William Shakespeare
THE COMPLETE WORKS
Second edition
Edited by Stanley Wells, Gary Taylor, John Jowett, and William Montgomery
1,344pp. Oxford University Press. £30 (US $40); paperback, £16.99.
0 19 926717 0
William Shakespeare remains Britain’s best-selling author. One online bookseller offers nearly 20,000 titles bearing his name, and although they include much ephemera, probably more than a thousand separate editions are currently on offer. Several established series compete for those wanting individual plays. The Arden Shakespeare is over halfway through its third series, while the Penguin and the New Cambridge are already issuing updated versions of editions published not long ago. In an age when less Shakespeare is taught at school, those who discover him later in life will need more help. It is timely to ask, what do the punters get for their money, and which edition should they choose?
Bearing the prestigious imprint of Oxford University Press, and beautifully produced, the second edition of The Complete Works, edited by Stanley Wells, Gary Taylor, John Jarrett and William Montgomery, might seem an obvious recommendation. First published in 1986, it entered the huge American text-book market in 1997 when Norton, unwilling to generate their own Shakespeare, bought in the Oxford text. That seemed a strange decision at the time, given the many substantial criticisms that the Wells–Taylor edition had received, criticisms which the intervening twenty years have only intensified. Its original appearance was ill-starred, for (apparently owing to financial constraints) the first edition appeared with no commentary, although one had been prepared. Users had to wait until 1988 for its Textual Companion to appear, a huge and expensive volume which contained the editors’ justification for the many drastic changes they had made to the received text. For its second edition, the Oxford editors have made some welcome changes. It now includes the whole text of Sir Thomas More, to which Shakespeare contributed one scene and an additional speech, with helpful indication of the probable co-authors. It also adds the anonymously published Edward III, for which Shakespeare wrote three or four scenes. Now, at last, George Peele is acknowledged as the co-author of Titus Andronicus, although his contribution is understated.
These are welcome additions, as are an excellent essay, “The Language of Shakespeare”, by David Crystal, and helpful suggestions for “further reading” by Susan Brock.
But otherwise the volume has had minimum revision. A laughably brief “User’s Guide” has been added, which still refers us to the Textual Companion (out of print) for “information” about the editors’ decisions. No further explanations are provided. There are no notes, and the highly selective Glossary omits many difficult words. In the fluent and wide-ranging General Introduction by Wells, errors pointed out by reviewers have not been corrected: Shakespeare’s mother is still said to have died in 1609 (for 1608), and Francis Meres supposedly never mentioned Shakespeare’s narrative poems. In an interview, Wells admitted that he had written the introductions to individual plays in a morning, and it shows. While they include good thumbnail descriptions, as on the “sophisticated erotic comedy” of Venus and Adonis, or Shylock as the “first great comic antagonist”, Wells lapses too often into vapid appreciations, as on “the rapt wonder of the antiphon of recognition” in Twelfth Night, the “glorious celebration” of the imagination in A Midsummer Night’s Dream, “the rapt and impassioned poetry of the lovers” in Romeo and Juliet.
It is regrettable that the Oxford editors did not take this opportunity to make their Complete Works more helpful. Symptomatic of their rather aloof attitude to readers is their extraordinary decision not to add a Preface. Normally, in revising such a major enterprise as this, editors would be expected to describe any changes they have made, and reflect on the original edition’s reception, perhaps responding to criticism, or at the very least printing a list of the major reviews it had received. But it seems as if Wells and Taylor would like to bury those reviews. In 1990, they contributed a rather peeved essay, “The Oxford Shakespeare Re-Viewed by the General Editors”, to the now sadly defunct journal Analytical and Enumerative Bibliography. In it they complained that the first edition had “been characterized/caricatured in many reviews on the basis of one or two conspicuous features”, confirming to them that “it is easier for reviewers to fume about a couple of conspicuous innovations than to work their way diligently through three long volumes”. This condescending dismissal is transparently unjust to the scholars – some of the leading Shakespeare editors and textual critics of our time – who produced long and carefully considered estimates. They included two senior editors of rival Complete Works, David Bevington and G. B. Evans, and such vastly experienced scholars as R. A. Foakes, E. A. J. Honigmann, G. Walton Williams and Thomas Berger, among others. (Details of these reviews can all be found in the excellent online World Shakespeare Bibliography.)
The most notorious of the “conspicuous innovations” by the Oxford–Norton editors was a poem of seventy-two lines beginning “Shall I dye”, which Taylor ascribed to Shakespeare amid much publicity. Of the many peculiar features of this poem, the ambitious rhyme-scheme (in which six of the eight lines in each stanza have an internal rhyme, two of them ending with a double rhyme) is partly responsible for its jerky movement and banal poetic diction.
In a dreame, it did seeme
but alas dreames doe passe / as doe shaddowes
I did walke, I did talke
with my love, with my dove / through faire
meadowes.
This poet set himself a technical challenge which he was comically unable to fulfil. The only evidence connecting this feeble poem with Shakespeare is that the scribe of the Bodleian copy (made in the 1630s) wrote his name at the end; the other copy, at Yale, is unattributed. Taylor, however, at first view “felt it in my guts that it was Shakespeare”, and was convinced that “I [had] found the literary equivalent of Sleeping Beauty, a nameless poem awakening from the ancient sheets in which it had lain undisturbed for centuries . . .”.
Taylor set out his evidence for the ascription in the TLS (December 20, 1985), but the same issue contained a vigorous refutation by Robin Robbins, showing that Taylor’s argument was based on isolated verbal parallels between the poem and Shakespeare, ignoring negative evidence. Robbins easily demonstrated that other writers of the period had employed its stock phrases. Undeterred, Taylor and Wells published the lyric in a section euphemistically called “Various Poems”, an attribution which prompted many disbelieving readers to parody. When I surveyed the response to Taylor’s ascription in a book called “Counterfeiting” Shakespeare (2002) – one title notably absent from Susan Brock’s “Further Reading” – I found that none of the many scholars who had commented on the poem supported Taylor’s attribution. To reprint it in the second edition of Shakespeare’s Complete Works is a strange rebuff to the notion of scholarly consensus. In a paper given to the London Forum for Authorship Studies last year Taylor confessed that he wasn’t sure that Shakespeare wrote the poem, dismissed some of his critics, and pointed out some resemblances with Barnaby Barnes’s poetry. Yet in public the Oxford editors continue to defend one of the most ridiculous ascriptions to Shakespeare ever made.
The inclusion of this spurious poem (the 1997 Norton edition also found room for John Ford’s Funeral Elegy) was typical of a high-handed treatment of Shakespeare’s text. A feeling of superior historical correctness made them change three of the play titles given in the 1623 First Folio edition by Shakespeare’s fellow actors John Heminge and Henry Condell. The Oxford editors also knew better than Shakespeare, renaming his Falstaff in 1 Henry IV as “Oldcastle” (but allowing the received name to stand in the other three plays in which he appears), and correcting the Bard’s Italian: Brabantio and Gratiano in Othello became “Brabanzio” and “Graziano”, a change damaging the acoustic properties of Shakespeare’s verse. (Norton accepted the gratuitous Italian, but declined to rename Falstaff.) Supposedly giving us Shakespeare’s Complete Works, the Oxford–Norton editors chose to include material by other hands, dating from the 1620s and the 1980s. It is well known that two songs for the witches in Macbeth also occur in The Witch by Middleton, and that the Folio only quotes their opening lines. The Oxford editors, however, inserted the complete songs and took the liberty of describing the play as being “(Adapted by Thomas Middleton)”. They confidently declared that Measure for Measure was “Adapted by Thomas Middleton” (no parentheses now), and Taylor has included both Shakespeare plays in the forthcoming Oxford Middleton, an enormous boost to that dramatist’s oeuvre. The editors’ evidence (only published in 1993) for Middleton’s hand in Measure for Measure mostly concerns Act One Scene Two, where several stylistic features, and some dramaturgical loose ends, suggest a revision by Middleton in about 1621. While accepting their attribution, I find it perverse that the Oxford–Norton editors should have printed Middleton’s revised scene in their text, knowing that it was “made for Shakespeare’s company after his death”, and consequently relegating Shakespeare’s briefer and wittier original to an appendix called “Additional Passages”. But this textual waste bin should really be called “Passages Deducted by the Oxford Editors”. Answering critics’ objections to this procedure, Taylor disingenuously responded that deleted passages “have not been thrown away; they have simply been moved to a different place”, a remark which recalls Mary Douglas’s anthropological definition of dirt as “matter out of place”.
For Pericles, the Oxford–Norton editors took an even greater liberty, inserting into Shakespeare’s text passages from George Wilkins’s heavily plagiarized novella, The Painfull Adventures of Pericles Prince of Tyre . . . as it was lately presented . . . . (1608). By versifying Wilkins’s turgid prose, Taylor refashioned himself as a co-author. Such an intervention, properly acknowledged, might be legitimate for a theatre production, to conceal the deficiencies of this badly mangled text. But for the Oxford editors to insert the product of Taylor’s all-unable pen into “Shakespeare’s Complete Works”, without any typographical indication, shows a quite stunning disregard for the integrity of the canon. (The Norton edition at least documents the inserted passages.) There will probably never be a third edition, but if there were it would need major changes.
The worst instances of the Oxford editors’ high-handed treatment of Shakespeare’s text concerned plays which have come down to us in two editions, an earlier Quarto (a single play, of about thirty pages) and the collected Folio. In some cases, such as Troilus and Cressida, the differences are slight, the Folio adding a Prologue not found in the 1609 Quarto. (Unaccountably, the Oxford editors leave out Pandarus’ closing speech, although it is found in both substantive texts.) But others, such as Hamlet, have major differences. Where Shakespearian Quartos include lines not found in the Folio, or omit lines found there, editors have traditionally printed a text containing “all of Shakespeare’s Hamlet”, so to speak. But the Oxford editors, impressed by some arguments against “conflated texts”, decided to follow “the more theatrical” of the two versions – that is, the shorter. Since the Quarto of Hamlet is about 3,800 lines and the Folio “some 230 lines shorter” (although it “includes about 70 additional lines”), they argue that it represents Shakespeare’s revision, for theatrical purposes. Having chosen the shorter text, they then relegated sixteen snippets from the authentic 1604 Quarto, totalling 247 lines and including Hamlet’s soliloquy “How all occasions do inform against me”, to the “Additional Passages” waste bin. But at 3,535 lines the Folio text is still far too long for performance, and would have had to “lose” another 800 to 1,000 lines to fit the performance slot agreed with the civic authorities, between 2 and 5 o’clock in the afternoon. In their 1990 retrospective, Wells and Taylor admitted that they had made a “hopelessly confusing” presentation of the “complex” text issue in Hamlet, and now wish they had presented two separate texts. But they were concerned about the “bulk” of the edition, and it “was one of the last plays we edited; we were tired”.
Twenty years on, the Oxford editors’ reasoning for preferring the “theatrical text” seems obsolete. Behind it is their assumption that Shakespeare was a theatre person, whose work only achieved its final form on the stage. Any cuts in a play, they believe, must have had his approval, as an author who was, uniquely, a sharer both in the King’s Men and in the ownership of the company’s theatres. But anyone can see the flaw in this argument, that Shakespeare the sharer may have assented to cuts that Shakespeare the author regretted. As Alfred Hart showed in the 1930s, Shakespeare and Jonson were the only two Elizabethan dramatists who regularly wrote plays longer than the average length of 2,500 lines.
Despite the Romantic belief that Shakespeare only wrote for the moment, scorning the permanence of print, the fact that forty-nine Quarto editions of his plays appeared during his lifetime, more than those of any other dramatist, suggests that he was not wholly indifferent to print. These and other gaps in the received picture led Lukas Erne to publish a timely book, Shakespeare As Literary Dramatist (2003), challenging the “either/or” division and arguing that Shakespeare wrote both for the stage and for a reading public, with the longer versions designed for readers. Erne’s book has earned its favourable reception, and many scholars now see the folly of basing major editorial decisions on the theory that Shakespeare was essentially a theatre dramatist.
The Oxford editors’ stubborn adherence to their theatrical paradigm caused considerable damage to Shakespeare’s text, nowhere more so than with King Lear. In 1976, Michael Warren suggested that the 1623 Folio text of King Lear was a revision of the 1608 Quarto made by Shakespeare himself, the revisions being so great that the Folio version was essentially a different play. In 1983, Warren joined Taylor in editing a book of nearly 500 pages called The Division of the Kingdoms: Shakespeare’s two versions of “King Lear”. Believing this new orthodoxy to be definitive, in 1986 Taylor and Wells included in the The Complete Works two separate texts, The History of King Lear (Quarto) followed by The Tragedy of King Lear (Folio). (The Norton edition covered its bets by adding a third, conflated text.) In the second edition, which arranges the plays in chronological order (a good idea, if difficult to carry out precisely), the Folio King Lear is made to follow The Winter’s Tale and precede Cymbeline, which would suggest that Shakespeare went back to Lear in 1610, cut 300 lines, added another 100, and so created a new play. This theory attracted much adverse criticism from the outset, and a number of recent studies have demolished it.
If one consults a parallel-text edition, such as the helpful one by René Weis (1993), for page after page the two versions are practically identical. As Taylor himself conceded, the Folio-only passages include no new narrative material, no major incidents, no new scenes, only “alterations here and there”. It adds bridging passages, clarifying local points, but any in-house writer could have provided these. Only one speech in the Folio text seems to have been rewritten, but in such a clumsy way that, as Peter Blayney suggests, it looks as if two slips containing additions at different points have been treated as one. Three-quarters of the variations between Quarto and Folio consist of cuts: one whole scene, a conversation between Kent and a Gentleman, is omitted, and the astonishing scene where mad Lear, sheltering in a hovel from the storm, “tries” the absent Goneril and Regan, is shortened. The Folio also cuts numerous passages in the closing scenes. However, Lear’s role is unchanged, those of Edgar and Albany are slightly shortened and there are both cuts and additions in the Fool’s part. The obvious explanation is that this is standard theatrical surgery, mostly made to shorten playing time. For the exponents of a “bi-textual” Lear, however, no change is fortuitous, every cut is purposeful, and was carried out by Shakespeare.
But, as one scholar objected, we may think that we can recognize “a genuine Shakespeare line, but how does one recognize a genuine Shakespearean cut?” No one can tell who made these cuts, or when. They may have been made after Shakespeare retired to Stratford in 1612, or after his death in 1616. Some of them are carried out so badly that he cannot have been responsible. The cuts in Act Three Scene One make a hash of the scene (as Dr Johnson scathingly observed, “in the folio the messenger is sent, he knows not why, he knows not whither”), losing information essential to the action. As for the mock trial, where mad Lear’s denunciations of his daughters seem to us uncannily accurate, attributing to them vices that he hasn’t seen but we have, the reviser omitted his accusation of Goneril left in its follow-up, which begins “Then let them anatomize Regan”. The reviser also retained two speeches by Kent and Edgar responding compassionately to Lear’s hallucinations, which seem unmotivated following the cuts. Although Taylor haughtily dismissed any notion “that the mad trial was removed for vulgarly theatrical motives”, other advocates of a bi-textual Lear justified the cuts on the rather feeble suppositions that Shakespeare “streamlined” the scene, having noticed in rehearsals that it didn’t work, or was worried that his audience might be losing their concentration. To equate “theatricality” with cuts that simply shorten the playing time is to overvalue “action” at the expense of substance, the definition of character, the articulation of moral and spiritual crises, the audience’s emotional involvement, and much else that this harrowing scene delivers. Robert Clare, an academic who also works in the theatre, has challenged the two-text theory over the supposed “theatrical” improvement of cutting the trial scene. He showed that seven productions by the Royal Shakespeare Company and National Theatre between 1962 and 1993 all included it, even Nicholas Hytner’s 1990 RSC production, which used the Folio text. But Hytner’s Lear, John Wood, objected before rehearsals began that cutting the trial omitted an essential stage in Lear’s journey, leaving the scene lacking rhythm and focus – so the full text was performed.
Setting aside what Taylor calls a dozen or so “converts” to bi-textualism, the growing consensus among Shakespeare scholars, including at least seven who have edited King Lear, is that the variations between the two texts are not so great as to constitute two separate plays; that the alterations are theatrical, not authorial; and that the play loses more than it gains. Recent research by Richard Knowles, editor of the New Variorum edition, has highlighted other features of the Folio text which point to the intervention of an editor (perhaps the scribe who made a transcript from the theatre’s prompt book). Between 1608 and 1623, someone standardized Shakespeare’s old-fashioned undifferentiation of the relative pronouns “who”, “which” and “that”, to reflect changing usage, and regularized the metre of over a hundred verse lines to produce regular iambics. Given the remarkable metrical freedom of Shakespeare’s late style, it is inconceivable that he would have bothered to tidy up an extant play in this manner.
The lessons to be learned from this attempt to reify two King Lears are salutary. Most important, the challenge to received belief was not based on bibliographical or textual evidence. The bi-textualists’ claims rested on dubious aesthetic arguments concerning character, dramatic structure and theatricality. As Andrew Gurr observed, editors usually invoke aesthetic criteria when the mechanical evidence derived from manuscripts and printing practices is inconclusive. But the bi-textualists deployed aesthetic arguments as “the primary evidence for two equally authoritative versions of King Lear, and the mechanical evidence is being reassessed to see if it supports” the aesthetic arguments, a topsy-turvy procedure. Moreover, the two-text converts ignore the mechanical evidence, an extremely surprising oversight given that the 1980s also brought us Blayney’s remarkably detailed study of King Lear in the printing house. For Nicholas Okes, the 1608 Quarto was the first play text he had confronted, and it was marred by poor preparation of the manuscript, bad printing and proofreading, and some idiosyncrasies. Having made Shakespeare’s name dominate the title page in larger type than ever used before, Okes also had the concluding “FINIS” set in large type, reducing the space available on the final page. Blayney showed that Okes regularly left the outside pages of a Quarto blank “so as to protect the print when folded or stitched copies were stored without wrappers”. This meant that the compositor setting King Lear ran out of space for the final page and had to take some desperate steps, setting verse as prose and introducing many errors. Such a banal mechanical cause can produce effects which would support any number of aesthetic interpretations.
The other lesson to be learned from the demise of the two-text theory is that scholars should always consider alternative explanations. Much of the reasoning used by Taylor and his fellow converts was characterized by intense promotion of their case, constantly quoting each other’s work, using heavily-weighted terminology, and either ignoring contrary arguments or sarcastically dismissing them. In a paper given in 1988, Taylor asserted his credentials, having spent “ten years of polemical participation” in editing Shakespeare, which resulted in “the most thoroughly researched and documented edition . . . published this century”. Having proudly aligned himself with “the two-text revolutionaries”, who apparently did “most of the substantive new work” on Shakespeare’s text in the 1980s, Taylor felt justified in denouncing as “reactionaries” whoever disagreed with them. According to Taylor, in criticizing the Oxford edition “the venerable Frank Kermode leads a chorus of gerontocrats”, and Taylor also waved away the adverse comments of David Bevington and other scholars as signs of “emotional resistance” to the truth, which “are not capable or worthy of ‘refutation’”. Taylor’s adversarial rhetoric harks back to the eighteenth century, to the vicious ad hominem attacks on Lewis Theobald and Edward Capell made by rival editors. Taylor’s so-called revolution sounds more like the end of an old age than the beginning of a new one. In the Textual Companion he spoke on behalf of the other Oxford editors in “looking forward to our obsolescence”: it has already happened. The Norton editors might reconsider their options: do they really want to devote 250 pages to three texts of Lear?
Of the other American one-volume texts, the Riverside edition by G. B. Evans (Houghton Mifflin, second edition 1997) remains the best buy. It has helpful notes on the page, full textual collations, and introductions by Anne Barton and Frank Kermode which remain fresh and illuminating. The rival Longman edition by David Bevington (4th edition, 1997) also has useful notes and contextual information. The only English competitor is the one-volume Arden Complete Works (revised edition, 2001). It is handsomely printed, with reliable texts from the second and third series, and a much bigger glossary than Oxford’s, but with no notes. The market is open for a selectively annotated edition, which would give the ordinary reader some much-needed help with Shakespeare’s language.
