the Gravedigger. Every child was required to memorize the Protestant catechism in school and Shakespeare certainly knew it. At least one published catechism called itself a card.

    Luther had introduced into the existing Catholic catechism the form of a series of questions and answers to ensure the reader understood the material, and that is appropriate to the question-and-answer context of the allusion in Hamlet. In Othello, the Clown uses the word "catechism" in the context of a question-and-answer exchange with Desdemona that is like the one in Hamlet, especially in the extended punning of "lie" meaning to lodge and to tell a falsehood. There are a couple of other allusions to the wording of the catechism in Hamlet: "pickers and stealers", and "the dead [and] . . . the quick".

    Daniel Kaczyński claims that "Do ye make an alehouse of my lady's house . . . ?" in Twelfth Night is an allusion to Jesus's cleansing of the temple (removing the money lenders) recounted in the gospels of Matthew, Mark, Luke, and John ('A Source for Malvolio's 'Do ye Make an Alehouse of My Lady's House' (Twelfth Night II.iii.87-88)', N&Q 262[2017] 278-81). The link is in the phrasing "make a [something] house of my [person's] house" as in "make not my fathers house an house of marchandise" (John 2.16). Moreover, this passage from John was traditionally read on Twelfth Night, the 6th of January. It suits Malvolio's self-importance to see himself as something of a Christ-like figure.

    In Hand D of Sir Thomas More, More refers to the rebels' "momtanish [moritanish?] inhumanity" (the manuscript reading is unclear), and in his edition Jowett rendered this as "mountainish", reporting that "mountainish" appears nowhere in EEBO-TCP. But Kristine Steenbergh has found it in Richard Harvey's An Astrological Discourse . . . upon the notable conjunction . . . of Saturn and Jupiter of 1582 ('A Possible Source for Sir Thomas More's 'Mountainish Inhumanity'', N&Q 262[2017] 281-3). Oddly enough, a search of ProQuest's EEBO-TCP on 13 October 2019 for this review does not find that occurrence but does find "mountainish" in Nicholas Byfield's The Rule of Faith (STC 4233.3, Qq7r).

    Steenbergh has also found "mountainous" near to "inhumanity"--well, separated by only 13 words--in Samson Lennard's English translation of the Italian Thommaso Buoni's Problems of Beauty and All Human Affections of 1606. It turns out that Steenbergh does not use the ProQuest EEBO-TCP dataset but rather the one at Early Modern Print based on Phase One of the EEBO-TCP open-access project, which for "mountainish" finds the Harvey and Byfield examples mentioned above. Steenbergh thinks that Lennard's translation influenced Shakespeare's writing of the scene in Sir Thomas More and hence that the play scene is no earlier than 1606, or perhaps 1605 if Shakespeare had "access to Lennard’s work in process" (p. 282). 

    Céline Savatier-Lahondès finds some loose thematic parallels between the events in Macbeth and earlier Celtic mythology, including the idea of walking forests ('The Walking Forest Motif in Shakespeare's Macbeth--Origins', N&Q 262[2017] 287-92). Brian Vickers finds a source for Sonnet 129 ("Th' expense of spirit in a waste of shame") in Robert Greene's pamphlet Never Too Late of 1590, which he quotes (Brian Vickers 'A New Source for Sonnet 129', N&Q 262[2017] 292-4). In support of the connection there are a few verbal parallels ("perjured", "full of", "lead to . . . hell"), and both works contrast love with lust and have a run of seven negatives.

    In Antony and Cleopatra the sounds of hautboys under the stage is said to be that of Hercules leaving Antony, but in Plutarch the god leaving him is Bacchus. Why did Shakespeare change the god? According to Shawn Normandin the answer is that Bacchus strongly connoted wine to early moderns, and the idea of wine leaving Antony might bathetically suggest that the sound of hautboys is that of Antony urinating or vomiting ('Replacing Bacchus in Shakespeare's Anthony and Cleopatra IV.iii', N&Q 262[2017] 294-7).

    After tracing the imagery of various kinds of games in Antony and Cleopatra, Caroline Baird, noting that Antony complains to Eros that Cleopatra "has | Packt Cards with Caesars", offers a new explanation for the image ('Board Game Squares, Face Cards, and Chess in Antony and Cleopatra', N&Q 262[2017] 297-300). To pack cards meant to shuffle them falsely to give oneself an advantage. Editors since Nicholas Rowe have emended "Caesars" to "Caesar" to mean that Cleopatra has shuffled the cards to favour Caesar, but Baird thinks the plural is correct: Cleopatra has packed her deck with kings, referring to Antony himself and her previous lovers Pompey and Julius Caesar. In French cards, and possibly English ones in Shakespeare's time, the King of Diamonds was known as Julius Caesar.

    Shakespeare's preferred bible was the Geneva Bible and he shows familiarity with its marginal explanatory notes. John W. Harris claims a number of allusions across various plays, but it is not clear if he is claiming to be the first to spot them or just the first to notice that they all come from the Geneva Bible and its marginal glosses (''Written in the Margent': Shakespeare's Metaphor of the Geneva Bible Marginal Notes', N&Q 262[2017] 301-4). Hamlet's "edified by the margent" is an allusion to such glosses, since the Geneva Bible in particular claims that edification is what its notes are for. Harris finds that other references to learning from the margins, made in Romeo and Juliet and The Rape of Lucrece, were also inspired by the Geneva Bible.

    James Harding-Morris has discovered that the verses on the east end of the Stanley monument in the church at Tong, Shropshire, appear in almost identical wording on the tomb of Sir William Gostwick (who died in 1615) in the church of St Lawrence in Willington, Bedfordshire ('A Shakespearean Epitaph in a Bedfordshire Church?', N&Q 262[2017] 304-5). Did Shakespeare supply his verses to both, or was the one at Tong simply copied at Willington? Maybe the east-end Tong verses are just stock elegy writing, not Shakespeare's, since only the west-end verses have the thematic and verbal connections to sonnets 55 and 81. On the other hand, Milton's elegy to Shakespeare (used in the preliminaries to the 1632 Folio) seems to echo the east-end and west-end Tong verses, which would be odd if Shakespeare did not write them both. (The echoes to the former are not strong, however.)

    Why do some plays have two titles, asks Andrew Gurr ('Double Play Titles', N&Q 262[2017] 331-6)? Sometimes the second one clarifies the first, but sometimes it seems to represent an entirely different way of thinking about the work. Shakespeare has just two examples: Twelfth Night, or What You Will, and Othello, The Moor of Venice. The latter of course need not be an alternative title but just a continuation and clarification of the first word. Gurr lists some non-Shakespearian examples too. How come John Manningham in his diary entry of 2 February 1602 used exactly the double title, "Twelfth Night, or What You Will", that the Folio catalogue would use two decades later? Perhaps it was written on playbills. Gurr ends by listing all 40 two-title plays published between 1590 and 1660.

Books Reviewed

Craig, Hugh and Brett Greatley-Hirsch Style, Computers, and Early Modern Drama: Beyond Authorship. Cambridge University Press [2017]. 320 pp., £44.62, ISBN 978-1107191013

Crawforth, Hannah, Elizabeth Scott-Beaumann and Clare Whitehead, eds. The Sonnets: The State of Play, The Arden Shakespeare State of Plays Series. Bloomsbury [2017]. 308 pp., £22.83, ISBN 978-1350094857

Neill, Michael and David Schalkwyk, eds. The Oxford Handbook of Shakespearean Tragedy, Oxford Handbooks. Oxford University Press [2016]. 992 pp., £34.29, ISBN 978-0198820390

Shakespeare, William A Midsummer Night's Dream, ed. Sukanta Chaudhuri, The Arden Shakespeare. Bloomsbury [2017]. 400 pp., £8.99, ISBN 978-1408133491

Shakespeare, William Cymbeline, ed. Valerie Wayne, The Arden Shakespeare. Bloomsbury [2017]. 440 pp., £8.99, ISBN 978-1904271291

Shakespeare, William King Edward III, ed. Richard Proudfoot and Nicola Bennett, The Arden Shakespeare. Bloomsbury [2017]. 432 pp., £8.99, ISBN 978-1903436387

Shakespeare, William The Comedy of Errors, ed. Kent Cartwright, The Arden Shakespeare. Bloomsbury [2017]. 312 pp., £8.99, ISBN 978-1904271246