Gabriel Egan . com

Gabriel Egan
Globe Education
Shakespeare's Globe
London SE1 9DT

26 February 2002

    Year's Work in English Studies 2000: Shakespeare: Editions and Textual Studies

    Six major critical editions appeared this year. For the Arden third series: edited by Gordon McMullan, edited by Edward Burns, and edited by Giorgio Melchiori; for the Oxford Shakespeare: edited by John Jowett, edited by Jill Levenson, and edited by Stanley Wells. Of these, Melchiori's The Merry Wives of Windsor was not received in time to be included in this survey and will be reviewed next year. Several New Cambridge Shakespeare editions were published in 1999 and 2000 and these will be reviewed together in next year's survey.

    McMullan's introduction to Henry VIII runs to nearly 200 pages, nearly half of which is a "Cultural History" constituting virtually a monograph on the play's meanings since its first performance. It should surprise no-one, observes McMullan, that the play (usually subtitled All is True) is only a partial slice of the truth: only 2 of the 6 wives are shown, and the play foregrounds engagements with the political truth of its own time and of Henry's time. Until recently, because "the watchword of criticism was 'unity'" (p. 4), those who liked the play tended to argue that it is all by Shakespeare, and those who disliked it blamed collaboration. Since the mid-1970s, however, it has been possible to read the play without the critical strait-jacket of 'unity', and to see its contrary impulses ("at once celebratory and cynical about display" p. 5) and its representation of history as inherently a contradictory narrative.

    Of the performance that burned down the first Globe playhouse, McMullan writes that "It is described in one of the reports as a 'new' play" but then he goes on to quote two reports in which it is said to be new (pp. 9, 58-59). McMullan thinks it had previously been performed at the Blackfriars because of the irresistible resonances of using the very hall where Katherine's trial had taken place, and because Henry mentions Blackfriars explicitly (2.2.139). If so, perhaps the big spectacles in the Folio text were added when it moved to the Globe which, unlike the Blackfriars, had room for them (p. 10). The second recorded performance (in which it is called "K. Hen. 8") was at Globe in 1628, sponsored by the Duke of Buckingham to bolster his popularity; his doing this makes sense only if one sees that he got the play's political irony. McMullan's account of the stage history documents the increasing attention to spectacle in the eighteenth- and especially the nineteenth-century productions, which entirely subordinated the words. John Downes's claim that Thomas Betterton had his Henrician acting instructions from William Davenant, who had them from John Lowin, who had them from Shakespeare, suggests that Lowin (not Richard Burbage) performed Henry in 1613, setting a precedent for productions which show "Henry in the prime of his years" (18n3). Burbage was 45 to Lowin's 37, and perhaps that is indeed enough of a difference.

    The modern stage history of the play McMullan begins at 1916, when Herbert Beerbohm Tree's pre-war production toured the United States and was deemed a relic of a bygone age. Even in the twentieth century the play was never done in modern dress and was still used as propaganda. The attention on 'truth' and 'authenticity' shifted from the period in which the play is set (which the nineteenth-century spectacles wanted to recover) to the period in which it was first performed. The BBC TV version is, ironically, low-budget and 'inward' rather than spectacularly expansive, thus breaking with the stage tradition. Terence Gray's 1931 production was irreverent towards the history and debunked it, albeit in the name of a different kind of 'authenticity': the original performance effect. Gray's method was not to recreate the original staging, but to find a modern way of doing what the original did. The characters were dressed like playing card figures and behaved like marionettes, although for the final moment Gray effectively rewrote Shakespeare: the baby turns out to be a cardboard doll of Elizabeth 1 aged 60 that is thrown into the audience. Rather than find subversiveness within the play, Gray worked against it to be radical. Gray's alienation effect has influenced three subsequent Stratford production, including Tyrone Guthrie's in 1949 where the duchess of Norfolk sneezed noisily during Cranmer's address in honour of the baby Elizabeth (p. 48). For the 1996 Royal Shakespeare Company production, Greg Doran's awareness of his predecessors made him want the play's ceremony to be taken 'straight' in order to show its emptiness, rather than have it undermined before it was even seen. Doran undermined the spectacle subtly by providing the rainstorm mentioned at 1.1.90, and by a persistent whispering of courtiers even when they should be attentive. McMullan believes that audiences did not understand these devices, expecting either unironic celebration or else entire debunking (p. 55n1).

    Returning to the play's first contexts, McMullan gives a splendid reading of Henry Wotton's account of the burning of the Globe as an example of the familiar Reformation genre of comedia apocalyptica (pp. 60-1). The play is situated between "celebration of James's Reformation inheritance and the suggestion that that Reformation had never truly taken place". The important historical context of 1613 was that prince Henry, a great hope for a Reformed Europe, was gone and protestant hopes were transferred to his sister Elizabeth's marriage to Frederick, the Elector Palatine, the most prominent continental protestant ruler (p. 64). The baby Elizabeth would remind the audience of this Elizabeth. The 'truth' of the play (its title and Cranmer's final speech) relates to the "Truth is the daughter of Time" iconography with which Elizabethan associated herself (Truth being the true church liberated from her sister's Catholic influence). The theme of veritas filia temporis was revived by Dekker in The Whore of Babylon and by Middleton in his pageant The Triumphs of Truth. To many protestant Henry 8 was hardly a true protestant reformer because, prior to the break with Rome, he had persecuted protestant heretics. Because of his intemperance (in the senses of immoderation and keeping odd hours) Henry could not be 'Time', so Cranmer (who was to become a protestant martyr under Mary) is a surrogate Father Time to the baby Elizabeth in the final scene (pp. 67-87). Cranmer's reference to Saba (Sheba) in his speech might also reflect badly on Henry, whose extra-marital infidelity is like David's with Bathsheba. Henry was known to be sensitive about this parallel, especially because God's punishment of David is the death of his first child with Bathsheba. This would also have made James 1 uncomfortable, who might also have taken the loss of his son as a divine punishment for his "negligence of the godly cause". The final scene can thus be read as straight celebration of the royal dynasty, or as a criticism of kings Henry and James, and especially the latter for not having "fully understood the nature of his responsibilities towards the continuing Reformation" (pp. 88-93).

    Were all this not apparent to the original audience, the play is still unlike a normal history in that it refuses teleology: events are related sequentially but not causally. In a rare lapse into jargon, McMullan writes that the spectators of the Field of the Cloth of Gold are "interpellated by official ceremony" without making clear if he means 'interpellate' in the Althusserian sense or the obsolete archaic sense of 'interrupt'. In place of soliloquies in which the protagonists might reveal their real feelings, this play has the endless reporting of one person's words by another; rumour is all and truth becomes indistinguishable from opinion. When two characters provide almost entirely conflicting accounts of someone's life (as Katherine and her gentleman usher Griffith do about Wolsey), we have no stable truth to fix upon. The same happens with events such as the Field of the Cloth of Gold, making the play a powerful reflection on "the way in which truth is debated and established within a culture and particularly within that culture's conceptions of history" (pp. 94-106). Of the genre indeterminacy of the play, McMullan notes that it can be seen as masque-like in using an anti-masque in the penultimate scene with the Porter holding back the lower-class characters, before the masque-like christening. But although the play borrows the masque form, it refuses the monolithic kind of truth purveyed by it. The play has something of the Romance qualities of Pericles, The Winter's Tale, Cymbeline, and The Tempest: multiple-character focus, a supernatural moment (Katherine's vision), it ends with "a redemptive father-and-daughter tableau", and it juxtaposes linear with cyclical time (pp. 108-10). McMullan links these plays as 'late writing': "the idea that authors, towards the ends of their careers, tend to return to patterns associated with their early writings in order to rework them from the perspective of the experience of the intervening years". Of course the play is late Shakespeare, but also early Fletcher. Shakespeare and Fletcher were in the vanguard of new fashionable genre, romantic tragicomedy, which was to dominate the stage beyond the Revolution. Fletcher's early plays draw on or require an audience to be familiar with Shakespeare's plays, and Shakespeare returned the compliment with The Tempest which draws on Fletcher's Faithful Shepherdess; Henry VIII draws on Beaumont and Fletcher's The Maid's Tragedy. But since the chronologies of the Shakespeare and the Fletcher canons are uncertain, and since Roslyn Knutson has shown that company needs and competitive fashion, not individual creators' tastes, shaped genres, it is unwise to speak of Fletcher influencing Shakespeare or vice versa (pp. 112-16).

    One of the sources of Katherine's dream vision was, ironically, Holinshed's report of Anne Bullen's death-dream vision. Indeed, in a number of ways Katherine and Anne are linked by the play in a way which undermines a simple religious reading of the play. Another source of Katherine's dream vision appears to be Elizabeth 1's dream vision in Heywood's If You Know Not Me, which is of course a most strange connection between Bloody Mary's mother Katherine and Anne Bullen's daughter Elizabeth. In her distance from the excesses of Wolsey, Katherine is thus something of a Catholic reformer, nearer to Anne and her Lutherism than to Rome. The movement from reign to reign in England (Henry 8, Edward, Mary, Elizabeth), each of which it is hoped will provide political and religious stability, is thus likened to Henry's restless sequence of wives in search for a son and heir. At the end of his career Shakespeare returned to Roman 'New Comedy', with its elements of clandestine marriage to a lower-class person that turns out to be non-transgressive (as in The Winter's Tale), and the rediscovery of a person at the end who was thought lost and has special knowledge to contribute, as with Perdita and also with Marina in Pericles. Henry VIII is in this pattern: Henry marries Anne secretly with Wolsey providing the necessary paternal disapproval of the marriage. McMullan points out the parallels between Henry VIII and Plautus's Amphitryo which tells the story of Jupiter (in the likeness of Amphitryo) sleeping with Alcmena (Amphitryo's husband) and their offspring being the prodigy Hercules. In Henry VIII Henry, initially in disguise, sleeps with Anne and produces the prodigy Elizabeth. There was in fact a tradition of protestant reworkings of Roman comedy, and Henry VIII would have been seen in this light and not as loosely episodic, and many have since claimed. Clandestine sex and marriage is made proper by the outcome (Elizabeth), but--and this is the subversion--it remains tainted by the impropriety. The 'siege' of the penultimate scene is crucial: the specificities of places and clothing speak of 1613 not 1533. This linking of the Henrician to the Jacobean world "creates a dramatic space within which the outcome of the English Reformation is still very much at stake" (pp. 134-46). This extended and highly persuasive reading of the play show McMullan's extraordinary range of historical knowledge couple to an exemplary literary and dramatic sense; the quality of this interpretative work is much higher than in Jay Halio's Oxford Shakespeare edition of the play published in 1999.

    Concerning the textual history of the play, McMullan avoids the conventional (and currently controversial) terminology. Rather than calling the copy for the Folio text (our only authority) a promptbook, or even a playbook, he calls it "a score for a stage play" (p. 149). The F copy seems to have been scribal since it uses "ha's'" for "has" and round brackets, neither of which is a Shakespeare or Fletcher habit, and the long stage directions are not in theatre-speak but often drawn directly from Holinshed. McMullan demurs from the view of William Montgomery (for the 1986 Oxford Complete Works) that the direction "Trumpets, Sennet, and Cornets" (2.4.0.1) indicates duplication derived from theatrical annotation; rather, McMullan thinks F's copy might have been, but not necessarily was, used as "prompt copy" (p. 155). McMullan provides a useful table showing the order of composition (with folios and act.sc.line numbers keyed to his own edition) by compositors B and I (called B and A in Foakes's Arden² edition, but with the same sections assigned to each) and records that Charlton Hinman's machine collation showed no significant press variants for this play (p. 157). Regarding modernization, McMullan decided reluctantly to retain capitalization aristocratic of titles because it reduces confusion for the modern reader, if a little old-fashioned and not like the historians' practice. Indeed, for the purpose of modernization, what counts as archaism can be culturally conditioned: "comptroller" is not considered archaic by Americans and McMullan leaves it alone in the text.

    Discussing how the playwrights used (or, in McMullan's phrasing, "read") Holinshed, McMullan again uses the term "interpellate", but this time he glosses it to mean: "call[ing] into being in an apparently natural but in fact constrained way". McMullan says the play does this to the individual audience member, but the syntax obscures that McMullan is in fact making a substantial Althusserian claim that the play calls the audience into being rather than, say, the audience calling the play into being by creating a market for it or by giving it their attention (p. 166). McMullan reports that the Shakespeare scenes follow source (usually Holinshed) more closely than the Fletcher scenes, which interweave more texts (including the Bible); frequently there is subtle subversion as the material is used. When fallen Wolsey speaks of himself as like a little boy whose swam out of his depth buoyed up by "bladders" (3.2.360), some of the audience would be reminded of the prose chronicles which likened Wolsey himself to an inflated bladder fit to burst (p. 172). N. W. Bawcutt's publication of Henry Herbert's records showed that The Birth of Merlin was new in 1622 and hence not, as sometimes though, a source for this play. McMullan explains the unobvious suspicion that some of Henry VIII is not Shakespearian--it is after all in the 1623 Folio--by pointing out that around the same time (the mid-1960s) that E. A. J. Honigmann showed the inherent instability of the Shakespearian texts, G. E. Bentley showed that collaboration was the norm. Since the Stationers' Register indicates that Shakespeare and Fletcher collaborated on Cardenio, and the quarto of The Two Noble Kinsmen names them both as its authors, Henry VIII fell under suspicion. McMullan describes the Stationers' Register as the place "in which all plays to be printed were registered" (p. 186), but Peter Blayney has shown that this is not the case (Blayney 1997, 402-04). McMullan follows Jonathan Hope's division of the work--Shakespeare wrote 1.1, 1.2, 2.3, 2.4, 3.2a, and 5.1 and Fletcher the rest--and makes a case for Hope's "socio-historical linguistic" solution of the old stylometric problem that compositorial or scribal intervention in a printed text might be the origin of preferences (such as "ye" for "you" and "'em"  for "them") that have traditionally been the means of attribution. Hope uses alternates that scribes and copyists are unlikely to have interfered with, such as relative markers "that" and "which", and auxiliary "do". Of course, as McMullan admits, this is still not perfect: people change their styles over their lifetimes, and in different contexts (such as when collaborating). and, as Jeffrey Masten notes, all drama is impersonation of others' styles of speaking anyway (pp. 187-95).

    The Arden³ policy of marking with an asterisk footnotes which discuss deviations from copy is most useful for comparing editorial practice. With only one authoritative early text (F), McMullan's work with the text is naturally confined to creative emendation of what he thinks is erroneous. Mostly his choices are the same as in the Oxford Complete Works of 1986 and in Jay L. Halio's 1999 Oxford Shakespeare edition. At 1.1.219 and 2.1.20 McMullan names the accused plotter as "Gilbert Park" where F has "Gilbert Pecke" because McMullan thinks the underlying copy read "Perke", since this is what Holinshed and Hall have. (This was not really the man's name but a misreading of his occupation, 'clerk'). Halio thought so too and printed "Perke", but the Oxford Complete Works modernized this by dropping the final 'e' to make "Perk". McMullan thinks this is just a variant spelling of "Park", which he uses. Where McMullan is confident that Shakespeare is following his source, he uses it to emend. Thus at 1.1.218, 1.2.162, and 2.1.20 McMullan names "John de la Court" where F has (and both Oxfords kept) "Iohn de la Car"; the Surveyor's testimony comes from Holinshed and its spelling is thus preferred. Where it looks like the dramatists misread Holinshed, McMullan is happy to respect their intention (to follow Holinshed) rather than their act. Thus at 1.2.147-8 McMullan (like the two Oxfords) prints "Nicholas Hopkins. KING What was that Hopkins?" where F has "Nicholas Henton. | Kin. What was that Henton?" because the dramatists confused the place he came from (Henton) with his name. C. J. Sisson, however, argued that the man could also be called Nicholas Henton precisely because he came from Henton (Sisson 1956, 99). At 1.2.164 McMullan prints "Whom after, under the confession's seal" where F has "Whom after vnder the Commissions Seale". The sense demands the word 'confession' and McMullan thinks the four uses of the word 'commission(s)' previously in this scene caused the error, although he does not mention that compositor I set all four of them and is the presumptive cause; McMullan's table of the order of setting helps a reader discover this for himself. Where punctuation strongly effects meaning, McMullan is prepared to be bold. At 3.1.21 he has Katherine comment on the news that two cardinals are coming to see her: "I do not like their coming. Now I think on't, | They should be good men, their affairs as righteous" for F's "I doe not like their comming; now I think on't, | They should bee good men, their affaires as righteous". Halio followed Capell and cited Sisson's argument (Sisson 1956, 100-01) that if one follows F's stop after "coming", Katherine first says she does not like their coming, then reconsiders since they are after all cardinals, whereas the sense should be 'I don't like this, now I think about it'. Halio and the Oxford Complete Works used a comma after "coming" to avoid a hard stop. McMullan argues that precisely this strange emotional shift ('I don't like this. Oh well, I suppose it is alright since they are cardinals') is right for Katherine at this moment and foreshadows her succumbing to their pressure.

    McMullan's choices regarding stage directions are clearly informed by an understanding of the fluidity of the early modern stage and he is prepared to stretch logic to avoid straitjacketing the text. At 5.1.157 the Folio has "Enter Olde Lady. | Gent within. Come backe". For "dramatic economy" McMullan makes this interior gentleman be Lovell who has anyway to come on to be addressed by the king 12 lines later. So, McMullan prints "Enter Old Lady[; LOVELL follows.] | LOVELL (within) Come back!", which produces the mild absurdity of Lovell being onstage (he is included in the entrance direction) and yet speaking 'within'. Surely the solution was to have Lovell speak offstage and then enter. Likewise, for Henry and Butts to spy on the privy council McMullan prints "Enter the KING and BUTTS at a window above" (5.2.18sd) and for Henry's intrusion on the main stage "Enter KING, frowning on them" (5.2.147sd). Between these two entrance directions McMullan prints no exit direction, but rather footnotes that once the curtain is closed above--fully, not partly as other editors have it--the king can come down at any time ready for his surprise entrance to the council chamber. It surprises the privy councillors and the theatre audience too, since they thought he was above. One can see the point of this arrangement, but it is not theatrically consistent to have two entrances directions with no intervening exit, leaving the explanation to the footnote; those using McMullan's script need to know what he wants from his stage directions. Finally in this list of objections, McMullan prints "Do you take the court for Parish Garden?" (5.3.2) where F has "doe you take the Court for Parish Garden". Why not modernize to 'Paris Garden', unless one thinks that a speech impediment is being indicated by the unusual spelling? McMullan's footnote explains Paris Garden as the a bull- and bear-baiting arena near the Globe (otherwise known as the Beargarden), without mentioning that it could just mean the Liberty and Manor of Paris Garden where, in the early sixteenth century, public bearbaiting was held without an arena. Indeed, the Beargarden had the alternative name of Paris Garden precisely because of the association of this area of park with bear baiting, and not because it was situated in the Liberty and Manor of Paris Garden (which, in fact, the Beargarden was not). Bull baiting, on the other hand, was just a part of butchery, not a sport, and there is no reason to think it was done in front of spectators.

    In contrast to McMullan's outstanding Arden³ edition of Henry VIII is Edward Burns's edition of Henry VI Part One for the same series, the introduction to which is about half the length. It is traditional when editing a marginal play to make a case for its being more important than the reader might otherwise think. McMullan did this brilliantly, but Burns does not even try. One can also compare Burns's handling of the problem of multi-authored writing of a play for which we have only the Folio text with McMullan's thoroughly theorized attempt at the same. Of the copy for the Folio, Burns thinks that Heminges and Condell probably had 30-year old papers to work from because he is convinced the play had not be revived since its first performance in the early 1590s. (That the epilogue to Henry 5 refers to the events of Henry 6's reign, "Which oft our stage hath shown", makes revival of the first tetralogy to run alongside the second somewhat likely, as does the reprinting of The Contention of York and Lancaster/2H6 and Richard Duke of York/3H6 in 1600.) Burns thinks the reference to a Talbot play in Nashe's Piers Penniless is to 1 Henry 6, and that it is also the "harey the vj" which Henslowe records as new at the Rose on 3 March 1592; hence the play is a prequel to what we now call 2 Henry 6 and 3 Henry 6. As such, it ironizes the known outcomes contained in Parts 2 and Parts 3 by showing the grand ideas by which these 'heroes' lived their lives, and yet the events of Parts 2 and 3 come to pass. As with McMullan's reading, Burns thinks his play "an ironic meditation on what history is, and as such it constantly exposes the gratuitousness of the signs and symbols which allows to think we know history" (p. 6)

    Burns notes that refurbishment of the Rose theatre in 1592 was just before 1 Henry 6 premiered and the addition of a stage cover implies that the stage was permanent, whereas hitherto stages may have tended to be temporary to permit bear-baiting. In fact, there is no reason to suppose that animal baiting took place at any playhouse prior to the opening of the Hope in 1614 (Brownstein 1979). As for the suiting of play to venue, Burns notes that Titus Andronicus (which also played at the Rose) and 1 Henry 6 both open with a funeral procession, which suits the Rose's wide shallow stage. Indeed, Burns makes rather contentious assertions about how the stage shape at the Rose influenced the way the audience was addressed and where actors stood, and he cites John Astington's work on the Roxana and Messalina pictures (Astington 1990) as supporting his view that the Rose-style tapered stage was "a more standard stage shape, with a longer history, than had been assumed" (p. 12n1). In fact, Astington's articles comes to virtually the opposite conclusion, that these pictures are virtually useless as evidence for contemporary theatres. Theatre history is not Burns's long suit, relying as he does on Christine Eccles book on the Rose which is widely considered to be of little value (Eccles 1990). Burns makes the surprising claim that a company patron such as Lord Strange would "meet financial losses" incurred by the company (p. 18)--one would like to see the evidence for this--and asserts that "Southwark . . . was exempt from control by the city", which is not quite right. Liberties such as The Clink were exempt, because that is what being a 'liberty' meant, but these could be anywhere including inside the city walls. Burns is not quite in control of his adjectives: the 'end of an era' anxiety in the play, he argues, mirrored the same anxiety regarding "the loomingly predictable end of Elizabeth's reign" (p. 22). He means "looming and predictable", since one can hardly claim that it was predictable in a looming way as distinct from other ways of being predictable. Burns thinks Holinshed was used largely "as a quarry for juicy bits about Joan Puzel" (p. 22n1) which is rather too colloquial and moreover a bad metaphor: quarries are notoriously dry places.

    Burns is better when lining up binary opposites. In his reading, the play alters its sources to make a binary of fighters Joan Puzel and Talbot, and this segues into the binarism of Gloucester and York's struggle to control the child king and the country, which links the end of this play to the beginning of The Contention of York and Lancaster/2 Henry 6. The historical Joan always called herself "Jeanne la Pucelle", and 'pucelle' means nubile but also 'whore', especially when spelt 'puzel'. As a gender transgressor, Joan is also a puzel with a pizzle (penis) and a puzzle; she is not a single character but an embodied self-contradiction of saint and a witch (pp. 23-27). From this Burns's makes an excellent 8-term homology: French is to English as Catholic is to Protestant as Magical is to Rational as Female is to Male. This neat pattern is somewhat disturbed by Burns's assertion that the male motif of history is primarily the broken or constrained body (p. 39) which I would rather have thought was a Catholic fixation. As Burns notes, Joan's claim to be pregnant links her to the pregnant virgin Mary and her cousin Elizabeth (of the Visitation), and Talbot's opposition to her is Protestant opposition to the Catholic cult of the virgin. Of course, the historical Talbot was as Catholic as Joan. When Burns decides to survey the critical debate regarding a particular point he tends to produce big footnotes (for example 48n1 and 72n1) occupying as much as two-thirds of a page; these would be better integrated into the main text or else thinned.

    The red and white rose material in 1 Henry 6 is not present in 2 Henry 6 or 3 Henry 6 and this itself is evidence that 1 Henry 6 was written later; were it written first we should expect this material to be followed up in the later plays. However, the red/white rose distinction does appear in the opening-scene stage directions of the quartos of Richard Duke of York/3 Henry 6 ("with white [later red] roses in their hats") but not the Folio text of 3 Henry 6, which could be evidence that the quarto represents Richard Duke of York/3 Henry 6 as it came to be revised after 1 Henry 6 had been rewritten as a prequel, and that the Folio text was printed from authorial papers representing the play as it was originally conceived. Burns observes that no director has presented 1 Henry 6 on its own; rather it is always part of a cycle and usually chopped around to suit the larger pattern. Burns's dating of the play takes the usual form of an argument derived from Robert Greene's Groatsworth of Wit (published September 1592) allusion to a line in Richard Duke of York/3 Henry 6. If we accept that 1 Henry 6 was "new" at Henslowe's Rose on 3 March 1592 (because Henslowe's Diary labels it "ne") it cannot have been written before The Contention of York and Lancaster/2 Henry 6 and Richard Duke of York/3 Henry 6 if the last of these is to be available for Greene's allusion (which depends on knowledge gained by public performance) in September 1592. The six months between March and September are not enough time for all three plays to been written in the order 1, 2, 3 Henry 6, so 2 Henry 6 and 3 Henry 6 must already have been in existence (p. 70). Actually, the pressure of time is even greater than Burns imagines, since a plague closure from 23 June 1592 means that 3 Henry 6 would have to be in performance by then in order for Greene to be able to allude to it in September 1592. Supporting this conclusion, Burns notes the publication of The Contention of York and Lancaster/2 Henry 6 and Richard Duke of York/3 Henry 6 as a matched pair in 1594 and 1595, called "The First Part . . ." and "The True Tragedie . . ." on their titlepages.

    What if  the label "ne" in Henslowe's Diary doesn't mean 'new', or perhaps only new to his repertory because revived? Burns suggests that the higher cost of entrance for a "ne" play (reflected in higher income in the Diary) might be to cover the expenditure on a new license from the Master of the Revels, but I do not think we can suppose that audiences were concerned with the impresario's outgoings. If Henslowe's "ne" means only revival, we could imagine that 1, 2, and 3 Henry 6 were written as a trilogy, with 1 Henry 6 subsequently revised to make it performable on its own. However, as Burns observes, there was a strong a tradition of two-part plays, but few examples of trilogies. Also against the theory that the 1, 2, 3 Henry 6 was conceived as a trilogy is the fact that the preparations for the printing of the 1623 Folio including the first-time entry in the Stationers' Register on 8 November 1622 of "The thirde parte of Henry the sixte". This cannot be what we now call 3 Henry 6 because this had already been printed in 1595 as Richard Duke of York, so it is most likely 1 Henry 6 considered the third part of the series in order of composition, not in order of historical events. Millington entered "The firste parte of the Contention . . ." in the Register on 12 March 1594 and then printed The Contention of York and Lancaster/2 Henry 6 (1594) and Richard Duke of York/3 Henry 6 (1595), so presumably the single Register entry covered this pair of plays. On 19 April 1602 Millington transferred his rights in "The first and Second parte of HENRY the VJ^t[h]" to Pavier who later printed The Contention of York and Lancaster/2 Henry 6 and Richard Duke of York/3 Henry 6. All this suggests that what we now call 2 Henry 6 was originally the first part of a two-parter and 3 Henry 6 was its completion. Thus, as Burns writes, when "The thirde parte of Henry the sixte" was entered in the Register on 8 November 1622, it was really the prequel 1 Henry 6 (p. 72).

    As McMullan observed regarding Henry 8, those who like 1 Henry 6 tend to see it as by Shakespeare and as part of a planned sequence, those who do not see it as by him and others. In the former group was the Arden² editor Andrew Cairncross--presumably this is why he edited all three Henry 6 plays--while in the latter group was Arden¹ editor H. C. Hart. Burns agrees with Gary Taylor's division of 1 Henry 6 into Shakespeare and Nashe sections, and Taylor's view that several others no longer identifiable also had hands in it. Burns reads Robert Greene's attack as being about Shakespeare as a cheater in collaboration, someone who passes off as his own work containing others' "feathers", and points that the wider context in the Groatsworth of Wit is a story of a player rescuing a down-at-heel writer, which player is "both his saviour and a kind of devil figure, drawing him into further artistic degradation" (p. 79). How collaborators might parcel out a play is not clear: by act, by scene, or smaller units? There is evidence that a writer might be responsible for individual speeches, but Burns prefers to think of the units of authorial division in 1 Henry 6 being "strands of rhetorical action" (p. 81). There are, Burns asserts, two discernible strands, perhaps by different authors: the English/French conflict, and the Shakespearian breakdown of English unity (p. 83).

    Burns's characterizes his editing as "a broadly non-interventionist approach to the punctuation of F" (p. 90) and he avoids brackets because they are "inhibiting for actors"; one is tempted to respond that inhibited actors should learn punctuation. Sometimes Burns retains F's ambiguous punctuation because ambiguity is the point and likewise he does not always fix what others have seen as F's failure to supply necessary exit stage directions. For example, Burns has the gaolers stay on stage during Mortimer's death-chair interview with Richard in 2.5. For asides, Burns has invented his own editorial convention (some are "to the audience" and others "to him/herself" p. 98), which novelty could be accused of anachronism since self-communion appears to be a proscenium-arch technique impossible on the Elizabethan thrust stage where the audience can never be ignored. Burns thinks that Gary Taylor showed that "plays for the professional stage were not split into formal act divisions, nor performed with gaps between the acts, before about 1610" and that "Only academic and court performances bothered to follow classical precedent by splitting plays into five acts . . . " (p. 101) This is wrong: the professional indoor hall theatres always used act divisions and this practice spread to the open-air amphitheatre stages after about 1609, as Taylor argued.

    Burns's decisions regarding the play's famous cruces are mostly conventional, but he appears to be unaware of the important principle of praestat difficilior lectio ('prefer the more difficult reading') when faced with exotic words in his copy. Burns's prints "A base villain, to win the Dolphin's grace, | Thrust Talbot with a spear into the back" (1.1.137) where F has "A base Wallon", which editors usually modernize to Walloon. Burns argues that the ethnicity of the assailant is not in the sources and the Walloons were on the English side, and that the error is a likely misreading of a sequence of minims. But the point is that one should trust one's copy in the case of an unusual word since compositors tend not to invent exotica. Likewise at 1.3.29 Burns prints "How now, ambitious Humphrey, what means this?". where F has "How now ambitious Vmpheir, what meanes this?". Burns calls "bizarre" the Oxford Complete Works's choice of "vizier" for F's "Vmpheir", but since Gloucester, as Protector, is Viceroy, "vizier" is a good abusive word to hurl at him. Much more serious than individual choices for emendation is the mess that Burns makes of scene divisions. At 4.3.53 Burns decides not to start a new scene, something about which his predecessor Cairncross agreed but did not act upon, mentioning it in a note but preserving the traditional break. At this point Burns starts two numbering sequences, continuing with 4.3.54 in the marginal numbering and the running-titles, but also adding a marginal marker "[4.4]" to show that other editors start a new scene here. To match this marginal "[4.4]" Burns starts a second marginal numbering (4.4.1 onwards) using steps of 10 rather than 5, which runs in the same column as his own numbering system (4.3.54 onwards) until the real end of the scene at his 4.3.99. Then 4.4 runs normally until its line 55 when again Burns decides that a scene conventional scene break (after Talbot's "And soul with soul from France to heaven fly") is wrong. So, again, he adds the conventional number "[4.6]" in the marginal column while continuing his own number (4.4.56 onwards) in the running-titles and two numbering systems in the marginal columns. The same thing happens again at Burns's 4.4.112 (Talbot's "And, commendable proved, let's die in pride"), where other editors generally start 4.7, so again there are marginal marks for the traditional break, the traditional numbering, and Burns's numbering, while the running-titles follow Burns's numbering. Distinguishing the two line-numbering systems is not especially difficult once one realizes what is going on--although I can find the practice explained nowhere in the book--because the traditional ones are in square brackets and mostly (but oddly not for the "[10]", "[20]", and "[30]" of the traditional 4.7) in a smaller typeface. But distinguishing the two act- and scene-numbering labels is tricky: both are in square brackets, with the traditional ones being (counterintuitively) in a bolder typeface. The only point of retaining a traditional numbering system at all is ease of reference, and the multiple numbering systems used here to not achieve that. Indeed, in the absence of an explanatory note, readers may well assume that the entire numbering system is simply erroneous. An editor who believes he has corrected his copy's faulty breaks should adopt his correction entirely, not try to run two numbering systems in parallel.

    A problem of equal magnitude occurs with Burns's idiosyncratic naming of characters. In his Appendix One, Burns defends his naming of "Puzel" and "Dolphin" instead of 'Pucelle' and 'Dauphin' (the common modernizations), because, like Churchill's deliberate "Naazi" pronunciation, these mark the refusal of the English to speak a despised foreign language properly. Burns argues for the spelling "Dolphin" instead of the more usually modernized 'Dauphin' on the grounds that the latter is the French word for the aquatic mammal, and he wants to use a word that will invoke the playful resonances understood by the original audience as well as its connections with beast fable and heraldry: it was the symbol of the Comte de Vienne, a title sold to Philip 4 in 1349 and thereafter given to the heir to the French throne. For Joan's character, the big difference in spelling her name is between medial 'c' and medial 'z'. Taylor argued that the modernized form turns 'c' or 'z' into 'c' (so "Pucelle"), but Burns argues that since 'c' and 'ss' make the same sound, this would make Talbot's "Puzel or pussel" (1.4.106) into the nonsensical "Pucelle or pucelle", which is a repetition of one sound when of course it was spoken with a distinction (because it could mean virgin or whore) in Shakespeare's time. Burns thinks that keeping the 'z' also helps to make it sound like 'puzzle', and Joan is of course a puzzle to Talbot. A reviewer need not raise the obvious objections to Burns's reasoning, since the Arden³ general editors do it themselves in a remarkable note distancing themselves from Burns's choices (pp. 294-6). "Puzel", they note, is the minority form in F ("Pucel(l)" occurs twice as often) and using it "deprives the French characters of an intelligible French epithet for their saviour, Joan 'the Maid'"; instead they have to use a derogatory English term. The French word 'pucelle' (virgin) has the advantage to an editor that it could be used derogatorily or straight, so both sides could use it with their own meanings. Burns's rendering "Puzel or pussel" has no contrast: both are derogatory. Regarding Dolphin/Dauphin, Burns's use of the former in a modernized text makes the sounding of the 'l' compulsory, whereas of course the point is that it is optional, as in Walter/Water; Ralph/Rafe, and salvage/savage. If 'Dauphin' had been used by Burns, the editors remark, then Talbot's "Dauphin or dogfish" (1.4.107) would have been the only time it is pronounced "Dolphin", and hence an appropriately scathing comparison. The freedom to individual editors granted by the Arden² general editors (and apparently Stanley Wells is equally liberal with the Oxford Shakespeare) obviously allows the rejection of extraordinarily wise counsel.

    For the Oxford Shakespeare, John Jowett's edition of Richard 3 handles with wisdom and impeccable literary-dramatic style the most difficult editing task in the canon (). Jowett's edition is based on the 1597 quarto (Q1), "the text that seems closest to the play as it would have been staged" (p. 3). Before the formation of the Chamberlain's men in 1594, Shakespeare appears to have been with Strange's men, whose patron Ferdinando Lord Strange was descended from the play's Lord Stanley, and Shakespeare has altered Thomas More's Stanley to present him more favourably; was this done because Shakespeare was one of Strange's men? On the other hand, the ancestors of the Earl of Pembroke, who had his own players, also get adjusted favourably in Richard 3, so perhaps Shakespeare wanted to please two potential patrons. Or did he please two patrons at two different times and both alterations got into the Folio? Jowett addresses the same problem of distinguishing synchronic from diachronic evidence in a footnote which usefully summarizes what is known of Shakespeare's activities before joining Chamberlain's men: the performance of "harey the vj" at the Rose on 3 March 1592 suggests he was with Strange's (who were using the Rose at the time); the titlepage of the 1594 quarto of A Shrew names the players as Pembroke's men (but it might not be Shakespeare's play); and the titlepage of the 1594 quarto of Titus Andronicus names the players as "the Earle of Darbie, Earle of Pembrooke, and Earle of Sussex their Seruants", which could be a joint performance or else a "summary stage history" (p. 7n1). Jowett thinks that Richard 3 was written "as if for Strange's" and then "given finishing touches towards its close that make it suitable for the new Pembroke company", which makes 1592 the likeliest year of composition, it being when Pembroke's company was formed. The play was not  finished in time before the plague closure of 23 June 1592 (else Strange's men would have played it at the Rose, and Henslowe's Diary shows they did not), and it was as he was finishing it during the closure Shakespeare added the Pembroke material (pp. 7-8).

    Concerning Shakespeare's use of his sources, Jowett observes that by making Richard 3 the epitome of evil, his being done away with allows a line to be drawn under the factious middle ages and the Tudor age of Richmond (Henry 7) can begin with past strife finally buried. More's account of Richard 3 ends before his death and Jowett wonders if he started to see Henry 8 as a bad king and so was reluctant to bring to the story to its end because this would amount to "developing a specifically pro-Tudor polemic" (p. 15). Jowett finds a tension between the teleological and analogical readings of the events of Richard 3: the former sees the overthrow of Richard as inevitable, but the analogical reading (which reads the events as metaphors for the present) comes up hard against the contemporary injunction not to rebel against even a tyrant, as the Homily against Disobedience insists. Indeed, Richmond's claim to the throne was weaker than Richard's, and some political thinkers argued that while tyrants had to be endured, usurpers (like Richmond) could be overthrown. More reversed the crookedness of Richard he found in his source (John Rous), making the left (sinister) shoulder higher than the right, and the tension between the play's two explanations of Richard's deformity (too much, or too little, gestation) match the tension between the teleological and analogical readings. Jowett reads Richard 3 as a revenge play in that Richard kills those responsible for the stabbing of Prince Edward at the Battle of Tewkesbury. Of course, he too was responsible for that stabbing; but in this genre then the revenger often does have his own crime for which he must be punished (32-38). Aspects of Jowett's close reading of the play are truly inspired. The young prince Edward's concerns with the documentary record regarding Caesar's building of the Tower of London in 3.1 mark him out as a proto-Humanist like More, and Richard's murder of the prince can be read as a delaying of Renaissance humanist culture in England. The dream of Clarence is like the dream of Lady More's wife in Sir Thomas More, a play in which Erasmus's visits to England are celebrated. This might suggest that the victory of Tudor Richmond marks the end of medieval period and the start of the humanist Renaissance in England, but on the other hand his son Henry 8 was eventually to lock More in the same tower and then execute him. The link between building construction and the passing on of knowledge about it is the word "edify", which has both meaning (as in "succeeding ages have re-edified" 3.1.71). Richard's "So wise so young, they say, do never live long" (3.1.79), said of the young prince Edward, sounds like determination to kill the truth that the prince embodies, but it simultaneous admits in its "they say" that oral transmission--what the prince has been talking about--makes it impossible to entirely silence shared knowledge (pp. 56-58).

    Jowett sees no reason to think that Q's economy of roles derives from the needs of touring, "for the main elements in the doubling pattern are already apparent in F". Q does reorder the ghosts so that Lady Anne's ghost does not enter immediately after princes leave, necessary because one of the princes (presumably, the elder, Edward) must double with Lady Anne. Dorset and Grey only become two characters after Grey is arrested. F is confused on this point, and Q resolves it in a surprising way: instead of separating the characters, it more strongly indicates in the early scenes that these are two names for one man. Once they are separated (in 2.4), "doubling of a more routine kind was no doubt followed" (75-76). Jowett observes that William Hogarth's picture of David Garrick as Richard awaking in his tent looks like a bedroom scene, which acts as a reminder of the onstage bedroom scene murders in this, Colley Cibber's, adaptation of the play (86). In fact, it is likely that the original staging would also have made the bed/tent connection because of the nature of the stage property use for either (Egan 2000).

    Jowett is a world-leading bibliographer with a gift for explaining clearly and concisely what others obfuscate, and he deals with the play "In Print" in just 22 pages (pp. 110-32). Jowett's philosophical position on the difference between stage and page is a rather subtle nominalism: texts are only "representations of the play rather than the thing itself, and there is a real sense in which the play can exist only in representations of it". The play in its ideal form, then, does not really exist, but his is not the now-conventional materialism since Jowett does not privilege the early printings: even possession of the early manuscripts "would offer not so much 'the play' as versions of it" (p. 110). Q1 was printed from a previously unprinted manuscript, and F by close attention to another unprinted manuscript; all other printings are derivative. Jowett takes the view expressed most articulately by Peter Blayney (Blayney 1997) that companies had their plays printed as a form of advertising, and in this case the printing of Richard 2 and Richard 3 in 1597 (after The Contention of York and Lancaster/2 Henry 6 and Richard Duke of York/3 Henry 6 in 1594-5) displayed that the Chamberlain's men's principal dramatist specialized in history plays. The Stationers' Registry entry for Richard 3 on 20 October 1597 showed that it had ecclesiastical authorization: William Barlow, under whose 'hand' the play was entered, was chaplain to John Whitgift, the archbishop of Canterbury, and another under whose 'hand' it was entered was Thomas Man, Warden of the Stationers' Company. In 1596 only 40% of books printed had ecclesiastical authority, but all potentially controversial ones were. Of Richard 3 Q1, sheets A-G were printed by Valentine Simmes, and sheets H-M by Peter Short. Each subsequent quarto was printed its predecessor, except sheets C and E-M of Q5 which were printed from Q3. The number of quarto reprints (Q6 appeared in 1622) shows how popular the reading text was. Although there is evidence of manuscript consultation for some of the quarto reprint, it is hard to say if the small changes in later quartos are authoritative. Q3's titlepage advertises that the contents are "Newly augmented", and it has some new stage directions and is of value for emending stage directions, speech prefixes, and occasionally dialogue. Jowett observes that Andrew Wise and Matthew Law printed their Shakespeare history plays quartos in linked groups (such as Law's pairing of Richard 2 and its historical sequel 1 Henry 4 in 1608) but does not fully draw out the implication of this for recent historicist scholarship. It is not straightforwardly true that the Folio's organization of material forced the history plays onto a Procrustean bed for which their sprawling generic and thematic material is not suited; the organization was already in evidence from the quarto printings.

    In his discussion of the copy for Q1 and F, Jowett introduces a useful innovation in nomenclature that greatly aids clarity. Instead of the usual longwinded formulae such as 'the manuscript underlying Q [or F]'--so clumsy that some omit the first part and wrongly imply that a printing is identical with its copy text--Jowett uses the shorthand MSQ and MSF. In fact, the copy for Folio was not simply a manuscript but rather alternated between Q3 and Q6, with manuscript passages not present in these interwoven into the text. These passages show that this manuscript was not the one used to make Q1 (which lack them), and it is also the source for hundreds of Folio readings which differ from whichever quarto (Q3 or Q6) being used at that point. Where later quartos had introduced error, the manuscript used to make F often restored the Q1 reading, which shows how much MSQ and MSF had in common. Rather that demonstrating these claims with detailed examples, Jowett collects in his Appendices A-D the supporting collation evidence. MSQ and MSF show no signs of conforming to the simple New Bibliographical categories of 'foul papers' and 'promptbook', but MSF is earlier than MSQ and probably represents the text before it came to the theatre. MSF is longer (for an already long play) than MSQ but lacks the "dramatically incisive" 'clock' passage in 4.2 and has lots of repetitive wordy stuff easily cut without harm. It is hard to see, Jowett points out, how this redundant material could have got into F by revision, but easy to see how it might have been cut to make Q. MSF is also more profligate with characters, such as an unnecessary daughter for Clarence. Like Q1's reordering of the ghosts (so one boy can be a prince and Lady Anne), cutting this daughter saves a boy actor. Q1 also lacks a Folio passage which links Richard 3 with Richard Duke of York/3 Henry 6, which is more likely to have been cut to make Q than added to make F since "The play is far more likely to have grown towards greater independence of the events in the Henry VI plays", especially as it proved (as the frequent reprinting shows) to be a much more popular play. Q also saves personnel by conflating characters such as the Keeper of the Tower and Brackenbury; these make little difference to the size of the cast but increase dramatic intelligibility at the cost of historical accuracy. Again, it is less likely that characters in MSQ were split to make a more historically accurate MSF; in general F is historically more accurate than Q, and it has signs of authorial confusion over names created by punctuation in the source material.

    What was this MSF? Apparently is was not in Shakespeare's hand since it departs from his lexical preferences in using "ay" where Shakespeare and Q1 preferred "yea", "prithee" where he and Q1 preferred "pray thee" (as E. A. J. Honigmann noted) and "which" where he and Q1 preferred "that" (as Jonathan Hope noted). These preferences and the act/scene divisions suggest scribal transcript. MSQ does not seem like the New Bibliographers' idea of a 'promptbook', but Jowett agrees with the New Textualists that the old idea of a promptbook was too narrow and that such features as imperfect stage directions could be permitted in the theatre document. In any case, the cramped printing of Q1 (from which MSQ is conceptually extrapolated) might have necessitated throwing away exit directions which we notice it lacking. Jowett thinks that Q1 probably is not a memorial reconstruction (the usual theory being that the company found themselves on tour without their playbook), and quotes the moment in scene 3.2 (F4v of the quarto, TLN 1912-21 of the Folio) where Hastings whispers in a priest's ear. Here F and Q1 suddenly diverge right in the middle of scene otherwise well reported, and furthermore the relationship of the F/Q versions here is much more like revision that garbling: a clash of idiomatic language in F is avoided by alteration to the syntax in Q. Moreover, Q's differences save a speaking role, and put together these are "signs that the dialogue has been consciously modified" (p. 125). Admittedly, there are a few single-word slips of the kind best explained by Memorial Reconstruction, but Jowett points out that any act of transcriptions involves memory, and a scribe could have made these slips during copying. Most damagingly, there are is a a pattern of variation in the naming of Stanley/Derby which could not survive Memorial Reconstruction, as Jowett argues more fully in a note reviewed below.

    While unhappy with the classificatory criteria of New Bibliography, Jowett does believe we can distinguish texts nearer or further from performance and on this rests his preference for Q1, newly freed from the stigma of Memorial Reconstruction. Jowett acknowledges that were one to follow slavishly the Oxford Complete Works's logic of preferring performance over authorial writing one might have to accept a crude adaptation over its original, and drawing a line somewhere before this is a matter of "choice rather than law" (p. 129). Unlike later plays, the 'theatricalizing' of the Richard 3 script might not have involved Shakespeare because he was not yet a sharer in a playing company, but still Q1 is to be preferred because it "largely retains authorial texture", it has not been so corrupted as to be inferior to F, and indeed it has some preferable verbal variant readings which might be authorial. Once an editor has settled on Q1 as the control text, the difficult decisions concern how bad something has to be before it should be corrected; the dividing line is not what the author wanted (as it would be if one worked from F) but what could have passed in the theatre without 'correction'. Most of the emendations of Q1 used by Jowett are readings from F (on the hypothesis that the faulty transmission lost the authorial reading and we can recover it from there), but some will be readings which first appeared in later quartos because their printers (using earlier printings as their copy) spotted the manifest mistakes and fixed them. Rather than stake out all the arguments for each genuine Q1/F variant choice in his edition, Jowett prefers to make limited comment in his notes and "recognize that many alternative readings can legitimately coexist". Editing with Q1 as the control text makes for a play less melodramatic and less connected with the other histories (because less connected with the real history) and more of a free-standing "psycho-political drama about Richard's rise and fall", which is what the play was becoming in the theatre in the 1590s (pp. 131-32).

    The main innovation of this edition is a double collation between the body text and the commentary. The first records this edition's departures from Q1 and the second records records readings from F not adopted in this text; emendations by other editors are recorded only in the commentary. In Jowett's opinion many of the Folio readings can be considered valid alternative readings. In the Folio collation, wherever the rejected reading in F originated in a quarto reprint (Q2-Q6) this fact is recorded but only up the quarto that was the copy for F at that point (Q3 or Q6). In other words, where a quarto was the first occurrence of a reading which later appeared in F, that quarto is named unless it is later than the copy used for F, in which case the quarto in question could not be the cause of F's reading. In the Q1 collation, if F also has the rejected reading this fact is explicitly confirmed except in the two places (3.1.0-144 and 5.4.28-end) where F was printed from Q without consultation of the manuscript, for which one may assume that F has the same reading. Alterations to stage directions are marked by broken brackets only where they are significantly disputable, otherwise they are just recorded in the Q1 collation, but a marker about address printed after a speech prefix (such as "to Margaret" or "Aside") may be assumed to be editorial and if, unusually, Q1 has them also (as with the ghost scene) this is recorded (pp. 134-135). Given Jowett's procedures for a two-text play, there is little point in a reviewer going through the editor's choices regarding particular cruces because most of the decisions will have been made for Jowett his choice of copy, whereas for a single-text play one is thinking about the correction of manifest error and must be more inventive. An interesting mixture of editorial and critical impulses is registered in Jowett's preservation of Richard's claim that Richmond's army were "Long kept in Bretagne at our mother's cost" (5.5.53) rather than changing this to the historically correct "brother's cost", since the Duke of Burgundy, his brother-in-law, armed Richmond. The error originates in Holinshed and Jowett keeps it (as Shakespeare decided to) because it is a "Freudian slip" (p. 66n2) prompted by the desertion of the women Richard had previously been able to control with his rhetoric. One tiny flaw which could be corrected in reprinting is that the running-headers on pages 244 and 245 wrongly give the act and scene label 3.1 (it should be 3.2).

    Jowett prints seven appendices. Appendix A provides the texts of passages which were first printed in 1623. Most of these are the same as the ones so designated in the Oxford Complete Works, which agreed that they were "probably deleted for stage performance", but two of the short ones (Oxford Complete Works's 'Additional Passage B' beginning after 1.3.166 "RICHARD GLOUCESTER Wert thou not banished on pain of death? | QUEEN MARGARET I was, but I do find more pain in banishment | Than death can yield me here by my abode" and 'Additional Passage F' after 3.1.170 "BUCKINGHAM And summon him tomorrow to the Tower | To sit about the coronation") Jowett simply includes in the F collation as variant material not used. This shows the advantage of Jowett's double collation in cutting the number of 'additional passages' he has to enumerate separately in his already full appendices. Conversely, there are four F-only passages (comprising 23 lines of speech) which in Oxford Complete Works were incorporated into the body of the main text which Jowett has freshly reassigned as 'missing from Q because cut for performance'--of course, were they missing from Q because its printers merely failed to follow MSQ, Jowett would have printed them--showing his continuation of the Oxford resistance to conflation and the increasing fragmentation into distinct versions begun by the Oxford Complete Works's decision to print two King Lear plays. Some such decisions are explained in the longer textual notes in Appendix D. Appendix B lists the variants between Q1 and Q2-6 where F (but no intervening quarto) restored the Q1 reading. This is useful because, except where the corrections are obvious, these confirm Q1 readings since they must be caused by MSQ and MSF agreeing on these words. Jowett divides the list into sections according to whether Q3 or Q6 was the copy for F, as determined by Gary Taylor using the evidence of incidentals (Wells et al. 1987, 229-30), and naturally does not show the variants that appeared after the quarto which is copy for F at that point (Q3 or Q6). Appendix C records Jowett's changes to Q1's lineation, and here he mentions an Oxford typographical convention (common to the Arden³ but not the New Cambridge) whereby the editor indicates his view on whether a passage is prose or verse by placement of the speech prefix on the same line as the speech or on its own line above the speech respectively. Like typographical conventions regarding voiced and unvoiced suffixal '-ed', this sort of thing should be stated more often by editors; even graduate student readers can be unaware of them. Another convention (begun by Edward Capell) is to push right half lines which complete another's metrical unit, and where there is more than one way to do this (as with the cut-and-thrust between Richard and Lady Anne at 1.2.178-88), Jowett does not impose the flush-right convention.

    In Appendix D ('Longer Textual Notes') Jowett discusses such matters on act/scene breaks, for which he largely follows the Oxford Complete Works except that when Richmond goes into his tent in Act 5 (the beginning of the 'simultaneous staging' of both camps, which allows the ghosts to speak to Richard and Richmond), Jowett continues without a scene break. The division of source material makes Jowett think that five-act structure might have been in Shakespeare's mind even though the open-air theatres did not mark intervals until after 1609. For the Derby/Stanley shifts in stage direction and speech prefixes Jowett summarizes his Notes & Queries essay (reviewed below). Jowett explains the Dorset/Grey confusion in the early texts: initially these were two names for what Shakespeare planned as one character, but later he decided to have one son of the queen die by Richard and another survive. Thus in the latter half of F Shakespeare has them be different men, but the exigencies of performance made conflation of them desirable, at least as far as the actor is concerned (so, the two roles are doubled). Thus Jowett's surprising 'Persons of the Play' list describes Dorset and Grey as "treated as one figure in the early scenes". For his decision to have Richard in his opening soliloquy say "Plots have I laid inductious, dangerous," rather than Q3/F's "inductions", Jowett points out that "inductious" is a perfectly comprehensible Shakespearian coinage (along the same lines as 'conceptious' in Timon of Athens 4.3.188) which John Ford used in 1620. "Inductious" would not seem strange to us had not Q3 (1602) changed it to "inductions"; thus does editorial practice become self-validating as the 1602 alteration becomes traditional. Jowett discusses the oddness of Q1's labelling of the two orations by the Richmond and Richard to their respective armies, and records that John Dover Wilson noted that Richard's follows a strong exit-like couplet. Even though he remarks that they involve "a rather awkward transition from a small group of on-stage leaders to an 'army'" and "probably should be taken to address the theatre audience", Jowett does not cite Ralph Cohen's fine suggestion (Cohen 1998, 28) that in the preceding few lines Richard addresses the theatre audience as his army and, since they do not, cannot, follow him "hand in hand to hell" as he rushes off, he has to come back and "say more" than he has "inferred" (5.5.43). Appendix E prints relevant passages from Thomas More's History of Richard III, Appendix F reprints the "quick if slanted guide to the characters" given by the English Shakespeare Company in their The Wars of the Roses (1988-9), and Appendix G prints an A to Z of practitioners who have done significant work on the play since its first performance. The index is essentially a guide to the Commentary.

    Jill Levenson claims that her interest in Romeo and Juliet sprung from work with likeminded students who wanted "no part of its sentiment", and the splendid introduction to her Oxford Shakespeare edition of the play makes a powerful case for the play having political and social substrata inherited from its sources (Shakespeare 2000e). Levenson gives an amusing example of how rhetorical the sources novella could be: in Matteo Bandello's 1554 Novelle Juliet, on discovering who Romeo is, says "Now let us assume that he really loves me . . . should I not be reasonable and consider the fact that my father will never agree to it?" (p. 11). These source novella also "strain towards verisimilitude", giving explanations for such things as the use of rope ladders in Italy and Romeo's living arrangements in Mantua, and Levenson's fine survey of how these pre-texts shaped the play should disabuse students who, having seen the film Shakespeare in Love, think the play is semi-autobiographical. (A further misconception which animates Jenny Tiramani of the London Globe Theatre's clothing department: doublets were kept on at all times, not swung on and off like a modern jacket, despite Joseph Fiennes's usage.) Despite the problem of sentiment, Levenson deals in one section with 'Love, Death, and Adolescence', where she focuses on the psychosexual matter in the play with special attention to adolescent sexuality, transition to adulthood, and guilt about sex (pp. 16-30). While this is more than perfunctory, Levenson gets properly into her stride with the section on 'Patriarchy' (pp. 31-43) where she argues that the feud does the work of ideology, making everyone identify their allegiances--thus giving each an identity--and meld into a group. The family unit and the city-state unit come under stress, but endure. An important part of patriarchy is the control of masculine aggression, and Levenson argues that Elizabethans interested in the decorum of duelling would notice that the fights in Romeo and Juliet start properly but descend into chaos and accident. Shakespeare thus showed that the formalized rules governing violence, promulgated in duelling manuals, do not work (p. 36). At a different level of subversion, what Juliet's autonomy enacts--and she's much more active in this than Romeo--is a disruption of capitalism, since her clandestine marriage prevents Capulet's transmission of his wealth to a count, which would be a consolidation of it. I am not entirely sure why Levenson thinks this consolidation capitalist rather than feudal, and this part of the introduction the reviewer would liked to have seen see expanded (p. 40). Of course, whatever disruptions the lovers create are finally inconsequential; they are doomed to fail.

    In the section 'Style and Genre' (pp. 42-61) Levenson argues that the play consistently shows rhetoric failing to express the real conditions people experience; language, no matter how overblown, is inadequate. As a 'tragedy of romantic love' rather than of statesmanship the play was a significant innovation by Shakespeare, indeed virtually an oxymoron. But comic and tragic drama had never really been far apart, being made from the same sources and of the same length: "It was primarily the conclusion that made the difference". In Romeo and Juliet Shakespeare satisfied the demand for tragedy and comedy from different theatre patrons "with the blended essences of [the] two favourite genres" (pp. 51-52). This claim can be most usefully compared with Martin Wiggins's argument (Wiggins 2000, 102-22) that tragicomedy--another kind of blending, although Romeo and Juliet is not a tragicomedy--came into being around 1600. Also on matters formal, Levenson points out that the play put into the plot the essential elements of the sonnet tradition: the anguished lover, the unattainable lady, and the equating of love and war. The sonnet form in the 1580s and 1590s became the way in which political and financial desire was mediated: the lover (poet) seeking gifts from his woman (the patron), and in his unrequited love for Rosaline, Romeo begins the play as the archetypal Petrarchan lover. Levenson's section on 'Performance History (pp. 61-96) confirms that, beyond the titlepages of printed editions, no pre-Restoration performance is recorded. Shakespeare probably started Romeo and Juliet before the Chamberlain's men formed, so he did not know who would play it, and yet he was fairly demanding in his staging since the plays needs a balcony, a bed, and a tomb. Of course, one might not always get what one wants, and the first two quartos (1597 and 1599) are permissive in their stage directions, a feature no longer thought to be incompatible with their origins being documents used in the theatre. There are three moments when the locale changes without a clearing of the stage: in 1.4 the masquers walk around the stage to represent going to the Capulet ball; 2.1 starts in the street but becomes the Capulet orchard; and 4.4. starts in a room where preparations are being made for the wedding, but becomes Juliet's bedchamber. Levenson has rightly decided not to mark scene breaks at these points of locale-change. Surveying the stage history from the Restoration to the present, Levenson describes Thomas Otway's adaptation The History and Fall of Caius Marius which is set the story in Rome. The immediate political context of this adaptation was the Exclusion Crisis as Charles 2 tried to get his Catholic brother James barred from succession, and accordingly Otway's version moralizes against civil war. Correcting the common error that Colley Cibber originating the alteration, apparent even in the latest film version of the play, Levenson notes that Otway allowed the lovers a moment of conscious togetherness in the tomb before both dying. Quite properly, Levenson alludes to "soft-core pornographic" versions of the play without wasting time on such masturbatory aids as Troma Films's Tromeo and Juliet to which even some academics seem disturbingly drawn.

    Regarding the dating of the play, Levenson argues that this is not necessarily a singularity since the matter could have been reworked by Shakespeare until Q2 appeared in 1599 (pp. 96-103). The outer limits of composition are 1591 and 1596, but linguistic evidence suggests that it was first written probably in 1593 and came on the stage 1594-5. Q1's titlepage gives the performing company as "L. of Hunsdon his Seruants", which name they had only between 22 July 1596 and 17 April 1597, being the Lord Chamberlain's men before and after that. But this titlepage evidence does not preclude the possibility that the Lord Chamberlain's men acted it before 1596; by 1598 there were many allusions to the play. In a rare slip, Levenson writes that "The Shaxicon database on [sic] World Wide Web should help to refine the study of linguistic evidence for purposes of dating Shakespeare's plays: it charts the interrelation of rare words in Shakespeare's texts with contemporary works from around 1591 to 1616". If only Donald Foster had fulfilled his promise to publish his SHAXICON database on the Worldwide Web then others, including this reviewer, might be less sceptical of everything it is supposed to prove, including Shakespeare's authorship of the Funeral Elegy.

    In titling her section on the early printings of the play 'The Mobile Text' (pp. 103-25), Levenson puts herself with the New Textualists rather than the New Bibliographers. Q1 (1597) is, she notes, less than 80% of the length of Q2 (1599) and separating them are numerous variants; all seventeenth-century editions derived from Q2. Levenson treats these two printings as two witnesses to "distinct phases" in the play's sixteenth-century career. What she calls "millennial postmodern theory", which is sceptical of everything, rightly makes the play "part of a multivalent and dynamic process", and although they acknowledged the uncertainty at the heart of their work, the Enlightenment-inspired New Bibliographers nevertheless "misconstrued the randomness" of the textual evidence. (One might more charitably say that they did not see as much chaos as we postmoderns do.) At the end of twentieth century, as Levenson worked on the play, the New Bibliographical binaries (author/stage, good-quarto/bad-quarto, Memorial Reconstruction/foul-papers, promptbook/foul-papers, and touring/London) were breaking down. Books necessarily stabilize performance, but Levenson reminds the reader that our books are more stable than were theirs, which were often non-identical within a print-run. Q1 or Q2 were not entered in the Stationers' Register, a fact that longer excites suspicion, and collation of the 5 copies of Q1 shows no press variants,  which as Levenson points out, is not unusual in so small a sample. Q1 has fewer obvious errors than Q2, yet most bibliographers consider Q2 more authoritative.

    Sheets A-D of Q1 were printed by Danter, sheets E-K by Edward Allde (we can tell by recurrence of types, by running-title differences, and by printing conventions), and the work was done by formes (not seriatim) simultaneously in the two printing shops after casting off, probably at Danter's. Levenson points out that the raid on Danter's shop during Lent 1597 did not stop him printing so it cannot be used to date Q1, nor can the titlepage reference to performance by Lord Hunsdon's men because that could still be said after their name changed back. Thomas Creede's Q2's titlepage say it is "New corrected, augmented, and amended" and indeed it is more than 20% longer than Q1, has variants from Q1 in more than 800 of its lines, and some passages are totally different. There are few significant press corrections evident by collating the 13 extant copies of Q2, but many errors remain, apparently because of difficult copy. Two compositors, A and B, set Q2, A doing most of it and B helping at the end. Paul L. Cantrell and George Walton Williams have reconstructed from running-titles in the two skeleton formes the order and timetable by which the sheets were printed, but no-one has yet looked at Creede's other work to determine when in 1599 Q2 was printed. Q3, Q4, and Q5 were each a reprint their predecessor, although Q4 appears to have been informed by sporadic consultation of Q1. The Folio text is almost entirely derivative, being set solely from a copy of Q3 which was probably annotated by someone who knew the play in performance. Levenson herself collated the 7 extant copies of Q3 and found no press variants. There is no evidence that Q1 was an illicit publication, but the New Bibliographers called it 'bad' because of its shortness, its lack of a Stationers' Registry entry, and its alleged poor quality printing; none of these alleged deficiencies stands up to scrutiny. The "short quartos"--Levenson's less-judgemental term for those formerly called 'bad'--were disliked by the New Bibliographers because of the non-authorial influences found in them. By the mid-twentieth century Q1 was widely dismissed as a memorially-reconstructed pirating and Q2 was thought to have been printed mostly from Shakespeare's holograph with just the occasional bit of Q1 used as copy. Fifteen years ago, John Jowett's entry on Romeo and Juliet in the Textual Companion to the Oxford Complete Works placed him squarely in this tradition.

    Certainly one section of Q1 (1.2.53-1.3.36 in this edition) was used as copy for Q2, to judge from the fact that Q1's incidentals--such as the Nurse's speech being in italics--were closely followed, Moreover, Q1 appears to have been consulted elsewhere in the setting of Q2 and since this influence cannot be measured, cannot be distinguished from simple agreement between what Jowett would call MSQ1 and MSQ2, we reach what Levenson (somewhat exaggeratedly) calls "an impasse which blocks the search for copy and a stemma" (p. 117). As New Bibliography undergoes necessary correction by the New Textualists some well-washed conceptual babies will be discarded. Levenson asserts that there is "no contemporary evidence to verify that any actor(s) ever reconstructed a play memorially" (p. 118), which is strictly true for early modern England but not for Spain where, as Jesus Tronch showed, this was done and with precisely the textual corruption we should expect (2001). Peter Blayney hypothesized a non-piratical form of memorial reconstruction as actors made texts for friends by recalling their lines; since plays were routinely abridged for acting, this would make for short versions. The theory that the differences between Q1 and Q2 might be caused by authorial revision--either Q1 being a first draft enlarged to make Q2, or Q1 being a cut-for-pace version of Q2--cannot, Levenson insists, be excluded. Q2 has three moments of repetition which look just like second thoughts being printed alongside undeleted first thoughts, and this suggests its copy was authorial (pre-theatrical) papers since, as William B. Long showed, when theatre people interfered with a play manuscript it was to solve problems, and the repetitions in question cry out to be solved. On the other hand, the repetitions might represent revision well after original composition, or "may record different versions in different performances" (p. 123). In short, Levenson concludes, we cannot be certain of the copy for Q1 or Q2 and therefore cannot privilege one over the other. This view conditions the entire edition since if Q1 and Q2 "represent two different and legitimate kinds of witnesses to two different stages of an ongoing theatrical event" (p. 126) then Romeo and Juliet follows King Lear in becoming a play we can no longer consider as a single entity. Necessarily, then, Levenson edits both texts of Romeo and Juliet for this edition and prints both. Because of the tradition which takes Q2 as basic, she puts it first and prints Q1 with minimal apparatus. With both versions of the play present, Q1 appears in the Q2 collation "only when its readings bear significantly on the later text". Levenson has tried to interfere as little as possible in either text but has cut such things as the "potentially confusing duplications" of Romeo and the Friar's shared 'dawn' speech.

    Levenson delivers on her promise to trust Q2, so Mercutio describes Queen Mab drawn "Over men's noses as they lie asleep" (1.4.56), as Q2 has, not "Athwart men's noses" which is the more familiar Q1 reading. Likewise, at 2.1.86-7 Levenson prints "That which we call a rose | By any other word would smell as sweet" which reflects Q2 and eschews Q1's "By any other name", the familiar reading. The mortally-wounded Mercutio says "I am hurt. | A plague a both houses, I am sped" (3.1.90-1), which is Q2's reading, whereas Q1 has "A pox of your houses". Levenson resists the usual emendation "A plague a both your houses", which makes this phrase identical with what Mercutio says 10 lines later in his short speech about death, and it also regularizes the metre; she instead thinks that "irregularities suit the dialogue of a fight and its aftermath". Levenson keeps as much as possible to Q2's stage directions, so at the transition from the street to the Capulet party (1.4.112) she keeps "They march . . . forth with napkins", but she deletes Romeo's subsequent entrance since he never left the stage; explaining the staging possibilities here requires commentary that occupies more than 80% of the printed page. Sometimes traditional emendations provided the more interesting reading, such as at 2.1.39 where Levenson prints "An open-arse, or thou a popp'rin' pear" where Q2 has "An open, or thou a Poprin Peare" and Q1 has "An open Et caetera, thou a poprin Peare". Levenson might have trusted Q1 here, since "An open etcetera" makes sense: Mercutio will not name the open thing Juliet is to be. Having asserted a principle of minimal intervention, Levenson ought perhaps to have outlined at greater length the rationale of this emendation. The 'dawn' speech ("The grey-eyed morn . . . Titan's burning wheels" 2.2.1-4) Levenson gives to the Friar alone (as Q1 does) rather than to the Friar and Romeo (as Q2 has it). She argues that it is easy to see how Shakespeare might have written these lines for Romeo and than imperfectly deleted them and written a slightly different and improved version for the Friar, thus creating Q2's duplication. As C. J. Sisson pointed out, it is hard to imagine the opposite case of of Shakespeare writing these lines for the Friar and then retrospectively reassigned them to Romeo, since there would be no room (Sisson 1956, 154). Respecting Shakespeare's second thought of giving the lines to the Friar, Levenson uses the form of words given the Friar in Q2 for this speech.

    Actual mistakes by Levenson are hard to find. At her 3.1.122 (but really 121, she miscounted) Levenson has Romeo say of Tybalt "He gan in triumph and Mercutio slain?", which is Q2's reading. Levenson defends "gan" as either past participle (gone) or infinitive (to go). Levenson's Juliet, awaiting her lover, says "Spread thy close curtain, love-performing night, | That runaways' eyes may wink, and Romeo | Leap to these arms, untalked of and unseen" (3.2.5-7). Thus she uses Q2's reading ("runnawayes"), but having cited some alternatives and the argument about it her commentary does not tell the reader who are these runaways, their plurality being implied by the position of her possessive apostrophe. Sisson thought the correct reading "runaway's eyes" because the runaway is Romeo, whom Juliet (unaware that he is banished) fears might be too cautious and not come to her (Sisson 1956, 156). For the obvious problem regarding what flies may do at 3.3.40-43, Levenson follows the Oxford Complete Works's conjectures about the line "This may flyes do, when I from this must flie" being imperfectly deleted and rewritten as "Flies may do this, but I from this must flie" and this change, together with the associated reordering of lines and the composition of a new one ("They are freemen, but I am banished"), being misunderstood by the Q2 compositor. Levenson has Capulet described Paris as "youthful and nobly ligned, | Stuffed, as they say, with honourable parts" (3.5.180), turning Q2's "liand" into "ligned". Levenson's commentary says that "noble ligned" means coming from noble lineage, and that the silent 'g' makes also a possible pun on 'lined' which goes with 'stuffed' in the next line. Fair enough, but since "ligne" is an archaic spelling of straightforward modern word (OED line n.²), there seems no reason to retain the 'g' in a modernized text. When in Q2 Juliet says "Or bid me go into a new made graue, | And hide me with a dead man in his" (4.1.84-5), Levenson supplies the apparently missing final word with "tomb", as did the Oxford Complete Works, in preference to Q4's "shroud" and F's "graue"; Q1 has a different wording altogether but uses word "tombe". This would appear to be another case where Q2 might have been trusted, since there is no need for a final word: "in his" can refer back to the "grave" of the previous line. Likewise, Levenson has Paris say "Have I thought long to see this morning's face" (4.4.67) where Q2 has "thought loue", which Sisson defended as better than Q1's "long" when punctuated "Have I thought, love, to see" (Sisson 1956, 161-62); Levenson points out that the source uses "long" at this point. Finally, although Levenson is aware of Katherine Duncan-Jones's persuasive argument linking Juliet's autonomy with possession of her own knives (Duncan-Jones 1998), Levenson prints "She takes Romeo's dagger" (5.3.169) to kill herself. Perhaps Levenson was swayed by Capulet's assertion that "This dagger hath mista'en, for lo, his house | Is empty on the back of Montague, (5.3.202-3), but as Duncan-Jones argued this is just another example of the father's ignorance about his daughter.

    The titlepage of Stanley Wells's Oxford Shakespeare edition of King Lear say that he edited it "on the basis of a text prepared by Gary Taylor", which in this case means he started with the electronic text of the quarto ('History') version published on floppy disks by Oxford University Press in 1989. Wells's introduction is short (88 pages), in keeping with Oxford Shakespeare guidelines, and provides a useful summary account of the special two-text status of this play, the first for which a majority of scholars has accepted the principle that it cannot be considered as one thing. The original version of the play was written probably in 1605 and led to Q1 of 1608, and then it was revised, probably by Shakespeare, for revival in 1610 which led to the F text of 1623 which was printed from an annotated copy of either Q1 or the Q2 of 1619. An omission from the argument here is a statement of why anyone should accept that the alterations to the text represented in Q1 took place some years after an initial run of the play rather than as part of the preparations for first performance. It is because of this delay that Q1 and F are witnesses to two different versions of the play, rather than being merely witnesses to two stages in the genesis of a single play. The reading list at the end of introduction does not mention Gary Taylor and Michael Warren's The Division of the Kingdoms which convinced scholars that substantial revision separates Q1 and F, and although it appears in the larger reading list at the end of the book, the title running-headers to that section ("Offshoots of King Lear") hardly encourage readers to look there for further reading on textual matters. It is perhaps too obvious to Wells, but worth stating explicitly for most readers, that having established that the revisions seen in F were marked onto a copy of Q1 or Q2, these revisions could not have been made before Q1 was printed, thus not before 1608.

    The series policy has been to "base an edition on the text that lies closest to performance", which in this case would be F, but Wells decided to break with the policy because F is well-represented in other editions while Q1, a distinct version, has only recently been properly edited in its own right, and only two critical editions (by René Weiss and Jay L. Halio) have been published. Although, as observed above, the argument that about 5 year passed between the original performances and performances of the revised text is not outlined, Wells admits that some of the differences between Q1 and F are absences from MSQ1 (to borrow Jowett's useful labelling) which would have been rectified before the play's first run. Thus it is a "nice philosophical problem" (p. 8) to distinguish these things from the later major revisions. The nicety would be more apparent were the reader given at least a sketch of the reasons against attributing all the differences to this cause, such as the evidence adduced by Gary Taylor that F shows influence of Shakespeare's post-1605 reading matter and that its vocabulary is typical of his later work (Taylor 1983). Wells's approach is to accept Folio readings which are necessary to make Q performable (thus including the music cues), but not those readings which are unnecessary. Wells does not collate Folio variants nor print (even as appendices) the F-only passages, since the Oxford Complete Works provided these, but he does collate choices made by Weis and Halio with great assiduity that is usual since "the Quarto has only just entered the editorial tradition". Dating the initial composition, Wells's discusses the dependence on Samuel Harsnet's Declaration of Egregious Popish Imposture and on the chronicle history of King Leir which, to judge from its influence on his earlier plays, Shakespeare knew in the mid-1590s. Wells thinks that references to astronomical eclipses are of little use for dating the play since audience would always relate these to the last such events they remember.

    In the section 'Where the Play Came From' (pp. 14-31), Wells indicates King Lear's themes and character-types appearing in earlier plays: Titus Andronicus has an elderly mad tyrant, the Henry 6 plays explore division in the kingdom, and Constance in King John is "an enfeebled by eloquent grieving parent"; apparently the artistic gestation of King Lear was a long one. For the tangible sources, Wells traces the King Lear story in legend, starting with Geoffrey of Monmouth's Historia regum Britanniae which was not in print in Shakespeare's time but was circulating in manuscript, through Holinshed, Mirror for Magistrates, William Warner's Albion's England, Spenser's Faerie Queene, and most importantly of all, the play King Leir. The Edmund/Edgar subplot comes from Sir Philip Sidney's Arcadia, printed in 1590. Discussing how Shakespeare shaped the ideas, events, and persons from these sources, Wells's wide range of critical reading is apparent, from Leo Salingar to Marianne L. Novy and Jonathan Goldberg. It is arguable that King Lear is beginning to overtake Hamlet as the work for which Shakespeare is most known, but as Ann Thompson observes King Lear has no moments to match the synecdochical power of Yorick's skull or "To be . . .". Wells notes that the poetry of King Lear is subordinated to the dramatic effect, which is why so little of this play bears being quoted out of context, and I suppose this might explain the difference (p. 52). Wells handle so deftly the material of his stage history that it strains the reader not at all. Regarding the early performances, Wells thinks the ballad, which he reprints in full, "gives us what may well be unique eyewitness impressions of moments from the play as performed by Shakespeare's company" (p. 57). Wells charts the dominance of Nahum Tate's adaptation of King Lear, which diminished over the period when it held the stage (1681-1838) as more Shakespeare was put back. Wells defends Tate's play as doing to Shakespeare what Shakespeare did to King Leir. In the section 'Return to Shakespeare' Wells surveys nineteenth- and twentieth-century productions, with Harley Granville Barker's work as the watershed between an entirely readerly appreciation and the play's rehabilitation as theatrical work.

    The section 'Textual Introduction and Editorial Procedures' Wells, who has nothing to prove in this area, keeps to 8 pages (pp. 81-88). There are, he observes, numerous press variants in the 12 extant copies of Q1, but Wells argues against automatic acceptance of a reading in a corrected sheet since "the compositor may guessed" when he spotted an error rather than consulting the manuscript copy. The errors and the press variants suggest that the copy was hard to read, possibly authorial manuscript rather than fair copy, and it had not been through the theatre. In Wells's text Q1 is used for every reading unless it does not make sense, in which case all possible explanations of the error and all concomitant readings are considered, including those in F but without giving it special preference. Because the play was written before act intervals were observed in the open-air amphitheatres, Wells has simply numbered the scenes sequentially (as with the Oxford Complete Works), and not marked a new scene when Edgar enters while Kent sleeps in the stocks, nor another when Edgar leaves, Kent wakes up, and Lear enters. The collation records all substantive departures from Q1 (that is, those affecting meaning), with a selection of plausible editorial emendations not adopted. Variants where F departs from Q1 are not recorded, but adopted Folio readings are, like any other adopted readings, recorded. Wells usefully lists the places where he accepts Q1 readings which the Oxford Complete Works rejected, and where he imports readings from F or from one of the two critical editions of Q1. Regarding the punctuation, Wells aims to "increase comprehensibility for the modern reader without being over-prescriptive for the actor", which modest goal is a useful corrective to the absolutism of Ros King's claims about editorial intervention (reviewed below). Although Wells has been saying in public for some time that he is not convinced of the desirability of the practice, in this edition stage directions "whose content and/or placing are uncertain" and speech prefixes which are "disputable" are printed in broken square brackets

    For individual cruces there is little point comparing this edition to earlier ones, nor comparing Wells's choices with Sisson's, since as Wells points out Q1 has only recently been edited independently. However, a few choices deserve special attention. Wells has Kent say "Be Kent unmannerly | When Lear is mad" (1.136-7), using Q2/F's "mad" rather than Q1's "man" which makes sense and, as Peter Stallybrass argues (in an essay reviewed below), is perhaps better than "mad". At 1.176.1 the entrance of the King of France and Duke of Burgundy is accompanied by a musical flourish (as in F but not Q) because they are important men. However, for the exit of Lear and Burgundy at 1.256.1 Wells puts the "Flourish" in broken brackets, even though it too is marked in F. It is not clear how these things are different, unless of course Wells thinks royal exit directions are less likely overall to be marked with a music, or perhaps because the scene is somewhat disordered and leaving out this mark of Lear's importance could signal that. At 1.257-8 Wells has Cordelia say "The jewels of our father, with washed eyes | Cordelia leaves you", which is what Q and F have. Wells rejects Nicholas
Rowe's emendation, used in the Oxford Complete Works, to "Ye jewels of our father" made on the basis that the sense is "You jewels" and the manuscript probably had either "ye jewels" or "y^e jewels" which the compositor misread as the abbreviation for "the". Wells accepts that this might be true, but since Q1 also makes sense so he does not emend it. This is a tricky footnote for readers who do not know that the 'y' in such advertising signs as "ye olde tea-shoppe" stands for the letter thorn ŝ or Ŝ (so "the olde . . .") which was already by Shakespeare's time disappearing from manuscripts and almost universal represented by 'y' in print (OED Y (3)). At 2.125-6, when Edgar enters Edmund says "and out he comes, like the catastrophe of the old comedy, my cue is villainous melancholy", which is Q1's reading, instead of the more familiar "Pat: he comes" which is F's. Wells has the First Gentleman says that Lear "Strives in his little world of man to outscorn | The to-and-fro-conflicting wind and rain" (8.9-10). George Steevens's conjecture that Q1's "outscorne" was misreading of the manuscript's "outstorm" (followed in Kenneth Muir's Arden² and the Oxford Complete Works) is possible, Wells says, but "outscorn" also makes perfect sense. At 11.4 Kent says "Good my lord, enter" and, in the same metrical line, Lear responds "Wilt break my heart?" which is Q1's reading. F has "enter heere" which regularizes the metre, but as Wells observes on E. A. Abbott's authority "a missing syllable at the caesura is acceptable".

    The textual choice which will probably be noticed by most readers is Edgar's reciting "Child Roland to the dark town come" (11.65), from Q1's "darke towne" rather than F's "darke Tower". Wells admits himself tempted to follow the familiar F reading, but sticks to his principle that where Q1 makes sense he should follow it. John Jowett suggests a dark British town is here contrasted with the enlightened city of Athens ("Come, good Athenian" Lear says to Edgar immediately before this), but Wells was sufficiently unsure as to post a message on the SHAKSPER email discussion list to poll others' opinions. Nearly as noticeable, and based on precisely the same principle, is Wells's decision to have mad Lear say "It were a delicate stratagem to shoe | A troop of horse with fell" (20.173-4), using Q1's "fell" against F's "felt"; Wells points out that 'fell' is a perfectly good word meaning 'skin'. Finally, an example of punctuation being crucial to meaning is Lear's "This feather stirs. She lives. If it be so" (24.261), regarding the breathing of Cordelia. Q1 has "This feather stirs she liues, if it be so" and F has "This feather stirs, she liues: if it be so", either of which could be understood as two indicative statements (the feather does move, therefore she is alive), although Q1 lacks something between the clauses to indicate that they are separate. But also, Q1 or F could be read as one subjunctive statement: if this feather moves, then she must be alive. Wells admits both possibilities and says the choice is "open to the actor", but his period between "stirs" and "She" eliminates the subjunctive interpretation whereas a comma would have left the options more obviously open. On the other hand, it would create an error which North American students are most strenuously warned to avoid--the comma splice--if taken to be to one subjunctive statement. (Senior scholars are, of course, allowed such liberties: one seldom hears Terry Hawkes's famous dictum "Shakespeare doesn't mean, we mean by Shakespeare" being criticized for this grammatical fault.)

    Moving from editions to books about, or in support of, editing, Trevor Howard-Hill has revised and enlarged his , a repetitious title which illustrates that awkward terminology is one of the things that makes work in this area difficult. Most students do not discover until graduate work that 'bibliography' can mean more than just a list of books they have read. Howard-Hill's second edition of this book first published in 1972 was, he reports, declined by Clarendon Press because they thought it "premature" (p. v), but I would suggest rather that it is too late. This sort of printed bibliography has limited value to a scholarly community used to the electronic indices provided by the Modern Language Association International Bibliography and the World Shakespeare Bibliography. A spot-check failed to reveal any items in Howard-Hill's book which could not easily be found by a 'keyword' search in one or other of these. Increasingly such indices are adding evaluative descriptions of the items indexed--an area where Howard-Hill's book has obvious value--and in some cases (especially for recent work which was created electronically in the first place) the full texts of the items are also included in the database.

    The published proceedings of the conference "Ma(r)king the text" at Trinity College Cambridge in September 1998 has several excellent articles of general interest, but only one of direct relevance to this survey: Ros King's claim that modern editions spoil Shakespeare's metrics with by punctuation (2000). The sixteenth-century colon, King observes, was used not only to divide clauses but also to show that they are linked (a usage derived from the Hebrew psalms) and so editors should not, when confronted with "strings of clauses separated by colons", simply chop them up using periods. King looks at the setting to music of poems in the period, and from this concludes that ". . . what is most important for mid-sixteenth century prosody is the natural rhythm of the words". The problem for editors, of course, is that editors think punctuation a matter of sense and actors think of it as a matter of pausing. King believes that compositors were "first and foremost copyists" and so she thinks that they mostly got their lineation right, which view should be contrasted with Paul Werstine's (1984). King promulgates that common actorly view that ". . . the last word in any line is usually one of special importance that needs to be picked our or emphasized in some way", and of course if an editor has relined to the script the wrong word will be chosen. (One would like to see some evidence for this claim about the terminal word.) King thinks we should expect the silence around short lines to be filled with business or sound-effects, and urges editors not to simply settle for the choices of "the eighteenth-century poets who were his first editors".

    Another collection of essays this year is Andrew Murphy's , containing 7 essays of interest. Michael Steppat subjects Leah Marcus's book Unediting the Renaissance to 'discourse analysis' to show that it coerces the reader into alignment with unproven ideas (2000). Marcus's book is hardly worth such an effort, since its claims can be more easily dismissed, as Paul Werstine showed regarding Marcus's claim that Q and F Merry Wives of Windsor are independent versions separated by authorial revision (1999, 313n9). Steppat then applies the same analysis to an essay by Graham Holderness, Bryan Loughrey, and Andrew Murphy to rather more effect, showing that their application of Marxist terms about the value of plays makes it unclear how use-value relates to exchange-value, an all-too-common misunderstanding. Moreover, since Holderness, Loughrey, and Murphy's concern is for the labour that goes into performance (including that by non-authorial theatre people) and into printing (including that by scribes and compositors), Steppat spots that authorial intention has re-entered by the back door since even as part-sharer in that collective labour, the dramatist's efforts must be accounted for. Steppat points out that as general editors of Shakespearean Originals series of play reprints, Holderness and Loughrey claimed that these offered a "unique window on to the plays as they were originally performed" while at the same time insisting that these earliest texts are as far back as we can going without committing the error of trying to 'see through' the material object to something beyond it. As Steppat asks, why assume that the early texts are windows onto the theatre and nothing else?

    Peter Stallybrass's excellent essay argues that costumes, properties, and speeches, but not characters, are at the centre of early modern drama's production processes (2000). Shakespeare uses 'personal' names to indicate deprivation, that a person has lost their socially-ascribed name, and for him the important names are those given by function. Speech-prefix variation is not, as McKerrow claimed, a sign of authorial carelessness but often signals the point of the whole play, to reunite a personal name (such as Perdita) with a real social position (such as Princess). Names are better thought of as attached to properties (beard and dresses) rather than to actors; costumes not actors are at the centre of the early theatre's economics. Considering the permutations of one-to-many relationships in acting, Stallybrass repeats the error that more than one man played Demetrius in performances of Believe as You List, citing David Bradley as his authority. In fact Bradley, like C. J. Sisson, thought the practice highly unusual, and T. J. King has argued that the three names in the play manuscript come from three different men who played the part at different times (King 1992, 47). Stallybrass reads speech-prefix variation in relation to the plays' concern with social status in a number of Shakespeare works including Twelfth Night, Richard 2, Hamlet, and King Lear. There is always the danger that such interpretations misread randomness as art, especially where no early printing is consistent. Stallybrass defends Q1 King Lear's "Be Kent unmannerly when Lear is man" over Q2/F's "is mad", on the grounds that since Kent has notably left off Lear's title, is appropriate for him to refer to Lear's transition from monarch to mere man, and the word also suits Kent (whose name is a whole county) going from man to unmannerly. The modern concern for individuality and personal names is inappropriate, Stallybrass argues, for an understanding of how early modern drama was written. The manuscript of the play Sir Thomas More suggests that speeches were written then divided up between the main speaker and 'others', the 'others' being sorted out later.

    Laurie E. Maguire argues that editorial theory and practice follow 'grief theory', here grief at the loss of the writer's presence, upon which loss all literature is predicated (2000). The first half of Maguire's essay rather tediously describes funeral practices and likens editing to the tidying-up of the corpse (or corpus), and includes such banalities as "Life and death were closely linked in the early modern period". Poststructuralism at its silliest identifies everything as its own opposite, as when Maguire approvingly quotes Richard Lippert writing that a coffin "protects something precious at the same time its protection confirms loss", leading to Maguire's comments that "To use Derrida's formulation, the funeral, like the text, is the ultimate in 'presence' as 'generalized absence'". Derrida's point was that things are not merely their own opposites, but rather are self-contradictory in a productive, possibly Marxian dialectical, and fascinatingly unstable way. Maguire claims that the Renaissance was all about recovering the dead, the lost classical cultures, and although she admits that there was editorial/authorial intervention, in these textual resurrections "the living and the dead, comfortably co-exist" (p. 148). At this point Maguire could have invoked Derrida's zombie simile, since the classics were reanimated for distinctly presentist humanist motives and were not so much like Lazarus as the undead. Near its end Maguire's essay takes a remarkable turn for the better, arguing that twentieth-century denial of death--unmentionable and postponed by medical intervention--led to "untidying, unediting the body of the text" and hence the current denial of the finalized text. We now have the same flexible ideas about the end of life as the early moderns did: they allowed a corpse to be arrested for debt en route to its burial, and our machines can keep a human vegetable going indefinitely. New Bibliographical desire for one originary text parallels Freud's death drive, the desire to return to the inorganic state. Loss, Maguire concludes, is at the heart of literary writing because the author is always absent and the words stand in for him. We grieve for loss first by denying and idealizing, then by simultaneously grieving and celebrating, and finally by looking forward whilst also remembering; the point is "not to reduce these oppositions to singularity".

    Graham Holderness, Stanley E. Porter, and Carol Banks argue that the 1623 Folio of Shakespeare has much in common with the King James authorized version of the Bible (2000). The Shakespeare Folio and the King James Bible were alike in their print format (folio size, double columns, expensive paper), both represented that which was also available in oral performance as play and sermons, and both were supposed to provide definitive versions to oust inferior competition from the marketplace. (In fact, it is not clear that the Folio was intended to oust the quartos, which in any case continued to be reprinted.)  The new Bible and the 1623 Folio collected together what was fragmentary and monumentalized it to preserve a tradition, which meant choosing between competing existing versions of texts. Holderness, Porter, and Banks write that plays were "officially printed only when the theatres were temporarily closed, or when the company needed extra money, or if a particular play had ceased to draw the crowds profitably in performance" (p. 167), yet at the end of the preceding sentence they cite Peter Blayney's essay which specifically argued against these explanations and offered the new one that plays were printed for publicity purposes (1997). On the basis of this error the authors distinguish in F1 the 'old Testament' Shakespeare plays (those worn out ones already printed) from the 'new Testament' Shakespeare (the newer ones or older ones that still drew crowds, some not previously printed). Since their premise is wrong, this distinction is wrong. Such errors the authors repeat, writing that "It is generally assumed that the First Folio editors worked from such manuscripts [that is, foul papers, parts, and promptbooks] for all the plays in their collected edition . . ." and they go on to mention that in fact this was not so, some quartos were used as Folio copy. Of course, this faulty assumption is not generally made and it was known before New Bibliography began that quartos were used as Folio copy. Indeed, one of the founding steps of New Bibliography was to show that this did not matter, since only the good quartos were used and that Heminges and Condell's phrase "diuerse stolne, and surreptitious copies" referred to the other, bad, quartos. Since Holderness, Porter, and Banks footnote A. W. Pollard's Shakespeare Folios and Quartos a couple of sentences earlier (p. 175n34), one would expect them to know this: Pollard addresses it on pages 1 and 4.

    Errors abound in this essay: Charles Jasper Sisson loses his first name and becomes "Jaspar Sisson". The writers think that the Histories section of the 1623 Folio rearranged "the random, non-historical order in which they were performed" into a historically chronological sequence, but of course The Contention of York and Lancaster and Richard Duke of York were not randomly ordered. This fact the writers could have discerned simply by looking back over their own writing: on page 169 they give the Contention its full title of "The First Part of the Contention . . .". which label ("The First") clearly indicates historical non-randomness rather than random non-historicalness. Likewise it is self-evident that 1 Henry 4, and 2 Henry 4 (and possibly Henry 5) were historically ordered, and as argued above in the review of Edward Burns's Arden³ edition of 1 Henry 6, the printing and performance of the history plays shows a pre-Folio concern for historical orderliness. While it is true that the Folio strengthened the connections between the plays, and arguably imposed a teleological principle to the grand narrative so constructed, Shakespeare's histories can hardly be called "ten discrete stories" when so obviously partaking of the well-known two-part construction format. Holderness, Porter, and Banks think that the King James Bible and the Shakespeare Folio make a false unity of disparate materials, and now we must disintegrate them to release "from their authoritarian structures the many and varied utterances" of which they were made.

    Emma Smith traces scholarly desire to have something tangible as the ur-Hamlet, making it up if necessary (2000). Just as the ghost of Hamlet Senior haunts Hamlet, so the ur-Hamlet haunts Shakespeare's play, and Smith neatly summarizes the evidence that there was an ur-Hamlet existing in 1589-1595, the slight evidence that Kyd wrote it, and that it is related to a German play Der Bestrafte Brudermord. In a familiar pattern, Smith shows that the presence of the ur-Hamlet in Shakespeare's Hamlet was used to exonerate Shakespeare for that play's weaknesses: the bits critics did not like were taken to be Kyd's. Smith quotes bizarre 'reconstructions' of ur-Hamlet using the source story (Belleforest's account of Amleth) and Spanish Tragedy as guides, and observes that bibliography seems to need this old play as a justifying principle just as Hamlet needs to the ghost to justify what he does. Andrew Murphy's "Texts and textualities: A Shakespearean history" is a fairly standard anti-New Bibliography survey of Shakespeare editing and editorial theory, starting as is often the case with E. K. Chambers's 1924 British Academy lecture, which argued against multiple authorship and against John Dover Wilson's notion of 'continuous copy' for sullying the authorial purity of a manuscript capturing a single moment of a single man's work. Murphy articulates the fashionable view that Shakespeare's plays, indeed all plays of the period, are inherently collaborative. This claim is easily overstated: dramatists, not whole companies of actors, went to gaol for their plays, titlepages named Shakespeare as an dramatist in his lifetime, and accolades such as Francis Meres's were addressed to Shakespeare not the company. The afterword to the book is by Leah S. Marcus and called "Confessions of a reformed uneditor", which title (but not the essay itself) suggests that doing some editing has significantly altered the views Marcus advanced in her Unediting the Renaissance. Marcus refers to her co-edited text of Elizabeth 1's writings, which prints multiple versions of her speeches rather than trying to editorial produce something definitive. Marcus thinks that the Worldwide Web will help the move away from singularity, but notes that editors will have to acquire the technical skills for themselves since the technical specialists tend to get lured into better-paid commercial work. (This seems to be an allusion to the technical work of John Lavagnino, the general editor of the forthcoming Oxford Complete Middleton, to whom Gary Taylor elsewhere in this book rather ungraciously apportions the lion's share of the blame for that edition's delayed publication. Taylor's excuse that he lost his computer in a divorce settlement suggests third-world levels of technological poverty in American academia.) Like W. Speed Hill, Marcus thinks that collations are more ostentatious than practical and remarks that the extreme variations in versions of Elizabeth's 1's speeches would render full collation impractical, but nonetheless Marcus has had to print variants. Marcus acknowledges the editorial tension between wanting to not intervene and having for financial reason to modernize spelling and punctuation, and furthermore as a feminist she wants to raise the international profile of women's writing by making it as widely-readable as possible.

    Shakespeare Quarterly published eight articles on matters textual this year, and one review whose consequences make it worth reporting. Scott McMillin argues that New Bibliography has long, and E. A. J. Honigmann has recently, misrepresented the situation regarding Q1 Othello because in pursuit of authoriality and supposing a principle of textual economy, neither of which is reasonable (2000). Contrary to W. W. Greg and Honigmann, Q1 has theatrical features, most especially in its distribution of 'cuts', the 160 lines of the Folio text which it lacks. Greg thought that Q1 was printed from authorial foul papers because of its indeterminate and erroneous stage directions, whereas Alice Walker argued that it was printed from a theatrical manuscript made by a scribe who introduced things he remembered from performance. That Q1 was printed from foul papers, or a scribal copy of foul papers, was widely accepted in editions of the 1950s-90s, including the Oxford Complete Works and Honigmann's Arden³. Greg's argument was that Q1's vague and/or erroneous stage directions show its authorial origins, but he admitted that its omission of Folio lines is not random but rather seems theatrical; moreover, there are readerly features in Q1 (such as act divisions and literary stage directions) which suggest a copy made for a private patron. Greg needed to eliminate the theatrical-copy and private patron theories, so he supposed that the authorial papers had intended cuts marked on them, and these were obeyed by the printers, so that instead of being a theatre document, it is still an authorial document albeit by a man of the theatre. To eliminate the private-patron evidence Greg imagined that the copyist making this extra copy for a patron was the book-keeper himself, and thus the single author-to-theatre line of transmission is preserved. Of course, in reality the extra copy could have been made anywhere in the chain of transmission, but Greg was trying to limit the proliferation of texts in his New Bibliography.

    As Paul Werstine has long argued, New Bibliography always has to suggest the most economical lines of textual transmission so that the choice of copy remains binary: authorial papers or promptbook. Extra scribal copies for private patrons or for revivals are, wherever possible, eliminated as a possible source of copy and even Honigmann, who spends many pages discussing the habits of scribes and compositors, admits only one extra scribe: the one who copied the foul papers to make Q1's copy. This scribal copy Honigmann would only accept being "at one remove" from the author, whereas of course this oft-revived play (at court 1604, at the Globe 1610, at court 1612-13) could have generated descending 'trees' of multi-generational copies. Like Greg, Honigmann wanted to preserve the purity of the author-to-prompter genealogy. Honigmann excluded the possibility that Q1 is a 'bad' quarto (because it is unlike other bad quartos), so all that was left was "foul papers or a scribal copy of foul papers". McMillin asks a pressing question: what about theatrical copy for a revival, or a copy of that made for a private patron? (p. 72) The problem is with New Bibliography itself, which ignores some real possibilities. Honigmann's recent book on the texts of Othello (Werstine's review of which is discussed below) states that Q's omission of those 160 line cannot be due to cuts for performance nor due to Folio additions, yet also attempts to distinguish those lines which Honigmann thinks were Q cuts from those he thought were Folio additions. One of the reasons Honigmann gives for dismissing the idea that the 160 lines are cuts from Q is that they save only 8 minutes, which McMillin thinks is an under-estimate given that they contain the willow-song. Even if only amounting to 8 minutes of stage time, such a cutting in the right way might, McMillin argues, be worthwhile. Honigmann introduced what he thought was new evidence that Q1 is based on foul papers: it has some false starts. This view should be contrasted with Pervez Rivzi's claim that the Folio text has false starts mended in Q (1998). One of Honigmann's example is the double questioning of Emilia by Desdemona ("Wouldst thou do such a deed") in which the first answer ("I might doe it as well in the darke") is too jokey for the serious situation, so Shakespeare cut it and wrote a more appropriate one; the cut was overlooked by the printer. (An obvious objection here is that this example of double questioning occurs in Q and F, so it can hardly be used to argue that they had different kinds of copy.)

    McMillin points out that a hypothesis of multiple errors (a false start, an insufficiently marked cancellation, a printer who overlooked the cancellation) is awkward and that since Honigmann thinks Crane made from foul papers (or a copy of them) the transcript that lies behind F (which also has the putative false start) this hypothesis requires that Crane also missed the cancellation, which is unlikely. Honigmann has an answer to this: in parts F was set from Q. Another false start Honigmann finds is Cassio being said to be "almost dambed in a faire wife" (1.1.20, present in Q and F), which Shakespeare meant to cancel once he decided it were better Cassio were a bachelor. Honigmann lists some Q1 odd spellings, and a stage direction of the form "x driving in y", which he thinks are Shakespearian, but McMillin shows that that are either found elsewhere or merely odd and not indicative of copy. Because Q1 has "Enter Montanio, Gouernor of Cypres" (2.1.0), and this information is not available from the dialogue but only implied, Honigmann argued that this is a sign of the author to whom this occupation--never made explicit--matters. McMillin responds that anyone involved in costuming Montano would care, and this information could well be recorded on prompt copy. None of the things that Honigmann claimed indicate that the copy for Q was authorial papers (or an accurate scribal copy of them) are really persuasive, but tied to New Bibliography Honigmann had to choose between foul papers or promptbook, so he staked all on the former (p. 78). What of the copy for F? Honigmann decided that it was a scribal transcript of Shakespeare's revision of those same foul papers, and thus both printings are tied to foul papers. This, McMillin observes, realized New Bibliography's greatest hope of putting us back in touch with the authorial hand.

    McMillin believes that a